DESCRIBING AND EXPLAINING CROSS - NATIONAL PUBLIC OPINION ON CLIMATE CHANGE By Brian Kennedy A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Political Science Do ctor of Philosophy 2015 ABSTRACT DESCRIBING AND EXPLAINING CROSS - NATIONAL P UBLIC OPINION ON CLIMATE CHANGE By Brian Kennedy Climate change could be the most important international public policy challenge of the 21st century. Its consequen ces, from rising seas to more violent storms, could change economies and societies. And it is an international problem requiring an international solution. So understanding global public opinion on climate change is important to understanding how the inter national community might respond. But global public opinion on climate change remains under explored. This project tackles two questions. First, what explains cross - national public opinion on climate change? Second, how does American public opinion on clim ate change situate in the comparative context? Around the world, people are concerned about climate change but not alarmed. Although large majorities express that climate change is dangerous and serious, only small minorities see the environment and clima te change as issues more important than the economy, health care, or even pollution. Climate change is not a front - burning issue. There is important variation around the world in how people perceive the risks of climate change too. Those with more educatio n and on the political left are more likely to say climate change is a serious problem than those with less education and on the political right. Citizens of richer countries, however, are not more likely than citizens of developing countries to say climat e change is serious, so policymakers should not assume that publics in wealthy countries will accept greater sacrifices as part of a climate change policy solution. Compared to other wealthy countries, the percentage of Americans who say climate change is not serious and dangerous stands out. Americans are also unlikely to be willing to pay higher prices to address climate change and rank the environment and climate change as more important than other issues. What explains why Americans are comparatively l ess worried? The American public is far more ideologically polarized on climate change, with many Americans seemingly taking cues from political similar elites. The most politically engaged and knowledgeable Americans are the most divided by ideology and p arty affiliation. And in no other wealthy countries is the public so polarized on climate change. The deep polarization in the United States figures to be an obstacle to international policy addressing climate change. iv To Mom, Dad, and Craig th ank you for everything v ACKNOWLEDGMENTS I did not go to graduate school to study public opinion on climate change, the environment, and in year s ahead. I would first like to thank my dissertation committee chair, Charles Ostrom, for his support, guidance, and encouragement in the process. He was a constant source of confidence throughout my graduate school career. Chuck was completed committed t o this idea and the plan, providing faith I would also like to thank other facul ty members who at one time or another served on my committee: Cory Smidt, Mike Colaresi, Josh Sapotichne , and Paul Abramson. All were generous with their help and a pleasure to work with. Cory played an especially important role in my graduate school caree r, inspiring my strong interest in quantitative research methods and providing a supportive comment when I needed it most. Many other political science faculty members at Michigan State helped me over the past six years. Thanks to them all. I would like t o particularly thank Valentina Bali for allowing me to teach in and get involved with the Public Policy Program over the last two years. This opportunity helped me directly and indirectly finish. I was extremely lucky to have many great graduate student c olleagues during my time at MSU. Bob Lupton, Melanie Whitley, Masaaki Higashij ima , Chunho Park, Sung - Min Han, and many others contributed in so many ways without knowing it. I would also like to thank friends from outside my graduate school world for their support Matt, Matt, Amy, Paul, Sa rah, Joel, and Molly, thanks for vi the laughs and visits that were too far in between coursework, comprehensive exams, and dissertation writing. Luckily, they will be more frequent now. To my Davidson roommates Richardson, Garrett, and Chris thanks for all t he help over the years and allowing me the opportunity to attend two of your weddings and one Kiawah weekend during my graduate school years. It is hard to believe that we met 13 years ago this week. Thanks also to all my second Belkers, including the inco mparable Teddy, for taking my money in fantasy football and reminding me that I was always more than a graduate student. Last , thank you to my brother Craig for years upon years of direct encouragement. Indirectly, his great success is a great source of inspiration. And my parents provided so much support from so far away during this journey, beginning more than a quarter century ago in kindergarten. Thanks for encouraging me to keep going, from Davidson to DC to graduate school and back t o DC, through good and bad times. vii TABLE OF CONTENTS .............................. . x LIST OF FIGURE .......................... . x v i 1.1 Why climate c ... 1 1.2 2 1.3 .4 Chapter 2: How important do people consider climate change?.................................................................. .7 2.1 Introducti 2.2 Previous work describing cross - .......8 2.4.1 Climat 2.5.3 Strength of concern about climate cha 2.6 Climate chang .33 .38 ..40 3.2 Individual wealth and post - material 41 3.2.1 Post - 3.2.2 The affluence hypothesis.. 3.2.3 Alternative expla 3.3.2 Opposing the White 3.4 Left - .49 3.5 Education and engageme .51 3.7 Methodological approach and estima 3.6.1 Multilevel mode 3.6.2 Exploratory factor an Chapter 4: What ex plains cross - national climate change viii 4 ..72 4 .3.2 Independent variables at the ind ividual - 4 .3.3 Control variables at the indiv idual - 4 .3.4 Independent variables at the c ountry - 4.4 Explaining cross - national climate change 4.5 Conditional relationship between left - right political orientation and education/interest in 4.6 Robustness Check 90 4.7 Discussion and conclusio Chapter 5: Where does left - right political orientation matter 5.2 Why wealth, development, and communi st experience matter 5.5 Use of left - right political orientation acr 5.6 The varying influence of left - r ight political orientation on cli mate change risk perception..105 5.7 Discussion and conclusi Chapter 6: Why is the American public far more polarized on climate change than other wealthy 6.2 The elites: United States and the rest 7 6.4 Data and methodolo 6.5 The effect of left - right political orientation on climate 6.6 The conditional effect of political knowledge i 6.7 The conditional effect of interest in politics in the United States an ...133 6.8 Discussion and conclusi Chapter 7: How do individuals think about climate change and 7.2 Data, methodo logy, and measuremen 7.3 Relating climate change risk perception and risk perception of other environmental 7.3.1 Relating perceived dangerousness of climate change and perceived dan gerousness 7.3.2 Perceived dangerousness of climate chang 7.4 Relating climate change risk perception and 7.5 Climate change risk percept ion and ranking the environment and climate change as most 7.5 .1 Climate change risk perception and considering the environment a most important 7.5 .2 Climate change risk perception and climate change as the most important environmental pro 7.6 Discussion and conclusi ix 8.4 A general policy response the global public co 175 7 8 APPENDICIES . Appendix A: Data, survey questions, measurement, and v ariables ... Appendix B: Additional models, analysis and robustness c hecks for Chapter 4 Appendix C: Additional models, analysis and robustness checks for Chapter 5 235 Appendix D : Additional models, a nalysis and r obustnes Appendix E: Additional models, a n alysis and robustness checks for Chapter 7 .. 293 REFERENCES. x LIST OF TABLES Table 3.1: Theory and empirical expect Table 4.1: Theory, hypotheses, and empirical expectation Table 4.2: Random - intercept ordered logistic models without country - level variables predicting perceived Table 4.3: Random - intercept ordered logistic models with country - level variables predicting perceived Table 4.4: Theory, hypotheses, and empirical expectations, and results Table 5.1: Random - intercept binary logistic models predicting not usin g the left - Table 5.2: Random - intercept, random - coefficient ordered logistic models predicting perceived Table 7.1: Exploratory factor analysis of perceived dangero usness of environmental problems in all Table 7.2: Random - intercept binary logistic models predicting agreeing to pay higher prices to address climate chang Tabl e 7.3: Random - intercept, random - coefficient binary logistic models predicting considering the environment the most or next most importan Tabl e 7.4: Random - intercept, random - coefficient binary logistic models considering cli mate change the 6 Table A1.1: Participating countries and sample i nformation for the World Values Survey (2005 - 2009 ) .......................................................................... ................................................................................181 Table A1.2: Part icipating countries and sample i nformation for the International Social Science Programme Survey (2009 - 2011) Table A1.3: Pa rticipating countries and sample information for the Pew Global Attitudes Survey (2007) Table A1.4: Participating countries and sample information for the Pew Global Attitudes Survey (2008) Table A1.5: Participating countries and sample information for the Pew Global Attitudes Survey (2009 Table A1.6: Participating countries and sample information for the Pew Global Attitudes Sur vey (2010) xi Table A1.7: Chapter 1 Survey questions, measurements, and variables Table A1.8: Chapter 3 Survey questions, measurements, and variables Table A1.9 Cha pter 4 Survey questions, measurements, and variables Table A1.10: Chapter 5 Survey questions, measurements, and variables Table A1.11: Chapter 6 Survey questions, measurements, and variables Table A2.1: Random - inter cept multilevel linear models with random country intercepts and polynomial terms for age predicting perceived seriousness of climate change Table A2.2: Ordered logistic multilevel models with random country intercepts and polynomial ter ms for GDP per capita and HDI predicting perceived seriousness of climate change Table A2.3: Random - intercept multilevel ordered logistic models with individual - level interactions predicting perceived seriousness of climate change Table A2.4: Random - intercept multilevel linear models without country - level predictors predicting perceived seriousness of climate change Table A2.5: Random - intercept multilevel linear models with country - lev el predictors predicting perceived seriousness of climate change Table A2.6: Random - intercept multilevel ordered logistic regression models without country - level predictors predicting perceived seriousness of climate change on multiple imputation datasets Table A2.7: Random - intercept multilevel ordered logistic regression models with country - level predictors predicting perceived seriousness of climate change on multiple imputation datasets Table A2.8: Random - intercept multilevel linear models predicting perceived dangerousness of climate change Table A2.9: Random - intercept multilevel linear models with polynomial terms for age predicting perceived dangerousness of climate change Table A2.10: Random - intercept multilevel linear models with individual - level interactions predicting perceived dangerousness of climate change Table A2.11: Random - intercept multile vel linear models predicting perceived dangerousness of climate change on multiple - imputation datasets Table A3.1: Random - intercept , random - coefficient multilevel linear models predicting conc ern about environmental issues Table A3 .2: Random - intercept , random - coefficient m ultilevel linear models predicting perceived dangerousness of climate change Table A4.1: Ordered logistic regression models predicting p erceived seriousness of environmental problems in the United States xii Table A4.2: Ordered logistic regression models predicting perceived seriousness of environmental problems in Australia Table A4.3: Ordered logistic regression models predicting perceived seriousness of environmental problems in Canada Table A4.4: Ordered logistic regression models predicting perceived seriousness of environmental probl ems in Chile Table A4.5: Ordered logistic regression models predicting perceived seriousness of environmental problems in Italy Table A4.6: Ordered logistic regression models pre dicting perceived seriousness of environmental problems in Spain Table A4.7: Ordered logistic regression models predicting perceived seriousness of environmental problems in Switzerland .245 Table A4.8: Ordered logistic regression models predicting perceived seriousness of environmental problems in South Korea Table A4.9: Ordered logistic regression models predicting perceived seriousness of environment al problems in Norway Table A4.10: Ordered logistic regression models predicting perceived seriousness of environmental problems in Finland Table A4.11: Ordered logistic regressi on models predicting perceived seriousness of environmental problems in Germany Table A4.12: Ordered logistic regression models predicting perceived seriousness of environmental problems in Sweden Table A4.13: Ordered logistic regression models predicting perceived seriousness of environmental problems in Japan Table A4.14: Ordered logistic regression models predicting perceived seriousness of environmental problems in Mexico Table A4.15: Ordered logistic regression models predicting perceived seriousness of environmental problems in Slovenia Table A4.16: Ordered logis tic regression models predicting perceived seriousness of environmental problems in Turkey Table A4.17: Ordered logistic regression models predicting perceived seriousness of environmental problems in Poland Table A4.18: Linear regression models predicting perceived seriousness of environmental problems in the United States xiii Table A 4. 19: Linear regression models predicting perceived seriousness of environmental problems in Italy Table A4.20: Linear regression models predicting perceived seriousness of environmental problems in Spain Table A4.21: Linear regres sion models predicting perceived seriousness of environmental problems in Switzerland Table A4.22: Linear regression models predicting perceived seriousness of environmental problems in South Korea Table A4.23: Linear regression models predicting perceived seriousness of environmental problems in Norway Table A4.24: Linear regression models predicting perceived seriousness of environm ental problems in Finland Table A4.25: Linear regression models predicting perceived seriousness of environmental problems in Germany Table A4.26: Linear regression model s predicting perceived seriousness of environmental problems in Canada Table A4.27: Linear regression models predicting perceived seriousness of environmental problems in Australia Table A4.28: Linear regression models predicting perceived seriousness of environmental problems in Japan Table A4.29: Linear regression models predicting perceived seriousness of environmental problems in Mexico Table A4.30: Linear regression models predicting perceived seriousness of environmental problems in Slovenia Table A4.31: Linear regression models predicting p erceived seriousness of environmental problems in Turkey Table A4.32: Linear regression models predicting perceived seriousness of environmental problems in Chile Table A4.33: Linear regression models predicting perceived seriousness of environmental problems in Poland Table A4 .34: Linear regression models predicting perceived seriousness of environmental problems in Sweden Table A4 .35: Regression models predicting climate change attitudes in United States (interaction between political ideology and political knowledge xiv Table A4.36: Regression models predicting climate change attitudes in United States for liberals and conservatives Table A4 .37: Regression models predicting climate change attitudes in United States for liberals and conservatives interest in politics ins tead of political knowledge Table A4 .38: Binary logistic model predicting cause of climate change in United States for liberals and conservatives attention to politics instead of political knowledge Table A4 .39: Reg ression models predicting climate change attitudes in United States for liberals and conservatives interest in politics instead of political knowledge Table A4 .40: Binary logistic models predicting cause of climate change in United Sta tes for liberals and conservatives attention to politics instead of political knowledge Table A4.41: Ordered logistic regression models predicting perceived seriousness of environmental problems in the United States among those on the left Table A4.42: Ordered logistic regression models predicting perceived seriousness of environmental problems in the United States among those on the right Table A4.43: Ordered logistic regression models p redicting perceived seriousness of environmental problems in the Switzerland among those on the left Table A4.44: Ordered logistic regression models predicting perceived seriousness of environmental problems in Switzerland among tho se on the right Table A4.45: Ordered logistic regression models predicting perceived seriousness of environmental problems in Canada among those on the left Table A4.46: Ordered logistic regression model s predicting perceived seriousness of environmental problems in Canada among those on the right Table A4.47: Ordered logistic regression models predicting perceived seriousness of environmental problems in Norway among those on the left Table A4.48: Ordered logistic regression models predicting perceived seriousness of environmental problems in Norway among those on the right Table A4.49: Ordered logistic regression model s predicting perceived seriousness of environmental problems in Germany among those on the left Table A4.50: Ordered logistic regression models predicting perceived seriousness of environmental problems in Germany among those on the right Table A4.51: Ordered logistic regression models predicting perceived seriousness of environmental problems in Australia among those on the left Table A4.52: Ordered logistic regression models predicting perceived seriousness of environmental problems in Australia among those on the right xv Table A4.53: Ordered logistic regression models predicting perceived seriousness of environmental problems in Sweden among those on the left Table A4.54: Ordered logistic regression models predicting perceived seriousness of environmental problems in Sweden among those on the right Table A4.55: Ordered logistic regression models pr edicting perceived seriousness of environmental problems in Japan among those on the left Table A4.56: Ordered logistic regression models predicting perceived seriousness of environmental problems in Japan among those on the rig ht Table A4.57: Ordered logistic regression models predicting perceived seriousness of environmental problems in Turkey among those on the left Table A4.58: Ordered logistic regression models predic ting perceived seriousness of environmental problems in Turkey among those on the right Table A5.1: Random - intercept multilevel linear models predicting perceived dangerousness of environmental problems Table A5.2: Linear models predicting perceived dangerousness of environmental problems, United States .300 Table A5.3: Correlations with perceived dangerousness of climate change across countries Table A5.4: Exploratory factor analysis with promax rotation of perceived dangerousness of environmental problems 5 Table A5.5: Logistic regression models predicting considering the environment most important or next most important issue, by country Table A5.6: Logistic regression models predicting considering climate change the most important environmental problem, by country xvi LIST OF FIGURES Figure 2.1: Perceived seriousness or dangerousness of climate change in all countries .. Figure 2.2: Mean perceived dangerousness of environmental problems in all countries .. Figure 2.3: Most important issues facing country in all countries... Figure 2.4: Most important environmental problem in all countries. Figure 2.6: Mean perceived seriousness by country, Wor Figure 2.7: Mean perceived dangerousness by country, Internationa Figure 2.8: Willingness to pay to address climate ch Figure 2.9: Mean perceived dangerousness of environmental problems , Figure 2.10: Environment as most or next most important Figure 2.11: Most important environmental problems 33 Figure 2.12: Perceived worry about climate chang e in the United States, 1997 - Figure 2.13: Perceived seriousness of climate change in t he United States, 2006 - Figure 2.14: Mean perceived seriousness of climate change , by c ountry, 2007 - .37 Figure 4.1: Re lationship between wealth/development and perceived seriousness of climate change by Figure 4.2: Effects of individual - level variables o n predicting perceived se riousness of climate change... ............................................................................................................................. ... ....................... 87 Figure 4.3: Conditional effect of education/interest in politics on left - right political ideology o n predicting Figure 5.1: Percentage who do not answer the political left - right continuum questio n by country Figure 5.2a: Effect of left - right political ideology on predicting perceived seriousness of climate change across values of natural logarithm of mean GDP per cap Figure 5.2b: Effect of left - right political i deology on predicting perceived seriousness of climate change across values of mean HDI scor e, by 108 Figure 5.3: Conditional effect of left - right political ideology on predicting considering climate change very serious ac ross values of mean GDP per capita, mean HDI, OECD, and post - xvii Figure 6.1: Effect of left - right political ideology on predicting considering climate change very seriou s in Figure 6.2: C onditional effect of political ideology on predicting climate change opinions across values of political knowledge in the United State Figure 6.3: Effect of interest in politics on predicting considering climate change very serious for the left and right in select OECD countries Figure 7.1: Factor loading plot, Can Figure 7.2: Factor loading plot, United St Figure 7.3: Relatio nship between perceived seriousness of climate change and willingness to pay to Figure 7.4: Relationship between perceived dangerousness of climate change and considering the environment t he most or next most important issue 9 Figure 7.5: Relationship between perceived dangerousness of climate change and considering climate change the most important environmental probl 6 4 Figure 7.6: Conditional effect of perceived dangerousness of climate change on predicting considering the environment the most or next most important issue and climate change the most important environmental problem across values of GDP per cap Figure A4.1: Percentage of respondents to say climate change is very serious by political ideology and interest in politics, by OECD country Figure A5.1a: Effect of perceived dangerousness of climate change on c onsidering the environment the most or next most important issue for country across natural logarithm of mean GDP per capita by country Figure A5.1b: Effect of perceived dangerousness of climate change on conside ring the environment the most or next most important issue for country across natural logarithm of mean HDI by country Figure A5.2a: Effect of perceived dangerousness of climate change on considering climate change the most important environmental problem facing country across natural logarithm of mean GDP per capita by country Figure A5.2b: Effect of perceived dangerousness of climate change on considering climate change the most important environmental pro blem facing country across mean HDI by country 5 1 Chapter 1: Introduction 1 .1 Why climate change Climate change could be one of the most important international public policy issue s of the 21 st century. Scientific evidence supporting climate change has accumulated over the past decade and a strong scientific consensus that climate change is the result of huma n activity has emerged. About 97 percent of climate scientists agree on anthropogenic climate change ( Anderegg, et al. 2010 ). If climate change is not addressed in a meaningful manner , its consequences could be devastating for human life ( Giddens 2011 ). The Intergovernmental Panel on Climate Change of the United Nations (IPCC), in its 2007 report, claimed the effects of climate change already included glacial lakes, melting ice sheets, and m ajor changes in the Arctic and Antarctic ecosystems. It warned of future consequences including the flooding of coastal cities, resource - based wars, and more extreme, violent weather. Some of the damage could result in irreversible changes. Climate change of this magnitude requires a global response and major policy changes. A major policy initiative requires committed politicians and a supportive public. Public opinion is important to policy making, especially in democracies. V.O. Key considered government responding to ( Key 1961 ) . The public, both in the United States and a round the world, plays and play s an undeniably important role in shaping the policy response to climate change. Climate change is a g lobal problem requiring an international response. The global public could drive an appropriate response or impede such an appropriate response. Public opinion, including policy preferences and substantive attitudes, could determine the policy response to climate change. Understanding global public opinion is important to public policy on climate change, yet it is underexplored. We still do not have a clear picture as to ho w so many people around the world think about climate change and what influences their opinions. This project seeks to provide clearer picture of pu blic opinion on climate change in two ways: by studying cross - national public opinion and United 2 States publ ic opinion in comparative perspective. Public opinion on climate change will determine the available options to mitigate the consequences of climate change for policymakers in democratic countries (and even non - democratic countrie s), where policymakers are more likely to be responsive to the public. This project also contributes to the study of public attitudes on complicated environmental and science issues. Complex science issues other than climate ch ange, such as childhood vaccines and genetically modif ied foods and energy issues , such as nuclear, wind, and solar power, are often prominent topics in the news. And some people reject scientific consensus on all of these issues, including climate change. Why do some individuals in the United States and cros s - nationally reject the scientific con sensus on climate? Are those skeptical of anthropogenic climate change persuadable? These questions are also insights on climate change could extend to other science and environmental topics. 1 .2 Important i nsights This project focuses on the following questions : What do people think of climate change? How do people form opinions on climate change? And how does American public opinion on climate change compare to the rest of the world? Based on my analysis , this project comes to these conclusions. The public, both in the United States and around the world, express worry about climate change in the abstract. Large majorities in cross - nation al surveys consider climate change serious and dangerous. But, the global public a cross cross - national surveys is far more reluctant to make an economic sacrifice to address climate change and r anks different general policy priorities and environmental iss ues as more pressin g. Even among those most concerned about climate change, other issues and environmental problems are ranked as more important than climate change. For many Americans and those aro u nd the world, climate change is a vague, weak worry, ofte n trumped by immediate, everyday - life general concerns like the economy, health care, and more pressing environmental concerns such as local air and water quality . Why does the public in the United States and around the world express concern about climate change but also consider it 3 unimportant? Unlike many public policy issues, such as guns and even pollution, climate change is both complex and abstract. Climate change is a complicated issue and many citizens do not possess adequate information to form dec isions on it. Moreover, unlike local air and water pollution, climate change is not concrete ; there are not visible and unequivocal indicators or even long - term visible consequences . People cannot see or feel the consequences of climate change, even if the y sometimes think of climate change during a particularly miserable heat wave or late season hurricane. They are concerned with the consequences of climate change, but at the same time cannot completely understand these consequences, so climate change is l ess important than more immediate issues. In general, the global public is concerned but not alarmed about climate change . B oth individual characteristics and context influence climate change opinions . People more likely to pay attention to climate chan ge are also more likely to consider cl imate change serious and important on average across countries (the influence of engagement is more complex in the United States). M ore educated individuals say climate change is serious and dangerous and more likely t o rank both the environment and climate change important issues than less educated individuals . Context matters and does not matter in important ways. C iti zens in wealthier countries are not more concerned or more likely to be willing to pay more to addres s climate change , so the assumption that wealthier countries would be more willing to take on greater burdens in an international climate change policy response is problematic . But wealth and development does shape the importance of left - right political or ientation on concern about climate change. Climate change has become a left - right political issue in only wealthy countries, another possible obstacle to a meaningful policy response by the wealthiest countries . American public opinion on climate change is an outlier compared to other wealthy countries . Th e group of Americans who consider climate chang e not serious and dangerous is much larger than in other wealthy countries. Why is this group so large? The political and media elite are far more polarized o n climate change in the United Stat es than in other wealthy countries . The 4 Republican Party in the United States is the only political party in an affluent country to so outwardly reject climate change, consider it unserious, and reject policy response opt ions. And the mass public is aligned with ideological elites in the U.S. , so American public was far more ideologically divided than in other wealthy countries. And this ideological divide increased as engagement with politics and climate change increases. As engagement increased , liberals become more worried and conservatives less worried, a pattern not observed in any other wealthy country. 1 .3 Chapter o utline This sec tion briefly describes the remaining seven chapters in this project. Chapter 2 describe s the great variation in public opinion about climate change around the world. It shows that while publics around the world expressed a high level of concern on numerous survey s, this . People are generally extremely reluctant to support economic sacrifice for climate change, and considered other environmental problems like pollution and issues such as health care more important than climate change. In add ition, the American public stands out for its lack of concern about climate change. Across numerous s urv eys, far more Americans are unconcerned about climate change than in publics of similarly wealthy and developed countries. Chapter 3 then explores the literature on environmental concern generally and climate change specifically on the United States and c ross - nationally . The literature offers contradictory expectations for the influence of many important factors, such as religiosity, affluence, and post - materialism. Chapter 4 then tests these competing explanations for climate change risk perception using cross - national public opinion data, coming to three important conclusions. First, among country charac teristics, country affluence does not increase concern in either of the cross - national surveys; on average citi zens in wealthier countries are not more co ncerned about climate change than citizens in less wealthy countries. Second, education has a strong association w ith climate change risk perception , as more educated individuals considered climate change more serious than less educated individuals on aver age across countries. Policymakers crafting international agreements on climate change should not assume that 5 citizens of wealthy countries are willing to make greater sacrifices. Third, general left - right political orientation is associated with climate c hange risk perception, as those on the left were more concerned than those on the right on average. Chapter 5 explores the substantial cross - national variation in the association between left - right political orientation and climate change risk perception in more detail. It shows this variation can be explained by country - level wealth and development. Using multilevel cross - level interaction models, I show that among low and middle income countries, left - right pol itical orientation generally is not related to concern about climate change, but left - right political orientation significant ly and substantively predicts perceived seriousness of climate change for affluent and high ly developed countries, a finding that could be attributed to the closer connec tion between political ideology and positions on economic issues and climate change in wealthy countries. T hi s shows that climate change is a left - right political issue in many of th e wealthy countries in Western Europe and Anglophone countries . The left - right divide in these countries could be an obstacle to any meaningful climate change agreement. Even among wealthy countries, the left and right are far more divided on cli mate change in the U.S. than in other wealthy countries, so there is a distinct factor d riving public opinion in the United States. In Chapter 6 , I argue that elite polarization drives mass polarization in the United States, as the American public uses elites as an information shortcut on climate change . In the United States, unlike any other wealthy country, I show the ideological divide between the left and right increased as political engagement increases, evidence that those who are most engaged are most likely to pick up on their respective ideological cue. Chapter 7 first looks at the re lationship between opinions on climate change specifically and environmentalism generally . People around the world and in the United States generally related their worries about climate change to their worries about other possible environmental problems, m ost prominently the da ngers of air pollution. Then focus of this chapter shifts to the relationship between climate change attitudes and intended behavior. Even though increasing perceived dangerousness does increase the likelihood of both willingness to p ay to address climate change and ranking the environment 6 and climate change as import ant, the size of this effect is modest . Cross - nationally, of those individuals who consider climate change danger ous/serious, a small majority i s willing to pay higher pri ces to help address climate ch ange and a large majority rans issues other than the environment and environmental problems other than climate change as more important. Moreover, climate change r isk perception does not predict ranking the environment and cli mate change as most important in poorer and less developed countries, demonstrating that the very high public concern in many developing countries may be an illusion. Climate change is not a front - burning issue for both many worried individuals in rich cou ntries and many citizens of developing countries, two important constituencies that could be needed to support an international agreement . Chapter 8 is the conclusion, summarizing the key findings, including ways in which global public opinion acts as aid and an obstacle to a n international policy response. It also offers policy options likely to have public support and proposes key questions for future research. 7 Chapter 2 : How i mportant d o people consider climate c hange? 2 .1 Introduction Scholars of Amer ican public opinion on the environment and climate change can provide a detailed description of public attitudes and behaviors on climate change ( Marquart - Pyatt, et al. 2011 , McCright and Dunlap 2011 , Guber 2013 ). They c an answer important questions such as who believe s climate c hange is happening, wh o believe s climate change is caused by human behavior, what the level of concern is amon g Americans, and who support s climate change mitigation policies. The amount of survey data available to these scholars is increasing as climate change becomes a more p rominent public policy issue. Public polling organizations such as Gallup and Pew Research regularly ask respondents for their views on climate change. 1 The U.S, however, is only one country in the international community. And climate change is a global pu blic policy challenge requiring an international response. These scholars of Ame rican public opinion also c annot situate American public opinion in the cross - national context. Without comparison to other countries, it is hard to provide meaning to descript ion. Studying only American public opinion leaves important holes. The cross - national work on climate change public opinion is still at a nascent stage. Until the last decade, there were only a limited number of cross - national surveys which asked responden ts about climate change. These studies often covered only a handful of relatively homogeneous countries, and they mostly asked respondents about climate change risk perception, by capturing perceived level of dangerousness, seriousness, concern, and worry. As the scientific evidence that climate change is happening and the result of human activity, climate change has beco me a more pressing policy issue. And interest in studying cross - national public opinion on climate change has subsequently followed. The n umber of surveys asking questions about climate change and the number of countries covered has increased dramatically over the past few decades. These new comparative surveys provid ed a more 1 The Yale/George Mason climate change communication project also regularly surveys Americans for their opinions and attitudes on climate change. 8 detailed picture of global public opinion on climate change gener ally and specifically how the United States compares to other countries. This chapter, through detailed description, add s to the basic understanding of cross - national public opinion on climate change. What do mass publics around the world make of climate change? And how does American public opinion compare to other countries, especially other wealthy countries? This chapter answers these questions, using six different cross - national surveys from 2005 to 2011 to show that public opinion on climate change is complex. Many individuals in do zens of countries broadly seem to be weakly concerned about climate change ( Bord, et al. 1998 , Ansolabehere and Konisky 2014 ). What do I mean by weakly concerned ? I define weakly concerned individual s as those who say climate change is serious or dangerous when asked, but also stating they were unwilling to accept economic sacrifices to address cl imate c hange and also ranking other issues as more important than climate change. A cross these surveys, most respondents told interviewers tha t climate change is serious or dangerous, but far less responded that the env ironment and climate change are most importa nt issues. And individuals across countries are reluctant to make economic sacrifices for climate change. So many seem concerned but not alarmed about climate change. As for the United States, by situating American public opinion in the comparative context , this c hapter shows that Americans are uniquely unconcerned about climate change. A majority of the American public is also uniquely unwilling to sacrifice economically to address climate change. And only a comparatively small number of Americans consider climate change to be the most important environmental problem facing the United States. For activists and policymakers who care passionately about addressing climate change, the American public does not seem like a vehicle to use to achieve their goals, b ut instead an obstacle that must be overcome. 2 .2 Previous work describing global public opinion on climate c hange First, despite its limitations, I describe the initial work describing cross - national public opinion on climate change from the late 1990s an d early 2000s as it provided the foundations for future work . Dunlap Health of the Planet 9 change in six countries: Canada, the United States, Mexico, Brazil, Portugal, and R ussia. He found that the public s in these countries viewed global warming as a problem, but not as big or serious of a problem as other environmental issues, such as the depletion of the ozone and destruction of the rain forest. In addition, Dunlap reporte d that most people did not understand climate change in these six countries. Many respondents linked climate change with air pollution and chlorofluorocarbons (CFCs), the main cause of ozone depletion (Ozone depletion was a prominent environmental problem in the 1980s and 1990s). Few people linked climate change with its actual main cause, the burning of fossil fuels, which is not surprising considering that climate change was a nascent issue at the time. 2 1998 ) compared an original 1997 survey from the United States with the 1993 Health of the Planet cross - national surveys. Similar to Du nlap, they found that individuals across countries were similarly unconcerned about climate change and considered it to be less important than many other environmental issues. As the authors write, climate change was not - early studies described mass publics around the world as misinformed and unconcerned about the dangers of climate change. Brechin (2003) updated the work of Dunlap (1998) and Bord and his co - authors (1998), using multiple cross - national surveys to compare public opinion on climate change in the United States to other change compared to European countries, especially on level of support the Ame rican public ha d for the Kyoto P rotocol (2003, 107). In addition, Brechin showed that Americans were among the least likely to correctly identify the cau se of climate change compared with citizens of other developed nations. In 2001, in the Environics International study of 15 countries, 26 percent of Mexicans correctly identified fossil fuels as the cause of climate change, while o nly 15 percent of Americans were able to answer this 2 When the survey was administered, respondents were allowed to mention two causes when volunteering causes of climate change. 31 percent of respondents in Canada mention ed climate fossil fuel use; 20 percent of respondents in the United States; 5 percent of respondents in Mexico; 10 percent of respondents in Brazil; 14 percent of respondents in Portugal; and 23 percent of respondents in Russia. A larger percentage of resp ondents mentioned air pollution as a cause for climate change in every country: 43 percent in Canada; 39 percent in the United States; 34 percent in Mexico; 43 percent in Mexico; 49 percent in Portugal; and 36 percent in Russia. 10 question correctly. Of the highly developed countries, only respondents in France and Jap an were more misinformed about the cause of climate change than Americans. Brechin demonstrated that the American public opinion on climate change might be unique . Although these early studies provide a foundation for which to understand cross - national pu blic opinion on climate change, they still have serious methodological limitations. First, they often include d a limited sample of countries, and these countries are usually richer and more developed. So these cross - national studies did not give us more th an a brief outline of public opinion on climate change in the developing world. In addition, in some of these countries, due to cost constraints, urban areas were oversampled or sometimes urban areas were the on ly parts of the country sampled . Fortunately, since climate change has become a more important environmental issue, it has been the focus of numerous cross - national public opinion surveys in the last decade. 2 .3 Data The rest of this chapter builds on the early work describing cross - national public opinion with more recent cross - national public opin ion surveys, which cover a larger and more diverse group of countries. I use six cross - national public opinion surveys conducted between 2005 and 2011 to describe public opinion around the world and situa te American public opinion in the comparative context. 3 Specifically, this section will draw from the Global Attitudes cross - national surveys from the Pew Research Center (Pew Global Attitudes) in 2007, 2008, 2009, and 2010, the World Values Survey (WVS), which was administered between 2005 and 2009, and the International Social Science Programme Survey (ISSP), which was conducted between 2009 and 2011. More information on the data sources is included in Appendix A . I look at three important survey quest ions on climate change. First, all six cross - national surveys different. Risk perception, which broadly covers concern or worry, is an important measure of public 3 In section 2.6, which discusses climate change public opinion over time, I also use additional public opinion surveys from the United States. 11 opinion on climate change. 4 I also compare the perceived dangerousness and seriousness of climate change to the other environmental problems , such as air pollution, water pollution, and genetically modified crops. Second, respondents to the Pew Glo bal Attitudes surveys were asked about their willingness to pay to address climate change. This question captures willingness to move words of concern into action in the context of survey research, so it is a measure of intended behavior. Third, respondent s to the ISSP survey were asked to name what issues they consider to be the most important and next most important, with the environment as a response option, and what environmental problem what they consider to be most important for their country, with cl imate change as a response option. This question capture s how strongly individuals care about climate change, as it forces individuals to choose an issue or problem that they believe is most important. 2.4 Describing global public opinion This section l ooks at cross - national public opinion for all countries included in the surveys. The goal is to describe global public opinion on climate change , ignoring differences across countries , so this section is not comparative . 2 .4.1 Climate c hange risk p ercepti on On all six surveys, despite differences in question wording, respondents in dozens of countries expresses a high level of perceived seriousness or dangerousness of climate change. 5 The distribution of 4 Unfortunately, more recent cross - national surveys have not consistently asked questions to test the knowledge and understanding of r espondents on climate change or support for climate change mitigation policies. It would have been very helpful to have data on these questions. 5 Question wording effects are important to consider in survey research. In the four Pew Global Attitudes surve ys and the WVS, respondents were asked how serious climate change is. For the ISSP, respondents were asked how dangerous climate change is. These words, of course, have slightly different meanings. Individuals can consider a problem serious without viewing it as dangerous. Dangerous is a higher threshold, so it is not surprising that concern seems lower in the ISSP surveys than it the other surveys. The response categories are also different across the six surveys. Both Pew and WVS have four response categ The biggest difference, once again, is between the ISSP and other surveys. The nd of the distribution helps explain why the distribution for the ISSP question seems less skewed. Next, research on public opinion in the United States found question wording effects based on whether the Schuldt, et al. 2011 ). The earlier surv 12 responses is heavily skewed toward considering clima te change a seriou s or dangerous problem. Figure 2 .1 shows the distribution of perceived seriousness or dangerousness in each of the six cross - national studies from a variety of countries. 6 In all four Pew Global Attitudes surveys, a ma jority of respondent s consider rcent in all four surveys call in the bottom two response categories. In th e four Pew Global Attitudes surveys, between 2 and 4 percent and 7 and 10 percent, dependi ng on the survey year, respond that climate change is 7 The distribution of responses for the perceived seriousnes s of climate change for the World Values Survey i s similarly skewed. The World Values Survey asked respondents about their perceived seriousness about three environmental problems, including climate change, for the world as a whole in 44 countries. Only tw o p ercent of respondents consider consider climate change as a serious problem . Specifically, 30 percent say climate change risk perception is less skewed for the ISSP, which included 32 countries, as it included five response categories instead of just four, and asked about dangerousness instead of seriousness. The lower level of the dist ribution is similar to those of the other question s, as only 10 percent consider that climate change is Obviously there are also translation and cultural effects in cross - national research. Although these topics are important for research, they are not the focus of this project. 6 Pew Global Attitudes surveys include a question about respondent perception of the seriousness of climate change in 37 countries in 2007, 24 countries in 2008, 25 countries in 2009, 22 countries in 2010. 7 It seems reasonable to assume that risk perception is connection to responses to other questions on climate cha nge. For example, those who do not consider climate change serious or dangerous are likely to not take the possible consequences seriously and unlikely to support the implementa tion of mitigation policies, but there are no available questions on cross - nati onal surveys to test this assumption. 13 r esponse category, responses are more dispersed. Twenty - six perc ent of respondents say climate change Overall, the responses a cross these six cross - national surveys show the world is concerned about climate change. A m across all six cross - national surveys. These descrip tive findings sho uld be encouraging for climate change activists, as the public expresses that climate change is serious , at least in isolation when answering a survey question. Responses to other survey questions on climate change should be more worrying to policymakers and activists around the world who want to address climate change 14 Figure 2 .1: Perceived seriousness or dangerousness of climate change in all countries (Pew Global Attitudes Surveys 2007 - 2010, World Values Survey 2005 - 2009, International Social Survey Programme 2009 - 2011) 15 Questions: Pew Global Attitudes 2007, 2008, 2009 : Pew Global Attitudes 2010: topic, in your view is global climate change a very serious problem, somewhat serious, not too serious, or not a WVS: you consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very ISSP: climate change is extremely dangerous, very dangerous, somewhat dangerous, not very dangerous, or not dangerous at all for the environment? 2 .4.2 Willingness to p ay to address climate c hange The Pew Global Attitudes Project asked respondents in 2009 and 2010 if they ag ree or disagree . This question measures intended behavior. In 2009, 52 percent agreed with this statement, while 40 percent disagreed in the 25 co untries. And in 2010, 48 percent agreed and 43 percent disagreed in 22 countries. This cross - national descriptive finding should be far more worrying to climate change activists, as people seem reluctant to back - up their concern about climate change with a statement of a ction. Far more respondents are willing to say climate change is serious than state they are willing to sacrifice 2 .4.3 Strength of c oncern about climate c hange So far, I have shown t hat global public opinion on climate change is mixed: people express a high level of concern, but are divided over whether to sacrifice economically. This section adds to the mixed portrayal by looking at the strength of concern about clima te change. I fin d more evidence those overwhelming majorities who say climate change is serious or dangerous provides a misleading picture of public opinion. Compared to risk perception of other environmental problems, climate change risk perception is not as impressive. Small minorities express that the environment is their most important issue and/or climate change their most important environmental problem. So even though p eople around the world express a high level of concern in abstract , for most, climate change does not seem to be an important issue . First, although the skewed distribution of concern about climate change seems to show that the world i s concerned , this finding is viewed in a vacuum. It is not compared to concern about other 16 environmental problems. Bo th the WVS and ISSP asked respondents about their perceived seriousness or dangerousness of other environmental problems , including the various forms of pollution. Specifically, the WVS, in addition to asking respondents about lobal warming or the greenh ouse effect perceived seriousness for these three questions on a range from ze ro ( not serious at all ) to one ( very serious ). For all three environmental iss ues, mean perceived seriousness is ve ry similar. Mean perceived seriousness is highe r for climate change than it is an perceived seriousness for climate change is ecifically, mean perceived seriousness is 0.82 for climate change, compared with n the 44 included countries are similarly concerned about all three e nvironmental problems, they see water pollution as slightly more serious than climate change. 8 The ISSP asked re spondents in 32 countries about the perceived dangerousness of seven possible environmental problems, providing a more de tailed picture of global public opinion on concern about environment problems than the WVS. Figure 2 .2 shows mean perceived dangerousne ss, ranging from zero ( not dangerous at all ) to one ( extremely dangerous ). 9 Of the seven environmen tal problems, climate change has the fourth highest mean perceived dangerousness value. Publ ics in these countries express a high level of con cern about . But, individua ls in these countries consider caused by e. Like the WVS , citizens in these countries are 8 On the zero - to - one scale, response categories received the following values for the calculation of the mean: 0 not serious at all; 0.33 not very serious; 0.67 somewhat serious; 1.0 very serious. 9 On the z ero - to - one scale, response categories received the following values for the calculation of the mean: 0 not dangerous at all for the environment; 0.25 not very dangerous; 0.50 somewhat dangerous; 0.75 very dangerous; 1.0 extremely dangerous. 17 once again more concerned about water pollution than climate change. By comparing perceived dangerousness of climate change to perceived dangero usness of other environmental problems in dozens of countries, climate change does not stand out. Figure 2 .2 : Mean perceived dangerousness across environmental problems in all countries (International Social Survey Programme 2009 - 2011 ) Questions: In ternational Social Science Survey 2010: categories received the following values for the calculation of t he mean: 0 not dangerous at all for the environment; 0.25 not very dangerous; 0.50 somewhat dangerous; 0.75 very dangerous; 1.0 extremely dangerous. These small differences in means across make it difficult to determine what environmental pro blem is most important to citizens; citizens seemed to think that all environmental probl ems are dangerous and serious . To better understand what environmental issues are considered more important, respondents need to be asked to rank the importance of bot h general policy and specific environmental issues, rather than asking about the seriousness and dangerousness of each issue separately . The ISSP forced respondents to name the most important issues and problems their country and family face. This requires respondents to identify their most impor tant issue or issues, rather than expressing concern about 18 every issue and problem. First, ISSP asked respondents what general issue was the most important for their country today health care, education, crime, the environment, immigration, the economy, terrorism, or poverty. Figure 2 .3 shows the distribution of responses for both the most important and second most important issues. The global public seems divided on the most pressing issue. T he mod al response across countries is the economy. The environment ranks sixth out of the eight issues. Only about five percent of respondents say that the environment i s the most important issue facing their country, compare d with the one - in - four who mention the economy, 22 perc ent for health care, and 16 percent for education. When respondents were asked for their second most im portant issue, nine percent say the environment. This is once again a far smaller percentage tha n those who answer health care (20 percent), education (1 8 percent), and the economy (16 percent). Only about 14 percent of citi zens in these 32 countries rank the environment as one of t heir two most important issues. Figure 2 .3: Most important issues facing country in all countries (International Social Survey Programme 2009 - 2011) Questions: International Social Science Programme 2010 another response categ ory not included in the figure. 19 In the realm of specific environmental problems, how important is climate change? ISSP also asked respondents which environmental problem, fr om a list of nine, is the most important for their co untry and their family. Figur e 2 .4 shows the cross - national distribution on these two questions. These results provide further evidence t hat publics care more about pollution than climate change. Twenty - two percent cho o se air pollution as the most important environmental problem f acin g their country, compared with 15 percent for climate change. Fewer individuals cho o se both air pollution and climate change as the most impo rtant issue facing their family, but more consider still air pollution (20 percent) than climate change (12 percent ) to be the most important problem. There i s not a consensus in these 32 countries as to what is the most im portant environmental problem. Figure 2 .4: Most important environmental problem in all countries (International Social Survey Programme 2009 - 2011) Questions: International Social Science Survey Programme : Which problem, if any, do you think is most As can be seen, publics around the world seem to be sending a mixed message to policymakers on climate change. In isolation, cit izens around the world express great concern about climate change, but 20 compared to other env ironmental problems, their level of perceived seriousness or dangerousness is very similar. And the global public is more concerned about vari ous forms of pollution than climate change. Moreover, few people care so much about the environment that they rank it as their first or second most important issues. For many around the world, climate change is a soft public policy problem, an issue that is not as important as other big public policy problems, such as health care and the economy, or even pollution. Wh ile many may be concerned about climate change, substantially fewer are alarmed . 2 .5 Describing public opinion on climate change across countries So far, I have focused on global public opinion , describing the distribution of responses to survey questi ons across dozens of countries, intentionally masking differences in mass public opinion between countries. This section describes these differences between countries, making two important conclusions. First, there is great variation in mass public opinion on climate ch ange on all the above questions, risk perception, wil lingness to pay, and importance, between countries. Second, the United States compared to similarly democratic and developed countries stands out as an outlier on climate change and the env ironment . 2 .5.1 Climate c hange risk p erception across c ountries First, I look at mean perceived seriousness of climate change for each country for the four Pew Global Attitudes surveys from 2007 to 2010. This variable ranges from zero (not a problem) to o ne (very serious), so increasing value reflects increased perceived seriousness. Figure 2 .5 shows the mean perceived seriousness for each country, showing that important variation at the country - level in level of seriousness. In 2007 , the public in Jordan expresses the lowest level of perceived seriousness in all four surveys, with a mean of 0.64, while the public in Bra zil expresses the highest level of perceived seriousness, with a mean of 0.95. Brazil has the highest mean seriousness level in all four su rveys. 10 Of the 37 countries included i n 2007, the United States ranks 34 rd in its mean seriousness, ahead of only Russia, Egypt, and Jordan, thr ee countries with societies with little in common with the U.S. In 2008, the 10 In the 2007 - In the 2010 Pew Global At 21 public in China expresses the lowes t level of concern, w ith a mean of 0.69, and China i s the only country with a lower mean than the United States. In 2009, the United States has the lowest mean concern of 0.69, behind China, Russia, Indonesia, Kenya, and Jordan. In 2010, Pakistan i s the on ly country with a lower mean than the United States . 22 Figure 2 .5: Mean perceived seriousness , by country, Pew (Pew Global Attitudes 2007 - 2010) 23 Questions: Pew Global Attitudes 2007, 2008, 2009 lobal warming a very Pew Global Attitudes 2010: topic, in your view is global climate change a very serious problem, somewhat serious, not too serious, or not a On the zero - to - one scale, response categories received the following values for the calculation of the mean: 0 not a problem; 0.33 not too serious; 0.67 somewhat serious; 1.0 very serious. The vertical solid line represents the overall mean serious ness in all countries. The four Pew Global Attitudes surveys show varying levels of seriousness between countries. Does this pattern hold for the World Values Survey? The mean seriousness for e ach country is shown in Figure 2 .6. The variable once again ranges from zero (not serious at all) to one (very serious). And like the Pew Global Attitudes surveys, ther e is considerable variation between countries. Zambia has the lowes t mean of 0.63, while Turkey has the highest of 0.95, reflecting the wide variati on between mass publics in level of perceived seriousness of climate change. Here, the United States ranks 37 th in mean seriousness , in between South African and Indonesia. 24 Figure 2 .6 : Mean perceived seriousness , b y country , World Values Survey (World V alues Survey 2005 - 2009) Figure 2.6 continued - Question: WVS: Please tell me how serious you consider each of the following to be for the world as a whole. Is it very serious, somewhat On the zero - to - one scale, response categories received the following values for the calculation of the mean: 0 not serious at all ; 0.33 not very serious; 0.67 somewhat serious; 1.0 very serious. The vertical line represents the overall mean. The ISSP descriptive results, even with differences in question wording, shown in Figure 2 .7, reinforce that there is important variation in climate change risk percep tion and concern in the United States is comparatively low . Like above, this variable ranges from zero (not dangerous at all) to one ( extremely dangerous). Norway has the lowe st mean of 0.57, while Chile has the highest mean of 0.87. Out of the 32 countrie s, the United States ranks 29 th in mean perceive d dangerousness, ahead of the Latvia, Belgium, and the aforementioned Norway. 25 Figure 2 .7: Mean per ceived dangerousness, by country , International Social Survey Programme ( International Social Survey P rogra mme 2009 - 2011) Question: ISSP: extremely dangerous, very dangerous, somewhat dangerous, not very dangerous, or not dangerous at all for the environment? The resp onse categories received the following values for the calculation of the mean: 0 not dangerous at all for the environment; 0.25 not very dangerous; 0.50 somewhat dangerous; 0.75 very dangerous; 1.0 extremely dangerous. The vertical line represent s the overall mean. 2.5.2 Willingness to pay to address climate c hange across c ountries Next, I turn my attention to variation between countries in sta ted willingness to pay. Figure 2 .8 shows the percentage in each country that agreed and disagreed that p eople should be willing to pay higher prices in order to address global climate change percentages , represented by the line . So the longer the line, the greater the difference between the percentage who agree d and d isagree d in each country . The variation between countries is stark in both the 2009 and 2010 survey, as evidence d by the distribution of responses in Jordan and China. In both sur veys, Jordanians and Chinese have dramatically different answers to this ques tion. For th e 2009 survey, 14 percent agree, while 72 percent disagree in Jordan; th is compares to 88 percent agree an d eight percent disagree in China. For th e 2010 survey, 21 percent agree and 73 percent disagree in Jordan; 91 percent 26 agree and 6 percent disagree in China. 11 In both the 2009 and 2010 surveys , more disagree than agree in 11 countries, including the United States. In the United St ates, in 2009, 41 percent agree and 55 percent disagree that people should be willing to pay higher prices to add ress climate change. And in 2010 , 39 percent of Americans agree and 61 percent disagree . So this question shows that when citizens are asked to make a statement of action on climate change , publics in these countries are far more divided than when they are simply as ked about how serious or dangerous they consider climate change. And a majority of Americans are unwilling to sacrifice economically to help address the possible dangerous consequences of climate change. 11 sample should not be used to make conclusions about the entire c ountry. More than likely, if I had access to a national sample, the difference between agree and disagree would be much smaller. 27 Figure 2 .8 : Willingness to pay to addre ss climate chang e, by country (Pew Global Attitudes 2009 - 2010) Question: Pew Global Attitudes Survey 2009 and 2010: following statement. People should be willing to pay higher prices in order to ad do not add up to 100. The line segment represents the difference between agree and disagree. 2 .5.3 Strength of c oncern about climate c hange across c ountries This section compa res strength of concern between countries. Like the previous section looking at cross - national strength of concern in all countries , this section first compares climate change risk percepti on to risk perception of other environmental problems . Across countries, individuals express a similar amount of concern over environmental problems, including climate change. On the WVS, mean 28 perceived seriousness of water pollution, climate change , and b iodiversity is clustered between 0.67 ( somewhat serious ) and 1 ( very serious ) in 44 countries. Mean perceived d angerousness of all seven environmental problems in the ISSP generally is clustered between 0.5 (somewhat dangerous) and 0.75 (very dangerous ) in 32 countries, as show n in Figure 2.9 . This makes the seemingly high concern about climate change across countries and surveys seem less impressive, as many respondents call all possible environmental problems, from climate change to pollution to nucle ar power, as serious and dangerous. Climate change risk perception does not stand out compared with other environmental problems . 12 Second, pub lics around the world consider pollution more dangerous and serious than climate change. In thirty - three of 44 c ountries in the WVS, mean perceived seriousness of s the highest of the three included environmental problems. Climate change rank s first, ahead of water pollution , in only 10 countries. In another 15 countries, in cluding the Unit ed States, climate change ranks last with the lowest mean perceived s eriousness , behind both water pollution and the loss of species/biodiversity. On the ISSP survey, in only two cou ntries Japan and South Korea is the mean perceived dangero usness value for cli mate change greater than the six other environmental problems. 13 Publics are at least slightly more worried about four other issues. Air pollution from indus try had the highest mean perceived dangerousness in 12 countries, nuclear power in seven countries, pollution of water in six countries, and pesticides and chemicals in farming in five countries. In the United States, the danger ousness of climate change ranks sixth out of the seven problems, only ahead of genetically modified foods . T he American public considers various forms of pollution to be more dangerous than climate cha nge. By comparing climate change risk perception to risk perception of other environmental issues, the seemingly high level of concern in the Unite d States and aro und the world i s 12 For example, look at Chile and Turkey, the two countries with the highest mean concern about climate change in the ISSP surve y. Chileans had the highest concern about air pollution from industry (0.87) and lowest concern about genetically modified crops (0.78). Turks were most concerned about genetically modified crops (0.84) and least concerned about air pollution from cars (0. 79). In both countries, concern about climate change ranked in the middle, and the means for all seven environmental problems were clustered together at the high end of the range. In these countries, publics were generally expressed that all the included e nvironmental problems were dangerous; they did not consider climate change to be particularly dangerous. 13 In three countries Chile, Mexico, and Taiwan mean concern about climate change was the second highest. 29 not unique. In most countries, publics considered other environmental problems, including pollution, to be more dangerous. Figure 2 .9 : Mean perceived dangerousness about environmental problems, by country (International Social Survey Pro gramme 2009 - 2011) Questions: International Social Science Survey 2010: and categories received the followin g values for the calculation of the mean: 0 not dangerous at all for the environment; 0.25 not very dangerous; 0.50 somewhat dangerous; 0.75 very dangerous; 1.0 extremely dangerous. The legend shows how the colors correspond with specific environ mental problems. F inally, I look at the percentage in each country w ho considers the environment to be a most important i ssue and climate change to be the most important environmental problem. There is 30 considerable variation across countries in the siz e of the group of citizens who consider the environment and climate change more important than other issues and problems. W hen forced to make a choice only a small minority of respondents rank the environment and climate change as an important priorities. Figure 2.10 shows the percentage of citizens in each country who considered the environment to be the most or next most important issue facing their country in the ISSP survey. Thirty - one percent of Norwegians call the environment either most important or second most important issue, the highest percentage of all 32 countries. The countries with the highest percentage of citizens to consider the environment one of their most important issues are mostly in Europe. Twenty - nine percent of Danes, 28 percent of Swiss, 25 percent of Swedes, 24 percent of Fins, 23 percent of Belgians , and 21 percent of French name the environment. The United States ranks 15 th out of 32 countries, with about 11 percent of Americans naming the environment as one of their two most imp ortant issues. This was below the mean across all countries o f 14 percent, but Americans are slightly more likely than the British (8 percent) to consider the environment an important issue. 31 Figure 2.10 : Environment as most important or next most impor tant issue, by country (International Social Science Programme 2009 - 2011) Questions: International Social Science Programme 2010 y that chose the environment as the response category for either question. Figure 2 .11 shows the percentage of citizen s in each country who consider climate change the most important environmental problem for their country, compared to the three mostly co mmonly chosen environmental problems in all countries air pollution , water pollution , and using up our natural resources . Immediately, the large percentage of Japanese who name most important problem jumps out. Fifty - two percent of Japanese adults consider climate change the most important enviro nment problem. And Japanese are far more likely than citizens from any other country to name climate change. The five next most likely, and seven out of eight, are in countrie s in Europe. In Norway, Germany, Denmark, Spain, and Austria, about one - in - four name climate change as their Finland, about one - in - five name climate change. 14 This provides more eviden ce that European publics are the most concerned about 14 Canada (21 percent) is the only non - Europ ean country other than Japan in which at least 20 percent call climate 32 climate change because they more likely than non - important environmental problem . Only 10 percent of Americans consider climate change to be the Uni ted States most important environmental problem, which ranked 14 th . By comparing across environmental issues, in most countries, I provide further evidence that publics in these countries consider air pollution more important than climate change. In nine countries of the 32 countries, climate change is the modal response when respondents are most important environmental problem. In 20 countries, a clear majority, air pollution i s the most mentioned issue. And in 19 countries at le as t one - in - five respondents call most important environmental problem. The U.S. stand out s because it is one of only two countries in s the most common response. Twenty - eight percent of Americans name also stands out because it i s one of only two coun tries in which climate change i s the least com mon among the four popularly menti oned problems . 15 15 France i s the other country. 33 Figure 2.11 : Most important environmental p roblem s , by country (International Social Survey Programme 2009 - 2011) Question : International Social Science Survey Programme : Which problem, if any, do you think is most important 2 .6 Change in climate change risk p erception over t ime across c ountries So far, I have focused on describing public opinion on climate change across and between countries. Due to the increased interest in public opinion on climate change, public opinion o n climate change can also be described across time within countries. How has climate change risk perception in the United States and around the world changed over time? Public opinion surveys in the United States have shown that the public considers climat e change less risky than it did previously; fewer Americans are worried and consider it a serious problem ( McCright and Dunlap 2011 , Scruggs and Benegal 2012 , Guber 2013 ). Figure 2.12 shows the level of worry in the United States about climate change from 1997 to 2013, 34 Organization. 16 Although there ar e year - to - year fluctuations in worry between 1997 and 2007, the American public becomes slightly more worried about climate change. Then from 2007 to 2011 the p ublic became much less worried about climate change. There i s a meaningful decrease in the percentage ge; by 2011, on ly 25 percent worry centage of Americans who worry in this same period. In 2007, o nly 16 percent of Americans are 11, 29 percent of Amer icans are 16 Gallup asks respondents about their worry about environmental issues, including climate change every yea r from 1997 to 2013, except 1998 and 2005. Obtained from the Roper Center on Public Opinion. 35 Figure 2 .12 : Perceived w orry about climate change in the United States, 1997 - 2013 (Gallup Organization) Gallup y worry about this problem a great deal, a fair amount, only a little, or not at all. How much do you personally worry Pew Research Center also asked Americans about how serious they consider climate change every year from 2006 to 2013, documenting the same decline in concern as the Gallup surveys. In 2007 , 47 percent of Americans view it y 33 percent of Americans view climate change as a problem . Unlike Gallup, which fo und a slight increase in worry in 2012 and 2013, Pew documents a continued decrease in perceived seriousness of climat e change over those two years. 36 Figure 2 .13 : Perceived s eriousness of climate change in the United States, 2006 - 2013 (Pew Research Center) Notes: a problem? Is the United States unique in its declining concern about climate change, captured by two different survey organizations, in recent years? Although the United States stands out for its low level of concern compared to similar and neighboring countries, its decline in concern dur ing this time period is common. Pew includes 19 countries in all four Global Attitudes surveys from 2007 to 2010, a pe riod of great decline in perceived seriousness and worry about climate chang e in the United States. Figure 2.14 shows mean perceived seriousne ss in these 19 coun tries from 2007 to 2010 on the zero - to - one range used previously, with increasing values reflecting increased perceived seriousness. 17 In 12 of the 19 countries, including the United S tates, mean perceived seriousness declines over this t ime period. The United State s has the fifth greatest decline, between Japan and Great Britain. Mean perceived seriousness also decreases in Spain, France, Germany, and South Korea. So the decline in concern about climate change a t the end of 17 It is important to note that the wording of this question changed in 2010. In 2007 - 2009, the Pew Global Attitudes Project asked respondents about the seriousness be the result of the change in question wording. Between 2007 and 2009, mean pe rceived seriousness declined in 9 out of 19 countries, including the United S tates. Mean seriousness declines by a greater amount between 2007 and 2009 in Japan, China, Spain, and Turkey than in the United States though. 37 the last decad e is not uniquely American; it occurred among publics in many other wealthy countries , possibly driven by the global recession . 18 Figure 2 .14 : Mean perceived seriousness of climate change, by country , 2007 - 2010 (Pew Global Attitudes Surveys 2007 - 2010) Que stions: Pew Global Attitudes 2007, 2008, 2009 Pew Global Attitudes 2010: 18 The percentage of citizen s in F rance and Japan who view climate change as a very serious problem decreases by even more than in the United States. The pe rcentage of Americans who view climate c very serious decreases by 10 percent (from 47 to 37 percent) from 2007 to 2010. The percentage of respondents in Franc e and Japan, however, who view climate change very serious by 20 percent (from 68 to 48 percent in France and from 78 to 58 percent in Japan). 38 topic, in your vie w is global climate change a very serious problem, somewhat serious, not too serious, or not a - to - one scale, response categories received the following values for the calculation of the mean: 0 not a problem; 0.33 not too serious; 0.67 somewhat serious; 1.0 very serious . 2 .7 Discussion and c onclusion This chapter provides a detailed description, capturing a variety of aspects of cross - national public opinion on climate change by using six cross - national surveys and additional survey data from the United States. It demonstrated three important conclusions. First, although most people around the world say climate change is serious and dangerous , this worry about climate change is weak. In general, pe ople around the world conside r all environmental problems to be serious or dangerous. And people in most co untries are more concerned about the forms of pollution than about climate change. So climate change - n public opinion on climate change is unique. Across all six surveys, America ns are less concerned and more divided about climate change other wealthy countries. And on the other questions, which capture willingness to sacrifice economically to address cli mate change and considering the environment and climate change important is sues, the United States lags behind similarly wealthy countries. Third, the decline in concern about climate change in the United States a t the end of the last decade is not unique though. The perceived seriousness of climate change declines during this time period in many other similar countries, such as Great Britain, France, and Japan. This chapter shows that researchers need to rely on a variety of survey questions across countri es to understand cross - national public opinion on climate change. Simply relying on one questi on, such as risk perception , might lead a researcher to overstate how much p ublics are concerned about climate change . By using multiple questions across surveys, this chapter portrays cross - national public opinio n on climate change as complex. This chapter also shows that simply describing public opinion about climate change in one country without situating it in the comparative context could be misleading. By des cribing public opinion in the United States compa red to other countries, I show that U.S. public is less worried 39 about climate change than other countries. The remaining chapters in this project shift the focus to explanation. 40 Chapter 3 : Previous expla nations for climate change and environment al risk p erception s 3 .1 Introduction The previous chapter described opinions and attitudes on climate change cross - nationally. So far, though, I have not attempted to explain climate change opinions and attitude s. This chapter shifts the focus from description to exp lanation by discussing theoretical explanations and empirical findings from previous research on public opinion on climate change and the environment . It organizes this interdisciplinary work , drawing from sociology and psychology, in addition to political science, on public opinion on climate change and the e nvironment, focusing on the United States and cross - nationally. 1 For two decades, scholars have had vigorous disagreements about the relationship between individual and country characteristics wi th environmental concern . Scholars have debated both what factors matter and how they matter in explaining public opinion on environmental issues . What are the m ost important differences in previous resear ch? First, previous research has not come to a consensus as to what direction individual and country affluence influences opinions on climate change and the environment . Some scholars argue that individual and country affluence increase concern about clima te change and the environment , while others argue individual and country wealth has no effect or even a negative effect. Second, s om e scholars argue that religiosity increas es concern, while others argue for the opposite effect . Third, scholars also do not agree as to why the relationship between left - right political or ientation and environmental/ climate change attitudes in the United States is so strong and if this relationship extend s cross - nationally . Some argue this wide difference between the left and right is the result of economic positions, others hold that psychological orientation explain the difference, while a different group of scholars contend that elite opinion is what matters . Fourth, scholars disagree on the impact of education and environme nt on climate change and environmental concern. Some argue that education should have a positive, additive effect on concern about these issues , as increasing education 1 The discussion comes mostly from the literature on enviro nmental concern as theories explaining public opinion on climate change are still nascent. 41 and knowledge make s people more aware of the dangers of climate change and environmenta l degradation , while others argue that the influence of education is conditioned by political orientation, as more educated and knowledgeable individuals are more likely to turn to elites and engage in motivated reasoning to align and p rotect their strongl y - held political beliefs . I explain the theoretical approac hes in detail below. At the end of this chapter , I also introduce data sources and methodological approaches used in the remaining chapters to analyze cross - national public opinion on climate chang e. 3.2 Individual w ealth and p ost - materialism Prevailing wisdom holds that protecting and enjoying the environment is a luxury good only available to rich people and rich countries ( Baumo l and Oates 1979 , Dunlap and York 2008 ). For poor people, basic needs , such as food and shelter , take precedence over concern for the environment. According to this conventional wisdom, it makes logica l sense that there is a causal connection between individual and national wealth and environmental concern; as affluence increases, so should environmental concern. The causal mechanism linking level of wealth and environmental attitudes has been informed by two similar theories: the well - known post - materialism theory ( Inglehart 1977 , 1990 , Abramson and Inglehart 1995 , Inglehart 1995 , Inglehart and Baker 2000 ) and less - known affluence hypothesis ( Diekmann and Franzen 1999 , Franzen 2003 , Franzen and Meyer 2010 ). In the first theoretical approach, the connection between wealth and environment attitudes is indirect; an increase in affluence (along with generational shifts) leads to changes in fundamental values, which then re sults in greater environmental concern. In the second theoretical approach the affluence hypothesis the connection is direct; an increase in wealth directly causes greater environmental concern, with the intermediate step of value s change s unnecessary. 3 . 2.1 Post - materialism t heory Inglehart ( 1977 , 1990 , 1997 ) argues that the fundamental value change resulting from increased wealth and opportunity drives people to become increasingly concerned about the environment and supportive of environmental protection. Specifically, the shift from materialist values, empha sizing survivalist needs like food, water, and shelter, to post - materialist values, focusing on rights such as 42 freedom of speech, explains why wealthier indi viduals are more likely than poorer individuals to become environmentalists. The primary assumption is that individuals place primary importance on meeting their physical needs their most basic needs for survival ( Maslow 1954 ). In most pre - World War II societies, most people were not meeting these most basic needs. Food and shelter were scarce. After World War II, most of Europe and the United States experienced an economic boom, so these survivalist goods were not as scarce. Most individuals had their economic and physical secur ity needs met . With this increased affluence, individuals began to put greater empha sis on issues of rights and self - expression, including the environment. Inglehart ( 1977 , 1990 ) argues that the adoption of these post - material values by many individuals in the developed world had pro found consequences . These post - materialists became increasingly interested in their quality of life, such as the state of the environment, resulting in environmental movements in the United States a nd much of Europe (Inglehart and Baker 2000). So, for individuals, post - materialism theory posits that a rise in wealth leads to the adoption of postmaterial values, which then results in increased support for environmental protection. Post - materialism t heory was largely unchallenged until the early 1990s. Then a large number of environmental groups from developing nations participated in the 1992 Global Forum in Rio de J aneiro ( Fisher 1993 Health of the Planet survey finds evidence that citizens of developing and d e veloped countries both express worry about the environment. The publics of deve loping countries actually seem more concerned than the publics of developed countries ( Dunlap, et al. 1993 ). The level of concern for the environment in the developing world was much higher than post - materialism theory predicted ( Dunlap and Mertig 1995 ). Proponents of post - materialism theory ( Abramson and Inglehart 1995 , Inglehart 19 95 , Abramson 1997 ) responded to the challenges of the early and mid - 1990s by focusing on environmental conditions. Inglehart ( 1995 ) argues that mass public support for environmental protection is high in countries with major environmental problems. Countries with t hese environmental problems are usually so poor that they di d not possess the resources to respond. So peo ple in the developing world are often directly confronted with local environmental problems, such as stifling pollution and smog, raising their level of 43 concern about the environme nt. In addition, in countries sc oring high on his p ost - materialism index, he argues that more individuals have the resources to make financial sacrifices for the sake of environmental protection. These citizens of developed countries want to make the necessary financial sacrifices becaus e concern for the environment aligns with their core values. So this explanation of environmental attitudes can account for high levels of concern in both developing and dev eloped countries. Inglehart uses data from the 1990 - 1993 World Values survey to pro vide empirical support his revised theo ry. Inglehart and supporters of post - materialism theory have consistently found empirical support for the post - materialism theory ( Inglehart 1990 , Abramson and Inglehart 1995 , Inglehart and Welzel 2005 ). For example, Kidd and Lee ( 1997 ), using the 1990 - 1993 WVS, also provi de supportive empirical evidence, finding that individual s who are post - materialists are more likely to express concern about the environment than individuals who were materialists, regardless of the level of development of their country. How could post - materialism explain opinions on climate change , a specific environmental problem, rather than just broad environmental concern? Although the connection between post - materialism and broader environmental concern has been hotly debated and tested empirically , the relationship between post - materialism and climate change attitudes is far less explored. Inglehart and Welzel ( 2005 hese new politics issues, in dividuals with secular and self - expression values are more likely to see these topics as being about rights. And these rights must be expanded in order to protec t and create self - expression, the fundamental post - material value. For example, gay rights need to be expanded in order to protect the self - expression values of gays; a particular group of people should not have their rights restricted. Climate change could be considered new politics issues, because mitigating the consequences of climate change help s keep the world in a condition all can enjoy . Climate change is also clearly an environmental protection issue. Mitigating the consequences of climate change is crucial for maintaining the quality of the environment. Therefore, accordin g to post - materiali sm theory, I should expect a positive relationship between post - material values and climate change risk perception . 44 3 .2.2 The affluence h ypothesis The affluence hypothesis, which posits that the relationship between individual wealth and concern about th e environment is direct, is based on a simple assumption: there is a trade - off between economic interest and protecting the environment. Environmental quality is a public good in which citizens of a country have to contribute to maintaining, yet citizens a lso have an incentive to free - ride to not contribute ( Diekmann and Franzen 1999 ). The primary driver of determining if and how much citizens contribute is their level of wealth. Assume that the cost of protecting the environment is fixed. Therefore, as individuals become richer, the relative cost of protecting the environment lowers (protecting o be able to affor d to sacrifice in order to protect the environment, and for poor individuals, cost of environmental protection is just too great. Other immediate survivalist needs take precedent ( Gelissen 2007 ). Franzen (2003: 199), the primary proponent of the affluence hypothesis, summarizes this rational choice argument nicely: Standard economic reasoning suggest that the protection of the environment is not only a public good, but also a normal good, whose demand increases with income. Citizens in wealthier countries not only have a higher demand for a clean environment, but they also have less pressing economic problems and are therefore more willing and able to reduce their standard of living in order to devote more resources to global environmental protection. Several studies have found confirmatory evidence of for the affluence hypothesis, showing expected positive relationship between national affluence an d environmental concern ( Diekmann and Franzen 1999 , Franzen 2003 , Kemmelmeier, et al. 2002 , Franzen and Meyer 2010 , Franzen and Vogl 2013 , Kemmelmeier, et al. 2002 ). And in addition, t hese studies, using International Social Survey Programme (ISSP) cross - national survey data from the 1990, 2000, and 2009 - 2010 waves, find the as individual affluence increases, so does concern for the environment. 45 Climate change i s an environmental problem which could fit nicely into the affluence hypothesis framework. Maintaining the climate so all can enjoy it is a public good. Implementing cl imate change mitigation policies could be extremely costly, as mitigation policies could require both massive national and international investments and changes in individual lifestyle. Everyone need s to contrib ute by limiting and decreasing carbon emissio ns, yet everyone only receive the benefits of this sacrifice if mostly all contri bute . Since there is such a strong incentive to defect and not sacrifice, the ability to make economic sacrifices should be important. Therefore, rich people and countries cou ld be more willing to pay the associated costs and make the necessary sacrifices to maintain the climate. For poor people, the relative cost of all implementing climate change mitigation policies is more prohibitive, thus there is likely to less concern ab out climate change. 3 .2.3 Alternative e xplanation A number of scholars argue that affluence and post - materialism do not impact environmental attitudes. Instead, rich and poor individuals can both support environmental quality. Mass publics in both devel oping and developed countries worry about the environment; caring about the environment is a global phenomenon ( Brechin and Kempton 1994 , Dunlap and Mertig 1995 ). Rich and poor people and experience with or perception of the environment should really matter in the fo rmation of environmental attitudes. What drives experience with or perception of the environment? For many people, it is likely that local environmental quality plays a prominent role. And poor people seem more likely than rich people to experience harmful environmental conditions ( Brulle and Pellow 2006 ). Rich people should be more able to shield themselves from local environmental degradation. For example, rich peop le are more able to move away for the pollution, smog, and other harmful environmental conditions of major cities in the developing world. According to this logic, poor people should be very motivated to care about the environment because environmental con ditions have a greater impact on their lives . This motivation could help explain the many grassroots environmental organ izations which formed in the developing world in the 1980s and early 1990s in response to harmful pollution ( Dunlap and Mertig 1995 ). 46 Supporters of this alternative approach have found empirical evidence of no or even a negative relationship between post - materialism, affluence, and concern for th e e nvironment, providing support ( Dunlap and Mertig 1995 , Brechin 1999 , Gelissen 2007 , Dunlap and York 2008 , Knight and Messer 2012 ) . For example, Dunl ap and York ( 2008 ), using three rounds of the WVS, found no relationship between post - materialism and concern about the environment. Supporters of the alternative approach are especially critical of post - materialism theory and the alternative approach. One group of scholars go as far as to call the search for a relationship between post - How does the logic of this alternative theoretical approach extend to climate change specifically? Individuals can experience the conseq uences of climate change in the long - term; rising sea levels, changing weather patterns, and droughts could have major consequences, depending on where people live. Rich people should be able to more readily adapt to these changes. For example, they can ea sily move and make changes to their property and lif estyle for safety to minimize the risks of climate change . Those who are poor cannot minimize the risk of climate change to the same degree . Thus, because rich people can more easily adapt than poor peopl e to a changing climate, the poor could be more concerned about climate change than the rich. Therefore, the alternative approach predicts that there is no relationship or a negative relationship between individual wealth and risk perception . It would also predict a negative relationship at the country - level , as richer countries should be more able to adapt to the consequences of climate change . And finally it would predict no relationship or even a negative relationship between post - material values and ris k perception. 3 .3 Religiosity and religious identity Like the impact of wealth and post - material values, the theoretical expectations and empirical evidence on the relati onship between religion and both en vironmental concern and climate change risk percep tion is mixed. White ( 1967 ), in what become kn own as the White Thesis, offers the first prominent and important statement on the relationship between religion and environmental concern, arguing that Judeo - Christian values are inherently at odds with concern about the environment. She holds that 47 Christianity allows followers to separate their human experience from the experience of nature and the en humankind and nature, in which nature must be conquered and subdued. In addition, Christians should focus their effort and attention on the afterlife, rather than the condition of earth. White contends that according to religious teachings and values, C hristians should not care about the environment. Ever since White offered this controversial argument, scholars have both debated the theoretical expectations and explored public opinion data to better understand the relationship between religiosity, relig ious identity, and environmental concern. 3 .3.1 Supporting the White Thesis ment that Christians are not concerned about the environment empirically, in both cross - national and single - country studies (usually focused on the United States). The empirical findings are mixed. Many scholars have found some eviden ce that Jews and Christians are less concerned a bout the environment than those who are not part of the Judeo - Christian religious tradition, both in the Un ited States and cross - nationally ( Greenly 1993 , Schultz, et al. 2000 , Biel and Nilsson 2005 ). But a host of other studies finds little support for the direct connection between religiosity and environmental concern. For example, Greenly ( 1993 ) finds in the United States a negative relationship between being Christian and environmental concern, yet a positive relationship between being Catholic and support for environmental spending. In another study, using a cross - national sample of undergraduates, Schultz and his coauthors ( 2000 ) propose that the direct effect of being Jewis h or Christian on environmental concern might be weak and statistically insignificant when controlling for other important factors. Since scholars struggle to find direct empirical evidence in support of the White T hesis, they have instead focus on explor ing the impact of specific aspects of religiosity and religious identity. For example, some research shows a negative relationship between biblical literalism and environmental concern ( Eckberg and Blocker 1989 , Guth, et al. 1995 , Schultz, et al. 2000 ). Individuals who in terpret the Bible 48 literally are less l ikely to worry about the environment than those who do not, although this finding is primarily based on studies from the United States. How could the White Thesis, or the qualified White Thesis discussed in the previous paragraph, focusing on biblical lite ralism, help explain attitudes toward global warming? Barker and Bearce ( 2013 ) provide a more detailed theoretical explanation for the possible negative relationship between Christian religiosity and concern about climate change, focusing on Christian End - Times theology. Christian End - T imes theology holds that environmental conditions on earth must rapidly deteriorate before the Second Coming of Christ. Therefore, the consequences of global warming could be seen as the environmental deterioration that must take place immediately before the Second Coming. In other words, for believers in End - Times theology, climate chan ge and its consequences could actually not be a so urce of concern, but instead welcomed. Climate change could be a sign that Christ is returning soon. And belief in End - Times theology is especially prevalent among fundamentalist and evangelical Christians in the United States. Using public opinion survey data from the United States, Barker and Bearce find that belief in the Second Coming is associated with a lack of support for government action to combat global warming. Specifically, belief in the Second C oming is predicted to decrease the probability of supporting government action on climate change by more than 10 percent age points . Other scholars also find supportive evidence of a relationship between Christian conservatism and evangelism and climate cha nge attitudes in the United States. Wood and Vedlitz (2007) find a negative relationship between religious conservatism and concern for global warming. Smith and Leiserowitz (2013) find that American evangelicals are less likely to think climate change is happening, less likely to say climate change is caused by human activity, and less worried than American non - evangelicals. In the United Stat es at least, there is supportive evidence that both religiosity and religious identity influences attitudes on clim ate change. 3 .3.2 Opposing the White Thesis Other scholars offer different arguments opposing the White Thesis. Religiosity could have a positive relationship with environment concern, and Christians could be more likely to care about the 49 environment. For example, Christians might believe in protecting the environment because for them the , providing a theoretical explanation for a possible positive relationship . Or religiosity and religious identity are not just important for explai ning environmental concern; there are other factors that better explanation variation in environmental concern in the United States and cross - nationally. For example, Kanagy and Willits ( 1993 ), using survey data from the U .S. state of Pennsylvania, find a positive relationship between religious service attendance and behavior in support of the environment. Moreover, the White Thesis and its empirica l support might also be specific to the United States. Religiosity and religious identity might have a different influence, both in d irection and strength, on environmental and climate change attitudes , depending on context. Most empirical work on the rela tionship between religiosity an d environmental concern uses survey data from only the United States. Compared to other wealthy countries, the United States has a large percentage of evangelical Christians and conservative Christians , which makes it a uniqu e case . 3 .4 Left - right political o rientation Scholars of public opinion in the United States have long emph asized the important role left - right political orientation, as measured by self - described political ideology and partisanship, play s in the way Ame ricans make sense of the political world ( Campbell, et al. 1960 ). In the United States, political ideology and partisanship are not only closely related to vote choice, but they also exert influence over a host of issue positions and attitudes. 2 Perhaps no policy area has illustrated this ideological and partisan polarization in the United States than th e environment. Scholars highlight the important influence of ideology and partisanship on environmental attitudes in the United States ( Dunlap and Gale 1974 , Buttel and Flinn 1978 , Pierce and Lovrich Jr 1980 , Van Liere and Dunlap 1980 , Jones and Dunlap 1992 , 2 Democrats and Republicans have been shown to be divided over domestic policy ( Layman and Carsey 2002 , Layman, et al. 2006 , Bafumi and Shapiro 2009 ) and even on questions of foreign policy ( Berinsky 2009 ). And political ideology in the United States has a s trong influence on issue attitude in a number of policy areas, as liberal and conservatives have been shown to be divided on issues including welfare ( Gilens 2000 ), racial issues ( Sniderman, et al. 1984 ), and government spending ( Jacoby 2000 , Jacoby 2008 ). 50 Dunlap, et al. 2001 ) . 3 Liberals and Democrats ar e more likely to be concerned about the environment and support action to protect it than conservatives and Republicans ( Dunlap, et al. 2001 ). Unsurprisingly, more recent resea rch shows this relationship with views on the environment extends to views on climate change, inclu ding belief in climate change, risk perception, and support for mitigation policies ( McCright and Dunlap 2011 , Egan and Mullin 2012 , Guber 2013 , Krosnick and MacInnis 2013 ). This rese arch shows that liberal s and Democrats are more likely to b e concerned about climate change , believe it is caused by human activity, and support mitigation policies than conservatives and Republicans. Why would left - right political orientation have a strong influence on climate change risk perception s ? One of the more prominent explanations argues that c limate change attitudes could be closely tied to economic attitudes ( McCright and Dunlap 2011 ). Action to combat climate change requires both increased economic costs and direct government action. This necessary respons e is deeply ideological, as it supports fundamental liberal policy id eas and r ejects conservative ones . Conservative economic ideology holds a fervent belief in the power of unrestricted free markets. To guarantee free markets, conservatives typically support individual freedom, the protection of property rights, and small government. Liberals, on the other hand, hold that free markets lead to inefficient, unequal outcomes. With this notion of conservative and liberal economic ideology, broad environmental protection is government regulation of markets to protect the environ ment, a public good. After all, environmental quality something all individuals can enjoy. Liberals do not believe that the market can protect the environment, because a firm in the free market has an incentive act in a way that benefits the firm but hurts environmental quality. But this explanation for the relationship between left - right political orientation and climate change opinions is context specific to many wealthy countries in which left - right political orientation is 3 Al though while scholars found evidence of the strong relationship between political ideology and concern for the environment, they showed the relationship between partisanship and concern for the environment was much weaker. Pierce and Lovrich (1980) argued that this weak finding was because the environment was not a prominent issue tied to partisan issue agendas. More recently, partisanship has been to have a much stronger influence on environmental concern than previously thought ( Dunlap, et al. 2001 ), showing that environmental attitudes are more closely tied to partisanship. 51 closely tied to economic ideo logy. I n this context, left - right political orientation should be related to how people think about climate change. But in other contexts, left - right political orientation is not related to economic ideology. In these other places , cleavages such as ethnic ity and religion might be important to shaping political identities, not left - right politics. Left - right political orientation should not be related to climate change opinions here . So there should be a great deal of variation acros s countries in the stren gth of the relationship between left - right political orientation and climate change attitudes if economic positions play a role in the explanation . 3 .5 Education and e ngagement Previous scholarship also has diverging expectations about the effect of educa tion, knowledge and information on environmental concern. Science communicators sometimes hopefully assume a positive relationship between these factors and broad conce rn about the environment; as individual s becomes more educated, informed, and knowledgea ble about environment (factors which should be closely related), individual s become more concerned about the environment. Other scholars criticize this approach as overly simplistic, contending that the effects of the education, information, and knowledge could be conditional. Two similar approaches , popular in political psychology, emphasize the effect of education, awareness, and knowledge is conditional on left - right political orientation, so the effect of these factors on c limate change attitudes is not uniformly positive for the left and the right. On e approach contends that educated and political aware individuals are more likely engage in motivated reasoning and so their beliefs and opinio ns stay in line with their broader political view s ( Lodge and Taber 2013 ). This approach lines up with system jus tification t heory in pyschology ( Feygina, et al. 2010 ) , as those on the right engage in motivated reasoni ng to protect the status quo. A second, very similar, approach argues tha t political engaged individuals are more likely to be aware of and use elite cues in forming political opinions ( Zaller 1992 ). 3 .5.1 The deficit m odel Scientific communicators often assume that if the mass public is presented with more factua l information about a scienc e issue, they better understand it and adopt the opinions and attitudes of 52 scientific experts ( Bauer, et al. 2007 ). So science communicator s often emphasize incre asing knowledge about science issues through media and educational programs, an approach known as the deficit model of science communication ( Bauer, et al. 2007 ). In the case of climate change, science communicators would hope that increasing factual knowl edge about climate change increase s concern. For example, widely disseminating the fact that the earth is warming as a result of human behavior should lead to an increase in concern in climate change; individuals should consider climate change more serious or dangerous with this information . In this way, the deficit model assumes that people are informal Bayesian update rs: they update their prior belief based on their evaluation of the acquired new information. If the evaluation is positive, then the belief or opinion will be revised upward; if the evaluation is negative, then the belief or opinion is revised downward. So the deficit model would predict a positive relationship between science knowledge and concern: as knowledge increases, concern about climate change should also increase. After all, additional information about climate change and its possible consequence s , such as rising sea levels and more violent storms, should cause a rational individual to revise their level of concern about climate change upward. So, according to the deficit model, to increase public concern about climate change, more information and facts need to be released. In addition, education should increase public concern about climate change, as the better educated should be more aware of climate change facts and information. 3 .5.2 Information processing and motivated r easoning Some s cholars of public opinion on climate change focus on how individuals interpret and process information , as climate change is an issue in which most individuals are exposed to new, complex information ( Marquart - Pyatt, et al. 2011 ). Resea rch in political psychology, which looks at how individuals process such information, finds that those on the right are more likely to engage in system justification motivatio n than those on the left ( Jost, et al. 2003a , Jost, et al. 2003b , Jost, et al. 2008 ). 4 4 S ystems justification theory is similar to dominant social paradigm theory. Individuals who subscribe to the dominant social paradigm have been found to less supportive of environmental protection than those who do not (Dunlap and Van Liere 1984). 53 Individuals with system justification tendencies are motivated by the desire to protect the status quo, which they find stable and secure. When faced with a threat to the current system, these individuals respond by becoming inten sely protective of the current system, even if the current system threatens their well - being in the short and/or long - term. Climate change represents a clear threat to the curren t system . 5 If those on the right find acknowledging climate change threatening because it changes the status quo, then they are likely to engage in mo tivated reasoning when processing new information on climate change. Work in social psychology indicates that individuals engage in motivated reasoning in all sorts of areas, including politics ( Kunda 1987 , 1990 ). When individuals process information about an issue such as climate change, th ey use a filter defined by predispositions. Prior beliefs serve as a powerful filter, so individuals engage in biased information processing that reinforces these prior beliefs ( Wood and Vedlitz 2007 ) . How is this related to system justification? When individuals are confronted with new information, like scientific information on climate change, those who refuse to change their opinions are protecting the status quo and engaging in system justification. ideology. Research in public opinion has shown political ideology has a strong influence on information processing, often biasing opinions and attitudes ( Popkin 1991 , Sniderman, et al. 1993 ). Politic al ideology is one powerful filter allowing people to more easily sort through new information. New information that supports the prior belief is accepted unchallenged and new information that challenges the prior belief is discarded. So new information is not evaluated rationally, and individuals will process new information differently based on their ideological filter . The consequence of this is that exposure to new information has a heterogeneous effect on individuals. From a non - experimental perspective, individuals with the same level of knowledge on climate change d ifferent opinions based on their political ideology ( Hart and Nisbet 2012 ). 5 As Fey environmental problems appears to be threatening to the very foundations of social, economic, and political status 54 Citizens could be especially likely to engage in motivated reasoning Taber and Lodge 2006 , Lodge and Taber 2013 the United States. In order to engage in motivated reasoning on climate change attitudes and opinions, citizens need to have some information ab out climate change. In other words, citizens must be aware of the position they should take on climate change based on their broader political ideology and values . And therefore making them more aware of the ideological similarity or dissimilarity on this particular issue, the impact of the ideological perceptual screen should become strong er; individuals become more entrenched in the opinion. Motivated reasoning theory pr edicts that there is an interacti on between engagement and political ideology in explaining concern they become more aware of how climate change attitudes align with their ideology. T he positions of liberals a nd conservatives on climate change should diverge. 3 .5.3 Elite c ues M otivated reasoning and information processing argues that people turn to their partisan and ideological predisposition s generally. This greatly reduces the information costs for citizens to form opinions. But citizens have other possible avenues to lower the informational cost of forming opinions . Normal citizens simply do wish to invest in learning about complex issues like climate change, so turning citizens want to lower the informatio n cost of forming an opinion ( Downs 1957 ). Citizens also value consistency and hence align political attitudes and policy preferences to form a coherent belief system, even if they have difficulty forming such a consistent belief system ( Converse 1964 ). Individuals can form consistent political opinions in the most efficient way possible by relying on shortcuts or cues ( Lupia and McCubbins 1998 ) . Scholars of public opinion in the United States have argued that the most important heuristics for opinion formation are partisanship and political ideology ( Lodge and Hamill 1985 , Rahn 1993 ). 6 6 There are other possible sources for heur istics on climate change. I ndividuals could use daily experiences from their own daily lives (Popkin 1991). For example, in forming their opinions on climate change, recent research has 55 By following an ideologically similar political e lite, citizens allows their beliefs, opinions, and policy preferences to align with their broader political values. For example, an individual who self - identifies as a conservative would turn to a trusted conservative political elite or elites for an opini on on a certain issue, since their share the same political ideology, using the elite to decrease the informational costs associated with forming an opinion on politics. When asked a question on a political issue in a public opinion survey, citizens answer it by using the cue ( Zaller 1992 , Sniderman, et al. 1993 ). This allows citizens with little public policy knowledge and poorly developed political belief systems to hold consistent opinions and beliefs about politics. Scholars of public opinion in the United States have found that evidence that partisanship and political ideology serve as information shortcuts on many public polic y issues ( Zaller 1992 , Levendusky 2009 ). Many public opinion scholars would argue that climate change is an issue in which citizens are likely to turn to like - minded elites as a heuristic. Climate change is a complex scientific issue which most citizens are not directly involved with, so the information costs are high and desire to pay those shown that individuals use their daily experience with the weather. So , if an individual experiences a hotter summer than usual or a more violent storm than what is common, then the individual might be more likely to express belief that climate change is happening and concern about climate change. Of course, short - term weath er variability is poor evidence of climate change, as the weather changes caused by climate change are slower, yet recent studies in the United States have found that this experience with local weather matters in public opinion on climate change . Krosnick and his colleagues (2006) find that respondents who think that local temperature has increased recently are more likely to accept that climate change is happening. Egan and Mullen (2012) look at the effect of objective weather conditions, coming to a simil ar finding: an increase in local temperature has a significant impact on climate change beliefs. Finally, Donner and McDaniels (2013) find that national temperature changes also affect climate change attitudes. There is little evidence that individuals use scientists or experts to form opinio ns on climate change. Experts c ould be extremely useful on co mplex public policy issues requiring a lot of specialized knowledge and information. For example, even though individuals do not seem to use or value scientif ic information, they could still use the opinions of scientists as a heuristic in processing. After all, trusting the opinions of experts seems like a rational way to form opinions on an informational costly issue such as climate change. And there is a nea r scientific consensus that climate change is happening and dangerous. Yet, there is to be a disconnection between scientific experts on climate change and the mass public, at least in the United States ( Zajko 2011 ). Public opinion surveys in th e United States have shown that a majority of Americans do not believe that there is a scientific consensus on climate change. For example, the Yale Project on Climate Change Communication and George Mason University Center for Climate Change Communication have surveyed Americans since 2008 on whether respondent think that there is a scientific consensus that global warming is happening. In none of the nine national surveys have a majority of Americans said that most scientists think global warming is happe ning. At least in the United States, scientists do not communicate their findings on climate change in a way that the mass public can understand ( Zajko 2011 ). Thus, there is a high threshold to citizens aligning their belief systems with the sci entific findings. 56 information costs low. 7 So, on the particular issue of climate change, an individual who self - identifies as on the right would turn to the conservative elite for information, and an individual who self - identifies as on the left would turn to the liberal elite for information. In order for political elites to influence mass opinion on a pu blic policy issue, such as climate change, ideological and partisan elites must first take clear and divergent positions on a particular issue ( Zaller 1992 ) . If the left and right take the same position, then the vast majority of the public also take the same position. It is not enough, however, for elites to just take simply take divergent po sitions on ideology. These divergent pos itions then need to be broadcast . After all, citizens cannot follow like - minded eli tes if they do not know their position . Even if political elites take a divergent position on an issue and broadcast their divergent position, elite opinion does not have a homogeneous effect on public opinion. Zaller (1992) argues that political awareness determines how citizens respond to elite messages. Politically aware citizens are more able to bring a cue in line with thei r polit ical leanings . They are not learning directly from the cue, but they are more able to scrutinize the source to determine whether it fits with their ideology and party. Politically unaware people cannot do this because they are less likely to understand how the cue fits with their ideology and party. Therefore, political unaware cit i zens are more likely to accept an inconsistent cue, thus holding a position that does not align with their broader political ideology. And there is plenty of evidence that politi cal awareness conditions the influence of political ideology and partisanship on climate change opinions in the United States. Recent research has shown in predicting opinions and attitudes on climate change, the influence of education ( Hamilton 2008 , Hamilton and Keim 2009 , McCright 2011 , McCright and Dunlap 2011 ) self - described scientific knowledge ( Hamilton 2011 ), self - described knowledge about climate change ( Malka, et al. 2009 ), assessed scientific knowledge ( Hamilton, et al. 2012 ), and even income ( Bohr 2014 ) were conditional on left - righ t political orientation. 7 Moreover, considering that the consequences of climate change are long - term and individuals discount tend to discount the future, which is an additional disincentive for citizens to seek information ( Jacobs and Matthews 2012 ). 57 These empirical findings are powerful empirical evidence that educated and scientifically knowledgeable Americans engage in motivated reasoning or turn to political elites for cues. Sorting out the motivations is difficult with observational data only in the United States . It is not known if this conditional relationship between engagement and left - right political orientation extends to cross - national setting . And by comparing public opinion in the United States to public opinion in other cou ntries, it is possible to test whether motivating reasoning or elite cue - taking best explain why Americans are so polarized on climate change. Table 3 .1 summarizes the different theoretical explanations for environment and climate change risk perce ption f rom this section. 58 Table 3 .1 : Theory and empirical e xpectations Theoretical Perspective Concept Expectation Post - materialism Theory Post - materialism Individuals with post - material values should be more concerned about climate change than those with mate rial values . Alternative Approach Individuals with post - material values should be equally or less concerned about climate change than those with material values . Affluence Hypothesis Individual Wealth Rich individuals should be more concerned about c limate change than poor individuals. Alternative Approach Rich individuals should be equally or less concerned about climate change than poor individuals. Affluence Hypothesis Country Wealth/ Development Citizens in rich and more developed countries sho uld be more concerned about climate change than citizens in poorer and less developed countries . Alternative Approach Citizens in rich and more developed countries should be equally or less concerned about climate change than citizens in poorer and less developed countries . Supporting the White Thesis Religiosity More religious individuals should be less concerned about climate change than less religious individuals. Opposing the White Thesis More religious individuals should be equally or more conce rned about climate change than less religious individuals. Supporting the White Thesis Protestantism Protestants should be less concerned about climate change than non - Protestants. Opposing the White Thesis Protestants should be equally or more concer ned about climate change than non - Protestants. Left - right political orientation Self - described Political Ideology Individuals on the left should be more concerned about climate change than individuals on the right. Deficit Model of S cience Communicatio n Education, Political Interest More educated and politically interested individuals should be more concerned about climate change than less educated and politically interested individuals. Motivated Reasoning Theory /Elite cues Education, political intere st, and environmental knowledge conditioned on political ideology interest, the difference between the left and right in concern about climate change should increase . 59 3.6 Data This project draws from n umerous public opinion surveys to explain global public opinion , including both cross - national surveys (including many countries) and surveys conducted in the United States. Many of these surveys were used in the previous chapter to describe global public opinion on climate change. This section describes the different surveys, why they are used , and in what remaining chapters they are used. More detailed information on the data sources and measurement is available in Appendix A . Five public opinion surveys are used in the remaining chapters . The larges t cross - national survey used is the World Values Survey (WVS) third wave, conducted from 2005 to 2009, including countries from all regions and levels of development. The WVS asked respondents about climate ch ange in 44 countries. The WVS is used throughout the four main empirical chapters. The next larges t cross - national survey used is the International Social Survey Programme (ISSP), conducted in 32 countries mainly wealthy countries between 2009 and 2011. Th e ISSP included numerous questions on climate change and the environment. In addition to using it in Ch apter 1, I also use it Chapter 7 to explore the relationship between attitudes on climate change and the environment. I also use the 2009 and 2010 Pew Gl obal Attitudes study to explore the relationship between climate change risk perception and willingness to pay higher prices to address climate change in Chapter 7. Finally, I use the American National Election Studies from 2008 and 2012 to analyze the con ditional relationship between political ideology and political knowledge on clima te change attitudes in the United States in Chapter 6. 3 .7 Method ological a pproach and estimation s trategy The analysis relies on a variety of different approaches to test t he theoretical expectations and hypotheses. This section des cribes the statistical analysis , with is used throughout to analyze influence of individual - level and country - level chara cteristics. 3.7.1 Multilevel m odeling This project makes extensive use of multilevel modeling because survey respondents were nested within countries in the cross - national surveys. Random - intercept models are used to estimate the average 60 effect of individual characteristics across countries. Random - intercept models are also used to explore the importance of context, in which country - level variation in climate change opinions are modeled as a function of important country - level characteristics, such as GDP per capita. Next, multilevel random - inter cept, random - effect models are use d to explore variation in the effect of individual - level characteristics across countries. Finally, multilevel cross - lev el interaction models are used to explore how the effect of independent variables across context. Multilevel modeling is explained in mo re detail below. Standard linear and generalized linear models assume each observation is independent, but this assumption is problematic in situations in which individuals are clustered into groups, such as in cross - national public opinion surveys in wh ich respondents are grouped within countries. If clustering is not accounted for in the modeling strategy, the parameter estimates and precision of the estimates can be biased . 8 In order to provide unbiased estimates the independent variables and also prop erly estimate uncertainty, the clustering of respondents in countries needs to be modeled. In other words, the model needs to account for the fact that individuals in different countries are exposed to different contextual effects . And these different cont extual effects impact individual opinions and behavior. There are two possible approaches for modeling clustering: fixed effects and the random - intercept model (often called random - effects) ( Clark and Linzer 2013 ). To illustrate the different approaches, consider a standard regression model that does not account for clustering. Let i denote individuals and j denote the c lusters/groups. (1) In order to account fo r the clustering, the intercept must be allowed to vary between clusters. Fixed effects controls for the unexplained group - g j - 1 group dummy variables in the model. This allows for each country to have its specific intercept. So the model looks as follows: 8 This modeling approach is often called the pooling approach ( Gelman and Hill 2006 , Clark and Linzer 20 13 ). There are situations in which the pooling approach produce s unbiased estimates of the parameters. First, if the variation in the intercept estimate is just random noise that cannot be model. Or second, the not included variable s in the model which p redict the dependent variable are not correlated with the included variables. I n practice, these situations almost never arise, so the pooled model is usually the incorrect approach when modeling clustered data. 61 (2) This modeling strategy controls for country - level heterogeneity, which produces unbiased estimates of the average effects of the indepen dent variables across countries. B ut these estimates may be inefficient because of the additional number of parameters included . The random - intercept model is a compromise between poo ling all the observations together and ignoring clustering and the fixed effects model ( Gelman and Hill 2006 ). Here, the intercept is modeled by the normal distribution, defined by the overall/grand mean and the variance of the intercepts, . (3) ) Estimates for each can be recovered from the model , although the distribution does shrink the estimates of toward the grand mean, . The estimates of each intercept are clos er to the grand overall mean than in the fixed effects model. The amount of shrinkage is determined by the variance between countries, . 9 So, when using the random effects model, there is only one term in the model for the intercept, rather than j 1 parameters for the fixed effects model. This improves the efficiency of the estimates of the independent variables in the model. This increased efficiency comes with a trade - off of introducing possible bias into the estimates of the independent varia bles because the model does completely control for country - level heterogeneity. With the random - intercept model, since country - level variation is modeled, omitted variable bias could be a problem ( Bafumi and Gelman 2006 ) . With random - intercept framew ork, the variation in the country - level intercepts can be modeled as a function of country - level characteristics. T he importance of certain contextual factors can be tested in order to explain variation in the dependent variable , which is not captured by t he individual - level 9 The shrinkage of each estimated cluste r intercept to the grand mean is the reason why the random effects model is considered a compromise. When , then there is complete pooling (as all cluster intercepts will be at the grand mean estimate. When , then the grand mean has no effect, and there is no pooling ( Gel man and Hill 2006 ). As Gelman and Hill (258) show, when the fixed effects and random effects models are equivalent. And as finite and estimable variance, , whereas the fixed effec ts specification assumes the intercepts are dist ribution with infinite variation . 62 independent variables . Equation 4 shows a multilevel model with one country - level covariate to explain variation in the country - 00 01 to denote the effect of country - level which captures the residual level - two, country - 01, i which captures the level - one residual variance after controlling for the level - one, individual - lev ij. Prediction error in the random - intercept model is the result of not perfectly modeling the response variable at the individual level and not perfectly modeling the country - level variation ( Steenbergen and Jones 2002 ). (4) This model can then be extended to accommodate multiple level - one and level - two predictors, as k is a vector of level - one predictors an 0m is a vector of level - two predictors. These are the equations used to estimate multilevel models predicted perceived seriousness of climate change using th e WVS in the next chapter . (5) Multilevel models can be extended to allow the effect of an independent variable or variables to vary across countries. In the random - intercept models, the covariate effects are fixed or the same acro ss countries. A researcher, however, might expect for the relationship between an individual characteristic and dependent variable to be different across countries. Models that allow for both country - specific intercepts and country - specific independent var iable effects are called random - intercept, random - effect models. Equation 6 shows random - intercept, random - effect model specified with one individual - level covariate and no country - level covariates ( S njiders and Bosker 2012 , Stegmueller 2013 ): (6) 63 The term represents the average effect of the independent variable across all countries. The country subscript of shows that the independent variable has a different coefficient across count ries, rather than being fixed, as it is in th e random - intercept model. With the introduction of a random coefficient, the model has a more complicated variance structure. The random - intercept, random - coefficient model requires the estimation of three variance components: the variance of the intercept , the variance of the coefficient, and the covariance of the intercept and the coefficient. When estimating the statistical models with random effects in this project, I distinctly estimate the covariance, rather than setting the variance to zero, a strong assumption which nevertheless is available in estimating multilevel models in some statistical software packages. The variance - covariance matrix is as follows ( Stegmueller 2013 ): (7) By estimating a random - intercept, random - effect model, I can test to see if the variation in the effect of an important individual - l evel independent variable, such as left - rig ht partisanship, is significant countries. Then I can test to see if country - level context factors, such as wealth and development, explain the meaningful variation in the effect of the important individual - level variables. To explain this variation, I rely on cross - level interaction models ( Snjiders and Bosker 2012 ). In these models, both the country - specific intercept and the effect of the individual - level variable vary across countries, and the variation in the effect of the i ndividual - level variable is explained by a country - level variable. So the effect of the individual - level variable varies according to the count ry - level variable. In Chapters 5 and 7 , I test mor e complex theories to explain cross - national public opinion on climate change. I hypothesize that the effect of an individual - level characteristic, such as left - right political orientation, is conditional on a country - level factor, such as income. To test these hypotheses, I need to use cros s - level level interaction terms . Equation 8 shows a cross - level interaction model with one individual - level and one country - level variable. (8 ) 64 The country - level factor is included to explain variation in both coefficient and intercept estimates across countrie s. So a country - level variable that measures wealth, such as GDP per capita, would be included to explain differences across countries for the predictive effect of an individual - level characteristic, such as left - right political orientation, and also to ex plain variation in the level of a dependent variable, like perceived seriousness of climate change , across countries. The variance - covariance matrix is the same as in the random - intercept, random - coefficient model, with the variance of the intercept, the v ariance of the coefficient, and the covariance of the intercept and coefficient. The dependent variables are measured at different levels in this project. Som e variables, such as saying climate change is the most important environmental problem or not, ar e binary, some variables are ordinal, such as the perceived seriousness of climate change, and some variable are assumed to be continuous - like, such as the perceived dangerousness of climate change. Using a linear functional form to predict an ordinal or b inary dependent variable could be problematic for a number of reasons. Most important, a linear model could produce inaccurate standard errors, thus resulting in incorrect significance testing. In order to motivate both the binary and ordered models, I as sume the underlying attitude that motivates responses to these survey questions is unobserved, latent, and continuous ( Hedeker and Gibbons 1994 , Hedeker 2008 , Bauer and Sterba 2011 ). If I could observe this latent attitude, they would be cont inuous, and allowing for a linear functional form. This continuous u nderlying attitude is y i * . The observed categorical response is based on this unobserved attitude, and response categories are determined by in a certain range in the observed attitude. I link the latent model with the observed values of climate change ris k perception with a threshold model , in which the thresholds determine the observed values . (9 ) ) 65 This is the connection between the latent and observed variables in the ordered model for a dependent variable with four ordered choice categories. (10 ) Finally, I need to specify distributions for the individual - level variance and country - level variance . The individual - level variance in a binary and ordered random - intercept and random - intercept, random - effect models are fixed . I use the logistic distribution, so the individual lev 2 /3 for both the binar y and ordered models. And at the country - level , the variance is assumed to follow the normal distribution, just as in the multilevel model because the random effects are not impacted by the response function. I also need to clearly specify the meaning of the intercept in multilevel binary and ordered models in which the level of the dependent variable is assumed to vary across context. In a multilevel random - intercept linear model, the country - specific interce pts vary across countries. In the multilevel binary model, the country - specific thresholds vary across countries. And in the ordered multilevel model, the first threshold is allowed to vary. The other thresholds for the latent variable are determined by th e variation in the first threshold, so the distance between the thresholds in the latent variable model is the same across countries. The linear multilevel models are estimated using maximum likelihood estimation methods, in which the specification of the of the two disturbance terms is normal. The binary and ordered logistic multilevel models are estimated using adaptive Gauss - Hermite quadrature as the integration method. 10 3.7.2 Exploratory factor a nalysis I use explorat ory factor analysis in Chapter 7 to explore the dimensions of concern about environmental problems. Surveys often ask respondents a series of questions in an effort to more 10 The multilevel models were estimated using Stata 13. 66 precisely estimate latent attitude s, opinions, or evaluation s . In this case and many others, it is assumed that ther e is one or are two observed factors motivating the responses to survey questions. Exploratory factor analysis, unlike confirmatory factor analysis, makes no pri or assumptions about these laten t factors. Factor analysis is model of linear correlation among variables; it then extracts common factors which explain as much of the common variance shared by responses to the survey questions. One way to think of exploratory factor analysis is as multivariate regression, with the common factors as the independent variables explaining variables in the observed dependent variables. The goal is to determine and identify the latent attitudes that are motivating responses to specific questions ( Bartholomew, et al. 2008 , Mulaik 2010 ). Here, the goal is to explore latent attitudes toward environmental problems. 67 Chapter 4 : What e xplains cross - national climate change risk p erception? 4 .1 Introduction The previous chapter introduced competing explanations for variation in individual concern about the environment and climate change from previous research, both cross - nationally and in the United States. This research contends that the direction of the relationship between factors such as post - materialism, individual - wealth, country - affluence, and religiosity with climate change risk perception could be different. This chapter tests hypotheses from these competing explanations by using cross - national public opinion data from the World Values Survey (WVS) from 2005 - 2009. By using the WVS, I can include a diverse group of more than three dozen countries in the analysis. From the empiri cal analysis, I make three main conclusions. First, individuals across countries think of climate change as a political issue. On average, across dozens of countries, how individuals think about politics and identify politically, but not their attitudes on e conomics and social issues, are closely related to climate change risk perception . S econd, educated individuals are more likely to say climate change is serious than those less educated on average across countries; education has a substantively important influence on climate change risk perception. Third, citizens in very r ich and developed countries are not more worried about climate change than citizens in less rich and developed countries. Other approaches to the study of public opinion on the environm ent were not as strongly supported empirically. Religio sity and religious identity are not cl early related to perceived seriousness , providing little evidence in support of the White Thesis across countri es. Post - materialism is weakly associated with perce ived seriousness , showing that fundamental values changes are not strongly related to risk perception. And self - described income and socio - economic class, e ven if imprecisely measured, do not influence climate change risk perception. Finally, the ass umptio n that younger people are more c oncerned about climate change i s not supported. Instead, across countries, the relationship between age and the perceived seriousness of climate change seems to be curvilinear. Younger and older individuals are less likely t o say climate change is serious than those in middle age 68 4 .2 Hypotheses From the discussion in the previous chapter, which introduced explanations for both individual - level and country - level variation in concern about climate change, I introduce the follo wing hypotheses. This discussion of previous theoretical and empirical work yields a number of conflicting, competing expectations. Table 4 .1 summarizes the theoretical perspective, the resulting hypotheses, and empirical measures. Table 4 .1 also provides expected directi on of the empirical expected based on the hypotheses. First, on the relationship between post - materialism, wealth, affluence, and climate change risk perception, the theoretical approaches on explaining attitudes toward environmental prote ction predict the relationship could be positive, zero, or even negative. According to post - materialism theory, I expect a positive relationship between individual post - materialism and concern about climate change because of the values changes: H1: Individ uals with po st - materialist values consider climate change more serious than those with survivalist values (post - materialism theory) . According to the alternative approach to explaining th is relationship, there is no relationship or even a negative relation ship between post - materialism and concern about climate change. H2: Individuals with post - materialist values do not consider climate change more serious than those with survivalist values (alternative approach) The affluence hypothesis posits that the rel ationship between wealth and concern is direct, rather than indirect; therefore, I should expect a positive relationship between individual affluence and saying climate change is serious . H3: Rich individuals consider climate change more serious than poor individuals (affluence hypothesis) . The affluence hypothesis also holds that country wealth and level of development is positively related to perceived seriousness . H4: Individuals in rich and mor e developed countries consider climate change more serious than individuals in poorer and less developed countries (affluence hypothesis) . 69 The alternative approach, however, predicts wealth and perceived seriousness of climate change are not positively related, both at the individual - level and country - level. Firs t, individual wealth could not be associated or even negatively associated with perceived seriousness of climate change: H5: Rich individuals do not consider climate change more serious than poor individuals (alternative approach) . The alternative approach also holds that citizens of affluent countries are not more likely to say climate change is a serious problem than citizens of less affluent countries. H6: Individuals in rich a nd more developed countries do not consider climate change more serious than i ndividuals in poorer and less developed countries (alternative approach) . Next, the relationship between religiosity and climate change risk perception could be either positive or negative. Previous empirical and theoretical work, based on the White thesi s and focused on the United States, predicts a negative relationship: H7: More religious individuals consider climate change less serious than less religious people (supporting the White Thesis) . This previous work building on the White Thesis also predict s that Protestants are less likely to be concerned about the environment and less likely to say climate change is serious. H8: Protestants consider climate change less serious than non - Protestants (supporting the White Thesis). Or it could be that the Uni ted States is exceptional because of its relatively large group of evangelical and conservative Christians, and cross - nationally, the religiously devout are not more likely to say climate change is serious. H9 : More religious individuals do not consider cl imate change less serious than less religious people (opposing the White Thesis) . And a cross - national study could also find that Protestants are actually about as likely or even more likely to see climate change as a serious problem. 70 H10: Protestants do n ot consider climate change less serious than non - Protestants (opposing the White Thesis) . The discussion on the relationship between political ideology, partisanship and concern about climate change yield s a number of hypotheses. First, based on previous e mpirical research on public opinion on the environment and climate change in the United States , I expect left - right po litical orientation to predict perceived seriousness of climate change . H11: Individuals on the left consider climate change more seriou s than individuals on the right. Finally, I also explore the relationship between political knowledge and engagement and education and perceived seriousness of climate change . The relationship could be direct, as posited by the deficit model. So as an indi education and political engagement increases , he or she also becomes more likely to say climate change is serious. d political interested increases , he or she considered climate ch ange more serious (th e deficit model) . Or it could be that the relationship between knowledge and concern about climate change is conditional on political ideology and/or political partisanship, as predicted by approaches that emphasize motivated reasoning and elite cues. Acco rding to motivated reasoning theory, only sophisticated individuals are able to use their political ideology and partisanship to form an opinion on an issue such as climate change. And proponents of the elite cues approach contend only sophisticated indivi duals are able to utilize cues from ideologically similar political elites. Therefore, according to these two approaches, the difference between the left and right in perceived seriousness of climate change increases as level of education, interest in poli tics, and knowledge about the environment increases. nterest in politics increases , the difference between the left and right in perceived seriousness of climate change increases (motivated reasoning and el ite cues) . 71 Recent empirical research on public opinion in the United States has found empirical evidence in support of this hypothesis. Here I test to see if this finding is also support ed in the cross - nationally . Table 4 .1 : Theory, h ypo theses, and emp irical e xpectations Theoretical Perspective Concept Variable Hypothesis/Hypothesis Empirical Expectation Post - materialism Theory Post - materialism Post - materialism H1: Individuals with po st - materialist values consider climate change more serious than tho se with survivalist values . + Alternative Approach H2: Individuals with post - materialist values do not consider climate change more serious than those with survivalist values . 0 or - Affluence Hypothesis Individual Wealth Socioeconomic Class Income D ecile H3: Rich individuals consider climate change more serious than poor individuals . + Alternative Approach H5: Rich individuals do not consider climate change more serious than poor individuals . 0 or - Affluence Hypothesis Country Wealth/ Developm ent GDP per capita Human Develo p ment Index (HDI) H4: Individuals in rich and mor e developed countries consider climate change more serious than individuals in poorer and less developed countries. + Alternative Approach H6: Individuals in rich a nd more developed countries do not consider climate change more serious than individuals in poorer and less developed countries. 0 or - Supporting the White Thesis Religiosity Attendance of religious services H7: More religious individuals consider cli mate change less serious than less religious people . _ 72 Opposing the White Thesis H9 : More religious individuals do not consider climate change less serious than less religious people . + Supporting the White Thesis Protestantism Protestant Identity H8: Protestants consider climate change less serious and dangerous than non - Protestants . - Opposing the White Thesis H10: Protestants do not consider climate change less serious than non - Protestants . + Left - right political orientat ion Self - described Political Ideology Placement on the left - to - right continuum H11: Individuals on the left consider climate c hange more serious than individuals on the right. - Deficit Model of S cience Communication Education, Political Interest Level of education Political interest Self - described environmental knowledge level of education, political interest, and knowledge increases, he or she considered climate c hange more serious. + Motivated Reasoning Theory /Elite cues Ed ucation, political interest, and environmental knowledge conditioned on political ideology Level of education Political interest Left - right political orientation level of education, political i nterest, and knowledge increases , th e difference between the left and right in perceived seriousness of climate change increa ses. The difference between l eft and right increases as education and political interest increases 4 .3 Measurement 4 .3.1 Dependent v ariable The dependent variable captures individual perception of the seriousness of climate change using the WVS. Specifically, as mentioned in previous chapters , WVS asked respondents: environmental problems in the world as a whole. Please tell me how serious you c onsider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious, or not serious at all? Global wa 73 4 .3.2 Independent variables at the individual level Measurement is ex pla ined in detail in Appendix A and Table 3.1 provides the variables used to measure concepts introduced in the previous chapter . The following variables are used in different model specifications. Post - materialism: In the WVS, post - materialism is a twelve it em - index measuring values, running from - 2.5, equated with materialist values, to 2.5, equated with post - materialist values. Left - right political orientation: I include a self - described political ideology to measure left - right political orientation. In th e WVS, respondents were asked to place themselves on a range from one to 10, in which one indicates far left and 10 indicates far right. 11 Economic attitudes : Q uestions are included to measure opinions on income inequality and private enterprise. Responde nts were asked to place their opinion on the appropriate level of income also asked to place their opinion of private enterprise on a range from one to 10; 1= private ownership of business and industry should be increased ; 10= government ownership of business and industry should be recoded so individuals who took positions on 11 Self - placement on a single dimension of ideology, from left to right, is often used in public opinion research, bu t it can be problematic. First, individuals struggle to place themselves in many individuals across countries on the left - right continuum because there is wide cross - national variation in the use and mean of the left - right distinction in politics. The vari ation in the use of the left - right continuum on the WVS is explored in more detail in Chapter 4. Second, responses to the self - placement question are clustered in the middle categories. Of those respondents who place themselves on the left - to - right - contin uum in all 42 countries (respondents in China and Malaysia were not asked to describe their political ideology), 39 percent cho o se one of the two middle categories, so identifying as self - described moderates. In only three of the countr ies Italy, Vietnam a nd Jordan i s one of two middle categories out of the 10 categories not the modal response. This could show that many citizens across these countries consider themselves moderates. Or this finding could be the result of social desirability bias because resp ondents think choosing a category in the middle show s they are moderates, and believing this answer is socially de sirable. After all, ideologically extremists are often viewed with suspicion. So self - placed ideology may be underestimating the va riance in l eft - right political orientation . 74 the economic issue to the left of center were given negative numeric values and those with positions right of center were given positive numeric values. 12 Social values index: I also include a measure of position on socia l is sues , which combines answers to questions about whether homosexuality, abortion, divorce, and euthanasia are never justifiable, always justifiable, or somewhere in between. 13 This measure captures the social dimension of left - right political orientation , with negative numeric values reflecting views on the left on social issues and positive numeric values capturing vie ws on the right on social issues . Education: I include an ordinal measure of education level with values ranging from no formal education to primary education to secondary education to attending and completing university education. According to previous research, I expect greater education to incre ase to likelihood of saying climate change is a serious problem . Interest in politics: I also include an ordinal measure of interest in politics, a four category variable ranging from not at all to very interested in politics. Higher values are associated with greater interest in politics. According to previous research, I expect increased interes t in politics to increase the probability of saying climate change is a serious problem . Individual Wealth: T o test the affluence hypothesis at the individual level, I measure two different measures self - described socio - economic class and self - described i ncome decile. 14 12 These two questions and other questions on government welfare and the role of competition do not seem to capture the same latent attitude on economic issues. They are weakly correlated with each other and do not map ont o a common factor in an exploratory factor analysis. Even in the United States, these variables were not capturing the same latent economic attitude ( Budge, et al. 2001 ). 13 This measure is adapted from Benoit and Laver 2009 . Unlike the questions on economics, responses to these social questions were unidimensional. Across all countries, responses to these questions map onto the same common factor, which explains 57 percent total variance in the four variables. All four measures were highly correlated with the common factor. Abortion was the most related with the common factor (0.83), followed by divorce (0.78), gay (0.75), euthanasia (0.66). In addition, specifically in the United States, the four measures mapped onto one common facto r, with the common factor explaining 56 percent of the variance in the measures. The four measures are also highly correlated with the common factor, with abortion once again most closely related (0.88), followed by divorce (0.73), euthanasia (0.70), and g ay (0.66). 14 These measures of individual affluence can be problematic though. First, they were self - described, and individuals might not accurately answer these questions, either because they do not know their relative position in society or due to socia l desirability bias. Fewer respondents than expected answer that they were in the upper class or the upper 75 - described economic class with categories ranging from lower class to upper - middle class/upper class. Second, I use their self - described income decile, from one, the first decile, to ten, the top decile . R eligiosity and Religious Identity : I measure religiosity with two variables. First, I measure attendance of religious services, by capturing how often respondents say they attend religious services. Second, I also include a measure capturing the self - descr ibed importance of God for the models using the WVS data. 15 I also measure for the effects of religious identities. Considering the White Thesis focuses generally on Christian identifies and specifically on Protestantism, I measure for the effect of identif ying as a Protestant, Catholic, or Evangelical. I also measure for the effect of measure of identifying as a Muslim. By including measures of both religiosity and religious identity, I am testing for applicability of the White Thesis on climate change risk perception. 4 .3.3 Control v ariables at the individual - level I include age and gender as control variables. I would ex pect younger people to be most likely to say climate change is a serious issue based on cohort replacement theory, in which younger c ohorts are more concerned about environmental issues than older cohorts because of changes in values ( Inglehart 1977 , Inglehart and Welzel 2005 ). One recent cross - national study using the WVS found that the effect of income deciles, and more respondents than expected answer they were in the middle class or middle income deciles, illustrating that social desirabili ty almost certainly biases answers to these questions. In the WVS across 43 midd not answer the question. Becau se so few respondents identify eople who consider themselves high income. And when asked on the WVS to place themselves into their income decile, only about four percent of respondents placed themselves into the one of the top two deciles. Me anwhile, about 40 percent place themselves in to one of the three middle deciles . So when respondents are asked about their income or class, some understate their place in society, which could lead to misleading results. lth in their country, by asking respondents to compare themselves to others within their country, not to others in the world as a whole. So their self - described wealth among all of the surveyed countries might be different. It might be asking too much of respondents to compare their affluence to the whole world, however, and this might lead to further response bias. 15 Across the 44 countrie ordered response category; this distribution is heavily skewed. 76 age on risk perception of climate change was nonlinear, with those in middle age more concerned than the young and old ( Kvaløy, et al. 2012 ). To allow for the possibility of a nonlinear relationship, in both cross - national surveys, age is included in categories: under the age of 30; 30 - 3 9; 40 - 49; 50 - 59, 60 - 69, and 70 and older. As for the effect of gender, research based on the United States has found that women are likely than men to be concerned about the environment ( Eckberg and Blocker 1989 ) and climate change ( Kellstedt, et al. 2008 ). 4 .3.4 Independent variables at t he country - level The primary independent variable of interest at the country - of development. 16 country affluence and development with two variables: mean GDP per capita between 2005 and 2009 and mean Human Development Index (HDI) score between 2005 and 2009. While GDP per capita measures country wealth directly, HDI, which encompasses life expectancy and education in addition to income, pro predicted positive relationship between country affluence and concern based on the affluence hypothesis. In addition to these two variables measuring a co - level factors in the multilevel models. Many of these variables have not been included in previous cross - national studies on climate change public opinion. First, I test for the effect of CO 2 emissions in a country, measuring mean CO 2 emissions in tons per capita from 2005 to 2009. Previous research has yielded somewhat divergent expectations for the effect of CO 2 emissions. Franzen and his co - authors ( Franzen 2003 , Franzen and Meyer 2010 ) find a positive relationship between CO 2 emissions and broad concern for the environment . Sandvik ( 2008 ) find a weak, negative relations hip between CO 2 emissions and aggregate country - level concern about climate change. They posit that the inconsistent empirical findings on the effect of CO 2 emissions could be the result of different samples of countries. Next, while previous work on cros s - national public opinion on climate change has looked at the impact of CO 2 emissions, the demand - side of the fossil fuel industry, it has largely ignored the impact the 16 Appendix A provides more information, including data sources, on the country - level variables. 77 supply of energy could have on public opinion on climate change. Suppliers in the foss il fuel industry arguably have more at stake in maintaining the status quo global energy economy. Widespread concern about climate change could lead to global policy changes , which hurt s fossil fuel supplying countries. For example, policy changes could le ad to increased demand for alternative forms of energy and decreased demand for traditional energy. Therefore, perceived seriousness of climate change could be lower in countries in which fossil fuel is produced. To measure the impact of oil and gas produc tion, I use the natural logarithm of the value of the mean of per capita oil and gas production in each country between 2005 and 2009 ( Ross 2013 ). I also include two additional country - level factors that could be related to polity score between 2005 and 2009. It could be that as a country becomes more democratic, its citizenry becomes mor e include to consider climate change serious . Second, I also explore the possible effect of that perceived seriousness of clima te change is greater in countries with less income inequality. 4 .4 Explaining cross - national climate change risk p erception Before jumping into the results and analysis of the multilevel models, I fi rst present preliminary evidence that the publics perceived dangerousness of climate change does not increase as a country - level wealth and development increased. Figure 4.1 shows the relationship between GDP per capita, HDI, and aggregate perceived seriou s ness . Figure 4.1 provides evidence that the relationship between wealth and c limate change risk perception i s not positive at the country - level. The fifth round of the WVS includes a broad sample of countries, including less developed and poorer countries . With this broader sample of countries, the evidence suggests that c ountry wealth and development i s not positively related to aggregate perceived seriousness. Although the linear regression line h as a positive slope, slope estimates for GDP per capita an d HDI are substantive weak and not significantly different from zero. 17 And the loess curve suggests that the percentage of those who answer that climate change is 17 Both H DI ( p= 0.307) and the natural logarithm of GDP per capita ( p =0.221) are not statistically different from zero in a simple linear regression. 78 problem decreases, increases and then decreases again across values of GDP p er capita and HDI, evidence to suggest that the publics of very rich countries are not more worried about climate change than citizens of middle - income countries. 18 The results of this aggregate - level analysis show that wealth and development is not positiv ely associated with perceived seriousness of climate change, contrary to the expectations of the Affluence Hypothesis. Figure 4 .1: Relationship between wealth / development and perceived seriousness of climate change, by country (World Values Survey 200 5 - 2009 ) Notes : The scatterplot shows the country - level relationship. The black line is linear best fit, while the grey line is the loess curve. Next , I pr esent the results of models predicting individ ual perceived seriousness of climate change . For these models presented in the text, I use an ordered logistic functional form to predict risk perception of climate change, so the individual - level variance is fixed and not estimated as part of the model. The latent variable formula is presented below wit h a vector of individual - level characteristics 18 To test this nonlinear relationship, I ran polynomial regression models with squared and cubic terms pred icting the proportion of citizens in each country to view climate change as very serious. All three HDI terms were statistically different from zero (all three p <0.01), while the terms for the natural logarithm of GDP model are not statistically different from zero ( all three p = 0.18) . 79 without any country - level characteristics, since the first set of statistical models do not include any country - level predict ors . (1) The estimates for random - intercept multilevel models allowing the first cut point to vary across countries without country - level va riables are presented in Table 4.2. Figure 4 .2 su mmarizes the predictive effects of the individual - level predictors. I use these models to explore the relationship of the individual - level variables with perceived seriousness of climate change before turning to models with country - level covariates . 19 Firs t, the association between post - material values and perceived seriousness of climate change is statistically significant, but the substantive association is quite small. Individuals with post - material values (post - mat erialism=1.5) are only about four perce nt more likely than those with more materialist values (post - materialism= - 1.5) to consider climate change very serious. So the adop tion of post - material values does not drastically inc rease the likelihood of considering climate change a serious problem . 20 Second, how indi viduals think about politics is closely related with worry about climate chang e. Across countries, an individual who identifies with the left is more likely to say climate change is serious than an individual who identifies with the right . Specific ally, an individual on the r ight (political ideology=8) is predicted to be 5.5 percent less likely to consider climate chan ge very serious than an individual on the left (political ideology=3) [95 percent confidence interval: - 0.072, - 0.028]. And M odel 3 provides evidence that the being on the right has a stronger effect than being on the left. Model 3 shows that the coefficient estimate for self - described political ideology on the rig ht (political ideology=7 - 10) is more than double the absolute siz e of self - described political ideology on the left (political ideology=1 - 4). 19 The subst antive effect and statistical significance of the coefficient estimates for the individual - level variables are very similar in both the random effects and multilevel models. 20 Unless explicitly mentioned, Model 6 in Table 4 .2 is used to calculate the predicted probabi lities and first differences with the other covariates held at their means and modes . 80 Specifically, an individual on the right is predicted to be a little more than about five percent less likely to consider climate ch ange very serious than a moderate (political id eology=5 - 6), whi le an individual on the left is predicted to be about two percent more likely to consider climate change very serious compared to a moderate . So on average across countries, being on the rig ht has a stronger influence on climate change risk perception than being on the left. But how people think abou t economic and social issues is not associated with perceived . Both measures of views on economic issues are not statistically significant. And the social issues index does ha ve its expected sign, but it is not statistically significant at conventional levels in either Model 4 or M odel 5. Education and poli tical interest have their expected relationship with perceived seriousness of climate change, as an increase in both is a ssociated with an increase in perceived seriousness . The coefficient e stimates for both variables are statistically significant, but their substantiv e impact i s vastly different. An individual whose political interest moves from not very interested to very interested has a resulting increase in the predicted probability of viewing climate change as very serious by only abo ut four percent . But an individual with a university degre e are predicted to be about 11 percent more likely to consider climat e change very serious than a n individual who completed secondary school but did not pursue additional university or vocational education [95 percent confidence interval: 0.074, 0.141]. Education has the strongest substantive predictive influence on perceived seriousness of the indivi dual - level variables in the model. The relationship between religiosity and religious identity and the perceived seriousness of climate change is more complex . The coefficient esti mate for religious attendance i s negative, but not statistically different from zero. The negative, statistically significant influence of Protestant identity in some model specifications i s evidence supportive of the White Thesis. But the association is quite weak. Specifically, a Protestant is predicted to be about three percen t less likely to consider climate change very serious than an individual who do es not self - identify as Protestants and th e estimate is very uncertain (95 percent confidence interval: [ - 0.064, - 0.001]. As shown in Model 1 in Table 4 .2 , the coefficient estim ates 81 for Catholic, Evangel ical, and Muslim identities are not significantly different from zero compared to individuals who do not identify with these religions. 21 Finally, the most important finding in op position to the White Thesis is the strong positive association between the importance of God and perceived seriousness of climate change. This showed a possible disconnection between religiosity and spirituality in the formation of attitudes on climate change. One does not necessarily need to attend church to consider God very important, so it seemed that more spiritual individuals are more likely to say climate change is serious , while those who were formally religious are less likely to say climate change is serious . Self - described wealth i s measured in two ways in the models analyzing the WVS self - identified class and income decile, and both measu res of individual affluence are not related to perceived seriousness of climate change, refuting the Affluence Hypothesis at the individual - level. Self - d escribe d socio - economic class i s estimated using a series of dummy variables, with the reference category as middle class. Lower class, working clas s, and upper class a re all not significantly diff erent from middle class in the probability of saying climate chang e is very serious . Moreover the classes were not significantly different from each other. 22 The coefficient estimate for self - described income decile in Model 2, in which respondents place themselves in deciles compared themselve s to others in their country , i s not statistically different from zero. So even though individual wealth i s not measured precisely, the Affluence Hypothesis at the individual - leve l i s not supported in the analyses . 23 As for the two control variables, wom en are more likely to say clim ate change is serious than men in only two model specifications, and the predicted substantive effect is weak, as women a re only about 1.5 percent more likely to consider climate change very serious than men. The finding for age i s 21 Using an F - test, I compared the coefficient estimates for the religious identities, and they are not significant ly different from each other at a standard level. Protestant is not significantly different from Evangelical ( p =0.556), Catholic ( p =0.223), or Muslim ( p= 0.079). Catholic is not significantly different from Evangelical ( p =0.869) or Muslim ( p =0.087). And Eva ngelical is not significantly different from Muslim ( p= 0.208). 22 Using an F - test, I compared the coeff icient estimates. Lower class is not significantly different from working class ( p =0.974) or upper class ( p =0.927). And working class is not significant ly different from upper class ( p =0.86 ). 23 In models predicted perceived dangerousness of climate change using the International Social Survey Programme data, self - placed income decile is not a statistically significant predictor. 82 particularly interesting and deserves further discussion. In all five models, individuals younger than 30 are less inclined to consider climate change very serious than those between the age of 40 and 49, the reference category once again in all models. Specifically, an individ ua l younger than 30 is predicted to be about three percent less likely to view climate change as very serious compared to 40 to 49 year - o lds. Individuals 70 and older a re less inclined to say climate change is serious t hose 40 to 49 years old, but the coeffi cient estimate for individuals 70 and older is not statistically different from zero. 24 This finding suggests that the relationship between age and concern is not linear, with perceived seriousness not decreasing with age, but instead curvilinear, with conc ern peaking in middle age. 25 24 A statistical modeling including terms for both age and age squared is presented in Appendix D (Table A4.2). Both terms are statistically significant at conventional levels, providing further evidence of a nonlinear relationship between age and climate change risk perception. 25 The cut points for underlying latent perceived seriousness of climate change in the random - intercept, ordered logistic regression model should be discussed bri efly. First, the cut points are all significantly different from each other, suggesting that fou r response categories are distinct and meaningful. Second, the three cut points are also almost equally spaced out. 83 Table 4 .2: Random - intercept ordered logistic multilevel models without country - level variables predicting perceived seriousness of climate change (World Values Survey 2005 - 2009) (1) (2) (3) (4) (5) (6) Political ideology - 0 .0474 - 0.0462 - 0.0483 - 0.0473 (from left to right) (0.0133) (0.0131) (0.0128) (0.0126) L eft 0.1012 (ideology=1 - 4) (0.0436) Right - 0.2042 (ideology=7 - 10) (0.0675) Social issues index - 0.0266 - 0.0270 (0.0154) (0.0157) Income equality - 0.0012 - 0.0027 (0.0101) (0.0100) Private enterprise 0.0061 0.0104 (0.0104) (0.0091) Post - materialism index 0.0715 0.0638 0.0652 0.0619 0.0634 0.0660 (0.0221) (0.0225) (0.0220) (0.0230) (0.02 47) (0.0221) Interest in politics 0.0602 0.0655 0.0643 0.0286 0.0503 0.0628 (0.0206) (0.0208) (0.0193) (0.0216) (0.0224) (0.0200) Education level 0.0900 0.0960 0.0928 0.0838 0.0856 0.0928 (0.0137) (0.0143) (0.0140) (0.0152) (0.0158) (0.0140) A ttend a nce of religious services - 0.0185 - 0.0209 - 0.0237 - 0.0262 - 0.0249 - 0.0240 (0.0145) (0.0151) (0.0142) (0.0143) (0.0149) (0.0142) Importance of God 0.0457 0.0459 0.0462 0.0471 0.0487 0.0461 (0.0086) (0.0082) (0.0084) (0.0082) (0.0094) (0.0085) P rotest ant - 0.1614 - 0.1646 - 0.1464 - 0.1053 - 0.1280 - 0.1443 (0.0820) (0.0802) (0.0740) (0.0687) (0.0790) (0.0747) C atholic - 0.0406 - 0.0403 (0.0483) (0.0476) E vangelical - 0.0452 - 0.0492 (0.0410) (0.0480) M uslim 0.0755 0.0641 (0.0936) (0.0964) Lower class 0.0321 (0.0713) Working class 0.0236 (0.0335) Upper class 0.0217 (0.0412) F emale 0.0799 0.0591 0.0684 0.0697 0.0713 0.0698 (0.0357) (0.0351) (0.0357) (0.0357) (0.0353) (0.0358) Age under 3 0 - 0.1013 - 0.1319 - 0.1185 - 0.1327 - 0.1162 - 0.1180 (0.0442) (0.0460) (0.0458) (0.0480) (0.0533) (0.0461) Age 30 - 39 - 0.0135 - 0.0166 - 0.0074 - 0.0239 - 0.0287 - 0.0089 (0.0324) (0.0336) (0.0320) (0.0368) (0.0387) (0.0323) Age 50 - 59 0.0179 0.0184 0.0190 - 0. 0218 0.0079 0.0191 (0.0340) (0.0341) (0.0334) (0.0354) (0.0376) (0.0332) Age 60 - 69 0.0062 0.0115 0.0141 0.0118 0.0173 0.0131 (0.0465) (0.0522) (0.0463) (0.0504) (0.0531) (0.0465) Age 70 and older - 0.0529 - 0.0432 - 0.0504 - 0.0669 - 0.0492 - 0.0512 (0.0 489) (0.0528) (0.0478) (0.0527) (0.0596) (0.0479) 84 Self - described income decile - 0.0032 (0.0120) Not serious at all | - 4.3201 - 4.3520 - 4.1069 - 4.0465 - 4.3952 - 4.3304 Not very serious (0.2233) (0.2422) (0.19 00) (0.2025) (0.2335) (0.2114) Not very serious | - 2.5436 - 2.5626 - 2.3319 - 2.2918 - 2.5864 - 2.5558 Somewhat serious (0.1678) (0.1729) (0.1440) (0.1569) (0.1704) (0.1579) Somewhat serious| - 0.5880 - 0.6016 - 0.3868 - 0.3821 - 0.6276 - 0.6118 Very serious (0 .1309) (0.1350) (0.1112) (0.1156) (0.1309) (0.1236) Variance component Country - level variance 0.4143 0.3849 0.4056 (0.4095) 0.4219 0.4075 (0.1013) (0.0967) (0.0991) (0.0981) (0.1049) (0.0997) N umber of respondents 38753 37189 39973 39068 32855 3 9973 Number of countries 40 39 41 40 39 41 AIC 69509 66923 71532 71526 59865 71554 Note s : Cell entries are ordered logistic regression coefficients with standard error in parentheses . Coefficients estimates that can be distinguished from zero at the con ventional level statistical significance (two - tailed p < 0.05) are in bold . Listwise deletion of missing values. Four countries are excluded from Model 1 because survey questions were not asked in these countries. In China, the self - placement political ide ology question was not asked. In Malaysia, the self - placement political ideology and religious attendance questions were not asked. Also, in Morocco, the rel igious attendance question was not asked. Finally, in Mexico, the question about socio - economic cla ss was not asked. In Model 2, five countries are excluded: China, Malaysia, Morocco, Jordan, and Argentina. In Jordan and Argentina, the income decile question was not asked. In Model 3, China and Malaysia are both excluded because the self - placement polit ical ideology question was not asked in either country. In Morocco, the religious attendance question was not asked. In Model 4, in addition to excluding to Morocco, Peru, Egypt and Malaysia are excluded because the social issues questions were not asked , so respondents from four countries were excluded . In Model 5, in addition to excluding Morocco, Per u, Egypt, and Malaysia, China i s excluded because self - described political ideology was re - introduced into the model , so respondents from five countries were excluded . In Model 6, China, Malaysia, and Morocco are excluded. Next, I turn briefly to the random intercept multilevel models predicting perceived seriousness of climate change, with particular focus on the country - level predictors. The random - inter cept models with no country - level predictors presented in Table 4 .2 show significant variation in perceived seriousness of climate change across countries. The results of the models presented in Table 4.2 include country - level predictors in an effort to ex plain this variation. Equation 3 shows the latent variable speci fication with a vector of indivi dual - level variables and a vector of country - level variables. (2) 85 None of the included country - level variables, however, have a statistically signi ficant influence on climate change risk perception , and the estimated country - level variation in models without country - level predictors is nearly identical to the country - level variation in models with country - level predictors. 26 This finding provides furt her evidence that richer and more developed countries were not more concerned about climate change than poorer and less developed countries. And it also shows that othe r possible country characteristic do not explain country - level variation. 27 26 In addition, in a different model specification not shown here, GDP per capita, instead of the natural logarithm of GDP per capita, i s not statistically different from zero when included as a country - level predictor . 27 In addition, I estimated multilevel models predicting perceived seriousness with polynomial terms for GDP per capita and HDI , shown in Appendix B . These models show the nonlinear relationship between development and climate ch ange risk perception, in which perceived seriousness increases as countries as countries go from low to medium development, but then declines when countries become very rich and development. For an a verage individual across countries in a country with an HDI of 0.5, the predicted probability of considering climate change very serious is about 42 percent. This predicted probability increases to 72 percent for when HDI is 0.8, a country at a pretty high level of development. But when HDI is 0.9, capturing a highly developed country, an average individual has a predicted probability of considering climate change very serious of 64 percent. This suggests that concern about climate change might decline in t he richest countries. 86 Table 4 .3: Random - intercept ordered logistic multilevel models with country - level variables predicting perceived seriousness of climate change (World Values Survey 2005 - 2009) (1) (2) (3) (4) Political ideology - 0.0473 - 0.0473 - 0.0473 - 0.0473 (from left to right) (0.0126) (0.0126) (0.0126) (0.0126) Post - materialism index 0.0658 0.0659 0.0658 0.0658 (0.0221) (0.0221) (0.0221) (0.0221) Interest in politics 0.0629 0.0629 0.0629 0.0630 (0.0200) (0.0200) (0.0200) (0.0200) Education level 0.0926 0.0926 0.092 7 0.0926 (0.0140) (0.0140) (0.0140) (0.0140) Attendance of religious services - 0.0238 - 0.0238 - 0.0238 - 0.0238 (0.0142) (0.0142) (0.0142) (0.0142) Importance of God 0.0463 0.0462 0.0462 0.0462 (0.0085) (0.0085) (0.0085) (0.0085) Protestant - 0.1446 - 0.1444 - 0.1447 - 0.1445 (0.0747) (0.0747) (0.0745) (0.0747) Female 0.0696 0.0696 0.0696 0.0696 (0.0359) (0.0359) (0.0359) (0.0359) Age under 30 - 0.1174 - 0.1174 - 0.1175 - 0.1174 (0.0463) (0.0463) (0.0464) (0.0463) Age 30 - 39 - 0.0087 - 0.0086 - 0.0087 - 0.0087 (0.0324) (0.0324) (0.0324) (0.0324) Age 50 - 59 0.0189 0.0189 0.0189 0.0189 (0.0331) (0.0331) (0.0330) (0.0331) Age 60 - 69 0.0125 0.0125 0.0126 0.0125 (0.0464) (0.0464) (0.0464) (0.0464) Age 70 and older - 0.0520 - 0.0520 - 0.0519 - 0.0519 (0.04 78) (0.0478) (0.0478) (0.0478) Country - level variables Natural logarithm of mean GDP per capita 0.0490 0.0625 (0.0671) (0.1080) Mean HDI 0.4235 0.7148 (0.7037) (1.2740) Mean polity score 0.0020 0.0038 (0.0270) (0.0259) Mean GINI coe fficient 0.0057 0.0065 (0.0132) (0.0118) Natural logarithm of CO 2 emissions per capita 0.0018 - 0.0190 (0.1279) (0.1489) Natural logarithm of oil and gas rents per capita - 0.0079 - 0.0074 (0.0289) (0.0303) Not serious at all| Not very se rious - 3.8997 - 4.0246 - 3.5828 - 3.5932 (0.6437) (0.5786) (0.9964) (0.9490) Not very serious| Somewhat serious - 2.1251 - 2.2499 - 1.8081 - 1.8185 (0.6639) (0.5903) (1.0034) (0.9504) Somewhat serious| Very serious - 0.1810 - 0.3058 0.1359 0.1256 (0.6439) ( 0.5701) (1.0017) (0.9454) Variance component Country - level variance 0.4017 0.4020 0.3995 0.3991 (0.0941) (0.0945) (0.0943) (0.0941) N umber of respondents 39973 39973 39973 39973 Number of countries 41 41 41 41 87 AIC 71555 71555 71563 71563 Note s : Cell entries are ordered logistic regression coefficients with standard error in parentheses. Coefficients estimates that can be distinguished from zero at the conventional level statistical significance (two - tailed p < 0.05) are in bo ld . Listwise deletion used for missing values. China, Malaysia, and Morocco are excluded due to missing variables at the individual level. Figure 4 .2 : Effects of individual - level variables o n predicting perceived seriousness of climate change (World Value s Survey 2005 - 2009) Notes: The error bar represents the 95 percent confidence interval. The first differences for non - dichotomous variables were calculated using the 90 th percentile and 10 th percentile values of the individual - level variables . The first differe nce for dichotomous variables i s calculated using the two categories. For the WVS, model 6 in Table 4.2 i s used and the other variables were held at their means and modes. For more information, see footnote. 28 28 The specific values used in the calculation were as follows. For the ISSP variables: income gap the difference between agree strongly and disagree; environmental knowledge the difference between - 1.5 and 1; partisanship the differ ence between left party and right party identification; education the difference between completing university and no formal education/primary education; age 70 and older the difference between 40 to 49 years old, the reference category, and 70 years old and older; female the difference between female and male; private enterprise the difference between disagree and agree strongly; age younger than 30 the difference between 40 to 49 years old and younger than 30 years old ; protestant the difference between non - Protestants and Protestants; post - materialism index the difference between 1 (post - materialist values) and - 1 (materialist values); attendance of religious services the difference between never attending and attending once a week. For the WVS variables : level of education the difference between university education and incomplete primary school; importance of Go d the difference between - 2.5 and 4.5; political ideology the difference between 2 and 8; post - materialism index the difference between - 1.5 and 1.5; interest in politics the difference between very interested and not at all interested; 88 4 .5 Conditional relationship between l eft - right political orientation and education/interest in politics The above analyses present additive models, but do not allow me to test all of the proposed hypotheses. Specifically , I cannot test the hypotheses that posit that education and interest i n politics condition the relationship between left - right political orientation and climate change risk perception. To test for this, I estimate a series of models predicting perceived seriousness of climate change including interaction terms at the individ ual - level, and allow the country - level intercepts to vary. The results shows political ideology i s conditioned by self - described political interest on average , but not education, on average across countries . Interest in politics ha s a conditional effect o n left - right political o rientation, as shown in Figure 4 .3, using the WVS models. Among those who say they know nothing at all about politics, an individual on the le ft (political ideology=1 - 4) is predicted to be about four percent more likely to consider climate change very serious than an individual on the right (political ideology=7 - 10), an insignificant different. But among those most interested in politics, an individual on the left is predicted to be more than 1 0 percent more likely than an individual on the right to view climate change as very serious. 29 So left - right polit ical orientation only influences perceived seriousness of cli mate change among those who are at least moderately interested in politics ac ross countries. Specifically, among individu als on the left, inc reasing interest in politics has a significant effect on perceived seriousness of climate change, while for those on the right the increasi ng interest in politics does not have a statistically significant effect on the perceived serious ness of climate change . protestant the difference between non - Protestants and Protestants; attendance of religious services the difference between never and more than once a week; age younger than 30 the difference between 40 to 49 years old, the reference category, and younger than 30; female the difference between female and male; age 70 and older the difference between 40 to 49 years old and older than 70. 29 For those individuals o n the left, the predicted probability of answering very serious increased by 10.5 percent moving from the least interested to the most interested; among those on the right, this change was only four percent and not statistically different from zero. 89 But increasing education increased perceived seriousness of climate change on both the left and right . T he divide between the l eft and the right also increases only slightly as education increases . Among the least educated, an indi vidual on the left is predicted to be five percent more likely to consider climat e change very serious than an individual on the right. 30 And among the most educated, this pre dicted difference only increases to about eight percent. Among those on the right, the predicted probability of considering climate cha by 13 percent moving from no formal education (the lowest level of education response category ) to a university degree (the lowest level of education response category) ; this change is 16 percent among those on the left. So in creasing education is associated with greater concern about climate change for both liberals and conservatives on average a cross countries. Figure 4 .3: Conditional effect of ed ucation/ interest in politics on left - right political ideology o n predicting perceived seriousness of climate change (World Values Survey 2005 - 2009) Notes: The predicted probabilities are calculated for a 40 - 49 year old female with the other variables at their means or modes. The dashed lines represent the 95 percent confidence interval of the predicted probabilities. These results presented evidence both in favor and against motivated reasoning and use of elite cues cross - nationally . How individuals describe d their interest in politics con ditions the influence of left - right political orientation on climate change risk perc eption. For individuals who are not at all interested 30 Thi s predicted difference is not statistically significant. The large confidence interval reflected the uncertainty of the predicted difference due to the lack of respondents in the cross - national sample with no formal education. 90 in politics, individuals on the left and the right are statistically equally likely to say climate change is very ser ious. But among individuals who a re very interested in po litics, those on the left view climate change as s ignificantly more serious than those on the right. But education does not condition the influe nce of ideology. Education increases perceived seriousn ess of climate change political orientation, on average across countries. 31 This shows that the finding using public opinion surveys in the United States that education conditions the influence of left - right political orientat ion on climate change attitudes does not exte nd globally ( Hamilton 2011 ). 4.6 Robustness c hecks To check results of the analysis, I conducted numerous other analyses as robustness chec ks. The results of these analyses , many of whic h can be found in the Appendix B , are virtually identical to the results presented above, providing strong support for the research findings. First, in the analysis above, I tested the hypotheses using only o ne cross - national public opinion dataset the World Values Surve y. To make sure the results are not the result of the specific use of this particular survey , I also analyze survey data from the International Social Survey Programme (ISSP). The results are n early identical. Left - right politi cal orientation, as measured by political party affiliation and education are strong predictors of perceived dangerousness of climate change a t the individual - level. There i s one important difference though. Among country - level factors, wealth and development, as measur ed by GDP per capita and HDI, i s actually significantly and negatively related to perceived dangerousness of climate change . So individuals who live in wealthier countries say climate change is less dangerous than those who live in middle - income and wealthy countries. 32 Second , in the above analyses, I a ssume an ordered logistic functional form for the WVS; the proportional odds assumption in the ordered logistic models could be problematic. The results of th e 31 This finding suggests that recent studies which found evidence of a conditional effect of education on ideology and partisanship in the United States might just be unique to the context of the United States and not extend cross - nationally. This is explore d in more detail in Chapter 5 . 32 The results of analysis of the I nternational S ocial S urvey P rogramme were included in Appendix B , but a detailed discussion of the results is not included. This discussion can be provided by the author. The author can be contacted at brkennedy84@gma il.com . 91 analysis, however, are not sensitive to functional form. I estimated two additional models first, a linear model predicting perceived seriousness and second, a binary logistic model predicting saying climate change is a very serious problem . The substant ive results are the same as in the models presented above in the text. Third , I use list wise deletion to deal with missingness in the stat istical models in the text . Models predicting perceived seriousness of climate change using both the WVS and ISSP su ffer from thousands of missing observations; the models using listwise deletion could produce less efficient and possibly biased estimates of the effects of the independent variables ( King, et al. 2001 ). To overcome this problem of missing data, I produced mul tiple imputation datasets for the WVS and ISSP using Amelia II ( Honaker, et al. 2011 ). Models on the multiple imputation data sets produce the very similar results as the list wise delet ion models presen ted above. The models using the multiple imputation datasets are presented in the Appendix. 9 2 Table 4 .4 : Theory, hypotheses, and empirical e xpectations , and r esults Theoretical Perspective Concept Variable Hypothesis/Hypothesis Empirical Expectatio n Result Post - materialism Theory Post - materialism Post - materialism H1: Individuals with po st - materialist values consider climate change more serious and dangerous than those with survivalist values + + statistically significant Alternative Approach H2: Individuals with post - materialist values does not consider climate change more serious than those with survivalist values . 0 or - Affluence Hypothesis Individual Wealth Socioeconomic Class Income Decile H3: Rich individuals consider climate change mor e serious than poor individuals . + Not statistically significant estimates for socioeconomic class and income decile Alternative Approach H5: Rich individuals do not consider climate change more serious than poor individuals . 0 or - Affluence Hypothes is Country Wealth/ Development GDP per capita Human Develo p ment Index H4: Individuals in rich and mor e developed countries consider climate change more serious than individuals in poorer and less developed countries. + Not s tatistically significant for both the nat ural logarithm of GDP per capita and HDI Alternative Approach H6: Individuals in rich and more developed countries do not consider climate change more serious than individuals in poorer and less developed countries. 0 or - Supporting the White Thesis Religiosity Attendance of religious services H7: More religious individuals consider climate change less se rious than less religious individuals. _ Not statistically significant Opposing the White Thesis H9 : More religious individuals do not consider climate change less se rious than less religious individuals. + 93 Supporting the White Thesis Protestantism Protestant Identity H8: Protestants consider climate change less serious than non - Protestants . - - ; statistica lly significant in some model specifications Opposing the White Thesis H10: Protestants does not consider climate change less serious than non - Protestants . + Left - right political orientation Self - described Political Ideology Self - described political p artisanship Placement on the left - to - right continuum H11: In dividuals on the left consider climate change more serious than individuals on the right. - for political ideology ; - for right party member; + for left party member - ; statistically significan t for right party +; statistically significant for left party Deficit Model of S cience Communication Education, Political Interest, and environmental Knowledge Level of education Political interest level of education and politica l interest increases, he or she considers climate change more serious . + +; statistically significant for level of education. Education increased seriousness for left and right. Motivated Reasoning Theory /Elite cues Education, political interest, and en vironmental knowledge conditioned on political ideology Level of education Political interest Placement on the left - to - right continuum H13: As an education and political interest increases , the difference between the left and righ t in perceived seriousness of climate change increases. The difference between left and right should increase as education, political inte rest increase. The difference between left and right did not substantively increase as level of education increased. The difference between left and right substantively increased as interest in politics increased. 94 4.7 Discussion and c onclusion This chapter tests explanations for variation, both at the individual and country - level, for climate change risk perception with cross - national survey data. Many of the factors deemed important in previous research, either cross - nationa lly or in the United States, do not help explain climate change risk perception on average across more than three dozen countries. Instead, the explanation of what matters in how people think about climate change globally focuse s on only a few factors. Table 4 .4 summarizes the theoretical expectations, hypotheses, and results from this chapter. The Affluence Hypothesis, which might explain en vi ronmental concern broadly, does not explain climate change risk perception specifically. Richer people are not more likely to say climate change is serious than poorer people. Citizens in high income countries do not consider climate change more serious an d dangerous than citizens in middle - income and low - income countries. Similarly, post - materialism does not strongly influence risk perception. Although it is found to have a statistically significant relationship in the models predicting perc eived seriousne ss , the substantiv e impact of post - materialism is weaker than many individual - level factors. Finally, for mal experience with religion does not explain climate change. Attendance of reli gious services does not predict perceived seriousness of climate change . If these broad and somewhat competing theories of environmental concern do not explain public opinion about climate change, what does matter? Climate change, unlike other environmental issues, has become a political issue in many parts of the world, inc luding in the rich countries such as the United States. Therefore, it is not surprising that left - right political orientation influences perceived seriousness of climate change. Self - described political ideology has a statistically significant and substant ively meaningful impact on perceived seriousness of climate change . In dividuals on the left consider climate change more serious than individuals on the right . And the analysis also shows that being on the right has a stronger influence on climate change r isk perception than being on the left. In addition, there is some empirical evidence in support of the deficit model of science communication. More educated individuals are also more likely to say climate change is very serious than 95 less educated individu als , regardless of their politics. But there i s also evidence in support of motivated reasoning and elite cues, as the effect of self - described interest in politics is conditioned by left - right political orientation. For those on the left, increasing inter est in politics and self - described significantly increases perceived seriousness , but those on the right, interest in politics does not predict perceived seriousness . Still, this analysis leave s important questions with specific focus on the relationship between left - right political orientation and climate change risk perception . First, even though the relationship between left - right political orientation and climate change risk perception is statistically significant globally, the strength of the associa tion could vary across countries. Is left - right political orientation more in shaping climate change opinions in some countries than in others? Second, what are those on the left more likely to think climate change is serious than those of the right? Theor ies are offered in the previous change, but this chapter does not explore these questions empirically. These two questions are tackled in the next two chapters. 96 Chapter 5: Where does left - right p o litical orientation matter across c ountries? 5 .1 Introd uction The previous chapter focuses on showing the average association between individual and country characteristics and climate change risk perception. On average across these countries, left - right poli tical orientation is associated with perceived serio usness of climate change; those on the left consider climate change more serious than those on the right. By only looking at the average effect of these independent variables of interests, the important variation in both the strength and direction of the r elationship between the independent variables of interest and cl imate change risk perception is masked. This chapter shifts the focus from estimating average effects to focusing on explaining variation across countries. Most prominently, t he strength of th e relationship between left - right political orientation and climate change risk perception varies across countries. What explains this variation? In this chapter, I show that a f wealth and development impacts the strength the relationship between political ideology and climate change risk perception. In low and m iddle - income countries, left - right political outlook does not climate change risk perception , but in high - income countries, left - right political outlook has a substantively strong predictive effect . I explain this by arguing that left - right political orientation is more prevalent and closely tied to economics in richer countries. This chapter first explains why wealth and development is connected to the influence of left - right pol itical orientation on climate change and the environment. I emphasize that political ideology should be more closely tied to economic attitudes in rich countries than in poor countries. Second, I show that left - right political orientation is more widely us ed in rich countries than in poor countries. Citizens in highly developed countries are more likely to be able to place themselves on the left - right continuum than citizens in less developed countries. Third, I show the effect of left - right political orien tation is conditional on wealth/development and communist history for climate change risk perception . The left - right divide on climate change and the environment almost exclusively exists in rich and wealthy countries. And in a few very affluent countries, the left and right are deeply divided on climate change 97 5 .2 Why w ealth, d evelopment , and communist e xperience m atter What could explain country - level variation in the influence of left - right political orientation on concern about the environment and clima te change risk perception? Why could left - right political orientation matter in some contexts and not others? I argue in the this section that two important factors country - level economic affluence/development and historical experience with communism condi tion the influence of left - right political orientation on perceived seriousness of climate change . Researchers have investigated why citizens of some countries are more likely to say climate change is serious than citizens of other countries . Scholar s have particularly focused on investigating the influence of country - level affluence and development, but the empirical findings have not been conclusive. Many studies have shown that citizens in richer co untries a re more concerned than citizens in poorer countries about the environment generally ( Diekmann and Franzen 1999 , Kemmelmeier, et al. 2002 , Franzen 2003 , Franzen and Meyer 2010 , Franzen and Vogl 2013 ), while other scholars have shown a negative relatio nship between country - wealth and concern about both the environment and climate change ( Gelissen 2007 , Sandvik 2008 ) and finally others have posited no relationship ( Kvaløy, et al. 2012 ). Chapter 4 finds no relationship using the World Values data. In addition, I found a negative relationship between country - level affluence and perceived dangerousness of climate change using the International Social Survey Programme data. Thes e inconclusive findings led scholars to increasingly focus on the conditional relationship country wealth and development on individual environmental attitudes ( Marquart - Pyatt 2012 , Nawrotzki 2012 , Pampel 2014 ) 1 . This chapter follows this new focus. I argue that these disparate findings show that country - level wealth and development does not have a clear additive effect on concern about the environment and climate change. Instead, the influence of country - level affluence and development is conditional on left - right political orientation. Why? The salience of the influence of left - political orientation on climate 1 For example, Pampel (2014 ) shows that the influence of individual socio - economic status on environmental concern is conditional on country - level wealth/development. In lower income countries, socio - economic status is weakly associated with envir onmental concern, but in high income countries, socio - economic status is strongly associated with environmental concern. 98 change attitudes should vary across contexts. In wealthy and highly de veloped countries in Western Europe and North America, left - right political orientation is very important for how citizens think about politics. Political parties are organized along ideological lines connected to economic issues ( Budge, et al. 2001 ). In other contexts though, citizens may think about politics by examining performance or using more fundamental identities, such as religion or ethnicity. In these countri es, political parties are not necessarily organized along ideological lines, but instead might be organized to promote regional, ethnic, or religio us interests. So the connection between left - right political orientation and opinions on the environment and climate change should be weaker. In addition, left - right political orientation also has different meanings across contexts. Simply put, I expect those on the right in the rich countries to be generally in favor of expanding the free market, while those on the left be in favor of greater limitations on the economic market. In other words, those on the right seek to protect and expand capitalism, while those on the left attempt to restrict and limit capitalism ( Jacques, et al. 2008 , McCright and Dunlap 2011 , Dunlap 2013 ). Environmental quality and mitigating t he consequences of climate change are public goods. Clean air and water are benefits everyone can enjoy, and rising seas and more violent storms caused by climate change are likely to aff ect everyone, albeit some individuals could be more affected than oth ers. Those on the left do not believe the economic market s can protect the environment and mitigate the consequences of climate change, because they believe firms h ave an economic incentive to denigrate the environment and damage the climate. So the left i n these rich countries call s for market intervention by governments. For those on the right, they fervently believe the market can address the problems caused by climate change or reject the possibility of consequences to preserve their belief in capitalis m. So it is unsurprising that those on the left and the right are deeply divided on questions of policy regarding the environment and climate change in rich countries such as the United States . Policy determines the way the government intervenes in the mar ket. For example, in the Unit ed States, there is a deep partisan and ideological divide on climate change mitigation policies, which require government intervention ( Krosnick and MacInnis 2013 ). How does this left - right divide extend to concern about the 99 environment and climate change in rich countries in Western Europe and North America? The public is likely to align their policy positions with overall issue attitudes, so policy preferences on climate change should match with risk perception . Research in political psychology has found that conservatives a re more likely to engage in system justification motivation than liberals, and justifying the capitalist system requires dismissing the threat of climate change ( Jos t, et al. 2003a , Jost, et al. 2003b , Jost, et al. 2008 ). 2 The connection between left - right economic position and concern about the environment and cl imate change is less clear outside of the Western Europe - North America (and Australia and New Zealand) context. Attitudes toward economic issues might not shape what people think about climate change and the environment. In less wealthy countries, environm entalism should not be as closely tied to an anti - business, anti - capitalism position. For example, environmentalism may be connected to protecting agriculture in more rural societies ( Nawrotzki 2012 ). And in po orer countries, people may be more likely to directly experience environmental degradation ( Inglehart 1995 , Brulle and Pellow 2006 ), and this could more influence their attitudes toward the environment. Finally, specifically, in post - communist societies, those who describe themselves as on the left might be pro - business and pro - capitalism, the opposite of th e relationship in Western European and North American countries ( Sabbagh 2005 , Nawrotzki 2012 ). So in these countries, the left should not be more likely to be concerned about the environment or think climate change is more serious than the right . Instead, it could be the self - identified right sees climate change as more serious than the left or there is no relationship in these societies. Research has found a different the relationship between left - right political orientation and environmentalism is different in post - communist countries. 3 I expect these findings about the varying influence of left - right political orientation to extend to climate change risk perception. 2 Systems justification theory is similar to dominant social paradigm theory. Individuals who subscribe to the dominant social paradigm have been found to less supportive of environmental protection than those who do not (Dunlap and Van Liere 1984). 3 Van Hiel and Kossowska ( 2007 ) show a negative relationship between identifying with the right and environmentalism in Ukraine, no relationship in Poland, and a substantively strong relationship in Belg ium. And Nawrotzki ( 2012 ) shows that the that there is a statistically significant difference between the right and the left in 100 5 .3 Hypotheses The previous sectioned argued left - right political orientation has different meanings and level of salience across countries. I argue this shows that the left - right divide is the primary political cleavage in rich countries, as politics is fought over economic and social policy. But in middle and low income countries, both political cleavag es other than the left - right divide are important and the left - right divide has dif ferent meanings . First, if left - right political orientation is more salient in rich countries than in poor countries, then I expect individuals in rich countries to be more likely to identify with left - right political orientation than individuals in poor countries. In contexts in which left - right political orientation does not impact politics, many citizens should not know how to make use of the left - right continuum. H1: Cit izens are more likely to be able to place themselves on the left - right continuum in richer countries than in poorer countries. Second, if left - right political orientation has a different level of salience to politics across contexts, determined by country - level wealth and income, then its influence concern about climate change should vary across countries. If left - right political orientation is less prevalent in poorer countries, then I expect that left - right political orientation should not influence perc eived seriousness of climate change in these contexts. But in richer countries where left - right orientation is more salient to politics, I expect that the left should be more concerned about climate change than the right. H2: Individuals on the left and in dividuals on the right are more divided on climate change risk perception in rich an d developed countries than in poorer and less developed countries. The divide between the left and the right should increase as wealth and development increase, an d the difference between the left and right should only be statistically significant in rich/developed countries. Moreover, left - right political orientation has different meanings across contexts. Specifically left - right political orientation might not cl early connect to the environment and climate change in post - willingness to pay to address environmental problems in countries without a c ommunist history, but the relationship is not statistically different from zero in post - communist countries. 101 communist countries. Therefore, I do not expect left - right political orientation to be related to climate change risk perception. H3: In post - communist countries, the left and the right are not di vided climate change, while in countries without this experience with communism, the left is more likely to say climate change is serious than the right . So I expect the left should not be significantly different from the right on perceived dangerousness o f climate change in post - communist countries, and in countries which were not part of the communist bloc during the Cold War, the left should be significant different from the right. 5 .4 Measurement There are two dependent variables included this chapter . First, to measure use of left - right political orientation, I capture whether a respondent could place himself or herself on the left - right continuum . Respondents were asked on the WVS: would you place your views on this scale, gener coded by Second, climate change risk perception was captured by using t he question from the WVS used in T here are two important independent variables in this chapter. First, I measure left - right political orientation in the WVS by using self - described political ideology. Respondents were asked to place their views on their on a range from one to 10, with one being on the far left and 10 being on the far right (this measure is also used as the dependent variable for H1 ). The second set of important independent variables captures country - level characteristics. I measure country affluence with the natural logarit hm of mean GDP per capita from 2005 to 2009 and whether the country is a member of the Organization of Economic Cooperation and Development (OECD), an association of the wealthiest countries, or not. I also measure n Human Development Index (HDI) score from 2005 to 2009. 102 Finally, I measure whether a country is post - communist, capturing whether the country experienced a transition away from communism after the fall of the Soviet Union or not. The st atistical models a lso include control variables, including age, gender, level of education, and other possible confounding factors, such as interest in politics, depending on the dependent variable and specification. The model specifications at the individual - level are very si milar to those used in Chapter 4 . More information on the measurement of variables is available in the Appendix A . 5.5 Use of l e ft - right political o rientation a cross c ountries I first test H1 , which posits that the ability to use the left - right continuu m increases with increasing country - level wealth and development. Unsurprisingly, the ability of respondents to place themselves on the left - right continuum varies across countries. Figure 4.1 shows the percentage of people in each country who could not pl ace themselves on the left - right continuum in the World Values Survey (WVS). The WVS asked respondents to place themselves on a le ft - right range, so those who do not answer the question are coded as being unable to use the left - right continuum. Many indivi dual s across all countries struggle to place themselves on the left - 4 Figure 5 .1 also clearly shows the cross - national variat ion. In only about one - in - four countries are more than 90 percent of respondents able to place themselves on the left - to - ri ght continuum. In this group is the United States, in which only ir political ideology. But in 24 of the 42 countries, a majority of countries, less tha n 80 percent of respondents are able to place themselves. In seven countries, Ukraine, Georgia, Romania, Ghana, India, Morocco, and Jordan, at least 40 percent of respon dents were unable to place themselves on the left right continuum. And in India, Morocco, and Jordan, a majority of respondents answer - right political orientation in politics is more meaningful to citizens in some countr ies than citizens in other countries; political ideology is more important in some contexts than others. 4 In China and Malaysia, respondents were not asked to identify their ideology on the left - to - right continuum, so these countries are unfortunate ly excluded from much of the analysis. 103 Figure 5.1: Percentage who do not answer the political left - right continuum question, by country (World Values Survey 2005 - 2009 ) This variation in the percentage of those who did not know their political ideology in each the left - right continuum as a function of demographic variables age, gender, ed ucation, and self - described economic class and at the country - level, the natural logarithm of GDP per capita or HDI, measures of country - level wealth and development. Since use of the left - right continuum is a binary variable, I use a logistic functional f orm. The latent variable model is shown in equation 1. (1) The results of the multilevel models are presented in Table 5 .1. The findings at the individual - level are not surprising. Less educated individuals, women, members of the lower and working class (compared the middle class, which is the reference category), individuals younger than 39 and 70 and older (compared to 40 - 49 year ol ds, the reference category) are significantly les s likely to be able to place themselves on the 104 left - right continuum. Those more educated and in the upper class (co mpared to the middle class) are the groups significant more likely to be able to place themselves on the left - right continuum. Table 5.1 : Random - intercept binary logistic models predicting not using the left - right continuum (World Values Survey 2005 - 2009 ) (1) (2) (3) (4 ) Level of education - 0.1713 - 0.1711 - 0.1711 - 0.1711 (0.0143) (0.0143) (0.0143) (0.0143) Female 0.4345 0.4346 0 .4346 0.4346 (0.0397) (0.0397) (0.0397) (0.0397) Lower class 0.2798 0.2794 0.2795 0.2794 (0.0635) (0.0635) (0.0635) (0.0635) Working class 0.1102 0.1103 0.1104 0.1102 (0.0481) (0.0481) (0.0481) (0.0481) Upper class - 0.2366 - 0.2365 - 0.2366 - 0.2366 (0.0493) (0.0493) (0.0493) (0.0493) Age younger than 30 0.1688 0.1683 0.1683 0.1684 (0.0437) (0.0438) (0.0438) (0.0437 ) Age 30 - 39 0.1069 0.1067 0.1067 0.1067 (0.0366) (0.0366) (0.0366) (0.0366) Age 50 - 59 - 0.0501 - 0.0499 - 0.0499 - 0.0499 (0.0439) (0.0439) (0.0439) (0.0439) Age 60 - 69 0.0791 0.0797 0.0796 0.0795 (0.0631) (0.0631) (0.0631) (0.0631) Age 70 and older 0.1765 0.1772 0.1772 0.1770 (0.0758) (0.0758) (0.0758) (0.0758) Country - level Natural logarithm of mean GDP per capita - 0.229 7 (0.1084) Mean HDI - 1.7251 (0.9454) OECD country - 0.7536 (0.4392) Constant - 1.8892 0.1222 - 0.6515 - 1.5942 (0.2335) (0.9882) (0.6810) (0.2974 ) Variance components Country - level standard deviation 2.0376 1.8892 1.9353 1.885 4 (0.7845) (0.7757) (0.7629) (0.7891 ) N umber of respondents 57574 57574 57574 57574 Number of countries 41 41 41 41 A IC 51566 51565 51566 51565 Note s : Cell entries are ordered logistic regression coefficients with standard error in parentheses with l istwise deletion . Bold indicates two - tailed p <0.05. China, Malaysia, and Mexico not included because of omitted questions in these countries. The results of most interest are the country - level measures of wealth and development. The country - level varianc e component is statistically significant in Model 1, which does not include any country - level covariates, showing that there is meaningful variation between countries in the ability to use the left - right continuum, not surprising given the amount of variat ion shown in Figure 4.1. In Models 2 105 and 3, which include either natural logarithm of GDP per capita or HDI as a country - level covariate, the country - level variance decreases. The coefficient estimate for the natural logarithm of GDP per capita is negative and statistically significant (two - tailed p - value = 0.032). Specifically, an average individual in a country with a GDP per capita of about $1100 had a predicted probability of 24 percent to be unable to use the left - right scale; for the same profile in a country with a GDP per capita of about $36000 the predicted probability was 11 percent. 5 So individuals a re significantly more likely to be able to place themselves on the left - right continuum in rich countries than in poor countries, and this difference is substantively meaningful. Even though the estimate for HDI is not statistically significant at the conventional 0.05 alpha, the two - tailed p - value of 0.068 suggests that this coefficient estimate is highly unlikely if HDI had no effect. 6 And finally, i n dividuals in OECD countries are about 10 percent less likely to not be able to use the left - right continuum than individuals in non - OECD countries, although this difference is again not statistically significant at the conventional level (95 percent confi dence interval: [ - 0.230, 0.016]). Still, this offers supportive evidence in support of H1 . Citizens in wealthier are more likely to use the left - right continuum. The findings from this section suggest left - right political orientation is more used and salie nt in rich countries than poor countries. 5.6 The varying influence of left - right political orientation on climate change risk p erception The previous section shows that the use of left - right continuum varies across countries according to level of wealth and development. I argue previously that this finding should influence the use of left - right political orientation on climate change attitudes. This section shows that influence on left - right political orientation on climate change risk perception is condi tional on country - level wealth and development. In poorer countries, left - right political orientation does not predict perceived seriousness of climate change , but in the richest countries, the divide between the left and the right is significant and in 5 The profile used to calculate t he predicted probabilities is a mean educated 40 - 49 year old female in the middle class. 6 The null hypothesis of no effect can be rejected in a one tailed test (0.034<0.05). HDI has a modest substantive effect on the use of the left - right range. At HDI 0.65, an average individual had a predicted probability of not being able to use the left - right continuum of about 20 percent; at HDI of 0.9, an average individual has a predicted prob ability of not being able to use the left - right scale of 13 percent. The same profile is used here as above. 106 so me countries, substantially important. And left - right political orientation does not predict perceived seriousness of climate change in post - communist countries, but left - right political orientation does have a significant association in countries that did not have this experience. Here, I test H2 and H3, which hypothesize that left - right political orientation only influences concern about climate change in rich countries and countries which are not post - communist. First, I informally test the hypothese s by running regression models predicting saying climate change is a very serious problem, controlling for oth er possibly explanatory factors and demographics such as age, education, and gender , for each country . 7 Out of 41 countries, in 16 countries, a mi nority, left - right political orientation is , as expected , a statistically significant and negative predictor of climate change risk perception. In these c ountries, an individual who places himself or herself on the left is more likely to think climate chan ge is very serious than an individual who places himself or herself on the right. And the overwhelming majority of these countries are wealthy and highly developed. The estimated difference between individuals on the left and those on the right is statisti cally significant for the nine richest , as measured by mean GDP per capita, and 10 most developed , as measured by mean HDI, countries included in the WVS. These differences are substantially important t oo. Individuals on the left are about 10 percent more likely than those on the right to consider climate change very serious in Germany, Norway, and Sweden, 13 percent more likely in Canada, 20 percent more likely in Spain, 25 percent more likely in Switzerland, and 48 percent more likely in the United States . In many North American and Western European countries, the public is meaningfully divided on climate change. Among poorer and less wealthy countries, general left - right political orientation does not predict climate change risk perception in most countri es. Climate change is not a left - right political issue in these countries. The lo ess line of best fit in Figure 5.2 illustrates this nicely. The line is flat and hugs zero for most values of the natural logarithm of mean GDP per capita and mean HDI. This s hows that for low and middle income countries, there is no predicted difference between individuals on the left and on the right 7 Coefficient estimates from multilevel models could be misleading, as the estimate could the result of a few highly influential countries. Individ ual country models check to make sure the results is consistent with the hypotheses. 107 in the likelihood of saying climate change is very serious . Only at the highest values of the natural logarithm of GDP per capi ta and HDI does the loess line increase rapidly, capturing the strong and substantially meaningfully difference between the left and the right among the richest and most developed countries. This informal analysis of the ISSP and WVS provides strong eviden ce in support of H2 ; general left - right political orientation only influences concern about climate change in affluent and developed countries. Figure 5 .2a: Effect of left - right political orientation on predicting perceived seriousness of climat e change across values of natural logarithm of mean GDP per capita , by country (W orld V alues S urvey 2005 - 2009 ) 108 Fig ure 5 .2 b: Effect of left - right political ideology on predicting perceived seriousness of climate change across values of mean HDI score , by coun try (W orld V alues S urvey 2005 - 2009 ) Notes: The blue line is the OLS line of best fit and the blue region is its 95 percent confidence interval. The red line is the loess line of best fit. And the vertical error bar is the 95 percent confidence interval f or the estimated difference for each country. I now turn to the multilevel models focus ing on the interaction terms to test the hypothesis that left - right political orientation only influences climate change risk perception in wealthy and highly develop ed countries. The results are shown in Figure 5 .3 visually and Table 5 .2 . The difference between the left and the right in the predicted probability of considering climate change a very serious problem fo r the world as a whole increases as country - level we alth and development increases . Specifically, country - level wealth and development does not influence climate change risk perception among those on the right. Controlling for competing explanations , an individual on the right has a predicted probability of about 60 percent of considering climate change very serious, regardless where she lives (wealthy or less wealthy country). An individual with the same profile expect with political ideology on the left in a country with a GDP per cap ita of about $1100 als o has a predicted probability of 58 percent of viewing climate chan ge as very serious. So there is no t a significant difference between the left and right on perceived seriousness of climate change in poor countries. But an individual on the left in a coun try wit h a GDP per capita of 109 $36000 has a predicted probability of answering th at climate change i s very serious of 71 percent. In very rich countries, an individual on the left is 13 percent more likely to consider climate change very serious than an indi vidual on the right, controlling for other possible factors and demographics, a statistically significant and substantively important difference. In another way of showing that the relationship between left - right political orientation and climate change ri sk perception only exists in rich countries, in non - OECD countries, the left is not significantly more likely to consider climate change very serious than the right . But in OECD countries, an individual on the left was 13 percent more likely than an indivi dual on the right with the same profile to consider climate change very serious, a statistically significant difference. These findings show that a substantively important divide between the left and right only exists in rich countries. In addition, left - r ight political orientation seems to only influence climate change risk perception in countries without a communist history. In post - communist countrie s, an individual on the right i s actually more two percent more likely to consider climate change very ser ious than an individual on the left with the same profile, but this estimate i s not statistically different from zero. But in countries without such experience with communism, which included a diverse sample of rich and poor countrie s, a citizen on the lef t i s seven percent more likely to consider climate change very serious, a statistically significant difference , than an individual on the right . 8 The analysis from the WVS supports H3 and H5 . The effect of left - right political orientation on perceived seri ousness of climate change is not constant across countries. Instead, it is conditioned by development and historical experience. 8 Considering the large effect political ideology had in the United States in the WVS analysis, as a robustness check, I tested the hypotheses on a sample of 40 countries , excluding the United States. The results are virtually identical; the conditional effect on political ideology of the GDP per capita, HDI, OECD, and post - communist is just slightly weaker. For example, when GDP per capita is set $36,000, the predicted di fference between left (ideology=3) and right (ideology=8) in considering climate change very serious is 10 percent instead of 11 percent in the analysis above. 110 Table 5 .2 : Random - intercept, random - coefficient ordered logistic models predicting perceived seriousness of climate change ( World Values Survey 2005 - 2009) (1) (2) (3 ) (4 ) (5 ) (6 ) Political ideology - 0.0473 - 0.0518 0.2604 0.1315 - 0.0088 - 0.0651 (from left to right) (0.0126 ) (0.0159) (0.0852 ) (0.0611 ) (0.0149 ) (0.0181 ) Post - materialism index 0.0660 0.0539 0.0 531 0.0534 0.0531 0.0 538 (0.0221 ) (0.0221) (0.0220 ) (0. 0220 ) (0.0220 ) (0.0221 ) Interest in politics 0.0628 0.0582 0.0577 0.0578 0.0580 0.0583 (0.0200) 0.0193 (0.0193 ) (0.0193 ) (0.0194 ) (0.0193 ) Level of education 0.0928 0.0932 0.0929 0.0929 0.0930 0. 0932 (0.0140 ) (0.0142) (0.0142) (0.0142 ) (0.0142 ) (0.01 42 ) Attendance of religious services - 0.0 240 - 0.0194 - 0.0191 - 0.0191 - 0.0192 - 0.0195 (0.0142 ) (0.0136) (0.0136 ) (0.0136 ) (0.0136 ) (0.0137 ) Importance of God 0.0461 0.0495 0.0499 0.0498 0.0499 0.0 494 (0.0085 ) (0.0088) (0.0088 ) (0.0088 ) (0.0088 ) (0.008 8 ) Protestant - 0.1 443 - 0. 129 3 - 0.1293 - 0.1289 - 0.1291 - 0.1294 (0.0747 ) (0.0691) (0.0692 ) (0.0692 ) (0.0692 ) (0.0692 ) Female 0.0698 0.0623 0.0617 0.0619 0.0618 0.0623 (0.0358 ) (0.0348) (0.0348 ) ( 0.0348 ) (0.0348 ) (0.0348 ) Age younger than 30 - 0.1180 - 0.1137 - 0.1132 - 0.1133 - 0.1137 - 0.1140 (0.0461 ) (0.0477) (0.0479 ) (0.0479 ) (0.0477 ) (0.0477 ) Age 30 - 39 - 0.0089 - 0.0020 - 0.0014 - 0.0016 - 0.0018 - 0.0021 (0.0323 ) (0.0332) (0.0325) (0.0325 ) (0.0324 ) (0.0323 ) Age 50 - 59 0.0019 0.0182 0.0177 0.0176 0.0177 0 .0186 (0.0332 ) (0.0332) (0.0332 ) (0.0332 ) (0.0333 ) (0.0333 ) Age 60 - 69 0.0131 0.0191 0.0191 0.0188 0.0191 0.0197 (0.0465 ) (0.0472) (0.0470 ) (0.0469 ) (0.0470 ) (0.0472 ) Age 70 and older - 0.0512 - 0 .0325 - 0.0323 - 0.0328 - 0.0325 - 0.0319 (0.0479 ) (0.0467) (0.0464 ) (0.0464 ) (0.0253 ) (0.0477 ) Country level variables Natural log of mean GDP per capita 0.2758 (0.0799 ) Mean HDI 2.0327 (0.8421 ) O ECD country 0.8074 (0.2530 ) Post - communist country - 0.5273 (0.2340 ) Interaction terms Political ideology X Natural log of GDP - 0.0356 (0.0098 ) Political ideology X Mean HDI - 0.2549 (0.0855 ) Political ideology X OECD coun try - 0.1039 (0.0296 ) Political ideology X Post - communist 0.0799 (0.0208 ) Not serious at all | Not very serious - 4.3304 - 4.3306 - 1.9116 - 2.8664 - 3.9952 - 4.4195 (0.2114) (0.2158) (0.7445 ) (0.6698 ) (0.23594 ) (0.2292 ) 111 Table 5.2 ( Not very serious | somewhat serious - 2.5558 - 2.5469 - 0.1281 - 1.0829 - 2.2117 - 2.6359 (0.1579) (0.1713) (0.7534 ) (0.6755 ) (0.2029 ) (0.1909 ) Somewhat serious | Very serious - 0.6118 - 0.5894 1.8293 0.8745 - 0.2543 - 0.6784 (0.1236) (0.1466) (0.7544 ) (0.6709 ) (0.1889 ) (0.1662 ) Country - level variance 0.4075 0.7879 0.5892 0.6519 0.6148 0.7468 (0.0997 ) (0.2265) (0.1737 ) (0.2027 ) (0.1774 ) (0.2295 ) Political ideology variance 0.0092 0.0060 0.0071 0.0065 0.0083 (0.0044) (0.0028) (0.0033) (0.0032) (0. 0043) N umber of respondents 39973 39973 39973 39973 39973 39973 Number of countries 41 41 41 41 41 41 AIC 71554 71331 71318 71325 71321 71331 Notes : Cell entries are ordered logist ic regression coefficients with standard error clustered by country in parentheses with list wise deletion . Bold indicates two - tailed p <0.05. China, Malaysia, and Mexico are not included because of omitted questions in these countries. 112 Figure 5 .3: Conditional effect of left - right political ideology on predicting considering climate change very serious across values of GDP per capita, HDI, OECD, and post - communism (World Values Survey 2005 - 2009) Notes: The black line/points represent the estimated difference between left and right in the predicted probability of considering climate change very serious, and grey area/error bars represent the 95 percent confidence interval of the estimate . The figures w ere produced using models presented in Appendix D . Specifically, left is defined as political ideology=3 and right is defined as political ideology=8. Post - materialism index, interest in politics, level of education, attendance of religious services, and importance of God at set to their overall means, and every other 113 variable in the model is set to zero exc serious at all; not very serious; somewhat serious; very serious. 5 .7 Discussion and c onclusion This chapter builds on the literature on comparative public opinion on climate change with focus on how the influence of individual characte ristics vary by country factors ( Gelissen 2007 , Sandvik 2008 , Franzen and Meyer 2010 , Bre chin and Bhandari 2011 , Kvaløy, et al. 2012 , Nawrotzki 2012 , Franzen and Vogl 2013 ). The pr evious chapter shows that left - right political orientation predicts perceived seriousness of climate change on average across countries by this finding masked important v ariation in both the size of this influence across countries. Controlling for competin g explanations, I show in this chapter that the difference between the left and the right on the perceived s eriousness of climate change is positive and statistically significant in only a minority of countries across cross - national surveys. But in some co untries, such as Cana da and Switzerland, the left is substantively and significantly more inclined to see climate change as a serious problem than the right. What explains this variation in the influence of left - right political orientation across countrie s? Using cross - level interaction models, I find that country - level wealth/development and historical experience with communism are two important contextual factors. I show that in poor and less developed countries, the difference in conce rn between the lef t and right i s not statistically different from zero, but in very rich countries, the left i s significantly likely to consider climate change serious than the right . And I also show that in post - communist countrie s, the difference in perceived seriousness is not statistically significant, while in the diverse group of countries without historical experie nce with communism, the left is significantly more inclined to say climate change is serious than the right. These findings illustrate that in the very rich countries of Western Europe and North America, the lef t and right seem to be divided on climate change. For policymakers and activists who support national and international policy to mitigate the consequences of climate change and protect the environmen t, this chapter offers both a positive and a 114 negative. The good news is that in many less wealthy and post - communist countries, climate change is not close ly tied to how individuals perceive their political ideology . In many of these countries, citizens ge nerally expr ess that climate change is serious . But it is open to debate just how deep concern about these issues is in these societies. Climate change might not be a front - burning issue in this context. The bad news is that in many of the very rich count ries, there is a clear political divide on climate change. T he right in these countries is far less likely to consider climate change serious than the left. In many Western European and North American countries, the environment and climate change has becom e politically politicized. These rich countries not only need to make the biggest policy and behavioral chan ges, considering they are responsible for an important amount of global carbon emissions , but they are also necessary leaders for an international climate agreement. Yet in these rich countries, the publics seem far from unified in support for such policy action, and the lack of concern on the right is a major obstacle. This chapter introduced another puzzle though. The divide between the left and right i s great in many wealthy cou ntries, but as shown in Figure 5 .2, this divide i s far greater in the United States. The United States stands out as an outlier; the public is far more polarized on climate change in t he United States. Why? Chapter 6 tackl es this question. 115 Chapter 6 : Why is the American public far more polarized on climate c hange than other wealthy c ountries? 5.1 Introduction In early January 2014, in the midst of extremely cold weather throughout the United States, Senator James I nhofe, a conservative from the state of Oklahoma, took to the Senate floor to give a published The Greatest Hoax: How the Global Warming Conspiracy Threatens Your Future. And In 2010, after the Climategate scandal, he called for criminal investigations of climate scientist (Nature 2010). And in 2003, as the chair of the U.S. Senate Commit tee on the Environment and Public Works, he asked if climate change might be 1 Inhofe, however, is hardly alone among conservative elites in Republican Party in the United Sta tes to deny climate change. During the 2012 Republican presidential p rimary campaign, Rick Santorum travesty of scientific research that was motivated by those who, in my opinion, saw this as an opportunity to create a panic and a crisis for government to be able to step in and even more greatly control your Hooper 2012 ). Out of the entire Republican primary field in 2012, only Jon Huntsman, the moderate, accepted climate change, and he was never a competitive candidate. 2 Politicians are not the only members of the conservative elite to speak out against global warming in the United States. The conservative media is also full of influential individuals who are climate change deniers. In April 2013, Rush Limbaugh, the most influential conservative radi o host in the United States, 1 Inhofe was appointed as Chair of Environment and Public Works Committee in the United Stat es Senate again in January 2015. 2 Lawrence 2011 ). 116 Robbins 2014 ). And Fo x News and the Wall Street Journal , both owned by Rupert Murdoch and News Corp, have become prominent outlets for climate change deniers to attack climate science to create uncertainty ( Robbins 2014 ). - called climate change debate in the United States ( Begley 2007 ). How much does this matter to mass public opinion though? How does the conservative and Republican political and media e lite affect mass public opinion on climate change in the United States? The previous chapter in this dissertation showed that individuals on the left are more worried about the environment in general and climate change specifically than individuals on the r ight in rich countries. And previous research on public opinion in the United States has shown that liberals and Democrats are more concerned about environmental problems ( Dunlap, et al. 2001 , Hamilton 2011 ; McCright 2011 , McCri ght and Dunlap 2011 ) and climate change ( McCright and Dunlap 2011 , Guber 2013 ) than conservatives and Republicans. Is the left - right divide on the environment and climate change in the United States just typical of rich and developed countries? Or is it exceptional? This chapter shows that both the elite and mass - level left - right d ivide are far greater in the United States than the elite and massive divides are in other wealthy countries polarized on c limate change in the United States than they are in other countries or even on other environmental problems. Second, the consequences of this elite polarization for mass public opinion in the United States are explored, motivating a series of hypotheses. T hird, I test these hypotheses by situating the relationship between left - right political orientation and climate change risk perception in the United States in the comparative context. The analysis shows that the ideological polarization of public opinion on climate change in the United States is truly unique. Left - right political orientation has a much stronger effect on the perceived dangerousness of climate change in the United States than its effect in other countries. In addition, in the United S tates, among liberals, increasing political knowledge an d interest in politics increases perceived seriousness of climate change , bu t among conservatives, increasing political knowledge an d interest in politics decreases perceived seriousness . Americans not inte rested in 117 politics are not divided on climate change, but th e most interested Americans are extremely divided. In no wealthy country does knowledge and interest have this conditional effect on left - right political orientation. Political ideology impacts pu blic opinion in the United States unlike anywhere else in the Americans seem to be turning to ideologically similar elites on climate change. 6 .2 The e lites: United States and the rest of the w orld In 2010, then British Foreign Secretary William Hague declared in a speech in New York City, st underpins our security and prosper Brownstein 2010 ). Hague is a prominent leader in the British conservative party. A similar comment from a member of the conservative political elite in the United blicans in this country are coalescing around a uniquely dismissive position on climate change. The GOP is stampeding toward an absolutist rejection of climate science that appears unmatched among major political parties around the globe, even conservative the Pew Center on Global Climate Change (now the Center for Climate and Energy Solutions) told no party - wide view like this anywhere in the wo rld that I am where denialism is part of the climate change debate. In the U.S., political elites are vastly more polarized on climate change than in any oth er country in the world, including other rich countries. Although there is no clear quantitative evidence to support this proposition, this section support s this argument through thorough deep anecdotal evidence from previous research and media reports. T Begley 2007 ) in the United States made up of a variety of different groups, including fossil fuel corporations, conservativ e think tanks, politicians, media, and contrarian scientists ( Dunlap and McCright 2011 ). Social scientists ( McCright and Dunlap 2003 , Dunlap and McCright 2011 , Dunlap 2013 , Dunlap and Jacques 2013 , Elsasser and Dunlap 2013 ) and journalists ( Mooney 2005 , Begley 2007 , Klein 2011 ) h elp us better understand how the denial machine works, 118 emphasizing its extremely close connection with conservative and Republican political elites. The fossil fuel industry, including prominent groups such as Exxon Mobil and American Petroleum Institute, corporations represented by the U.S. Chamber of Commerce, and conservative foundations, like the multitude of groups run by the Koch family, give money to conservative think tanks to attack climate scie nce. Conservative think tanks, such as the Heritage Foundation and Heartland Institute, then support this work by hiring contrarian scientists to lead studies, which invariably question climate science ( Dunlap and McCright 2011 ). Analysis shows that studies with questio n or refute climate science are almost exclusively funded by conservative think tanks ( Dunlap and McCright 2011 ). Conservative think tanks also publicize their work in the media and distribute it in conservative media, thus making it accessible to the public. Contrarian scientists use the complexities of climate science as a way of creating dou bt that climate change is both the result of human activity and harmful for the environment . The strategy employed by climate change deniers is similar to the one used in the middle of the twentieth century after scientists discovered the major health risk s posed by tobacco: generate uncertainty ( Oreskes and Conway 2010 ). To create unc ertainty, climate change skeptics use contrarian scientists to attack scientists and their work ( McCright and Dunlap 2003 , Dunlap and McCright 2011 ). For example, the denial machine used basis of emails, to manufacture uncertainty. The goal of this work is to convince the policymakers and the public that a scientific consensus on climate change does not exist. Actors who are p art of the denial machine share a strong commitment to conservative political ideology ( Dunlap and McCright 2011 ). Dunlap and McCright (2011) argue conservative elites see merely acknowle dging climate change as real as a major threat to capitalism. Any policy response to climate change would require increasing government regulations. Denial machine actors came together to not as an ecological prob lem but as a problem for unbridled Du nlap and McCright 2011 ). 119 How could the conservative elite denial machine reach mass public opinion? To reach conservatives in the public, the denial machine relies on conservative media and politicians to spread the message. In addition to Rush Limbaug h, whose radio show reaches millions of Americans each day, the denial machine relies on Fox News and the Wall Street Journal , both owned by Rupert Murdoch. 3 The Wall Street Journal opinion and editorial pages have become forums for contrarian scientists t o attack climate change and create doubt. These op - eds and editorials are not used to present new scientific research, but instead to attack research done by mainstream climate scientists. For example, immediately Intergovernmental Report on Climate Change fifth assessment report in late 2013, Matt Ridley, who does not have any background in climate science, wrote an op - ed in the Wall Street Journal ver the next He argues that a warming of two degrees Celsius would results in positive economic and environmental externalities. Climate scientists di sagree with him and accuse him of misreading their work. Ridley might not be a household name, but George Will and Charles Krauthammer, two prominent conservative columnists, are well - known because their syndicated columns are widely circulated in newspa pers around the county. And both are committed climate change deniers. Both have compared climate science to a religion, claiming that climate change is a progressive plot with the aim of restricting the freedoms of Americans ( Dickinson 2010 ). In 2010, in the midst of the supposed Climategate scandal, assertions that everything, including a historic blizzard, supposedly confirms and nothing, not even the 2010 ). In news stories about climate change, the viewpoints of the denial machine are often included in an attempt to create su pposed balance ( Boykoff and Boykoff 2004 , Boykoff 2011 , 2013 ). The denial 3 Phil Plait, a journalist for Slate in the Daily Mail , the Wall Street Journal , the Financial Post , and/or anything owned by Rupert Murdoch and News Corp? If 120 machine broadly and contrarian scientists specifically are vastly overrepresented in conservative media outlets. For example, in reaction to th e recent IPCC, one report showed that 50 percent of those quoted in the Wall Street Journal Greenberg, et al. 2013 ). This compared with 29 percent in the Los Angeles Times , 17 percent in the Washington Post , and 12 percent in Bloomberg News. The New York Times , Associated Press, and Reuters did not quote any climate change deniers over this period. There is no media outlet, however, that plays a more prominent role in the denial machine than Fox News . Content analyses have shown that Fox News covers issues and events, like the Iraq War and 2008 presidential election in a way that is far more supportive of conservative viewpoints than other news networks, such as CNN and MS NBC ( Aday, et al. 2005 , "The Color of News: How Different Media Have Covered the General Election" 2008 ). On climate change specifically, Fox News presdents news coverage that emphasizes the lack of consensus and uncertainty on climate change, while CNN ha s presented coverage tending to emphasize that climate change i s real and happening ( Feldman, et al. 2012 ). The anecdotal evidence of climate change denial is str ong . For example, one Fox News cooling caused not by greenhouse gases but by fluctuations in solar energy Gertz 2013 ). And prominent Fox News political c ommentators follow . In 2013, Sean Hannity proclaimed that the Robbins 2014 ). In the study discussed above, 69 percent of guests on Fox News no denial coverage on CNN. Of the Fox News guests, 73 percent have no background in climate science ( Greenberg, et al. 2013 ). Conservative and Republican politicians, sometimes by joining Fox News as guest s or being quoted in the conservative media, also spread the message of uncertainty about climate change. Inhofe and Santorum may be extreme examples, but most conservative politicians in Washington are climate deniers. The liberal think tank Center for Am erican Progress tracked climate change denialism among members of the 113 th U.S. Congress using public statements ( Germa in, et al. 2013 ). They found that 58 percent of the Republican caucus rejected climate science. In 2011, the National Journal attempted to 121 survey every Republican senator and representative in Congress. Of the 65 who responded on the record, 19 said tha change is a not causing the planet to warm, and five responded that climate change was not the result of human activity ( Dave nport 2011 ). In January 2015, only five out of 54 Republicans voted yes in a sense of the US Senate that climate change is real and human activity contributes significantly to it, while every Democrat voted yes ( Davenport 2015 ). 4 For Republican senators and representatives, denying climate change is very common. 5 Yet climate change denialism seems like an even more dominant position among the Republican and conservative elite because the climate change deniers are the most eager to talk about it, while those not in the denial camp are silent. In the 2011 National Journal survey of Republican senators and conve Davenport 2011 ). John Boehner, the Speaker of the taking a clear position on climate change could crack the caucus in two and Davenport 2011 ). 6 The climate change deniers are extremely vocal in their at tacks on climate science. In the same article, Inhofe sought out reporters multiple times to make sure his positions on climate change were noted. When Darrell Issa, a powerful conservative representative, was asked if he thought climate change was causing 4 In a different vote, 15 out of 54 Republicans voted yes that climate change is real and human activity contributes (Davenport 2015) . 5 In a survey of members of Congress in 2006 by the National Journal, called the Washington Insiders polls, 95 percent of Democrats agreed that man - made climate change is a proven fact, while only 23 percent of Republicans agreed. In 2007, the gap between Republicans and Democrats widened, with 98 percent of Democrats saying that warming temperatures are the result of human activity compared to only 13 percent of Republicans. This was another survey of senators and representatives in which most Republicans in Congress did not participate. 6 The fossil fuel industry and business interests have long provided crucial campaign funding to Republi can Party politicians in the United States. Between 1998 and 2010, the fossil fuel industry gave more than $200 million to Republicans, about 75 percent of its total contributions (Davenport 2011). And in the 2010 elections, business interests known to que stion climate science, such as the U.S. Chamber of Commerce, spent $105 million. Conservative politicians have clear fi nancial interests to deny and doubt climate change . 122 ( Davenport 2011 ). Th is constant noise from the conservative media and conservative politicians makes the elite conservative position clear on climate change in the United States: doubt or denial of the science, no or little concern about possible consequences, and a rej ection of mitigation policies . Conservative and liberal political elites are clearly polarized in the United States, and this sends a clear signal to the mass public. Are elites as divided in other countries on climate change by ideology and party? On this ques tion, there is little research, but the answer seems to be no. The climate change denial machine is still uniquely American. Still, U.S . conservatives are spreading their denial machine. to other countries, and there is evidence of some possible elite pola rization in other countries. According to Dunlap and McCright ( 2011 ), climate change denialism is strongest in the United Kingdom, Canada, and Australia, all countries that have strong free - market conservative parties and active conservative think tanks. These are also Anglophone countries, so the denial message from the U.S. does not have to overcome the language barrier. In Canada and Australia, the fossil fuel industry has also helped promote climate change denialism. As for the rest of the world, conservative think tanks are taking advantage of their internat ional network to spread their message. Still, climate change denialism is a very recent phenomenon outside the United States, so it is unclear how divided elites are outside the United States. 7 Beyond the U.S., U K, Canada, and Australia, the denial machine is extremely weak or does not exist, so political elites may not be divided on climate change. Recent research also shows that climate change denialism is more prominent in the U.S. media than in other similar countries. Boykoff ( 2011 ) compares the United States to the United Kingdom, findi ng that climate change doubt and denial is far more common in the United States. Painter and Ashe ( 2012 ) i n a media analysis of Brazil, China, France, India, the UK, and U.S. , finds that climate change 7 McCright and Dunlap (2011: 156) warn of the consequences of the international effor ts of conservative think future of climate science and especially for effective international action and policy making to deal with the rea l ity be interesting to track the effects of this effort on elite and mass public opinion on climate change around the world in the future . 123 denialism is limited to the UK and United States. Challenges in the media to a pol icy response to climate change are also limited UK and United States. This sec tion shows that climate change denialism is far more prevalent in the United States than in other countries. This denial machine has helped create elite polarization in the United States . The conservative media and politicians push a clear message, contrar y to the actual scientific research, that climate science is uncertain. In some other wealthy countries, such as the U nited K ingdom , Canada, and Australia, there evidence of a nascent climate change denial machine. In the rest of the world, this climate ch ange denial machine does not exist. This section has established there is cross - national variation in e lite opinion on climate change. The next section introduces hypotheses for the consequences of this variation in elite polarization on mass public opinio n across countries . 6 .3 Hypotheses The previous section argues that political and media elites are far more divided in the United States than in similar countries . If elites influence public opinion on climate change, the n there should be substantial diffe rence between the United States and these other countries in the effect of left - right political orientation on climate change attitudes. So I should expect mass public opinion on climate change to be uniquely polarized in the United States, reflecting elit e ideological polarization. And in other rich countries, left - right political orientation should have a much weaker effect on climate change opinions. H1: Left - right political orientation is more strongly related to climate change risk perception in the Un ited States than in other rich countries. It is also possible that in a few other countries, such as the UK, Australia, and Canada, with nascent denial machines on t he right, political ideology has a substantively meaningful impact on climate change public opinion. The association in these countries should be weaker than in the United States though, because the elite polarization is not nearly as strong in these countries. Second, as discussed above, even when elite polarization exists on an issue, not all citizens pick up on and use the cue. Only politically engaged citizens are aware of the elite divisions, so only these 124 citizens can use these elite divisions as an information shortcut to form their opinions on climate change. I n the presence of elite pol itical polarization on an issue, political knowledge and awareness should have a divergent impact on the left and right. Elites are polarized on climate change in the United States. So in this specific context, among liberals, political awarenes s and knowl edge should concern and probability of accepting climate science, while for those on the right, political awareness and interest should decrease concern and the probability of accepting climate science. This interactive hypothesis can be stated as follows: H2: In the United States, for liberals, political engagement is positively related to concern about climate change and the probability of agreeing with climate science, while for conservatives, political engagement is negatively related to concern about climate change and the probability of agreeing with climate science. Third, this conditional effect of political engagement on left - right political orientation should be specific to public opinion in the United States on climate change. I n other wealthy co untries, I do not expect this conditional relationship because political elites in these rich countries are not divided on climate change , so politically engaged citizens of these countries cannot use elite polarization as a cue . H3: The conditional relat ionship between political engagement and left - right political orientation, in which interest in politics is positively related to perceived seriousness of climate change for the left and interest in politics is negatively related to perceived seriousness f or the right, is unique to the United States. In other rich countries, this conditional relationship between political engagement and left - right political orientation on perceived seriousness of climate change does not exist. These three hypotheses, when t aken together, all test the theory that mass polarization reflecting elite polarization is unique to the United States. By drawing on numerous public opinion surveys, including just the United States and many additional countries it is possible to provide evidence to support or refute the argument that many Americans use elite opinion as a cue on climate change. 125 6.4 Data and m ethodology The hypotheses propose a distinctive explanation for public opinion about climate change in the United States. The factor s that explain public opinion in the United States on climate change are different from other countries and other environmental issues. In order to test these hypotheses, it is necessary to compare the United States to other countries while controlling for as many competing explanations as possible. The previous chapter showed that left - right political orientation mattered in mostly rich countries. Therefore, I compare the United States to other rich countries, so this chapter focuses on only OECD countries . By comparing the United States to other OECD countries, many contextual factors, like wealth, development, and level of democracy, can be held relatively constant. An important contextual difference across countries is the level of elite polarization on climate change. Although this important factor cannot be precisely measured, the United States has much stronger and clearer elite polarization on climate change than other OECD countries. I do not rely on multilevel modeling here. Work on cross - nationa l public opinion, in which the hypotheses proposed additive relationships across countries (random intercept models) or conditional relationships across countries (cross - level interaction models) use multilevel modeling. Here the factor elite opinion that shapes public opinion in the United States is different from other countries. Multilevel modeling borrows information across groups, yet to test these hypotheses, I do not want to borrow information, as the contexts are different. So to avoid borrowing inf ormation, I must produce estimates of relationships each country by estimating a statistical model for each country. I can then compare these effects across OECD countries to see how similar or different the effect for the United States is compared to othe r countries. This chapter draws from the American National Elections Studies (ANES) from 2008 and 2012 to test H2 and the World Values Survey to test H1 and H3 . 6.5 The effect of left - right political orientation on concern about climate change risk percep tion Liberals and conservatives are polarized on climate change in the United States, but does a similar level of polarization exist in other wealthy countries? I argue in H1 that left - right political orientation has a much stronger effect in the United S tates than other rich countries. To explore this 126 question and test H1 , I estimated ordered logistic models predicting climate change risk perception across 17 OECD countries while controlling for other competing explanations and demographic factors using t he WVS. The results show that left - right political orientation has a much stronger influence on c limate change risk perception in the United States than in other rich countries. Americans are far more ideologically polarized than the publics in other OECD countries. The relationship between left - right political orientation and climate change risk perception is much stronger in the United States than in o ther wealthy countries. Figure 6 .1 shows the predicted difference between an individual on the left (poli tical ideology=3) and right (political ideology=8) in the the effect of left - right political orientation on cl imate change risk perception is much strong er in the United States than in other wealthy and developed countries. An American on the left has a predicted probability of about 76 percent of considering climate change a very serious problem. An American on the right with the same profile has a predicted probability of just 28 percent of considering climate change a very serious problem. This pr edicted 48 percent difference i s about double the size of any other OECD country controlling for competing explanations and demographic factors. This predicted dif ference is 26 percent in Switzerland and 20 percent in Spain, the countries in which left - right political orientation had the next strongest effects. 8 At the other end of climate change risk perception , conserv atives in the United States are far more l ikel y to think climate change is not serious than conservatives in other OECD countri es. An American on the right is predicted to have a 10 percent probability of considering climate change not serious at all . The next high est predicted probability for an in dividual on the right considering climate change 8 The left - right divide in Switzerlan d is most like the one in the United States. S peci fically, a Swiss liberal has a predicted probability of considering climate change very serious of 62 percent, while a Swiss conservative i s predicted to be 37 percent likely to consider climate change very serious. Still, Swiss a conservative is about 10 percent more likely to call climate change very serious than American conserv atives. In Spain, citizens likely to call climate change a serious problem regardless of their self - described left - right orientation . An individual on the right is predicted to be about 64 percent likely to consider climate change very serious. And Spaniar ds on the left are overwhelmingly likely to view climate change as very serious, similar to the United States. A liberal is 85 percent likely to consider climate change very serio us. 127 Switzerland at 2.5 percent. Moreov er, an American conservative is 22 percent likely to consider climate change not very serious, another response category that reflects a lack of conc ern about climate change. A Swiss conservative is about 14 percent likely to say climate change is not very serious. And in the rest of the OECD countries , the predicted probability for each of these countries for a conservative viewing clima te change as n ot very serious is less than six percent. 9 Figure 6 .1: Effect of left - right political i deology on predicting considering climate change very serious , by OECD country (World Values Survey 2005 - 2009) Notes: The first di fference is the difference in probability of considering climate change very serious between left political ideology (ideology=3) and right political ideology (ideology=8), for 40 - 49 year old, non - Protestant females holding other variables at their overall mean for OECD countries. Statistical models used to produce the first differences are presented in Appendix D . 9 The above discussion is based on ordered logistic models predicting concern about climate change, controlling for competing factors and demographic explanations. These models are presented in the Appendix D. I also looked at relationship between le ft - right political orientation and climate change risk perception directly by not controlling for competing explanations. I n the United States, 58 percent of those on the right (ideology=7 - 10) call climate change not very serious or not serious at all, by far the largest percentage of OECD countries (and all countries). Just 12 percent of those on th e left (ideology=1 - 4) consider climate change not very serious or not serious at all. This 46 percent difference between those on the left and those on th e righ t in the United States is the largest in OECD countries (and all countries included in the WVS) . The next largest differen ce between the left and right i s in Switzerland. Finally, I also estimated linear models, instead of ordered models predicted perceiv ed seriousness of climate change. The coefficient estimate for the United States is about 2.5 stronger that the next largest estimate for Switzerland. The linear models are presented in Appendix D. So analyzing the direct relationship and collapsing catego ries to measure left - right political orientation produces nearly identical results: Americans are far more ideologically divided on climate change than citizens in other rich countries. In addition, I separately estimated models predicting perceived seriou sness of climate change using the Pew Global Attitudes from 2007 - 2010 for the United States, Germany, Great Britain, France, and Spain. In all four years self - described political ideology has its strongest predictive effect in the U.S.. 128 The above analysis shows that left - right political orientation has a particularly strong influence on climate change risk perception in the Unit ed States compared to other wealthy countries. Americans are comparatively more polarized in the United States. H1 is supported; the evidence suggests that political polarization is distinctly strong in the United States and could be aligned to elite opini on. 6.6 The conditional effect of political k nowledge in the United States If citizens are following ideologically similar elites in the United States on climate change, I expec t to also a conditional relationship between left - right political orientation and political knowledge and awareness. As hypothesized in H2, in the United States for those on the left , I expect that concern about climate change and belief in climate science to increase as en gagement with politics increases , and for those on the right , I expect concern about climate change and belief in climate change to decrease as political knowledge increases . I test this hypothesis using survey data just from the United States, specifically the 2008 ANES and 2012 ANES. I measure political engagemen t using a measure of political knowledge. Specifically I measure the number of questions a respondent can answer about basic political This direct mea sure of knowledge of basic political facts is used throughout the American public opinion literature as a measure of how closely citizens follow politics in the United States ( Zaller 1992 ). I expect citizens who closely follow politics to be more aware of elite opinions, and therefore more able to use elite cues to form opinions and answer sur vey questions. I find very strong evidence supporting H2 in both the 2008 ANES and 2012 ANES using many dependent variables to measure opinions on climate change . Across questions measuring belief in climate change and concern about climate change, polit ical know ledge has a positive correlation for liberals and a negative correlation for conservatives in the United States. The results of th ese models are shown in Figure 6.2 . First, in the May 2008 ANES wave, respondents were asked about their level of con cern about change. An average American who describe s himself or herself as all six 129 political k nowledge questions correctly is predicted to be 0.10 mor e concerned than someone who does not answer a single question correctly. For conservatives, concer n about climate change decr eases as political knowledge increases all six questions correctly was predicted to be 0.14 less concerned than an individual wi th the same profile who answers all six questions incorrectly. 10 Next, I look at the conditional effect of political information on three measures of belief in climate change in the 2008 ANES and 2012 ANES. The results are virtually identical those presented using concern about climate change above; political knowledge has a positive effect in belief in climate change for liberals and a negative effect for conservatives. First, in the May 2008 ANES wave, respondents were asked for the level of agreement on a 0 - 10 scale on the following four statements: ( 1) use continues, serious long - ientific evidence to support from zero to 10 measuring belief in climate change, with higher values capturing greater belief in climate change and its possible consequences. 11 Moving from the lowest value of political knowledge, not answering any of the political knowledge questions correctly, to the highest, answering all six questions correctly, increases the predicted value on this scale by slightly mo re than one for somewhat liberals, yet 10 These predicted differences, moreover, are statistically significant. For somewhat liberal, the 90 percent confidence interval of the estim ate of 0.1033does not include zero: [0.0169, 0.2033 ]. And for somewhat co nservative, the difference was s tatistically significant. Th e 90 percent confi dence interval of the estimate also does not include zero: [ - 0.2555, - 0.0159]. Moreover, I estimated regression models for both self - identified liberals and conservatives, predicting concern, shown in Appendix D . The coefficient estimate o f political knowledge is positive but barely not statistically significant for liberals (two - tailed p - value = 0.108 ) and the estimate for political knowledge is negative and statistically significant for conservatives ( two - tailed p - value = 0.052 ). I assume a higher alpha value of 0.1 because I am analyzing a subsample of the survey . 11 Two of the questions are reverse - coded so sca le is showing that the scale is reliable. In addition, in an ex ploratory factor analysis, the common factor explained 68 percent of the total variance and the common factor was strongly correlated with responses to all four questions. So there is strong evidence that the climate change belief scale is capturing latent belief that climate change is the result of human activity and its consequences are serious. 130 decreases the value by more than three for somewhat conservatives. Both of these differences were significantly different from zero. In both the 2008 and 2012 ANES, respondents were also asked about the cause of clima te change. I estimated models for both survey years predicting whether respondents cho o result of human activity . In 2008, for somewhat liberals, moving from the lowest to highest value of political knowledge increases the predicted probability of believing climate change is caused by human activity by about 50 percent, a statistically significant and substantially importa nt change. Liberals who answer all six questions incorrectly a re predicted about 30 percent likely to say climate change i s caused mostly by things people do, while liberals who answer all the questions correctly are predicted to be about 78 percent likely to consider climate change caused by human activity. For somewhat conservatives, the predicted probability of viewing climate change as the re sult of human activity decreases by 18 percent moving from the lowest to highest values of political knowl edge. Somewhat conservatives who answer a ll six political kn owledge questions correctly are predicted to be only 21 percent likely to say climate change is caused mostly by things people do. 12 I n the 2012 ANES, respondent are once again asked about the cause of climate change, and the results on t his question are very similar. Among those who did not answer any political knowledge questions correctly, the predicted probability of both liberals and conservatives considering climate change the result of human activity i s about 30 percent. For l iberal s who answer all five political knowledge questions correctl y, the predicted probability is about 80 percent, a statistically significant increase; for som ewhat conservatives who answer all five questions correctly, the predicted probability of answering t hat climate change i s caused s only about 10 percent, a 12 In the regression models for just the liberal and conservatives samples , political knowledge had a statistically significant and positive effect among libera ls and a negative but not statistically significant effect among conservatives (two - tailed p - value=0.452). The results ar e shown in Appendix D . 131 statistically significant decrease. 13 So among the most political knowledgeable Americans, the predicted difference in accepting that climate change is the re sult of human activity is about 70 percent, a massive difference capturing the deep ideological polarization on climate change in the United States. 14 So the findings from two public opinion surveys in the United States show that political knowledge conditions the influ ence of left - right political orientation on climate change opinions. For liberals, greater political k nowledge and awareness increases their concern about climate change and acceptance of climate science, while for conservatives, greater political knowledg e and awareness decreases their concern about climate change and acceptance of climate science. 15 This is strong evidence that Americans are using elite cues to form their climate change attitudes. Less knowledgeable and aware conservatives and liberals hav e a very similar level of concern about climate change and belief in climate science. This group of citizens is not divided on climate change because they are unlikely to be able to use elite cues to form opinions. But more politically knowledgeable and aw are citizens more able to ident ify and use elite cues are in line with ideologically similar elites. This group of citizens is deeply divided. There is evidence in support of H2 . 13 In a regression model predicting considering climate change the result of human activity among liberals, p olitical knowledge had a significant and positive influence. In the model for conservatives, political knowledge had a negative influence, but it was not statistically significant (two - tailed p - value=0.136). Re sults in Appendix D. 14 I also estimated model s using measures of self - described interest in politics in the 2008 ANES and self - described attention to politics in 2012 instead of the measure of political knowledge. The results are nearly identical. These results are presented in Appendix D . 15 Statis tical analysis using a measure of political awareness rather than political knowledge produces virtually the same results. 132 Fig ure 6 .2: Conditional e ffect of political ideology on predicting c lima te change o pinions across values of political knowledge in the United States (2008 and 2012 A merican N ational E lections S tudies ) Notes: Predict ed values and probabilities are calculated for a 40 - 49 year old female holding edu cation and biblical literalism at their mean values. Statistical models presented in Appendix D . 133 6.7 The conditional effect of interest in politics in the United States and wealthy countries In the United States, for liberals, political knowledge and awa reness has a positive effect on concern about climate change and in belief climate change science, while for conservatives, political knowledge and awareness has a negative effect on these same attitudes. However, it could be that I am not capturing the ef fect of elite opinion. Instead, more politically knowledgeable Americans might be more likely to be aware of how th eir social and economic positions align with the ir climate change opinions . To explore this possibility, I now compare the condition al relati onship between political engagement and left - right political orientation in the United States to other wealthy countries. If this conditional relationship exists across other wealthy countries, then it could the result of more politically aware individuals matching their political ideology and opinions on climate change. If this conditional relationship does not exist across other wealthy countries, then this specific finding in the United States can likely be attributed to citizens using elite polarization as a cue in only the United States. Unfortunately , the WVS does not provide a direct measure of political knowledge like the one used in the previous section (or even a measure scientific and/or environmental knowledge). So instead use a difference me asure of political engagement : self - described interest in politics. Self - described political interest is open to possible biases in responses, such as a respondent overstating or understating his or her interest for social desirability, but it should provi de a measure as to how closely an individual fol lows politics and elite discourse about politics. An individual who is says he or she is very interested in politics should be more likely to know where elites stand on issues than an individual who says he o r she is not very interested in politics. In the 2008 ANES, self - descr ibed interested in politics is positively and linearly correlated to political knowledge ( r =0.29 ) and in the 2012 ANES, self - de scribed attention to politics i s positively and linearly co rrelated to political knowledge ( r= 0.26). I estimated ordered logistic regression models predicting perceived seriousness of climate change for those on the left and those on the right in each OECD country using the WVS . The most important 134 independent var iable is interest in politics. 16 I also control for socio - economic factors, age, gender, self - described class, and education. The results of th ese models are shown in Figure 6.3. The United States i s the only country in which political interest has positive and statistically significant association with perceived seriousness for those on the left, yet political interest has a negative and statistically significant association with perceived seriousness for those on the right . For liberals, moving from low to h igh political interest increases the probability of considering climate change very serious by about 30 percent; for conservatives, moving from low to h igh political interest decreases this probability by 15 percent. As shown, in a few other countries po litical interest does have a significant effect on perceived seriousness of climate change among individuals on the left. In Sweden and Canada, unsurp risingly, political interest has a positive, significant, and strong substantive association with perceive d seriousness of climate change for citizens on the left. Specifically, an individual on the left with high political interest is about 22 percent and 31 percent more likely to consider climate change very serious than an individual on the left with low po litical interest in Sweden and Canada respectively. But in Sweden and Canada, there is not a significant relationship between political interest and concern for those on the right, as the 90 degree confidence interval for the estimate on the right includes zero in both countries. 17 16 Political interest is collapsed this way in order to overcome small sample sizes in each cell and also show patterns across countri es. I o nly include countries which have at least 25 respondents in ea ch cell, so eight countries are dropped from the analysis. 17 But as shown in Figure 6.3, in Canada, the 68 percent level of the estimate for those on the right does not include z ero. The climate change denial machine exists in Canada, so this is evidence that it might be working among the public on the right in Canada. 135 Figure 6.3 : Effect of interest in politics on predicting considering climate change very serious for left and right in select OECD countries (World Values Survey 2005 - 2009 ) Notes: The predicted difference in the probability of considering climate change very serious between very interested and not at all interested in politics for a 40 - 49 year - old female who identifies with the middle class, holding education at its mean. The thick error bar repr esents the 68 percent confidence interval , and the thin bar the 90 percent confidence interval . The statistical models used to produce this figure are presented in Appendix D . The analysis in this section shows that self - described general interest in poli tics has a unique conditional effect on left - right political orientation in the United States, supporting H3 . Americans on the left with high political interest are significantly more likely than those with low political interest to say climate change is v ery serious. But for those on the right, the effect is the opposite; those who are more interested in politics are less likely to say climate change is very serious. This conditional relationship i s not found in any other OECD countries. Americans seem to be suing ideologically similar elites to form opinions on climate change. It seems more engaged Americans are more likely to follow ideologically similar elites on climate change. 136 6.8 Discussion and c onclusion This chapter started with one main question in an effort to situation American public opinion on climate change in the comparative context. Why is American public opinion so polarized on climate change? I argue that Americans are likely turning to political and media elites as a shortcut in fo rming opinions on climate change. Unlike anywhere else in the world, American politi cal elites are deeply divided on the possible risks from climate change and also more fundamentally on climate science. Conservative and Republican opinion leaders, such as Santo rum, Limbaugh, and Will, call climate change on climate change, citizens could turn to ideologically similar elites, rather than attempt to gather and process the tremendous amount of information needed to form opinions on this issue. What evidence do I have that the American mass public in the United States is following political elites? First, I showed that Americans are far more polarized on climate change than in other wealthy countries; left - right political orientation has by far its strongest effect on perceived seriousness of climate change in the United States. Specifically, using the World Values Survey, a liberal is predicted to be 48 percent more likely to consider climate change very serious than a conservative in the United States, while the next largest differenc e between the left and right is 26 percent in Switzerland. Second, using public opinion surveys from the United States, I show that the influence of left - right political orientation on climate change op inions is conditioned by political knowledge. For liberal s, political knowledge increases concern about climate change and the probability of accepting climate science, while for conservatives, political knowledge decr eases concern about climate change and the probability of accepting climate science. For Americans with little kn owledge about politics, there i s no ideological divide on climate change. But Americans with great knowledge about politics are deeply polarized on climate change. Third, I show tha t this interactive relationship between left - right political orientation and engagement with politics is unique to the United States. Although in some other wealthy countries increasing interest in p olitics increased perceived dangerousness of climate chan ge for 137 individuals on the left, b ut in only the United States does interest in politics decrease perceived dangerousness for those on the right. This analysis provides evidence that aware and knowledgeable Americans turn in to their ideologically similar elites for cues in forming opinions on climate change . Politically aware conservatives watch Fox News , read the Wall Street Journal , and follow these expressed opinions on climate change, denying and dismissing its possible dangers. American public opinion is polarized not because everyone is divided on climate change, but because the most engaged Americans are deeply ideologically divided on climate change. In the United States, this analysis provides a bleak picture for climate change communicators and educators. This analysis r econfirms that the assumption that increasing awareness leads to increased concern found in other research should be treated with skepticism ( Guber 2013 ). For conservatives and Republicans, increased awareness actu ally seems to make it less likely climate change is considered serious . Prominent climate change activists, such as former Vice President Gore and billionaire Tom Steyer, are aggressive in their push for publicity and media attentio n. As Guber (2013) noted, increased focus on climate change activism might have the unintended boomerang effect of further alienating conservatives and Republicans from climate change science and action. Moreover, it also raises concerns about the relation ship political elites have with science in the United States. Conservative and Republican elites are rejecting an overwhelming scientific consensus on climate change. American conservatives and Republican are following individuals who do not know the scien ce; political elites are not supporting s cience, they are undermining it. This conservative rejection of science is similar to the elite movement which rejec ted the health risks of tobacco, and t his elite movement in support of tobacco delayed the proper p ublic health response for decades ( Oreskes and Conway 2010 ). American society paid a tremendous cost for this delay, and might once again pay a tremendous cost for inaction on climate change. But there is a positive for climate change activists. This chapter does not find that American conservatives are innately unconcerned about climate change. They are not unconcerned for fervent 138 ideological reasons. Instead, they are climate change deniers because they adopt the opinions of similarly conservative ideological elites. So when conservative and Republican elites realize the dangers of climate change and become concerned, we should expect the mass public to follow. The last three chapters focused on explaining climate change risk perception. The last chapter, instead of looking at the influences of climate change risk perception, shifts to concentrating on the consequences of climate change risk perception. Specific ally, it focuses on how perceived seriousness and dangerousness of climate change relates to environmental attitudes broadly and intended behaviors. 139 Chapter 7 : How do individuals think about climate c hange and environmentalism ? 7 .1 Introduction Cory Gardner, a Republican runn ing for the United States Senate in Colorado in 2014, said in a ( Siddiqui 2014 ). Scientists would call carbon emissions a pollutant as it is very harmful to human health due to the consequences of climate change. But the problem of carbon emissions which cause climate change is different than a ir pollution caused by sulfur oxides, nitrogen oxides, toxic metals, and smog from industry and automobiles. For the public, it would be natural to separate climate change from air pollution. Climate change is a global environmental problem which cannot be easily observed and does not clearly and directly affect lives, while air pollution is a local environmental problem which can be directly observed. People can see smog and feel the effects of poor air quality, but they cannot see and feel climate change. 1 Yet Gardner at least this one time combined air pollution and climate change together as the same environmental problem. And Gardner is not alone. Americans and people around the world generally tend to think about climate change and air pollution togeth er, even though there is reason to separate them onto distinct dimensions. This chapter shows that concern about climate change was closely related to concern about air and water pollution. But many in the global public, including a sizeable majority in th e United States, see both climate change and nuclear power as dangerous , even though nuclear power is one possible policy solution to address climate change. Next, this chapter looks at the influence of climate change risk perception on attitudes and behav iors requiring greater commitment to climate change. Specifically, I analyze the relationship betw een climate change risk perception and intended willingness to pay to address climate change and then the relationship between 1 Even though people cannot directly experience climate change, people can experience local temperature fluctuati ons. And recent research has shown these local temperature fluctuations influence climate change attitudes in the United States ( Egan and Mullin 2012 , Donner and McDaniels 2013 ) 140 risk perception and considering the environment generally and/or climate change specifically important issues. This section comes to three conclusions, one descriptive and two analytical. F irst, even among individuals who think climate change is serious , a substantia l minority across c ountries are unwilling to alter their behavior to address the consequences of climate change. Across about two dozen countries in two survey years, about four - in - ten of individuals who view clima unwilling to agree to pay hig her prices to address climate change. And most of those who consider climate change a dangerous problem are unwilling to rank the environment as more important than other issues and climate change as more important than other environmental proble ms. Of tho se who express that climate change is dangerous, only about one - in - five call the environment one of their most important issues or climate change their most important environmental problem. Second, c itizens in richer countries are not more likely to suppo rt economic sacrifices than citizens in poorer countries , an obstacle for any international clim ate change agreement requiring richer countries to make greater economic sacrifices . T hird, the influence of climate change risk perception on considering the environment and climate change important varies across countries. Specifically, cl imate change risk perception has a far stronger influence on ranking the environment and climate change as important in very rich and developed countries than in less wealthy and developed countries. The findings in this chapter show that global public opinion on climate change presents both an opportunity and a challenge for policymakers seeking to address climate change. The opportunity for leaders who want to shape and mov e public opinion is to relate climate change to other more discernable environmental problems. Individuals seem to think of environmental problems in a collective way, rather than as individual issues. Therefore, policymakers could take advantage of this b y framing climate change in the broader environmental context. Instead of warning of consequence s of climate change separately , 141 the y could emphasize the positive externalities of possible mitigation policies for other environmental problems . Mitigation pol icies improving air quality, address ing worries about air pollution, reducing our use of oil and coal, and tackling concerns about using up our natural resources could be very popular with the public. Broadening the frame of climate chang e as an issue coul d make the global public more supportive of mitigation policies . 2 But the challenge is daun ting: individuals who consider climate change serious or da ngerous do not necessarily see climate change as alarming . Many of those w ho say climate change is serious or dangerous seem unwilling to go beyond a statement on a survey. For most of those who say climate change is serious or dangerous, it is not an immediate, top - of - the - head issue. Many of those saying climate change is serious or dangerous are also unwilli ng to make economic sacrifices and most considered other issues and environmental problems more important. Most observe rs assume that a public obstacle to a policy solution is the group unconcerned by climate change, but this chapter shows that another obs tacle could be this larger group for whom climate change is concerning but not as important as other issues. For policymakers who want to address climate change, cross - national public opinion is tough to make sense of, as the public seems worried about cli mate change , yet consider s other issues more important. And for policymakers who do not want to address climate change, these findings are a reas on to continue to focus on more publicly pressing issues. 7 .2 Data, methodology, and m easurement Since the Wor ld Values Survey (WVS 2005 - 2009) asks only one question about climate change, this chapter draws from three additional cross - national public opinion surveys the 2009 Pew Global Attitudes Study, 2010 Pew Global Attitudes Study, the International Social Surv ey Programme ( ISSP 2009 - 2011) to analyze how individuals think about climate change. These cross - national surveys are also used in Chapter 2 to describe global public opinion. The first section examines the relationship between climate change risk percepti on and risk perception of other environmental problems, with 2 Ansolabehere and Konisky ( 2014 ) make a similar argument in the context of the United States. I argue here that this argument extends cross - nationally. 142 specific focus the dimensionality of risk perception. The ISSP asked about the dangerousness of seven pesticides and chemicals used in farming nuclear power stations modifying the genes in certain crops ries and corresponding numeric values for t he analysis in this chapter are In the second section, I examine the influence of c oncern about climate change on willingness to pay to address climate change, considering the environment an important issue, and considering climate change an important environmental problem. So the perceived seriousness or dangerousness of climate change is the independent variable of interest, and there are three dependent variables in the statistical models. First, the Pew Global Attitudes surveys in 2009 and 2010 ask ed respondents whether they agree or disagree with the following statement: ld be willing to pay higher prices in order to address global climate chang e analysis, agree is coded as one and all other responses are coded as zero, so I am modeling the probability of agreeing to pay higher prices to address climate change. S econd, the ISSP a sked respondents which issue is most important and also next most important for thei r country. Response options are h ealth care; education; crime, the environment; immigration; the economy; terrorism; poverty; none of these; and se. To measure considering the environment an important issue, choosing the environment as the most important issue or next most important issue for their country is coded as one and all other response combinations are coded as zero. Third, the ISSP also a sked respondents which enviro nmental problem they think is w a ir pollution; chemicals and pesticides; water shortage; water pollution; nuclear waste; domestic waste disposal; climate cha nge; genetically modified foods; using up our natural resources; none of these; and To measure considering climate change an important environmental problem, choosi ng climate change is coded as one an d all other responses are coded as zero. T he most important independent va riables in the analysis are perceived seriousness of climate 143 change in the Pew Global Attitudes surveys and perceived dangerousness of climate change in the ISSP. In the 2009 Pew Global Attitudes Survey, respondents were ask ed about the level of seriousness of global warming slightly different across the two Pew Global Attitude s surveys . The ordered response options and corresponding nu meric values for both years are would be, with the response options and numeric values discussed in the previous paragraph. The first section, which looks at relating climate change to other environmental problems, uses both multilevel modeling and exp loratory factor analysis. The random - intercept multilevel models , which allow for clustering of survey respondents within countries , are used to produce estimates of the strength and direction of the relationship between perceived dangerousness of climate change and perceived dangerousness of other environmental problems , controlling for demographic factors. I also estimate linear models to predict the strength and direction of the relationsh ip between perceived dangerousness of climate change and perceived dangerousness of other environmental problems in the United States. The goal of exploratory factor analysis is to explore the underlying dimensions of the perceived dangerousness of seven environmental problems ( Bartholomew, et al. 2008 , Mulaik 2010 ). The second section, which looks at the rel ationship between climate change risk perception and other opinions and intended behaviors on climate change, uses multilevel modeling to estimate the average effect across countries and also explore how this effect varies across countries. Since all three dependent variabl es in this section willingness to pay to address climate change, considering the environment an important issue, and viewing climate change as an important environmental problem are dichotomous, I assume a binary logistic functional form in estimating the models, so t he individual level variance is fixed and not estimated. Multilevel random - intercept models model are used to estimate the average effect of individual - level independent variables, including the perceived seriousness or dangerousness of climate change, across countries. To explain country - level heterogeneity in the random - 144 intercept multilevel models, I include measures of country - level wealth and development, GDP per capita and Human Development Index (HDI) score. Finally, I introduc e cross - leve l interaction terms in random - intercept , random - effect models, in which allow both the constant and the estimated effect of perceived seriousness of climate change to vary across countries. 7 .3 Relating c limate change risk perception and risk perception o f other environmental problems This chapter first looks at the relatio nship between perceived dangerousness of climate change and perceived dangerousness of other environm ental problems. Is risk perception of some environmental problems more c losely relate d to climate change risk perception than others? Is there a lat ent attitude that sparks worry about environmental issues? This section shows that there is little evidence that people think about climate change differently than other environmental problems, including the various forms of pollution; individuals seem to think about environmental problems in the same way. In most countries, including the United States, many individuals who water and air pollution are dangerous also say climate change is dangero us 7.3.1 Relating perceived dangerousn ess of climate change and perceived dangerousness of pollution Worry about climate change is related to worry about other environmental problems, and in particular, worry about air and water pollution. I first explore the relationship between perceived dangerousness of climate change and perceived dangerousness of other environmental problems. Controlling for demographic factors, perceived dangerousness of climate change is si gnificantly correlated with perceived dange rousness of all six other included environmental problems . Specifical ly, an individual who considers cli predicted to consider all six problems between 0.16 and 0.19 more dangerous on the zero - to - han an individual who considers In the United States, perceived dangerousness of climate change is also strongly related to perceived dangerousness of other environ mental problems. An American who views climat the six other problems as between 0.13 and 0.2 more dangerous on the zero - to - one range than an American who considers Climate change 145 risk perception is most strongly correlated with concern about risk perception of air pollution from cars and industry in the United States, showing that many Americans worried about climate change are also worried about air quality. By exploring t he dimensionality of environmental concern, I again show that risk perception of climate change and risk perception of pollution is closely related. While environmental concern is usually defined as including various attitudes, opinions, behavior on the en vironment, here the focus is on one particular dimension of environmental concern: risk perception of possible environmental problems, including climate change. 3 Most previous research on environmental concern has focused on public opinion in the United St ates, but here I look at both concern about environmental problems in the United States and cross - nationally. 4 To explore the dimensionality, I use exploratory factor analysis, a method to discover the underlying linear structure of risk perception of envi ronmental problems ( Bartholomew, et al. 2008 , Mulaik 2010 ) . Generally, risk perception of environmental problems is a uni - dimensional, and across most countries, the perceived dan gerousness of climate change is closely related to the perceived dangerousness of air and water pollution. I first conducted an explor atory factor analysis for the entire sample of 32 countries in the ISSP. The exploratory factor analysis of responses to the dangerousness of the seven environmental problems in 3 Environmental concern cons ists of beliefs, attitudes, intended behaviors, and actual behaviors. Dunlap and Jones ( 2002 Environmental concern could include belief in environmental and scientific facts, assessing the seriousness of environmental problems, willingness to sacrifice economically to protect the environment, and participation in recycling, limited electricity use, and environmental activism ( Van Liere and Dunlap 1980 , Carman 1998 , Guber 2003 ). 4 Re cent cross - national research on environmental concern has found evidence that it is a multi - dimensional concept ( Dunlap and York 2008 , Marquart - Pyatt 2008 , Knight and Messer 2012 , Marquart - Pyatt 2012 , Pampel 2014 ). Marquart - Pyatt ( 2008 ), using the 2000 International Social Science Programme (ISSP) cross - national survey, fi nds a environment and an intended behavior dimension, capturing how willing an individual is to make economic sacrifices for the environment. An d, in later work, Marquart - Pyatt ( 2012 ) adds another attitudinal dimension to environmental concern, capturing risk perception of environmental problems. Dunlap and York ( 2008 ) find five dimensions of environmental concern using the World Values Survey (WVS) wave 4: willingness to pay, environmental protection versus economic growth, support for the environmental movement, local environmental problems, and global environmental problems. Knight and Messer ( 2012 ) also find five dimensions of environmental concern analyzing wave 5 of the WVS. 146 all countries suggests one underlying dimension. 5 The common factor has an eig envalue of 2.8, explaining 40 percent of the all variance in the observed responses. 6 Individuals around the world tend to think about the risks from climate change like they think about the risks from other environmental problems; the same latent attitude is related to responses to seven environmental problems, including climate change. Table 2.1 shows the results of factor structure, or how correlated each variable is with the common, unobserved factor. There are differences in the correlations by environ mental problem with the common factor. Perceived dangerousness of air pollution from industry, air pollution from cars, water po llution, and climate change are more highly correlated with the common factor than the perceived dangerousness of other environm ental problems. This also suggests that the one factor model does a poorer jobs explaining variation in the perceived dangerousness of genetically modified crops and nuclear power. Each variable is at least moderately correlated with the common factor, so one factor is sufficient, but this result suggests that the structure might be more complicated within countries. Tabl e 7.1: Exploratory factor analysis of risk perce ption of environmental problems in all c ountries (International Social Survey Programme 2009 - 2011) Environmental Problem Standardized factor loading Climate change 0.64 Air pollution from cars 0.65 Air pollution from industry 0.71 Pollution of rivers, lakes and streams 0.65 Pesticides and chemicals in farming 0.68 Genetically modified c rops 0.58 Nuclear power stations 0.50 Proportion of variance explained 0.40 Notes: The factor analysis was fit with principle axis factoring. 7 To explore acros s country variation, I perform exploratory factor analyses for each individual count ry in th e ISSP to determine the structure of risk perception on environmental problems. This 5 I use a scree plot to determine the number of factors to include in the analysis. A scree plot shows the eigenvalues by the factors in order of extr identified. Another common decision rule is to include factors with eigenvalues greater than one. In this situation, both decision rules result in a solution with one c ommon factor for responses to the dangerousness of the seven environmental problems in all included countries. 6 The eigenvalue is a measure of the variance accounted in all variables by the common factor. 7 Using maximum likelihood produced virtually identical results . 147 analysis confirms the findings of the all country model, showing a latent attitude motivates climate change risk perception and risk perception of air and water pollution. For example, in Canada, perceived dangerousness of clima te change and perceived dangerousness of air pollution from industry, air pollution from cars, and pollution of rivers, lakes, and streams map on to the same underlying dimension. Figure 7 .1 shows t he results of an exploratory factor analysis for respondents in Canada with a varimax rotation, resulting in two uncorrelated common factors. The perceived dangerousness of climate change, like air pollution from cars and industry, is far more correlated w ith Factor 1 than Factor 2, while nuclear power and genetically modified crops are more correlated with Factor 2. Figu re 7 .1: Factor loading plot, Canada (International Social Survey Programme 2009 - 2011) In at least a few countries, such as the United States, climate change does not map onto either underlying dimension. Here one dimension is defined by underlying attitudes toward air pollution and the other dimension is defined by newer environmental problems genetically modified crops and nuclear power . Other problems, including climate change, form a cluster that does not map onto either dimension. Figure 7 .2 shows the results of an exploratory factor analysis for respondents in the United States, resulting in two orthogonal factors. But still, even in the United States, there is evidence that Americans modestly relate the risks from climate change to the risks from air and water pollution. 148 Figure 7 .2: Factor loadings plot, United States (International Social Survey Programme 2009 - 2011) This section argues that at the individual - level , climate change risk perception is closely related to risk perception of other environmental problems, and in particular risk perception of air and water pollution . Individuals who consider climate change dangerous also mostly also consider air and water pollution dangerous and vice - versa, in both the United States and cross - nationally. For policymakers, this shows that their focus should not necessarily be on only climate change. It seems that the global public would be more o r as welcome to broader policy addressing numerous environmental problems, including climate change and pollution. A single emphasis on climate change might actually be detrimental to addressing the consequences of climate change, as broader environm ental policy could not only address climate change, but also improve air and water quality and expand sustainable energy , positive externalities of climate change policy the global public would support . 7.3.2 Perceived dangerousness of climate change and nuclear power Nuclear power, since it does not emit CO 2 emissions, is considered by policymakers and scientists to be crucial to any solution addressing climate change. Yet public opinion has always been a major obsta cle to expanding nuclear power. Citizen s generally reject the construction of nuclear power plants in their local area and react negatively to accidents at nuclear power plants ( Ansolabehere and Konisky 149 2014 ). 8 In addition, perceived dangerousness of climate change is positive ly corre lated with perceived dangerousness of nuclear power plants . Across all 32 countries in the ISSP, the relationship between the perceived dangerousness of climate change and perceived da ngerousness of nuclear power is of similar strength as the relationship between the perceived dangerousness of climate change the perceived dangerousness of six other environmental problems. 9 Controlling for demographic factors, the positive relationship between clim ate change and nuclear power is slightly stronger the relatio nship between climate change and pesticides in farming, climate change and pollution from cars, and climate change and pollution from industry across all countries. And in the United States specifically, controlling for demographic factors, the relationshi p between nuc lear power and climate change is stronger than the relationship between pollution of water and climate change, pesticides/chemicals in farming and climate change, and genetically modified foods and climate change. Unsurprisingly, what drives this posi tive and substantially relationship between perceived dangerousness of climate change and perceived dangerousness of nuclear power is that many individuals find both to be dangerous. Across all co untries, of those who consider 54 percent also co nsider percent of Americans who view extreme nuclear power stations as 8 Research has shown that a nuclear disaster likely results in a decline in public acceptance and increase in perceived dangerousness of nuclear power. Previous studies examining public opinion after nuclear accidents at Three Mile in Pe nnsylvania and Chern obyl show that public becomes wary ( Rosa and Dunlap 1994 ). And research on after the Fukush ima nuclear power disaster finds that acceptance of nuclear power decli ned in Australia ( Bird, et al. 2014 ), Great Britain ( Corner, et a l. 2011 ), and Switzerland ( Siegrist and Visschers 2013 ) . The ISSP was completed between 200 9 and 2011, so some country surveys were either in the field or took place after the Fukushima nuclear power disaster. Specifically, the ISSP survey was in the field in Canada, Denmark, Slovenia, and Switzerland during the Fukushima nuclear power disaster. And the ISSP survey was field after Fukushima in Croatia, Finland, Israel, Latvia, Mexico, Norway, and Taiwan. 9 In every country the perceived dangerousness of climate change is positively correlated with the perceived dangerous of nuclear power. 150 10 Among those who see climate change as very risky, most also see nuclear power as very risky, both globally and in the United States . These individuals would likel y reject a solution to climate change that includes an expansion of nuclear power, and instead probably prefer cleaner sustainable alternatives, such as wind, solar, and water, even if these sources are ex pensive and unlikely to meet global energy needs . I f policymakers decide to include expanding nuclear power in any sol ution to reduce carbon emissions , they are going to have to change the minds of many of those most worried about climate change, which could be especially difficult. Public opinion is often seen as an obstacle to expanding nuclear power. This section shows just the difficulty of making policy to address climate change while also appeasing the public . 7 .4 Relating climate change risk perception and willingness to p ay Democracy should requ ire consistency between attitudes and behaviors. If policymakers cannot make sense of what the public wants, then it is difficult to make policy that reflects public opinion. The next two sections show how difficult it is for the policymakers in the United States and around the world to get a sense of public opinion on climate change. They could be concluding that the global level of concern about climate change is actually weak and lacking depth in many countries, and peop le care more about o ther issues . This could help explain why both domestic policy in many countr ies and international policy have yet to address the consequences of climate change by meaningfully reducing carbon emissions . This section looks at the relationship between climat e change risk perception and intended behavior on climate change. The Pew Global Attitudes surveys in 2009 and 2010 asked respondents if 10 In the United States, of those Ameri cans who said climate change is power stations are of the Americans who consider percent consider 151 order to addre 11 Citizens will most likely have to make economic sacrifices in any domestic and international policy to address climate change as sustainable energy is expected to be more expensive . Unsurprisingly, as perceived seriousness incre ases , so did the percentage of respondents willing to pay higher prices to higher prices in all countries in both survey years. But, importantly, a substantial minority of individuals who say climate change is serious, however, is unwilling to ma ke economi c sacrifices. Of those who view 2009 and 44 percent disagreed in 2010. And of those who view percent disagreed in 2009 and 35 percent disagreed in 2010. Even in this group, there is not widespread support for economic sacrifices to address climate change in all included countries . How does this pattern vary across countries? Figure 7 .3 shows the rel ationship between perceived seriousness of climate change and willingness to pay across countries in the Pew Global Attitudes 2009 and 2010 surveys. The results show that in most countries as perceived serious ness of climate change increases, the probability of being willing to pay higher prices to addre ss climate ch ange also increases . 12 But these results also show the reluctance of tho se who say climate change is serious to agree to make economic sacrifices. In seven countries in each survey year, less than a majori ty of individuals who consider cl willing to pay higher prices to address climate change. Less than a fifth of Egyptians an d Jordanians who say climate change is very serious are willing to pay, and ab out one - third of Mexicans who say climate change is very serious are willing to pay. 13 11 Note that the question wording does not address where the additional revenue from higher prices would go or specifically how much prices would go up . 12 In Egypt and Mex ico in 2009, a grea ter proportion of those who say climate change is not serious (choosing either "not serious at al l" or "not very serious") agree to pay to address cl imate change than those who say climate change is sample sizes in these groups, this is a surprising finding. 13 Even in those countries in whic h a majority of those who view clima willing to make economic sacrifices to a ddress climate change, there is still a substantial m inority unwilling to back up their Gr eat Britain and France there is not a clear consensus among citizens who say climate change is serious that they shou ld sacrifice economically to address climate change. In France, less than four - in - ten a nd six - in - ten of those who say climate change is to pay higher prices. In Great Britain, 70 percent and 74 perce nt in 2009 and 2010 resp ectively of those who consider cl imate change very serious agree to sacrifice economically, the largest percentage of any country. Even though the concerned segment of the 152 more Americans are unwilling to pay than willing to pay higher prices. For most of this group, which expressed at least some concern about climate change, t here seems to be little desire to sacrifice to address climate change. At least among Americans who view climate chang agree than disagree , b ut roughly one - third in b oth survey years still disagree with paying higher prices to addr ess climate change. British public is among the most willing to pay, a substanti al minority about three - in - ten and o ne - in - two of those very serious somewhat serious respectively does not agree to pay higher prices. Even in Western Europe among the most worri ed about climate change there i s not overwhelming support for paying higher prices. 153 Figure 7 .3: Relationship between perceived seriousness of climate change and willingness to pay to address climate change , by country (Pew Global Attitudes Project 2009 - 2010 ) 154 Notes: collapsed sample . Still, in Brazil and South Korea in both 2009 and 2010, fewer than 30 respondents say climate change i s estinian Territ ories in 2009 is also excluded from this figure. Next, I turn to multilevel models predicting willingness to pay higher prices as a function of perceived seriousness of climate change, gender, age, and at the country - level, the natural logarithm of G DP per capita and the Human Development Index, to measure country wealth and development. Equation 1 shows the estimated formula for considering willingness to pay as a latent attitude. 155 (1) Considering willingness to pay has two response options, the logistic distribution is used as a link function, so the individual - level variance is fixed. Table 7 .2 shows the results. These are random - intercept models, so each intercept varies across countries. 14 The models show that risk perception has a strong influence on willingness to pay across countries. The coefficient estimate for perceived seriousness of climate change is positive, showing that increased perceived seriousness is associ ated wit h increasing likelihood of stating a willingness to sacrifice economically . And its substantive strength is unsurprisingly great. In 2009, the difference in predicted probabilities on average across these two dozen countries of being willi ng to pay between those who say climate change i s 37 percent; it 20 10, this predicted difference i s 49 percent. Unsurprisingly, climate change risk p erception dramatically increases the probability of agreeing to pay higher prices to a ddress climate change. 14 More information on the var iables is included in Appendix A . The modeling strategy for this proje ct is addressed in detail in Chapter 3 . I use multilevel models to address for clustering within countries. I allow intercepts t o vary across countries in all models, thus allowing intercepts to be predicted by level - two covariates. 156 Table 7.2: Random - intercept binary logistic models predicting agreeing to pay higher prices to address climate change (Pew G lobal Attitudes Studies 2009 - 2010) 2009 2010 (1) (2) (3) (4) (5) (6) Climate change risk percep tion 0.5561 0.5561 0.5561 0.7948 0.7948 0.7948 (perceived seriousness) (0.0795) (0.0795) (0.0795) (0.0922) (0.0923) (0.0923) Female - 0.0727 - 0.0727 - 0.0727 - 0.0014 - 0.0015 - 0.0015 (0.0445) (0.0445) (0.0445) (0.0585) (0.0585) (0.0585) Age you nger than 30 0.0843 0.0846 0.0845 - 0.0955 - 0.0952 - 0.0953 (0.0480) (0.0480) (0.0480) (0.0676) (0.0677) (0.0677) Age 30 - 39 0.0760 0.0761 0.0761 - 0.0344 - 0.0342 - 0.0342 (0.0458) (0.0457) (0.0457) (0.0569) (0.0569) (0.0569) Age 50 - 59 0.0671 0.0669 0.067 0 - 0.0447 - 0.0449 - 0.0448 (0.0538) (0.0540) (0.0540) (0.0593) (0.0592) (0.0592) Age 60 - 69 0.1218 0.1214 0.1216 - 0.0178 - 0.0183 - 0.0182 (0.0692) (0.0694) (0.0694) (0.0726) (0.0726) (0.0726) Age 70 and older 0.0981 0.0976 0.0979 - 0.1289 - 0.1297 - 0.1294 (0.0913) (0.0915) (0.0914) (0.0736) (0.0737) (0.0736) Country - level variables Natural logarithm of GDP per capita 0.0866 0.1010 (0.1325) (0.1120) HDI 0.4726 0.7954 (1.1556) (0.9777) Constant - 1.3322 - 2.1231 - 1.6902 - 1.8353 - 2.7447 - 2.4283 (0.1984) (1.2584) (0.9160) (0.1851) (1.0089) (0.7378) Variance component Country - level variance 0.7539 0.7418 0.7499 0.7556 0.7395 0.7453 (0.2751) (0.2930) (0.2849) (0.2893) (0.3061) (0.3015) N umber of respondents 23858 23858 2 3858 22095 22095 22095 Number of countries 25 25 25 25 22 22 A IC 27451 27452 27453 24392 24394 24394 Note s : Cell entries are logistic regression coefficients with the standard error in parenthesis for a random coefficient, random intercept model . Listwi se deletion was used. Bold indicates two - tailed p <0.05. The Palestinian Territories were not included in the 2009 models. But the m odels also show that individuals who say climate change is a serious problem are far from guaranteed to also say they agr ee to pay higher prices to address climate change . A 40 - 49 year - old woman who considers cli mate change somewhat serious has a predicted probability of 43 percent in the 2009 survey and predicted probability of 44 percent in the 2010 survey of agreeing to p ay higher prices to address climate change on average across countries. And a 40 - 49 year old women who views cl imate change as very serious is predicted to be 57 percent in 2009 and 64 percent in 2010 likely to accep t higher prices . So the models confirm t he finding from above: it is far from a certainty that an individual who 157 considered climate change very serious also would agree to pay higher prices. Globally, those who think climate change is a serious problem are predicted to be divided over being will ing to pay to address it. In addition, the multileve l models show that there is variation across countries in probability of agreeing to pay higher prices to address climate change. But a coun not explain this variation. In models for both surveys , the estimates for the natural logarith m of GDP per capita and HDI are not statistically different from zero. Citizens in wealthier a nd more developed countries are not more willing to pay to address climate change than citizens in less wealthier and developed countries. We should not assume that citizens in rich countries would be more willing to make greater economic sacrifices as part of international mitigation police . The evidence from these two cross - national surveys suggests that citizens in rich countries would be as reluctant to make any economic sacrifices as citizens in poor countries, even though citizens in rich countries should be able to more easily afford economic sacrifices. This reluctance among citizens of richer c ountries could be a major obstacle for policymakers crafting a national policy and international agreements going forward. 7 .5 Climate change risk perception and the environment and climate change as most important i ssues Is perceived dangerousness of c limate change related to considering the environment an important issue and climate change the most important environmental problem? The analysis here shows this relationship is similar to the relationship between risk perception and willingness to pay di scussed in the previous section. Once again there is a positive relation ship; individuals who consider climate change dangerous are more likely to view the environment as one of their most important issues or climate change as their most important environm ental p roblem than those who consider climate change not dangerous. But even t hough concerned individuals are more likely to consider the environment and climate change important issues, large majorities of those most worried about climate change still ran ked other issues or other environmental problems as more important. This section provides further evidence that for many, alarm about climate change is not part icularly deep. 158 7 .5.1 Climate change risk perception an d considering the environment a most imp ortant issue What is the relationship between the perceived dangerousness of climate change and viewing the environment as one of the two most important issues? On the ISSP, respondents were asked to name their most important and next most important issu es from a list which included the environment. First, cross - nationally, unsurprisingly, perceived dan gerousness of climate change is positively related with ranking the environment as an important issue. In all 32 co untries, of those who consider climate c cent name the environment as their most or next most important issues, while 17 percent of those who consider climate ch the environment as their most or next most important issues. In all - in - five rank the environment as one of their two most important issues. For example, in the Uni ted States, among those who say climate ch ange is the economy (34 percent), health care (22 percent), education (19 percent), or terrorism (8 percent) than the environment (6 percent) as the most important issue for the United States. Another 11 percent in this group say the environment is issue. 15 Among the who are concerned about the risks of climate change , across both rich and poor countries, only a fraction rank the environment as their most or next most import ant issue, instead 15 For another example, in France, of those individuals who say climate change is very dangerous or extremely dangerous, more call the economy (27 per cent), education (20 percent), health care (17 percent), and poverty (16 percent) than the environment (12 percent) their most important issue. About 16 percent of French call the 159 Figure 7 .4: Relationship between perceived dangerousness of climate change an d considering the environment the most or next most important issue, by country (International Soc ial Survey Programme 2009 - 2011) Notes: larger sample sizes across the categories of perceived dangerousness. Still, in Argentina and Chile, fewer th an 30 respondents say c , so this category was dropped for these two countries. Next, I use multilevel models to predict the probability of considering the environment an important issue as a function of risk perception, partisanship, education, age, gender, and at the country - level, the natural logarithm of GDP per capita and HDI. Equation 2 shows the cross - level interaction specification for the latent variable model. Both the intercept and effect of perceived dangerousness of climate change are allowed to vary across countries, while all other covariates are fixed. The systematic differences in the intercept and effect of perceived dangerousness are predicted by the natural logarithm of GDP per capita or Human Development Index (HDI) score. And the covariance between the intercept and effect is estimated as part of the model. 160 (2) The results are presented in Table 7 .3. Models 1 - - intercept mode ls. Models 4 - 6 allow both the intercept and coefficient for dangerousness of climate change to vary across countries, in random - perceived dan gerousness of climate change is significantl y relat ed to the probability of saying the environment is one of the two most important issues facing his or her country. An individual who co nsiders climate predicted to be about seven percent more likely to say the environ ment is an important i ssue than an individual who considers climate change perception has the strongest statistical influence of all of the individual - level predictors included in the model, its substantive influence is surprisingly weak, co nsidering the expectation that those who see climate change as a large risk would be more likely to say the environment is a most important issue . The modest effect of the perceived dangerousness of clima te change across all countries does not tell the whole story though. Models 1 - 3 show that there is signi ficant variation in ranking the environment as an important issue across countries. And Model 4, which allows the coefficient estimate for perceived dan gerousness of climate change to vary across countries, shows that there is significant and meaningful variation in the influence of perceived dangerousness of climate change on considering the environment an important issue across countries. 16 Models 5 and 6 in Table 2.4 then model this variation in the effect 16 The Akaike Inform ation Criterion (AIC) for Model 4, the random - intercept, random - coefficient model, decreases by 84 compared Model 1, the random - intercept model, showing that allowing climate change risk perception to vary across countries provides superior model fit. 161 of perceived dangerousness of climate change across countries by estimating cross - level interaction models. Specifically, the models include interaction terms for perceived dangerousness of climate cha nge and the natural logarithm of GDP per capita in Model 5 and perceived dangerousness of climate change and HDI in Model 6. The results sho w that climate change risk perception has a str ong influence on ranking the environment as a most important issue in very rich and developed countries. Figure 2.7 shows the difference in predicted pr obability between those who say climate change is those who say capi ta. When GDP per capita is about $3,000, perceived dangero usness of climate change does not significantly predict But in a rich country, in which GDP per Specifically, w hen GDP pe r capita is $360 00, an individual who considers cli predicted to be 19 percent likely to rank the environment as important, compared to 11 percent for an individual who considers 162 Table 7 .3: R andom - intercept, random - coefficient binary logistic models predicting considering the environment the most or next most important issue (International Social Survey Programme 2009 - 2011) (1) (2) (3) (4) (5) (6) Dangerousness of climate change 0.2765 0.2771 0.2770 0.2162 - 1.2310 - 1.0878 (0.0399) (0.0399) (0.0399) (0.0441) (0.3469) (0.2986) Left party identification 0.1130 0.1126 0.1138 0.1082 0.1081 0.1092 (0.0680) (0.0677) (0.0677) (0.0670) (0.0669) (0.0669) Right party identification - 0 .2931 - 0.2944 - 0.2929 - 0.2759 - 0.2770 - 0.2757 (0.0674) (0.0672) (0.0673) (0.0671) (0.0669) (0.0671) Level of education 0.0575 0.0573 0.0568 0.0551 0.0552 0.0548 (0.0183) (0.0183) (0.0183) (0.0181) (0.0182) (0.0182) F emale 0.0606 0.0604 0.0604 0.0537 0.0533 0.0532 (0.0505) (0.0505) (0.0505) (0.0500) (0.0500) (0.0500) Age younger than 30 0.1296 0.1303 0.1307 0.1254 0.1257 0.1259 (0.0510) (0.0508) (0.0509) (0.0515) (0.0514) (0.0514) Age 30 - 39 - 0.0020 - 0.0016 - 0.0014 - 0.0061 - 0.0061 - 0.0061 (0.058 8) (0.0588) (0.0589) (0.0584) (0.0584) (0.0584) A ge 50 - 59 0.0853 0.0849 0.0847 0.0900 0.0902 0.0899 (0.0617) (0.0616) (0.0617) (0.0611) (0.0611) (0.0611) Age 60 - 69 0.0661 0.0654 0.0649 0.0775 0.0777 0.0770 (0.0600) (0.0600) (0.0600) (0.0594) (0.0593) (0.0593) Age 70 and older - 0.0108 - 0.0117 - 0.0128 0.0040 0.0040 0.0030 (0.0865) (0.0866) (0.0864) (0.0864) (0.0864) (0.0863) Country - level variables Natural logarithm of mean GDP per 0.7915 0.3521 capita (2005 - 2009) (0.1514) (0.1405) Mean HDI (2005 - 2009) 8.1174 3.4045 (1.6432) (1.2589) Cross - level interaction terms Climate change risk perception X 0.1465 Natural logarithm of mean GDP (0.0350) Climate change risk perception X 1.5590 Me an HDI (0.3548) Constant - 2.8437 - 10.6771 - 9.6571 - 2.6432 - 6.1340 - 5.5131 (0.1650) (1.5838) (1.4874) (0.1507) (1.4204) (1.0933) Variance components Country - level standard deviation 0.8354 0.3167 0.3753 0.4792 0.3891 0.4217 (0.1579 ) (0.0698) (0.0840) (0.1832) (0.1732) (0.1771) Dangerousness of climate change 0.0415 0.0235 0.0240 standard deviation (0.0116) (0.0079) (0.0085) N umber of respondents 37666 37666 37666 37666 37666 37666 Number of countries 30 30 30 30 30 30 A IC 28086 28060 28065 28002 27976 27981 Note s : Cell entries are logistic regression coefficients with the standard error in parenthesis for a random coefficient, random intercept model . Listwise deletion was used. Bold indicates two - tailed p <0.05. Israel and Taiwan are not included in the models because respondents were not asked for their partisan identification in these two countries. 163 7.5.2 Risk p erceptio n and climate change as the most important environmental problem This section looks at the relationsh ip between climate change risk perception and ranking climate change as the included in the ISSP, there is a strong positive relationship between climate change risk perce ption and ranking climate change as the most important environmental problem; as perceived dangerousness increases , so does the probability of considering climate change the most important environmental problem. In every cou ntry, individuals who consider c limate change dangerous are more likely to choose roblem than those who consider climate change not dangerous. 17 In the United States, of those who say climate change is or environmental problem, compared to 23 percent of those who consider Among those most worried about climate change, however , a large majority of individuals rank another environmental problem other than climate change as the most important for their country. Mor e individuals consistently name air pollution than climate change as the most important environment problem for their 18 Of those Americans who express that climate change is climate change a s the most important environmental problem . In this group, more consider 17 Of problem than those who say climate change Japan, this difference is 46 and 49 percent respectively. 18 In all 32 count ries, among those who consider their count problem. Sixteen percent rank environmental problems. Among those who call r th e most important environmental problem. In only two coun tries Canada and Japan does a ma climate change is ercent of individuals who call d 60 percent of those who call climate chang answer that climate change is 164 inconsistency in American public opinion on climate change. Even among those most worried about climate change, onl y about one - in - five ranks climate change as more important than other environmental problems, let alone more important than other general issue s. Even among those who say cl imate change is dangerous , there is little pressure on policymakers in the United States to actually do something on climate change. Figure 7 .5: Relationship between perceived dangerousness of climate change and considering climate change the most importa nt issue, by country (International Social Survey Programme 2009 - 2011) Notes: adequate sample sizes across the categories of perceived dangerousness . Still, in Argentina and Chile, fewer than 30 respondents these two countries. Next, I model the choosing climate change as the most important environme ntal problem as a function of climate change risk perception, political partisanship, education, age, gender, and at the country - level, the natural logarithm of GDP per cap ita and HDI, as shown in Table 7 .4. The model 165 specification is the same as presented previously in Equation 2. Once again, the intercept and effect of perceived da ngerousness of climate change are a llowed to vary across countries in the random - intercept, random - effect models . And the logistic distribution is used since the dependent varia ble is binary (climate change is an important environmental problem or not), so the individual - level variance is fixed and not estimated. Equation 3 shows the cross - level interaction specification for the latent model. (3) The results show once agai n that climate change r isk perception predicts assessing the climate change as the most import a predicted probability of about three percent of ranking climate cha m a predicted probability of 23 percent, a difference of 20 percent. 166 Table 7 .4: Random - intercept, random - coefficient binary logistic model s predicting considering climate change the most important environmental problem (International Social Survey Programme 2009 - 2011) (1) (2) (3) (4) (5) (6) Climate change risk perception 0.5693 0.5706 0.5703 0.5361 - 1.0488 - 0.8488 (0.0379) (0.03 81) (0.0380) (0.0438) (0.3070) (0.3813) Left party identification 0.0880 0.0843 0.0862 0.0801 0.0792 0.0804 (0.0543) (0.0537) (0.0533) (0.0538) (0.0534) (0.0530) Right party identification 0.0157 0.0108 0.0131 0.0258 0.0241 0.0257 (0.0481) (0.0482) ( 0.0480) (0.0465) (0.0466) (0.0463) Education 0.0711 0.0705 0.0703 0.0701 0.0701 0.0700 (0.0153) (0.0151) (0.0152) (0.0152) (0.0151) (0.0151) Female - 0.0440 - 0.0444 - 0.0443 - 0.0493 - 0.0500 - 0.0499 (0.0327) (0.0328) (0.0328) (0.0343) (0.0344) (0.0343) Age younger than 30 - 0.0323 - 0.0318 - 0.0315 - 0.0411 - 0.0410 - 0.0411 (0.0542) (0.0542) (0.0542) (0.0555) (0.0556) (0.0555) Age 30 - 39 - 0.0148 - 0.0144 - 0.0143 - 0.0199 - 0.0207 - 0.0207 (0.0613) (0.0613) (0.0613) (0.0626) (0.0627) (0.0627) Age 50 - 59 - 0.09 88 - 0.0994 - 0.0995 - 0.0975 - 0.0970 - 0.0973 (0.0479) (0.0479) (0.0479) (0.0475) (0.0474) (0.0474) Age 60 - 69 - 0.2328 - 0.2337 - 0.2339 - 0.2276 - 0.2269 - 0.2276 (0.0480) (0.0482) (0.0481) (0.0481) (0.0481) (0.0481) Age 70 and older - 0.3235 - 0.3244 - 0.3251 - 0.3123 - 0.3111 - 0.3119 (0.0905) (0.0907) (0.0906) (0.0912) (0.0913) (0.0911) Country - level Natural logarithm of mean 0.5898 0.0809 GDP per capita (0.1721) (0.1940) Mean HDI 5.6253 0.3195 (1.6226) (1.8792) Cross - leve l interaction terms Climate change risk perception X 0.1602 Natural logarithm of GDP per capita (0.0309) Climate change risk perception X 1.6547 Natural logarithm of GDP per capita (0.4446) Constant - 3.4433 - 9 .2831 - 8.1667 - 3.3164 - 4.1178 - 3.5954 (0.1441) (1.7924) (1.4380) (0.1391) (1.9442) (1.5850) Variance components Country - level standard deviation 0.5750 0.2925 0.3581 0.4801 0.4585 0.4777 (0.1218) (0.1016) (0.0835) (0.1641) (0.1597) (0.1625) Cl imate change risk perception standard 0.0427 0.0193 0.0215 D eviation (0.0133) (0.0079) (0.0073) N umber of respondents 37480 37480 37480 37480 37480 37480 Number of countries 30 30 30 30 30 30 A IC 28303 28285 28291 28238 28214 28221 Note s : Cell entries are logistic regression coefficients with the standard error in parenthesis for a random coefficient, random intercept model . Listwise deletion was used for missing values. Bold indicates two - tailed p <0.05. Israel and Taiwan were not include d in the models because respondents were not asked for their partisan identification in these two countries. 167 In Model s 1, 2, and 3 in Table 7 .4 show that countr y - level variation in choosing climate change as the most important environmental problem facin g their country is significant. Model 4 allows the coefficient estimate for perceived dangerousness of climate change to vary across countries in addition to the constant, and the variation of perceived dangerousness is statistically significant. 19 So the s trength of the relationship between the perceived dangerousness of climate change and answering that climate change is the most important environmental pr oblem varies across countries. Models 5 and 6 introduce cross - level interaction terms to capture the v ariation in influence the perceived dangerousness of climate change across values of country wealth and development. Figure 7 .6 shows the difference in predicted probability of considering climate change the most important environment problem between viewi ng influence of perceived dangerousness on ranking the environment important, this gap between those worried about climate change and not worried about c l imate change increases as country affluence increases about two percent when GDP per capita is about $3000. But when GDP per capita is about $36000, this predict ed difference i s about 13 percent. Climate change risk perception has a much stronger influence on ranking climate change the most important environmental problem in rich and developed countries than in poorer and less developed countries. Why is climate change risk perception more strongly related to ranking climate change as important in richer countries than in poorer countries? Individuals not w orried about climate change are ronment problem regardless of context. For those individuals who respond that climate change is probability for considering climate change the most important environmental problem facing their country is four percent wit h GDP per capita at about $3000 and only six percent with GDP per capita is $36000. This minor difference i s not statistically signi ficant. But among those who say climate change is 19 The AIC decreases by 65 decreased in Model 4 compare d by Model 1, suggesting that the random intercept and coefficient model provides a better fit than the random - intercept model. 168 the most importa nt environment problem increases from five percent with GDP per capita at $3000 to 20 percent with GDP per capita at $36000, a change of 15 percent. 20 Figure 7 .6: Conditional effect of perceived dangerousness of climate change on predicting considering the environment an important issue and climate change the most important environmental problem across values of mean GDP per capita (International Social Survey Programme 2009 - 2011) Notes: The first differences are calculated by holding educa tion at its female, 40 - 49 year - olds with no partisan identification and mean ed ucation. The first difference i s calculated: pr(very dangerous) pr(not very dangerous). The ribbon represents the 95 percent confidence interval. This section shows that cl im ate change risk perception has a varying influence across countries on ranking the environment and climate change as most important. Specifically, c limate change risk 20 To check the results of the cross - level interaction models, I estimated logistic regression models predicting considering climate change an important environmental problem for each of the 30 countries. The resu lts are presented in Appendix E , which plots the difference between very dangerous and not very dangerous in the predicted probability of considering climate change the most important environmental problem across values of HDI, a measure of wealth and development. The relationship is statistically significant for all 16 countries with a mean GDP per capita greater than $16000. And it the relationship is particularly stro ng in some wealthy countries. For example, an individual who says climate change is very dangerous has a predicted probability 22 percent and 21 percent g reater of ranking climate change as the most important environmental problem than an individual who s ays climate change i s not very dangerous in Japan and Norway, res pectively. The United States has one of the weakest associations of all the wealt hy countries. An American who says climate change is not very dangerous has a predicted probability o f ranking climate change as most important of less than one percent; the predicted pr obability for an American who says climate change i s very dangerous i s about four percent. 169 perception has a much stronger relationship with ranking the environment and climate chan ge as most important in rich and developed countries than in poorer and less developed countries. In poorer and less developed countries, citizens a re very unlikely to view the environment and climate change as more important than other issues, regardless of their level of concern ab out climate change. In Chapter 2, I show that many in developing countries say they are very worried about climate change, but this high level of worry might be misleading, especially in less wealthy countries. Citizens in these poorer countries, even if they respond that climate change is serious or dangerous, overwhelmingly care more about other issues and environmental problems. The relationship between climate change risk perception and ranking the environment and climate cha nge as most important is weak in the developing world. This finding is noteworthy considering climate change is likely to have disproportionate consequences for the developing countries, and publics in these countries do not seem to be putting the necessar y pressure on their governments to prepare. 7 .6 Discussion and c onclusion This chapter analyzes how individuals relate their climate change risk perception to their risk perception of other environmental issues and then to other attitudes and behaviors on climate change. On these questions, the American public closely mirrors the global public. First, this chapter shows that Americans and individuals around the world relate climate change to other environmental problems; individual s who think climate chang e is dangerous also think other environmental problems, such as air pollution are dangerous . And for citizens in most countries, the perceived dangerousness of climate change is strongly correlated with the perceived dangerousness of air pollution . I argue this finding presents a different path for policymakers seeking to address climate change. Instead of narrowly framing fighting climate change with a focus only on the mitigating the devastating consequences of climate change, they should frame climate ch ange as part of broader environ mentalism. As shown in Chapter 2 , individuals see air pollution as more dangerous and serious than climate change . People generally want to enjoy the environment around them and an environment that allows them to be healthy. Policymakers need to show how fighting climate change also helps individuals receive these benefits they can see and 170 feel, while also mitigating the damaging long - term consequenc es of climate change that they cannot see and feel. T his approach would be mor e effective communication strategy for policymakers seeking to rally the public. But this chapter also shows that perceived dangerousness of climate change is positively correlated with perceived dangerousness of nuclear power stations , even though nucle ar power could be an i mportant part of any policy response to reduce carbon emissions , both in America and cross - nationally. Specifical ly, most of those who consider climate change dangerous also consider nuclear power dangerous, so those most worried abou t cl imate change are also those most likely to reject any expansion of nuclear power. This part of the public could be a major obstacle, and they would need to be convinced that nuclear power is not dangerous. Next, this chapter explores the relationship between climate change risk perception and arguably more important attitudes and behaviors: willingness to pay to address climate change, ranking the environment as one of the two most important national issues, and ranking climate change as the most impor tant national environmental problem . In both the Un ited States and globally, among those who see climate change as a serious risk, there is a substantial minority that was not willing to pay higher prices , and large majorities consider issues other than th e environment and environmental problems other than climate change more important. This is alarming. Assuming that citizens who tell a survey interviewer that climate change is dangerous and serious deeply care about climate change are ready to accept sacr ifices is at best problematic . For many of these so - called concerned citizens, clim ate change is not a pressing issue , but instead a passive one . This also shows tha t the heated rhetoric coming from climate change activists might not even be convincing the ir own side. Scholars of American public opinion on climate change have argued that the impassionate rhetoric has done litt le to persuade those not worried about climate change ( Guber 2013 , Ansolabehere and Konisky 2014 ). This chapter also shows that the rhetorical app roach of those passionate about acting on climate change might not be convincing those already on their side, both in the Un ited States and globally . There is a small group in the United States and around the world that passionately cares about climate cha nge. The problem is that this group is not 171 big enough to influence policy. There is a larger group of people, who are concerned about climate change, but unwilling to go beyond expressing that it is a serious or dangerous problem . Climate change activists need a n ew communication approach , focusing on convincing t hese concerned but not alarmed individuals. Finally, this chapter shows that cl imate change risk perception has a weak relationship with these attitudes and intended behaviors in less affluent and developing countries. I n these countries, citizens are highly unlikely to consider the environment and climate change most important, regardless of their level of worry. So t he group of individuals who say climate change is dangerous but unwilling to t ake the next step is even larger in these relatively poorer countries than in rich countries. This findin g should is worrying, as the public in these countries might not support needed policy changes to address the consequences of climate change. 172 Chapte r 8 : Conclusion 8 .1 Introduction If climate change is the most important public policy challenge facing the world in the 21 st century, then understanding what the public, not only in the United States but also around the world, thinks about climate change is valuable. Public attitudes and policy preferences about climate change will shape domestic and international policy responses. Yet as this project shows through focusing on specifically on public attitudes on climate change, the state of cross - national public opinion on climate change is complex. What the global public thinks about climate change is not clear because of its multifaceted nature. Cross - national public opinion on climate change may both help and hinder a meaningful policy response. 8 .2 Obs tacles to p olicy How does global public opinion on climate change serve as an obstacle to climate change policy? I argue there are three primary barriers: the publics in the United States and other countries do not consider climate change an important iss ue or problem, citizens in wealthier countries are not more concerned about climate change or more willing to make economic sacrifices for climate change than citizens of poorer countries, and specifically many on the American right do not consider climate change serious and dangerous. Even though much of the world express es that climate change is serious and dangerous in the abstract, the global public considers other issues and environmental problems as more important . People around the world rank the eco nomy and health care as more important than the environment and environmental problems such as air and water pollution as more important than climate chang e, as demonstrated in Chapters 2 and 7. Even among those who say climate change is dangerous in the a bstract, only a small minority rank the environment and climate change as important. And citizens of developing countries are particularly likely to focus on other issues, as cl imate change risk perception has a weak relationship with rank ing climate chang e as important. 173 public cares more ab out. Neglecting climate change does not seem likely to alienate the public. In a likely international policy response, the wealthy countries of Western Europe and North America would have to lead and take on a greater share of the economic costs. These countries are the most powerful actors in international politics. Wealthy countries are also the greatest emitters of carbon d ioxide emissions, benefited most from fossil fuels in developing their affluent economies, yet developing countries figure to be most at risk for the damaging consequences of climate cha nge. But, as shown in Chapters 3 and 7 , cit izens in affluent countries are not more or less l ikely to say climate change is serious or more willing to pay higher prices to address climate change than citizens in dev eloping countries. And Chapter 5 demonstrates that the strength of left - right political orientation increases a s country affluence increases. In wealthy countries, lef t - right political orientation has a significant effect on perceived seriousness of climate change. So climate change has become more closely associated with political and economic ideology in wealthy countries . The United States is the second largest emitter of absolute carbon dioxide emissions, behind China, and second largest emitter of carbon dioxide emissions per capita behind Canada. In an international policy response that meaningfully addresses climate change, the United States will have to reduce its carbon dioxide emissions. Yet, most of the American right, both elites and the mass public, will not support such a step. Prominent American Republicans reject climate science and consider climate c hange unserious, and most o f the like - minded public seems to be using Republican elites as a cue . As Chapter 6 shows , the more engaged conservatives are in the United States, the less serious they consider climate change and the less likely they are to acc ept that human behavior causes climate change . Most conservatives in the United States will reject any policy designed to address climate change. Even though President Obama may have legal authority to reduce carbon dioxide emissions from power plants, his ability to go further is limited by conservative and Republican strongly - held opinions on climate change. It is hard to 174 imagine a major international agreement and policy response in which the United States plays a leading role. P ublic opinion on the righ t in the United States is a major obstacle. 8 .3 Aides to p olicy Even though public opinion on climate change is complicated, the public may also serve as a constructive force for a meaningful policy response. How? I argue tha t following three findings hel p show that the public would be supportive of meaningful policy: most people around the world are worried about climate change, greater engagement is associated with greater worry about climate change cross - nationally , and in the United States, Americans w ho say climate change is not serious or dangerous and reject human - caused climate change are in the minority . Across all cross - national surveys, an overwhelming majority of people across dozens of countries expressed that c limate change is serious and dang ero us in the abstract. As Chapter 2 demonstrated, across four Pew Global Attitudes surveys from 2007 to 2010, at least 87 p ercent of all respondents call c rcent call from 2009 to 2011. And 90 percent of respondents in the Internationa l Social Survey Programme call each country included in these surve ys large majorities say climate change is serious or dangerous . Only a small fraction of respondents in thes e cross - national surveys say climate change is not serious or dangerous . Even though climate change is not an everyda y, immediate problem for most people, most still express worry. Although the public might not demand a substantial policy response, they would likely suppo rt meaningful policy changes. Individuals most likely to be engaged with climate change are also the most likely to think climate change is serious and rank it as the most important envi ronmental problem. As Chapters 3 and 7 demonstrate , those with greater education are more likely than those with less education to say climate change is a serious problem, . And self - described 175 interest in politics increases perceived seriousness of climate change. Those with more education and greater interest in politics are probably the most likely to be closely following climate change. Those most engaged with climate change are a natural base of public support for policy change, and across many countries, those most engaged are most likely t o view climate change as important and concerning. The major exception, of course, is the United States. 96 Although the American public clearly stands out for its lack of concern, those who say climate change is not serious or concerning are in the minority in the United States. Less than t hree - in - ten Americans consider climate change not serious or dangerous in these cross - national surveys. And although conservatives are far less likely than liberals to view climate change as serious or dangerous, conservat ive opinions on climate change are not unifo rm. Many Americans who describe themselves as somewhat conservative express that climate change is ser ious, and moderate Americans are even more likely to say climate change is serious . An alliance of liberals, m oderates, and moderate conservatives might push a policy response in the future in the United States. And a response from the United State could ignite international policy changes. 8 .4 A Gener al policy response the global public would s upport Broadly, h ow should policymakers and climate change activists frame a policy response for the public? The international public does not seem to be strongly calling for policy change, so designing a public response to maximize public support around the world is impor tant. Politicians in democracies should be especially aware of public opinion on climate change. This project has general insights into what factors should play an important role in shaping domestic and international policy. First, policy should be tied broadly to environmentalism and no t specifically to climate change. People say other environmental problems, including local air and water pollution, are more serious and dangerous than about climate change. Moreover, as shown in Chapter 7 , the perceived d angerousness of 96 Researc h, such as Chapter 6 , has shown that the effect of engagement is conditional on ideology and partisanship in the United States. For liberals and Democrats, engagement increases concern and also the probability of accepting climate change is the result of human activity. For conservatives and Republicans, engagement decreases concern and the probability of accepting climate change is the result of human activity. 176 climate change is closely related to the perceived dangerousness of air and water pollution. Obviously a policy response to mitigate the consequences of climate change needs to focus on reducing carbon dioxide emissions ( Nordhaus 2013 ). But this policy response can also include other pos itive externalities for the environment. The policy response should not only address demands for improved local air and water quality, but also be framed so these improvements are consequential. For example, any policy response to climate change requires r educing carbon dioxide emissions from power plants, as power plants are a major source of carbon dioxide emissions. This r equirement, no matter what the type of policy, puts increased costs on power plants, which they then pass onto consumers. Instead of f raming these policy changes for power plants as designed to solely address climate change, policymakers and activists should also highlight how these changes would improve the local environment around the power plant too. 97 Consumers might believe they are receiving increased benefits for the increased costs. Policy changes to address climate change have a host of positive externalities for the environment. These externalities are important to the public and should not be ignored and deemphasized . Second, a vast majority of people across many countries signaled that they would not support any policy response that expands the use of nuclear power. Across 32 countries in the ISSP, about 85 percent Moreover, as Chapter 7 demonstrates , perceived dangerousness of cl imate change is positive related to perceived dangerousness of nuclear power plants; those who say climate change is dangerous are also likely to say nuclear power plants are dangerous . So any policy response will have to rely on the expansion of energy so urces other than nuclear power which do not produce carbon dioxide emissions, like wind and solar. Wind and solar are probably more popular , but also probably more difficult to develop on a large - scale. 97 Ansolabehere and Konisky ( 2014 ) argue that Americans are more support ive of government regulation to reduce carbon emissions than the so - called cap - and - trade plan , becaus e they associate government regulation with previous efforts to improve environmental quality, while cap - and - trade is associated only with climate change. 177 Third, cross - nationally, most likely people will be sensitive to prices and reluctant to accept major economic sacrifices to address climate change. The economy is the primary issue of consequence aro und the world. And, as Chapter 7 shows , many of those w ho consider climate change serious are unwilling to a gree to pay higher prices when answering a survey question. This is a problem because the energy technologies, such as wind, solar, and nuclear, which do not emit carbon dioxide emissions, are a lso likely to be more expensive than fossil fuels. How much of an economic sacrifice the public will accept is unclear and more research is needed in this area. In addition, policymakers will need to design ways to reduce costs for consumers, and technology will need to be developed and expanded to make clean energy sources more affordable. 8 .5 Future r esearch Although this project tackles many questions of public opinion on climate change, there were still questions left unanswered that could be tackled in future projects. Here I highlight two potential ave nues for return to build on this project. First, this project focused on substantive attitudes about climate change, such as risk perception, and largely ignored policy preferences. In this conclusion I infer what type of policy responses people would sup port through their substantive attitudes, but researchers should increasingly ask respondents about their policy preferences toward climate change and reducing carbon emissions . Climate change policies are even more complex and abstract that the problem of climate change itself, but even weakly formed policy preferenc es are important . After all, substantive attitudes, such as considering climate change serious or important, might not translate to support for certain policy preferences, like cap - and - trade or increased regulation. It would be helpful to know two dimensions of public support for policy. First, what type of policies would the public support? Policy responses include both mitigation policies designed to curb carbon emissions and adaptation polici es designed to deal with climate change. Mitigation policy could include stick - like measures such as regulation and taxes designed to make carbon emissions more expensive. Does the public support these measures? Mitigation policy could also include carrot approaches, such as tax breaks for energy sources that do not emit carbon emissions, such as water, 178 wind, solar, and nuclear, so - called clean energy industries. Does the public support these measures? How does the public evaluate and balance stick and carr ot approaches to mitigation policy? Second, any climate change policy will include increased costs for the public. What specific increased cost, if any, would citizens be willing to pay? Although scholars of public opinion on climate change in the United S tates are starting to focus on policy preferences toward climate change ( Krosnick and MacInnis 2013 ), researchers have not focused on policy preferences in cross - national surveys. Considering climate change requires an international policy response, it is important to know specifically what people around the world think of policy options. Second, this project and others have looked at factors that condition the relationship of ideology and partisanship with views on climate change. This project specifically looks at measures of engagement, such as knowledg e and interest, while others have examined education ( Hamilton 2011 , McCright 2011 ) and income ( Bohr 2014 ). I also compare this conditional relationship in the United States to other countries. But with few exceptions (see Kahan, et al. 2012 ), researchers have not explored the conditional and unconditi onal effect of science knowledge on left - right political orientation in the United States and cross - nationally. 98 Engagement with politics and engagement with science are different, so they might not even be related. Does science knowledge increase concern about climate change cross - nationally? Does it have a conditional effect, like political engagement, in the United States? Does a conditional effect exist in other countries? This type of research would improve understanding as to how people form opinions on complex science and environment issues in the United States and around the world. 8 .6 Public opinion and p olicy In January 2012, United States President Barack Obama called for action on climate change in his second inaugural address. 98 Kahan and his co - authors ( 2012 ) focus on the interaction between science knowledge and values, not science knowledge and partisanship/ideology, in the United States. 179 to the threat of climate change, knowing that the failure to do so would betray our children and future generations. Some may still deny the overwhelming judgment of science, but none can avoid the devastating impact of raging fires, and crippling drought , and more powerful storms. The path towards sustainable energy sources will be long and sometimes difficult. But America cannot resist this transition; But President Obama and other political leaders cannot go it alone. As V.O. Key obse rved more than 50 years ago, in a democracy like the United States, how and if w e respond to climate change is determined by public opinion. So understanding how the public, in the United States and around the world, thinks about climate change is importan t to understanding what could be the one of the most important policy challenge of the 21 st century. This project provides a detailed description a nd explanation of global public opinion on this challenge . 1 80 APPENDICIES 181 Appendix A : Data, survey que stions, measurement, and v ariables Data Sources This project draws from multiple public opinion surveys, conducted in multiple countries and in just the United States. Here I briefly describe each survey and also how the surveys are sued throughout the p roject. World Values Survey (2005 - 2009) (WVS) The World Values Survey (WVS) is cross - national survey that studies the beliefs, values, and opinions of people around the world. The WVS conducts surveys in a wide range of countries at every level of develo pment in every region of the world. The sample for each country designed to be nationally representative of the population 18 years and older. Face - to - face interviews are the primary mode of data collection, although phone interviews are sometimes used for remote areas. The minimum sample size is 1000. Weights based on demographics are used in some countries. This project draws from wave 5 of the WVS, conducted between 2005 and 2009. Wave 5 asked respondents in 44 countries for the perceived seriousness o f climate change. Table A1.1: Participating countries and sample i nformation for the World Values Survey (2005 - 2009) Participating country Fieldwork date Sample size Nationally representative sample? Availability of survey weights? Argentina June 24 J uly 20, 2007 1002 Yes Yes Australia September 14 - December 21, 2005 1421 Yes No Brazil November 1 December 26, 2006 1500 Yes Yes Bulgaria May 2007 1001 Yes Yes Burkina Faso March 16 - 25, 2007 1534 Yes Yes Canada February 14 April 6, 2006 2164 Yes Ye s Chile June 14 - 24, 2006 1000 Yes Yes China March 25 May 10, 2007 2015 A dults between the ages of 18 and 70 Yes 182 Cyprus February March, 2006 1050 No; adults between the ages of 18 and 70 Yes Egypt March 15 April 5, 2008 3015 Yes Yes Ethiopia March 26 April 6, 2008 1500 Yes Yes Finland August 28 October 10, 2005 1014 Yes Yes Georgia January 30 February 10, 2009 1500 Yes No Germany May 2 June 21, 2006 2064 Yes Yes Ghana February 19 April 4, 2007 1534 Yes Yes Ind ia December 2006 January 2007 2001 Yes; 18 of 28 states accounting for 97 percent of population No Indonesia June October 2006 2015 Yes Yes Italy May 10 October 20, 2005 1012 Yes; adults between the ages of 18 and 75 Yes Japan July 7 August 5, 2005 1096 Yes No Jordan May June 2007 1200 Yes Yes Malaysia September 20 November 15, 2006 1201 Yes No Mali March 16 March 22, 2007 1534 Yes Yes Mexico November December, 2005 1560 Yes Yes Moldova November 18 - 30, 2006 1046 Yes No 183 Table A1.1 ( Morocco September 15 October 7, 2007 1200 Yes Yes Norway March 5 May 31, 2007 1025 Yes No Peru December 7 - 15, 2006 1500 Yes Yes Poland December 2005 1000 Yes Yes Romania September 15 October 15, 2005 1776 Yes No Rwanda June 9 July 2, 2007 1507 Yes Yes Slovenia October November 2005 1037 Yes No South Africa November 22 December 20, 2006 2988 Yes Yes South Korea December 1 - 15, 2005 1200 Yes (20 years and older instead of 18) Yes Spain July 10 - 24, 2007 1200 Yes Yes Sweden Novembe r 23, 2005 February 17, 2006 1003 Yes Yes Switzerland April August 2007 1241 Yes, adults between the ages of 18 and 85 No Thailand June July 2007 1534 Yes No Trinidad September 14 - November 8, 2006 1002 Yes Yes Turkey January 28 March 5, 2007 1 346 Yes Yes Ukraine November 15 - 25, 2006 1000 Yes Yes United States September 19 - 29, 2006 1249 Yes Yes 184 Uruguay October 27 November 21, 2006 1000 Yes No Vietnam October 1 November 30, 2006 1495 Yes No Zambia March 30 April 14, 2007 1500 Yes Yes International Social Science Programme (ISSP) (2009 - 2011) A program of cross - national collaboration on surveys related to important topics in social science research. In 2009 - 2011, a module of questions about the environment was asked in participating countries. Some countries include survey weights so the sample is representative of the population. Countries with samples which are not weighted to the population are Argentina, Croatia, Denmark, Israel, Japan, Latvia, New Zealand, Norw ay, Slovenia, South Korea, Sweden, Switzerland, and Turkey. Countries which include survey weights are Austria, Belgium, Bulgaria, Canada, Chile, Czech Republic, Finland, France, Germany, Great Britain, Lithuania, Mexico, Philippines, Russia, Slovak Republ ic, South Africa, Spain, and the United States. All samples were designed to be nationally representative of adult citizens. All countries have a sample size of at least 928. Table A1.2 provide more information on the 2009 - 2011 ISSP. Table A1.2: Part icipa ting countries and sample i nformation for the International Social Science Programme Survey (2009 - 2011) Participating country Fieldwork date Sample size Nationally representative sample? Availability of survey weights? Argentina 2010 1130 Yes No Austria 2010 1019 Yes Yes Belgium 2010 1142 Yes Yes Bulgaria 2011 1003 Yes Yes Canada 2011 985 Yes Yes Chile 2010 1436 Yes Yes Croatia 2011 1210 Yes No Czech Republic 2010 1428 Yes Yes Denmark 2010/2011 1305 Yes No Finland 2010/2011 1211 Yes No 185 Table A1.2 France 2010 2253 Yes Yes Germany 2010 1407 Yes Yes Great Britain 2010 928 Yes Yes Israel 2011 1216 Yes No Japan 2010 1307 Yes No Latvia 2011 1000 Yes No Lithuania 2010/2011 1023 Yes Yes Mexico 2011 1637 Yes Yes New Zealand 2010 1172 Yes N o Norway 2011 1382 Yes No Philippines 2009/2010 1200 Yes Yes Russia 2010 1619 Yes Yes Slovak Republic 2009 1159 Yes Yes Slovenia 2011 1082 Yes No South Africa 2010 3112 Yes Yes South Korea 2010 1576 Yes No Spain 2010 2560 Yes Yes Sweden 2010 1181 Yes No Switzerland 2011 1212 Yes No Taiwan 2010 2209 Yes Yes Turkey 2010 1665 Yes No United States 2010 1430 Yes Yes Pew Global Attitudes Survey 2007 The Pew Global Attitudes Survey was administered in 47 countries in 2007. The survey was administer ed through telephone and face - to - face interviews. All of the country surveys were based on national adult samples except for Bolivia, Brazil, China, India, Ivory Coast, Pakistan, South Africa, and Venezuela. In these countries, the samples were disproport ionately or exclusively urban. African countries were not included in the analysis because respondents in these countries were not asked about the level of seriousness of climate change. 186 Table A1.3 : Participating countries and sample i nformation for the P ew Global Attitudes Survey (2007) Participating country Fieldwork date Sample size Nationally representative sample? Argentina April 13 - 23, 2007 800 Yes Bangladesh April 11 - 30, 2007 1000 Yes Bolivia April 14 - May 1, 2007 834 No, the sample was disproport ionally urban Brazil April 12 - May 5, 2007 1000 No, the sample was disproportionally urban Britain April 21 - May 6, 2007 1002 Yes, representative of households with telephones Bulgaria April 13 - May 7, 2007 500 Yes Canada April 16 - April 26, 2007 1004 Yes , representative of households with telephones Chile April 18 - 27, 2007 800 Yes China April 20 - 30, 2007 3142 No, the sample was disproportionally urban Czech Republic April 11 - May 4, 2007 900 Yes, representative of households with telephones Egypt April 9 - May 7, 2007 1000 Yes Ethiopia* April 27 - May7, 2007 710 Yes, except for area near Somali border 187 France April 13 - 18, 2007 1004 Yes, representative of households with telephones Germany April 16 - 30, 2007 1000 Yes, representative of households with telephones Ghana* April 25 - May 3, 2007 707 Yes India April 20 - May 17, 2007 2043 No, the sample was disproportionally urban Indonesia April 18 - 28, 2007 1008 Yes, except for extremely remote areas Israel April 20 - May 11, 2007 900 Yes Italy April 18 - May 23, 2007 501 Yes Ivory Coast* April 12 - 16, 2007 700 No, the sample was disproportionally urban Japan April 6 - May 23, 2007 762 Yes Jordan April 9 - May 7, 2007 1000 Yes Kenya* April 20 - 30, 2007 1000 Yes Kuwait April 15 - May 10, 2007 500 Yes Lebanon April 9 - May7, 2007 1000 Yes Malaysia April 13 May 9, 2007 700 Yes, except for extremely remote areas Mali* April 7 April 18, 2007 700 Yes 188 Mexico April 13 - 27, 2007 828 Yes Morocco April 20 - May10, 2007 1000 Yes Nige ria* April 23 - May 29, 2007 1128 Yes Pakistan April 18 May 10, 2007 2008 No, the sample was disproportionally urban and excluded areas of instability Palestinian Territories April 21 - 30, 2007 808 Yes Peru April 13 - 29, 2007 800 Yes Poland April 12 - 26, 2007 504 Yes Russia April 10 - 24, 2007 1002 Yes Senegal* April 14 - 19, 2007 700 Yes Slovakia April 11 May 6, 2007 900 Yes, representative of households with telephones South Africa* April 20 May 20, 2007 1000 No, exclusively urban South Korea April 9 - 24, 2007 718 Yes Spain April 18 May 15, 2007 500 Yes Sweden April 18 May 9, 2007 1000 Yes, representative of households with telephones Tanzania* April 21 May 14, 2007 704 Yes Turkey April 10 May 3, 2007 971 Yes Uganda* April 15 - 24, 2007 112 2 Yes Ukraine April 13 - 24, 2007 500 Yes 189 United States April 23 May 6, 2007 2026 Yes, representative of households with telephones in continental U.S. Venezuela April 22 May 21, 2007 803 No, the sample was disproportionally urban On another topic, in your view is global warming a very serious problem, somewhat serious, not too not asked in these countries. Pew Global Attitudes Survey 2008 Table A1.4 : Participating countries and sample i nformation for the Pew Global Attitudes Survey (2008) Participating country Fieldwork date Sample size Nationally representative sample? Argentina March 25 April 3, 2008 801 Yes Australia March 20 April 4, 2008 700 Yes Brazil March 20 April 8, 2008 1000 No, the sample was disproportionally urban Britain March 17 April 6, 2008 753 Yes, representative of households with telephones China March 28 April 19, 2008 3212 No, the sample was disproportionally urban Egypt March 19 April 7, 2008 1 000 Yes France March 31 April 8, 2008 754 Yes, representative of households with telephones 190 Germany March 25 April 9, 2008 750 Yes, representative of households with telephones India April 1 - 16, 2008 2056 No, the sample was dis proportionally urban Indonesia March 29 April 14, 2008 1000 Yes, except for extremely remote areas Japan March 19 April 13, 2008 708 Yes Jordan March 18 April 6, 2008 1000 Yes Lebanon March 19 April 7, 2008 1000 Yes Mexico March 18 - 31, 2008 8 05 Yes Nigeria April 8 - 21, 2008 1000 Yes Pakistan April 1 - 19, 2008 1254 No, the sample was disproportionally urban and did not include areas of instability Poland March 26 April 14, 2008 750 Yes Russia March 18 April 4, 2008 1000 Yes South Africa March 18 April 4, 2008 1001 Yes South Korea March 20 - 27, 2008 714 Yes Spain March 17 April 17, 2008 752 Yes Tanzania March 31 April 16, 2008 704 Yes 191 Turkey March 31 April 21, 2008 1003 Yes United States April 9 - 17, 2008 10 00 Yes, representative of households with telephones in continental U.S. Pew Global Attitudes Survey 2009 Table A1.5 : Par ticipating countries and sample i nformation for the Pew Global Attitudes Survey (2009) Participating country Fieldwork date Sample s ize Nationally representative sample? Argentina May 18 June 3, 2009 800 Yes Brazil May 18 June 14, 2009 813 No, the sample was disproportionally urban Britain May 26 June 9, 2009 754 Yes, representative of households with telephones Canada May 19 June 5, 2009 750 Yes, representative of households with telephones Chine May 22 June 10, 2009 3169 No, the sample was disproportionally urban Egypt May 24 June 11, 2009 1000 Yes France May 29 June 9, 2009 753 Yes, representative of households w ith telephones Germany May 25 June 6, 2009 751 Yes, representative of households with telephones 192 India May 22 June 13, 2009 2038 No, the sample was disproportionally urban Indonesia May 29 June 5, 2009 100 Yes, except for ext remely remote areas Israel May 18 June 16, 2009 1201 Yes Japan May 20 June 10, 2009 700 Yes, representative of households with telephones Jordan May 24 June 11, 2009 1000 Yes Kenya May 22 - 30, 2009 1002 Yes Lebanon May 20 June 3, 2009 1000 Yes Mexico May 26 June 2, 2009 1000 Yes Nigeria June 2 - 14, 2009 1000 Yes Pakistan May 22 - June 9, 2009 1254 No, the sample was disproportionally urban and did not include areas of instability Palestinian territories May 23 June 11, 2009 1204 Yes Pol and May 21 June 8, 2009 750 Yes Russia May 20 June 9, 2009 1001 Yes South Korea May 25 June 8, 2009 750 Yes Spain May 25 June 9, 2009 750 Yes, representative of households with telephones 193 Turkey May 20 June 15, 2009 1005 Y es United States May 27 June 10, 2009 1000 Yes, representative of households with telephones in continental U.S. (including cell phone only households). Pew Global Attitudes Survey 2010 Table A1.6 : Participating countries and sample i nformation for t he Pew Global Attitudes Survey (2010) Participating country Fieldwork date Sample size Nationally representative sample? Argentina April 13 May 4, 2010 803 Yes Brazil April 10 May 6, 2010 1000 Yes Britain April 15 May 2, 2010 750 Yes, representati ve of households with telephones (including cell phone only households). China April 9 - 20, 2010 3262 No, the sample was disproportionally urban Egypt April 12 May 3, 2010 1000 Yes, except for extremely remote areas France April 15 - 23, 2010 752 Yes, re presentative of households with telephones (including cell phone only households). 194 Germany April 15 - 30, 2010 750 Yes, representative of households with telephones (excluding cell phone only households) India April 9 - 30, 2010 2254 No, the sample was disproportionally urban Indonesia April 16 - 29, 2010 1000 Yes, except for extremely remote areas Japan April 9 - 26, 2010 700 Yes, representative of households with telephones (excluding cell phone only households) Jordan April 12 May 3, 2010 1000 Yes Kenya April 9 - 23, 2010 1002 Yes Lebanon April 12 May 3, 2010 1000 Yes Mexico April 14 - 20, May 1 - 6, 2010 1300 Yes Nigeria April 18 May 7, 2010 1000 Yes Pakistan April 13 - 28, 2010 2000 No, the sample was disproportionally urban Polan d April 9 May 8, 2010 750 Yes Russia April 7 May 1, 2010 1001 Yes South Korea April 11 - 23, 2010 706 Yes 195 Spain April 14 May 4, 2010 755 Yes Turkey April 12 April 30, 2010 1003 Yes United States April 15 May 5, 2010 1002 Y es, representative of households with telephones (including cell phone only households). Chapter 1 How Important Do People Consider Climate C hange? Pew Global Attitudes Survey 2007 Pew Global Attitudes Survey 2008 Pew Global Attitudes Survey 2009 Pew Global Attitudes Survey 2010 World Values Survey (2005 - 2009) (WVS) International Social Science Programme (2009 - 2010) (ISSP) Gallup Environment Polls (1997, 1999 - 2004, 2006 - 2013) Pew Center for the People and Press Surveys (2006 - 2013) Chapter 3 What Ex plains Cross - National Climate Change Risk Perception? World Values Survey (2005 - 2009) (WVS) Chapter 4 Where Does Left - Right Political Orientation Matter Across Countries? World Values Survey (2005 - 2009) (WVS) Chapter 5 Why is the American Public Fa r More Polarized on Climate Change than Other Wealthy Countries? World Values Survey (2005 - 2009) (WVS) American National Elections Study 2008 (ANES) American National Elections Study 2012 (ANES) Chapter 6 How Do People Think A bou t Climate C hange? Pew Global Attitudes Survey 2009 Pew Global Attitudes Survey 2010 International Social Science Programme (2009 - 2011) (ISSP) 196 Survey questions, measurement, and v ariables for all variables Table A1.7: Chapter 1 Survey questions, measurements, and variabl es Variable Name Description/Question Gallup Surveys (1997 - 2013) Worry about climate change please tell me if you personally worry about this problem a great deal, a fair amount , only a little, or not at all. How much do you personally worry about global warming? Not at all; only a little; a fair amount; a great deal. Int ernational Social Survey Programme (2009 - 2010) Dangerousness of climate change climate change is extremely dangerous, very dangerous, somewhat 0=not dangerous at all; 0.25=not very dangerous; 0.5=somewhat dang erous; 0.75=very dangerous; 1=extremely dangerous. Dangerousness of a ir pollution from cars In general, do you think that air pollution caused by cars is extremely dangerous, very dangerous, somewhat dangerous, not very dangerous, or not dangerous at all dangerous; 0.5=somewhat dangerous; 0.75=very dangerous; 1=extremely dangerous. Dangerousness of a ir pollution from industry In general do you think that air pollution caused by industry is extr emely dangerous, very dangerous, somewhat dangerous, not very dangerous, or not dangerous; 0.5=somewhat dangerous; 0.75=very dangerous; 1=extremely dangerous. Dangerousness of p o llution of water is extremely dangerous, very dangerous, somewhat dangerous, not very all; 0.25=not very da ngerous; 0.5=somewhat dangerous; 0.75=very dangerous; 1=extremely dangerous. 197 Dangerousness of p esticides and chemicals in farming And do you think that the pesticides and chemicals used in farming are extremely dangerous, very dang erous, somewhat dangerous, not very all; 0.25=not very dangerous; 0.5=somewhat dangerous; 0.75=very dangerous; 1=extremely dangerous. Dangerousness of nuclear power stations And do you think that nuclear power stations are extremely dangerous, very dangerous, somewhat dangerous, not very dangerous, or not dangerous at all 0.5=somewhat dangerous; 0.75=very dange rous; 1=extremely dangerous. Dangerousness of m odified genetic crops risk perception And do you think that modifying the genes in certain crops is extremely dangerous, very dangerous, somewhat dangerous, not very dangerous, or not dangerous at all for the dangerous; 0.5=somewhat dangerous; 0.75=very dangerous; 1=extremely dangerous. Most important issue Which of these issues is the most important for [COUNTRY] today? Health care; education; crime; the en vironment; immigration; the economy; Next most important issue Which is the next most important (issue)? Health care; education; crime; the environment; immigration; the economy; terrorism; poverty; none of Most important environment problem country Here is a list of some different environmental problems. Which problem, if any, do you think is the most important for [COUNTRY] as a whole? Air pollution; chemicals and pesticides; water sh ortage; water pollution; nuclear waste; domestic waste disposal; climate change; genetically modified crops; Most important environmental problem family Here is a list of some different enviro nmental problems. Which problem, if any, affects you and your family the most? Air pollution; chemicals and pesticides; water shortage; water pollution; nuclear waste; domestic waste disposal; climate change; genetically modified crops; using up our natura l Pew Research Center Global Attitudes Surveys (2007 - 2010) Seriousness of c limate change (2007 - 2009) On another topic, in your view is global warming a very serious problem, somewhat serious, not too serious, or n ot a problem ? 0=not a problem; 0.33=not too serious; 0.67=somewhat serious; 1=very serious Seriousness of climate change (2010) On another topic, in your view global climate change a very serious problem, somewhat serious, not too serious, or not a proble m? 0=not a problem; 0.33=not too serious; 0.67=somewhat serious; 1=very serious 198 Willingness to pay to address climate change (2009 - 2010) Please tell me whether you agree or disagree with the following statement. ng to pay higher prices in order to address global Pew Research Center for the People and the Press Surveys (2006 - 2013) Seriousness of climate change In your view, is global warming a very serious problem, somewhat serious; not too serious; or not a problem? Not a problem; not too serious; somewhat serious; very serious. World Values Survey (2005 - 2009) Seriousness of climate change ase tell me how serious you consider each of the following. Is it very serious, somewhat serious, not very serious, or not serious at all? Global warming or greenhouse effect. 0=not serious at all; 0.33=not very serious; 0.67=somewhat serious; 1=very serio us. Seriousness of loss of species/biodiversity tell me how serious you consider each of the following. Is it very serious, somewhat serious, not very serious, or not serious at all ? Loss of plant or animal species or biodiversity. 0=not serious at all; 0.33=not very serious; 0.67=somewhat serious; 1=very serious. Seriousness of pollution of water tell me how serious you consider each of the following. Is it very serious, somewhat serious, not very serious, or not serious at all? Pollution of rivers, lakes and streams. 0=not serious at all; 0.33=not very serious; 0.67=somewhat serious; 1=very serious. 199 Table A1.8: Chapter 3 Survey questions, measurements, and variables Variable Name Description/Question Minimum Value Maximum Variable World Values Survey Variables Seriousness of climate change s a whole. Please tell me how serious you consider each of the following. Is it very serious, somewhat serious, not very serious, or not serious at all? Global warming or greenhouse effect? 1=not serious at all; 2=not very serious; 3=somewhat serious; 4=ve ry serious 1 4 Post - materialism 12 item index: People sometimes talk about what the aims of this country should be for the next ten years. On his card are listed some of the goals which different people give top priority. Would you please say which one of these you, yourself consider the most important? And which would be the next most important? First list : a high level of economic growth; making sure this country has strong defense forces; seeing that people have more say about how things are done a t their jobs and in their communities; trying to make our cities and countryside more beautiful. Second list: maintaining order in the nation; giving people more say in important government decisions; fighting rising prices; protecting freedom of speech Third list: a stable economy; progress toward a less impersonal and more humane society; progress toward a society in which ideas count more than money; the fight against crime - 2.5 2.5 200 Political ideology /left - right political orient ation generally speaking? 1=far left; 10=far right 1 10 Right Self - positioning on the above left - right continuum : 7 - 10 0 1 Left Self - position ing on the above left - right continuum: 1 - 4 0 1 Social issues index Please tell me for each of the following actions whether you think it can always be justified, never be justified, or something in between. - 4.5=never justifiable; 4.5=always justifiable H omosexuality Abortion Divorce Euthanasia - 4.5 4.5 Income inequality How would you place your views on this scale? 1 means you agree completely with the statement on the left; 10 means you agree com pletely with the statement on the right; and if your view fall somewhere in between, you can choose any number in between. 1=incomes should be made more equal; 10=we need larger income differences as incentives for individual effort. Recoded to be zero - cen tered at the midpoint. - 4.5 - 4.5 Private enterprise How would you place your views on this scale? 1 means you agree completely with the statement on the left; 10 means you agree completely with the statement on the right; and if your view fall somewhere in between, you can choose any number in between. 1=private ownership of business and industry should be increased; 10=government ownership of business and industry should be increased. Values revers ed and recoded to be zero - centered. - 4.5 4.5 Interest in politics How interested would you say you are in politics? - 1.5=not at all interested, - 0.5=not very interested, 0.5= somewhat interested, 1.5=very interested - 1.5 1.5 201 Level of e ducation What is the highest education level that you have attained? - 3.5=no formal education - 2.5=incomplete primary school - 1.5=complete primary school - 0.5=incomplete secondary school: technical/vocational type 0.5=complete secondary school: techni cal/vocational type 1.5=incomplete secondary: university - preparatory type 2.5=complete secondary: university - preparatory type 3.5=university - level education - 3.5 3.5 Attendance of religious services Apart from weddings and funerals, about how often do yo u attend religious services these days? - 3=never practically never; - 2=less often; - 1=once a year; 0=only on special holy days; 1=once a month; 2=once a week; 3=more than once a week - 3 3 Importance of God is scale to - 5=not at all important; 5=very important. - 5 5 Protestant Do you belong to a religious denomination? 1=Protestant ; all other response categories. 0 1 Catholic Do you belong to a rel igious denomination? 1=Catholic ; 0=all other responses categories. 0 1 Evangelical Do you belong to a religious denomination? 1=Evangelical ; 0=all other responses categories. 0 1 Muslim Do you belong to a religious denomination? 1=Muslim ; 0=all other res ponses categories. 0 1 Lower class People sometimes describe themselves as belonging to the working class, the middle class, or the upper or lower class. Would you describe yourself as belonging to the 1=lower class; 0=all other response categories. 0 1 Working class People sometimes describe themselves as belonging to the working class, the middle class, or the upper or lower class. Would you describe yourself as belonging to the 1=working class; 0=all other response categories. 0 1 Middle class People sometimes describe themselves as belonging to the working class, the middle class, or the upper or lower class. 0 1 202 Would you describe yourself as belonging to the 1=lower middle class; 0=all other response categories. Upper class People sometimes describ e themselves as belonging to the working class, the middle class, or the upper or lower class. Would you describe yourself as belonging to the 1=upper middle class or upper class; 0=all other response categories. 0 1 Income decile of incomes on which 1 indicates the in your country. We would like to know in what group your household is. Please specify the appropriate number, counting all wages, salaries, pensions and other i ncomes 1 10 Female Respondent sex by observation. 1=female; 0=male. 0 1 Age younger than 30 Self - identified age of respondent in years. 1= 29 and younger; 0=all other ages. 0 1 Age 30 to 39 Self - identified age of respondent in years. 1= 30 to 39; 0=all other ages. 0 1 Age 40 to 49 Self - identified age of respondent in years. 1= 40 to 49; 0=all other ages. 0 1 Age 50 to 59 Self - identified age of respondent in years. 1= 50 to 59; 0=all other ages. 0 1 Age 60 to 69 Self - identifie d age of respondent in years. 1= 60 to 69; 0=all other ages. 0 1 Age 70 and older Self - identified age of respondent in years. 1=70 and older; 0=all other ages. 0 1 Country - level variables GDP per capita Natural logarithm of mean GDP per capita from 2 005 - 2009 Source : World Bank 5.56 11.28 HDI Mean Human Development Index score, 2005 - 2009. A composite measure of developing capturing educational attainment, life expectancy, and income. Source: United Nations Development Programme 99 0.31 0.95 99 In late 2013, the United Nations Development Prog ramme (UNDP) adjusted the calculation of previous HDI scores as a result of newly available data. The data used here was collected before this adjustment, so currently available HDI scores from the UNDP are slightly different than the scores used here. 203 Table A Oil rents per capita Natural logarithm of mean oil and gas exports per capita by country in 2009 dollars, 2005 - 2009. Source : Ross 0 9.95 CO 2 per capita Mean metric tons of CO 2 emissions per capita, 2005 - 2009 Source : World Bank 0.46 29.07 Polity 2005 - 2009. Source : Polity Project - 7.0 10 GINI Mean GINI coefficient measuring inequality from 2005 - 2009 Source: Solt 23.02 64.73 204 Table A1.9 Chapter 4 Survey questions, measurements, and variables Variable Name Description/Question Minimum Value Maximum Variable World Values Survey Variables Seriousness of climate change a whole. Please tell me how serious you consider each of the following. Is it very serious, somewhat serious, not very serious, or not serious at all? Global warming or greenhouse effect? 1=not serious at all; 2=not very serious; 3=somewhat serious; 4=very serious 1 4 Post - materialism 12 item index: Peop le sometimes talk about what the aims of this country should be for the next ten years. On his card are listed some of the goals which different people give top priority. Would you please say which one of these you, yourself consider the most important? An d which would be the next most important? First list : a high level of economic growth; making sure this country has strong defense forces; seeing that people have more say about how things are done at their jobs and in their communities; trying to make o ur cities and countryside more beautiful. Second list: maintaining order in the nation; giving people more say in important government decisions; fighting rising prices; protecting freedom of speech Third list: a stable economy; progress toward a less im personal and more humane society; progress toward a society in which ideas count more than money; the fight against crime - 2.5 2.5 205 Political ideology /left - right political orientation generally speaking? 1=far left; 10=far right 1 10 Interest in politics How interested would you say you are in politics? - 1.5=not at all interested, - 0.5=not very interested, 0.5= somewhat interested, 1.5=very interested - 1.5 1.5 Level of e ducation What is the highest education level that you have attained? - 3.5=no formal education - 2.5=incomplete primary school - 1.5=complete primary school - 0.5=incomplete secondary school: technical/vocati onal type 0.5=complete secondary school: technical/vocational type 1.5=incomplete secondary: university - preparatory type 2.5=complete secondary: university - preparatory type 3.5=university - level education - 3.5 3.5 Attendance of religious services Apart fr om weddings and funerals, about how often do you attend religious services these days? - 3=never practically never; - 2=less often; - 1=once a year; 0=only on special holy days; 1=once a month; 2=once a week; 3=more than once a week - 3 3 Importance of God ow important is God in your life? Please use this scale to - 5=not at all important; 5=very important. - 5 5 Protestant Do you belong to a religious denomination? 1=Protestant ; all other response c ategories. 0 1 Female Respondent sex by observation. 1=female; 0=male. 0 1 Age younger than 30 Self - identified age of respondent in years. 1= 29 and younger; 0=all other ages. 0 1 Age 30 to 39 Self - identified age of respondent in years. 1= 30 to 39; 0=all other ages. 0 1 Age 40 to 49 Self - identified age of respondent in years. 1= 40 to 49; 0=all other ages. 0 1 206 Age 50 to 59 Self - identified age of respondent in years. 1= 50 to 59; 0=all other ages. 0 1 Age 60 to 69 Self - id entified age of respondent in years. 1= 60 to 69; 0=all other ages. 0 1 Age 70 and older Self - identified age of respondent in years. 1=70 and older; 0=all other ages. 0 1 Country - level variables GDP per capita Natural logarithm of mean GDP per capita from 2005 - 2009 Source : World Bank 5.56 11.28 HDI Mean Human Development Index score, 2005 - 2009. A composite measure of developing capturing educational attainment, life expectancy, and income. Source: United Nations Development Programme 100 0.31 0.95 OECD country Member of the Organization for Economic Cooperation and Development (OECD) or not. 0 1 Post - communist country Experienced a transition from communism in the early 1990s or not. 0 1 Table A1.10: Chapter 5 - Survey questions, measurements, a nd variables Variable Name Description/Question Minimum Value Maximum Variable American National Elections Study 2008 Concern about climate change 0=totally unconcerned; 0.25=largely unconcerned; 0.5=mildly unconcerned; 0.75=concerned; 1=very concerned (May 2008 wave) 0 1 100 In late 2013, the United Nations Development Programme (UNDP) adjusted the calculation of previous HDI scores as a result of newly available data. The data used here was collected before this adjustment, so currently available HDI scores from the UNDP a re slightly different than the scores used here. 207 Climate change belief scale what extent you agree or disagree with that statement. Please use a ze ro - to - 10 scale, where 0 means you completely disagree and 10 means you completely agree. 1. We are already in the first stages of global warming and climate change. 2. If the present rate of coal and oil use continues, serious long - term environmental damage will occur. 3. The dangers of global warming are being over emphasized for political reasons. 4. There is not enough scientific evidence to I recoded responses to statements three and four so higher values capture gr eater belief in climate change. Then responses to these four questions were used to create a scale measuring climate change belief. (May 2008 wave) 0 10 Cause of climate change caused mostly by thi ngs people do, mostly by natural causes, or about equally by things people do and by equally by things people do and by natural causes; 1=most by things people do. (February 2008 wave) 0 1 Political id eology When it comes to politics, would you describe yourself as liberal, conservative, or neither liberal nor conservative? 1=very liberal; 2=somewhat liberal; 3=closer to liberals; 4=neither (moderate); 5=closer to conservatives; 6=somewhat conservative; 7=very conservative (February 2008 wave). 1 7 Political knowledge Respondents were asked six questions about their knowledge of politics: 1. Do you happen to know how many times an individual can be elected President of the United States under current laws ? Correct: 2. 2. For how many years is a United States Senator elected that is, how many years are there in one full term of office for a U.S. Senator? Correct: 6. 3. How many U.S. Senators are there from each state? Correct 2. 4. For how many years is a member o f the United States House of Representatives elected that is, how many years are there in one full term of office for a U.S. House member? Correct: 2. 0 6 208 5. According to federal law, if the President of the United States dies, is no longer willing or able to serve, or is removed from office by Congress, the Vice President would become the President. If the Vice President were unable or unwilling to serve, who would be eligible to become president next? (The Chief Justice of the Supre me Court, the Secretary of State, or the Speaker of the House of Representatives / The Speaker of the House of Representatives, the Secretary of States, or the Chief Justice of the Supreme Court)? Correct: The Speaker of the House of Representatives. 6. What percentage vote of the House and the Senate is needed to override a Presidential veto? (A bare majority, two - thirds, three - fourths, or ninety percent / Ninety percent, three - fourths, two - thirds, or a bare majority)? Correct: Two - thirds. Respondents were gi ven one point for each correct answer , so the variable ranges from 0 to 6 ( February 2008 wave). Interest in politics going on in government and politics? (Extremely interested, very interested, moderate ly interested, slightly interested, or not interested at all)? 1=not interested at all; 2=slightly interested; 3=moderately interested; 4=very interested; 5=extremely interested. 1 5 Education What is the highest degree or level of school you have complet ed ? 1=no high school diploma; 2=high school professional, or doctoral degree) (January 2008 wave). 1 5 Biblical literalism Which of these statements comes clos est to describing your feelings about the Bible? 1= The Bible is the actual word of God and is to be taken literally; 2= The Bible is the word of God but not everything in it should be taken literally, word for word; 3 = The Bible is a book written by men and is not the word of God (February 2008 wave). 1 4 Female Respondent sex. 1=female; 0=male (January 2008 wave). 0 1 Age younger than 30 Self - identified age of respondent in years (January 2008 wave). 0 1 Age 30 to 39 Self - identified age of respond ent in years (January 2008 wave). 0 1 Age 40 to 49 Self - identified age of respondent in years (January 2008 wave). 0 1 209 Age 50 to 59 Self - identified age of respondent in years (January 2008 wave). 0 1 Age 60 to 69 Self - identifi ed age of respondent in years (January 2008 wave). 0 1 Age 70 and older Self - identified age of respondent in years (January 2008 wave). 0 1 American National Elections Study 2012 Cause of climate change [Do / Assuming it's happening, do] you think a rise in the world's temperatures would be caused mostly by human activity, mostly by natural causes, or about equally by human activity and by natural causes? 0=mostly by natural causes/about equally by human activity and natural causes; 1=mostly by human activity (Pre - election survey). 0 1 Political ideology Where would you place YOURSELF on this scale, or haven't you thought much about this? 1=extremely liberal; 2=liberal 3=slightly liberal; 4=moderate, middle of the road; 5=slightly conservative; 6=c onservative; 7=very conservative (Pre - election survey). 1 7 Political knowledge Respondents were asked five questions about their knowledge of politics: 1. Do you happen to know how many times an individual can be elected President of the United States under current laws? 2. Is the U.S. federal budget deficit, the amount by amount of money it collects, now bigger, about the same, or smaller than it was during most of the 1990s? 3. For how many years is a United States sen ator election, that is, how many years are there in one full term of office for a U.S. Senator? 4. What is Medicare? A program run by the U.S. health care; A program run by state governments to provide health care to poor people; A private health insurance plan sold to individuals in all 50 states; A private, non - profit organization that runs free health clinics. 5. On which of the following does the U.S. federal government currently spend the least? Foreign aid; Medicar e; National Defense; Social Security 0 5 210 Respondents were given one point for each correct answer, so the variable ranges from 0 to 5 (Pre - election survey). Attention to politics n government and politics? (Always; most of the time; about half the time; some of the time; or Never)? 1=never; 2=some of the time; 3=about half the time; 4=most of the time; 5=always. 1 5 Education What is the highest degree or level of school you have completed ? 1=no high school diploma; 2=high school professional, or doctoral degree) (Pre - election survey). 1 5 Biblical literalism Which of these statements c omes closest to describing your feelings about the Bible? 1= The Bible is the actual word of God and is to be taken literally; 2= The Bible is the word of God but not everything in it should be taken literally, word for word; 3 = The Bible is a book writte n by men and is not the word of God (Pre - election survey). 1 3 Female Respondent sex. 1=female; 0=male (Pre - election survey) 0 1 Age younger than 30 Self - identified age of respondent in years (Pre - election survey). 0 1 Age 30 to 39 Self - identified a ge of respondent in years (Pre - election survey). 0 1 Age 40 to 49 Self - identified age of respondent in years (Pre - election survey). 0 1 Age 50 to 59 Self - identified age of respondent in years (Pre - election survey). 0 1 Age 60 to 69 Self - identified age of respondent in years (Pre - election survey). 0 1 Age 70 and older Self - identified age of respondent in years (Pre - election survey). 0 1 World Values Survey Variables Seriousness of c limate change the world as a whole. Please tell me how serious you consider each of the following. Is it very serious, somewhat serious, not very serious, or not serious at all? Global warming or greenhouse effect? 1=not serious at all; 2=not very serious; 3=somewhat se rious; 4=very serious 1 4 211 Seriousness of loss of species/biodiversity world as a whole. Please tell me how serious you consider each of the following. Is it very serious, somewhat seriou s, not very serious, or not serious at all? Loss of plant or animal species or biodiversit y. 1=not serious at all; 2=not very serious; 3=somewhat serious; 4 =very serious. 1 4 Seriousness of pollution of water he world as a whole. Please tell me how serious you consider each of the following. Is it very serious, somewhat serious, not very serious, or not serious at all? Pollution of rivers, lakes and streams. 1=not serious at all; 2=not very serious; 3=somewhat serious; 4 =very serious. 1 4 Post - materialism 12 item index: People sometimes talk about what the aims of this country should be for the next ten years. On his card are listed some of the goals which different people give top priority. Would you please say which one of these you, yourself consider the most important? And which would be the next most important? First list : a high level of economic growth; making sure this country has strong defense forces; seeing that people have more say about how thi ngs are done at their jobs and in their communities; trying to make our cities and countryside more beautiful. Second list: maintaining order in the nation; giving people more say in important government decisions; fighting rising prices; protecting freed om of speech Third list: a stable economy; progress toward a less impersonal and more humane society; progress toward a society in which ideas count more than money; the fight against crime - 2.5 2.5 212 Political ideology In politica generally speaking? 1=far left; 10=far right 1 10 Right Self - positioning on the above left - right continuum : 7 - 10 0 1 Left Self - positioning on the above l eft - right continuum: 1 - 4 0 1 Political interest How interested would you say you are in politics? - 1.5=not at all interested, - 0.5=not very interested, 0.5= somewhat interested, 1.5=very interested - 1.5 1.5 Education What is the highest education level that you have attained? - 3.5=no formal education - 2.5=incomplete primary school - 1.5=complete primary school - 0.5=incomplete secondary school: technical/vocational type 0.5=complete secondary school: technical/vocational type 1.5=incomplete secondary: uni versity - preparatory type 2.5=complete secondary: university - preparatory type 3.5=university - level education - 3.5 3.5 Religious attendance Apart from weddings and funerals, about how often do you attend religious services these days? - 3=never practically never; - 2=less often; - 1=once a year; 0=only on special holy days; 1=once a month; 2=once a week; 3=more than once a week - 3 3 Importance of God i - 5=not at all important; 5=very important. - 5 5 Protestant Do you belong to a religious denomination? 1=Protestant 0 1 Lower class People sometimes describe themselves as belonging to the working class, the middle class, or the upper or lower class. Would you describe yourself as belonging to the 1=lower class; 0=all other response categories. 0 1 213 Working class People sometimes describe themselves as belonging to the working class, the middle class, or the upper or lower class. Would you describe yourself as belonging to the 1=working class; 0=all other response categories. 0 1 Middle class People sometimes describe themselves as belonging to the working class, the middle class, or the upper or lower class. Would you des cribe yourself as belonging to the 1=lower middle class; 0=all other response categories. 0 1 Upper class People sometimes describe themselves as belonging to the working class, the middle class, or the upper or lower class. Would you describe yourself as belonging to the 1=upper middle class or upper class; 0=all other response categories. 0 1 Gender Female=1 0 1 Age younger than 30 Self - identified age of respondent in years. 0 1 Age 30 to 39 Self - identified age of respondent in years. 0 1 Age 4 0 to 49 Self - identified age of respondent in years. 0 1 Age 50 to 59 Self - identified age of respondent in years. 0 1 Age 60 to 69 Self - identified age of respondent in years. 0 1 Age 70 and older Self - identified age of respondent in years. 0 1 Co untry - level variable OECD country Member of the Organization for Economic Cooperation and Development (OECD) or not. 0 1 214 Table A1.11: Chapter 6 Survey questions, measurements, and variables Variable Name Description/Question Minimum Value Maximum Variable Pew Global Attitudes Surveys (2009 - 2010) Willingness to pay to address climate change Please tell me whether you agree or disagree with the higher prices in order to addres s global climate chang 0 1 Seriousness of climate change ( 2009) On another topic, in your view is global warming a very serious problem, somewhat serious, not too serious, or not a p roblem? 0=not a problem; 1=not too serious; 2=somewhat serious; 3 =very serious . 0 3 Seriousness of climate change (2010) On another topic, in your view global climate change a very serious problem, somewhat serious, not too serious, or not a problem? 0=no t a problem; 1=not too serious; 2=somewhat serious; 3 =very serious . 0 3 Age younger than 30 Self - identified age of respondent in years . 0 1 Age 40 - 49 Self - identified age of respondent in years . 0 1 Age 50 - 59 Self - identified age of respondent in years . 0 1 Age 60 - 69 Self - identified age of respondent in years . 0 1 Age 70 and older Self - identified age of respondent in years . 0 1 Female Respondent gender. 1=female; 0=male. 0 1 World Values Survey (2005 - 2009 ) 215 Variable Name Description/Question Minimum Value Maximum Variable Seriousness of climate change world as a whole. Please tell me how serious you consider each of the following. Is it very serious, somewhat serious, no t very serious, or not serious at all? Global warming or greenhouse effect. 0=not serious at all; 1=not very serious; 2=somewhat serious; 3 =very serious. 1 4 Seriousness of loss of species/biodiversity worl d as a whole. Please tell me how serious you consider each of the following. Is it very serious, somewhat serious, not very serious, or not serious at all? Loss of plant or a nimal species or biodiversity. 0=not serious at all; 1=not very serious; 2=somewha t serious; 3 =very serious. 1 4 Seriousness of pollution of water world as a whole. Please tell me how serious you consider each of the following. Is it very serious, somewhat serious, not very serious, or n ot serious at all? Pollution of rivers, lakes and streams. 0 =not serio us at all; 1=not very serious; 2=somewhat serious; 3 =very serious. 1 4 International Social Survey Programme (2009 - 2010) Most/next most important issue Environment Which of these i ssues is the most important for [COUNTRY] today? Which is the next important? Health care; education; crime, the environment; immigration; the economy; terrorism; poverty; none 1= = all other responses 0 1 216 Most important environmental problem for country climate change problems. Which problem, in any do you think is the most important f or [COUNTRY] as a whole? Air pollution; chemicals and pesticides; water shortage; water pollution; nuclear waste; domestic waste disposal; climate change; genetically modified foods; using up 1=climate ch ange; 0= other responses categories 0 1 Dangerousness of c limate change temperatu re caused by climate change is 0 =not dangerous a t all for the environment; 1=not very dangerous; 2=somewhat dangerous; 3=ve ry dangerous; 4 =extremely dangerous for the environment 0 4 Left party identification Political partisanship. 1=far left (1) or l eft (2) political party; 0=all other categories. 0 1 Right party identification Political partisanship. 1=r ight (4) or far right (5) ; 0=all other categories. 0 1 Level of e ducation Highest education attainment. - 2=primary education or no formal education; - 1 = intermediate secondary education; 0 =secondary education; 1=university incomplete; 2=university complete - 2 2 Female Respondent sex. 1=female; 0=male. 0 1 Age younger than 30 Self - identified age of respondent in years . 1=29 and younger; 0=all other ages. 0 1 Age 30 to 39 Self - identified age of respondent in years. 1=30 to 39; 0=all other ages. 0 1 Age 40 to 49 Self - identified age of respondent in years. 1=40 to 49; 0=all other ages. 0 1 Age 50 to 59 Self - identified age of respondent in years. 1=50 to 59; 0=all other ages. 0 1 Age 60 to 69 Self - identified age of respondent in years. 1=60 to 69; 0=all other ages. 0 1 Age 70 and older Self - identified age of respondent in years. 1=70 and older; 0=all other ages. 0 1 217 Data Sources American National Elections Study. 2012. Ann Arbor, MI: University of Michigan, Center for Political Studies. American Nation al Elections Study. 2008. Ann Arbor, MI: University of Michigan, Center for Political Studies. International Social Survey Programme Group. 2010. Environment III Data Version 1.0. Distributor: GESIS. Pew Forum on Religion & Public Life. 2006. Pew Researc h Center, Washington, DC, July. Pew Global Attitudes Project Survey. 2007. Pew Research Center, Washington, D.C. Pew Global Attitudes Project Survey. 2008. Pew Research Center, Washington, D.C. Pew Global Attitudes Project Survey. 2009. Pew Research Center , Washington, D.C. Pew Global Attitudes Project Survey. 2010. Pew Research Center, Washington, D.C. Pew News Interest Index Poll. 2007. Pew Research Center, Washington, D.C. January. Pew Research Center for the People & the Press. 2008. Pew Research Center , Washington, D.C. April. Pew Research Center for the People & the Press. 2009. Pew Research Center, Washington, D.C. October. Pew Research Center for the People & the Press. 2010. Pew Research Center, Washington, D.C. October. Pew Research Center for the People & the Press, Pew Forum on Religion & Public Life. 2011. Pew Research Center, Washington, D.C. November. Pew Research Center for the People & the Press. 2012. Pew Research Center, Washington, D.C. October. Pew Research Center for the People & the Pre ss. 2013. Pew Research Center, Washington, D.C. March. Polity IV Project. 2012. Political Regime Characteristics and Transitions, 1800 - 2012. - Social Science Quarterly 90(2): 231 - 242. SWIID Version 4.0, September 2013. World Bank. 2013. World Bank Development Indicators. World Values Survey Association. 2009. Wave 5: Official Data File. United Nations Development Programme. 2013. Human Development Report. 218 App endix B : Additional models, anal ysis and robustness checks for C hapter 4 Table A2.1: Random - intercept multilevel linear models with random country intercepts and polynomial terms for age predicting perceived seriousness of c limate change (World Values Survey 2005 - 2009) (1) (2) Political ideology - 0.0472 - 0.0469 (from left to right) (0.0127) (0.0127) Post - materialism index 0.0669 0.0665 (0.0220) (0.0220) Interest in politics 0.0644 0.0635 (0.0201) (0.0199) Le vel of education 0.0924 0.0919 (0.0140) (0.0140) Attendance of religious services - 0.0234 - 0.0235 (0.0144) (0.0143) Importance of God 0.0452 0.0457 (0.0084) (0.0083) Protestant - 0.1487 - 0.1441 (0.0753) (0.0747) Female 0.0709 0.0707 (0.0358) ( 0.0358) A ge 0.0017 0.0130 (0.0011) (0.0048) Age 2 - 0.0001 (0.0000) Not serious at all | - 4.2272 - 4.0008 Not very serious (0.2364) (0.2410) Not very serious | - 2.4523 - 2.2259 Somewhat serious (0.1755) (0.1739) Somewhat serious | - 0.5084 - 0. 2816 Very serious (0.1328) (0.1503) Variance component Country - level 0.4095 0.4088 (0.1005) (0.1001) N umber of respondents 39874 39874 Number of countries 41 41 A IC 71387 71377 Note s : Cell entries are ordered logistic regression coefficients wit h standard error in parentheses. Coefficients estimates that can be distinguished from zero at the conventional level statistical significance (two - tailed p < 0.05) are in bold . China, Malaysia, and Morocco were all excluded from the model s because relevan t survey questions were not asked in these countries. 219 Table A2.2: Ordered logistic multilevel models with random country intercepts and polynomial terms for GDP per capita and HDI predicting perceived seriousness of climate change (World Values Survey 20 05 - 2009) (1) (2) Political ideology - 0.0473 - 0.0471 (from left to right) (0.0126) (0.0127) Post - materialism index 0.0659 0.0660 (0.0221) (0.0221) Interest in politics 0.0631 0.0631 (0.0200) (0.0201) Education level 0.0927 0.0927 (0. 0140) (0.0140) Attendance of religious services - 0.0239 - 0.0240 (0.0142) (0.0142) Importance of God 0.0460 0.0459 (0.0085) (0.0085) Protestant - 0.1432 - 0.1402 (0.0741) (0.0741) Female 0.0698 0.0698 (0.0359) (0.0359) Age younger than 30 - 0.1178 - 0.1174 (0.0464) (0.0464) Age 30 - 39 - 0.0088 - 0.0087 (0.0324) (0.0324) Age 50 - 59 0.0189 0.0188 (0.0331) (0.0330) Age 60 - 69 0.0126 0.0126 (0.0465) (0.0464) Age 70 and older - 0.0517 - 0.0518 Country - level variables (0.0478) (0.0478) Natural log arithm of mean GDP per capita - 11.1538 (6.4951) Natural logarithm of mean GDP per capita 2 1.3911 (0.7745) Natural logarithm of mean GDP per capita 3 - 0.0562 (0.0303) Mean HDI - 89.4429 (19.5499) Mean HDI 2 148.0480 (32.4418) Mean HDI 3 - 76.6277 (16.9083) Not serious at all | Not very serious - 33.2017 - 20.7863 (17.8731) (3.6108) Not very serious| Somewhat serious - 31.4270 - 19.0117 (17.8465) (3.5883) Somewhat serious| Very serious - 29.4830 - 17.0676 (17.8536) (3.6050) 220 Ta Note s : Cell entries are ordered logistic regression coefficients with standard er ror in parentheses. Coefficients estimates that can be distinguished from zero at the conventional level statistical significance (two - tailed p < 0.05) are in bold . China, Malaysia, and Morocco were all excluded from the model s because relevant survey ques tions were not asked in these countries. Table A2.3: Random - intercept multilevel ordered logistic models with individual - level interactions predicting perceive d seriousness of climate change (World Values Survey 2009 - 2011) (1) (2) (3) (4) Left 0.0 889 0.1089 (political ideology=1 - 4) (0.0437) (0.0434) Right - 0.1951 - 0.1983 (political ideology=7 - 10) (0.0670) (0.0674) Political ideology 0.2596 0.2785 (left=1; right=0) (0.0725) (0.0714) Level of e ducation 0.0925 0.0922 0.0815 0 .0929 (0.0159) (0.0141) (0.0185) (0.0140) Interest in politics 0.0631 0.0081 0.0974 0.0542 (0.0194) (0.0224) (0.0239) (0.0384) Left X education 0.0165 (0.0143) Right X education - 0.0113 (0.0145) Left X interest in politics 0.1495 (0.0394) Right X interest in politics 0.0474 (0.0469) Political ideology X education 0.0284 (0.0184) Political ideology X interest in politics 0.1109 (0.0617) Post - materialism index 0.0638 0.0630 0.0827 0.0818 (0.0219) ( 0.0218) (0.0271) (0.0268) Attendance of religious services - 0.0236 - 0.0235 - 0.0276 - 0.0279 (0.0141) (0.0141) (0.0146) (0.0145) Importance of God 0.0465 0.0473 0.0428 0.0435 (0.0084) (0.0084) (0.0089) (0.0089) P rotestant - 0.1443 - 0.1471 - 0.1271 - 0.12 93 (0.0736) (0.0732) (0.0868) (0.0871) F emale 0.0678 0.0670 0.0857 0.0831 (0.0354) (0.0356) (0.0347) (0.0349) Age under 30 - 0.1190 - 0.1171 - 0.0827 - 0.0786 (0.0457) (0.0455) (0.0500) (0.0493) Variance component Country - level 0.3610 0.2760 (0.0832) (0.0710) N umber of respondents 39973 39973 Number of countries 41 41 AIC 71555 71544 221 Age 30 - 39 - 0.0074 - 0.0064 0.0390 0. 0411 (0.0319) (0.0320) (0.0418) (0.0417) Age 50 - 59 0.0184 0.0186 0.0785 0.0788 (0.0335) (0.0335) (0.0421) (0.0420) Age 60 - 69 0.0140 0.0146 0.0889 0.0888 (0.0463) (0.0463) (0.0528) (0.0530) Age 70 and older - 0.0505 - 0.0495 - 0.0475 - 0.0458 (0.0484 ) (0.0476) (0.0618) (0.0612) Not serious at all | - 4.1052 - 4.0979 - 3.7615 - 3.7562 Not very serious (0.1928) (0.1902) (0.2001) (0.2016) Not very serious | - 2.3305 - 2.3231 - 2.0499 - 2.0442 Somewhat serious (0.1464) (0.1430) (0.1563) (0.1556) Somewhat s erious | - 0.3855 - 0.3773 - 0.1781 - 0.1717 Very serious (0.1121) (0.1087) (0.1225) (0.1205) Variance component Country - level standard deviation 0.4056 0.4029 0.4033 0.4005 (0.0989) (0.0981) (0.0934) (0.0925) N umber of respondents 39973 39973 24255 2 4255 Number of countries 41 41 41 41 AIC 71531 71509 43945 43936 Note s : Cell entries are ordered logistic regression coefficients with standard error in parentheses. Coefficients estimates that can be distinguished from zero at the conventional level st atistical significance (two - tailed p < 0.05) are in bold . China, Malaysia, and Morocco were all excluded from the model s because relevant survey questions were not asked in these countries. 222 Table A2.4: Random - intercept multilevel linear models without c ountry - level predictors predicting perceived seriousness of climate change (World Values Survey 2005 - 2009) (1) (2) (3) (4) (5) (6) Political ideology - 0.0060 - 0.0057 - 0.0057 - 0.0059 (from left to right) (0.0017) (0.0017) (0.0017) (0.0016) Left 0.0083 (ideology=1 - 4) (0.0044) Right - 0.0298 (ideology=7 - 10) (0.0092) Social issues index - 0.0029 - 0.0032 (0.0017) (0.0017) Income equality - 0.0009 - 0.0012 (0.0012) (0.0011) Private E nterprise 0.0005 0.0012 (0.0012) (0.0010) Post - materialism index 0.0073 0.0065 0.0066 0.0065 0.0062 0.0067 (0.0026) (0.0027) (0.0026) (0.0027) (0.0029) (0.0026) Interest in politics 0.0069 0.0073 0.0076 0.0039 0.0060 0.0071 (0.0024) (0.002 5) (0.0023) (0.0025) (0.0027) (0.0024) Education level 0.0093 0.0101 0.0096 0.0084 0.0086 0.0096 (0.0015) (0.0015) (0.0015) (0.0017) (0.0017) (0.0015) Attendance of religious services - 0.0018 - 0.0022 - 0.0022 - 0.0017 - 0.0017 - 0.0022 (0.0015) (0.0016) (0.0015) (0.0015) (0.0016) (0.0015) Importance of God 0.0046 0.0047 0.0048 0.0044 0.0051 0.0048 (0.0010) (0.0010) (0.0009) (0.0009) (0.0011) (0.0009) Protestant - 0.0212 - 0.0225 - 0.0215 - 0.0174 - 0.0203 - 0.0215 (0.0082) (0.0081) (0.0072) (0.0061) (0.00 73) (0.0073) Catholic 0.0004 0.0008 (0.0061) (0.0062) E vangelical 0.0008 0.0005 (0.0045) (0.0053) M uslim 0.0087 0.0075 (0.0107) (0.0110) Lower class 0.0035 (0.0085) Working class 0.0042 (0.0040) U pper class 0.0025 (0.0044) Female 0.0123 0.0095 0.0108 0.0107 0.0106 0.0110 (0.0043) (0.0044) (0.0044) (0.0043) (0.0043) (0.0044) Age younger than 30 - 0.0153 - 0.0182 - 0.0166 - 0.0181 - 0.0166 - 0.0164 (0.0051) (0.0054) (0.0051) (0.0053) (0.0 057) (0.0051) Age 30 - 39 - 0.0034 - 0.0038 - 0.0027 - 0.0046 - 0.0048 - 0.0028 (0.0042) (0.0045) (0.0041) (0.0046) (0.0048) (0.0041) Age 50 - 59 - 0.0010 0.0008 - 0.0004 - 0.0049 - 0.0022 - 0.0005 (0.0044) (0.0045) (0.0043) (0.0046) (0.0047) (0.0043) Age 60 - 69 - 0 .0072 - 0.0049 - 0.0060 - 0.0063 - 0.0064 - 0.0063 (0.0063) (0.0069) (0.0062) (0.0066) (0.0072) (0.0063) 223 Age 70 and older - 0.0095 - 0.0069 - 0.0086 - 0.0107 - 0.0092 - 0.0088 (0.0061) (0.0068) (0.0061) (0.0067) (0.0074) (0.0061) Self - descri bed income decile - 0.0008 (0.0014) Constant 0.8544 0.8574 0.8320 0.8308 0.8600 0.8574 (0.0143) (0.0147) (0.0124) (0.0128) (0.0138) (0.0133) Variance components Country - level standard deviation 0.0695 0.0670 0.0686 0.0687 0.0703 0.068 8 (0.0095 ) (0.0096 ) (0.0094 ) (0.0087 ) (0.0095 ) (0.0094 ) Individual - level standard deviation 0.2250 0.2256 0.2251 0.2291 0.2259 0.2252 (0.0067 ) (0.0064 ) (0.0064 ) (0.0072 ) (0.0070 ) (0.0065 ) N umber of respondents 38753 37189 39973 39068 32855 39973 Num ber of countries 40 39 41 40 39 41 AIC - 3765 - 3430 - 3872 - 2935 - 3100 - 3840 Note s : Linear regression coefficients with standard error in parentheses. Coefficients estimates that can be distinguished from zero at the conventional level statistical signific ance (two - tailed p < 0.05) are in bold . China, Malaysia, and Morocco were all excluded from the model s because relevant survey questions were not asked in these countries. The dependent variable, perceived seriousness of climate change, ranges from 0 to 1: 0 not very serious; 0.33 not very serious; 0.67 somewhat serious; 1.0 very serious. 224 Table A2.5: Random - intercept multilevel linear models with country - level predictors predicting perceived seriousness of climate change (World Values Survey 2005 - 2009) (1) (2) (3) (4) Political ideology - 0.0059 - 0.0059 - 0.0059 - 0.0059 (from left to right) (0.0016) (0.0016) (0.0016) (0.0016) Post - materialism index 0.0067 0.0067 0.0067 0.0067 (0.0026) (0.0026) (0.0026) (0.0026) Interest in politics 0.0071 0.0071 0.0071 0.0071 (0.0024) (0.0024) (0.0024) (0.0024) Level of education 0.0095 0.0095 0.0095 0.0095 (0.0015) (0.0015) (0.0015) (0.0015) Attendance of religious services - 0.0022 - 0.0022 - 0.0022 - 0.0022 (0.0015) (0.0015) (0.0015) (0.0015) Importance of God 0.0048 0.0048 0.0048 0.0048 (0.0009) (0.0009) (0.0009) (0.0009) Protestant - 0.0216 - 0.0215 - 0.0215 - 0.0215 (0.0073) (0.0073) (0.0073) (0.0073) Female 0.0109 0.0109 0.0109 0.0109 (0.0044) (0.0044) (0.0044) (0.0044) Age younger t han 30 - 0.0163 - 0.0163 - 0.0163 - 0.0163 (0.0051) (0.0051) (0.0051) (0.0051) Age 30 - 39 - 0.0027 - 0.0027 - 0.0027 - 0.0027 (0.0042) (0.0042) (0.0042) (0.0042) Age 50 - 59 - 0.0005 - 0.0005 - 0.0005 - 0.0005 (0.0043) (0.0043) (0.0043) (0.0043) Age 60 - 69 - 0.006 4 - 0.0064 - 0.0064 - 0.0064 (0.0063) (0.0063) (0.0063) (0.0063) Age 70 and older - 0.0089 - 0.0089 - 0.0089 - 0.0089 (0.0061) (0.0061) (0.0061) (0.0061) Country - level variables Natural logarithm of mean GDP per capita 0.0094 0.0031 (0.0080) (0.01 15) Mean HDI 0.0934 0.0855 (0.0835) (0.1406) Mean Polity score 0.0018 0.0015 (0.0033) (0.0032) Mean Gini coefficient 0.0005 0.0007 (0.0015) (0.0013) Natural logarithm of mean CO 2 emissions 0.0063 0.0009 (tons per capita) (0.0136 ) (0.0157) Natural logarithm of oil and gas rents per capita - 0.0017 - 0.0014 (0.0028) (0.0029) Constant 0.7734 0.7884 0.7959 0.7627 (0.0774) (0.0685) (0.1146) (0.1064) Variance Components Country - level standard deviation 0.0673 0.0670 0.0668 0.0666 (0.0083 ) (0.0081 ) (0.0083 ) (0.0083 ) Individual - level standard deviation 0.2252 0.2252 0.2252 0.2252 (0.0064 ) (0.0065 ) (0.0065 ) (0.0065 ) 225 N umber of respondents 39973 39973 39973 39973 Number of countries 41 41 41 4 1 AIC - 3840 - 38340 - 3832 - 3832 Note s : Linear regression coefficients with standard error in parentheses. Coefficients estimates that can be distinguished from zero at the conventional level statistical significance (two - tailed p < 0.05) are in bold . Chin a, Malaysia, and Morocco were all excluded from the model s because relevant survey questions were not asked in these countries. The dependent variable, perceived seriousness of climate change, ranges from 0 to 1: 0 not very serious; 0.33 not very serio us; 0.67 somewhat serious; 1.0 very serious. Table A2.6: Random - intercept multilevel ordered logistic regression models without country - level predictors predicting perceived seriousness of climate change on multiple imputation datasets (World Values S urvey 2005 - 2009) (1) (2) (3) (4) (5) (6) Political ideology - 0.0468 - 0.0459 - 0.0503 - 0.0467 (from left to right) (0.0098 ) (0.0094 ) (0.0094 ) (0.0093 ) Left 0.0920 (ideology=1 - 4) (0.0368 ) Right - 0.0298 (ideolo gy=7 - 10) (0.0092) Social issues index - 0.0166 - 0.0131 (0.0119 ) (0.0012 ) Income equality 0.0223 0.0080 (0.0079 ) (0.0078 ) Private E nterprise 0.008 0.0102 (0.0073 ) (0.0074 ) Post - materialism index 0.0646 0.0617 0. 0610 0.0608 0.0537 0.0605 (0.0184 ) (0.0189 ) (0.0182 ) (0.0184 ) (0.0191 ) (0.0182 ) Interest in politics 0.0588 0.0613 0.0606 0.0465 0.0573 0.0610 (0.0181 ) (0.0182 ) (0.0182 ) (0.0192 ) (0.0190 ) (0.0176 ) Education level 0.0861 0.0944 0.0910 0.0862 0.0876 0. 0909 (0.0114 ) (0.0122 ) (0.0123 ) (0 .0125 ) (0.0128 ) (0.0123) Attendance of religious services - 0.0198 - 0.0231 - 0.0239 - 0.0307 - 0.0279 - 0.0240 (0.0111 ) (0.0116 ) (0.0109 ) (0.0105 ) (0.0109 ) (0.0109 ) Importance of God 0.0489 0.0504 0.0485 0.0475 0.0527 0.0 491 (0.0074 ) (0.0076 ) (0.0073 ) (0. 0070 ) (0.0076 ) (0.0074 ) Protestant - 0.1367 - 0.1500 - 0.1223 - 0.01179 - 0.1237 - 0.1215 (0.0628 ) (0.0646 ) (0.0584 ) (0.0602 ) (0.0597 ) (0.0589 ) Catholic - 0.0399 - 0.532 (0.0421 ) (0.0446 ) E vangelical - 0.0475 - 0.04 57 (0.0342 ) (0.0347 ) M uslim 0 .0633 (0.0789 ) Lower class 0.0055 (0.0567 ) Working class 0.0236 (0.0319 ) Upper class 0.0385 (0.0444 ) 226 Female 0.0658 0.0724 0.0593 0.0695 0.0710 0 .059 4 (0.0311 ) (0.0281 ) (0.0307 ) (0.0310 ) (0 .0311 ) (0.0309 ) Age younger than 30 - 0.0814 - 0.0867 - 0.0966 - 0.1179 - 0.0934 - 0.0964 (0.0340 ) (0.0360 ) (0.0361 ) (0.0602 ) (0.0384 ) (0.0364 ) Age 30 - 39 - 0.0116 - 0.0088 - 0.0109 - 0.0959 - 0.0118 - 0.0116 (0.0311 ) (0.0036 ) (0.0305 ) (0.0380 ) (0.0329 ) (0.0306 ) Age 50 - 59 0.0163 0.0180 0.0146 - 0.0101 0.0134 0.0150 (0.0337 ) (0.0334 ) (0.0334 ) (0.0330 ) (0.0339 ) (0.331 ) Age 60 - 69 0.0027 0.0094 0.0058 0.0209 0.0101 0.0052 (0.0048 ) (0.0485 ) (0.0472 ) (0.0491 ) (0.0491 ) ( 0.0471 ) Age 70 and older - 0.0703 - 0.0604 - 0.0691 - 0.0606 - 0.0578 - 0.0699 (0.0427 ) (0.0424 ) (0.0419 ) (0.0425 ) (0.0425 ) (0.0417 ) Self - described income decile 0.0004 (0.0102 ) Not serious at all | - 4.0074 - 4.0093 - 3.7913 - 3.7605 - 4.037 3 - 4.0201 Not very serious (0.1637 ) (0.1713 ) (0.1482 ) (0.1566 ) (0.1707 ) (0. 1602 ) Not very serious | - 2.3820 - 2.3885 - 2.1681 - 2.1158 - 2.4048 - 2.3968 Somewhat serious (0.1357) (0.1363) (0.1273) (0.1312) (0.1390) (0.1321) Somewhat serious | - 0.5719 - 0.5678 - 0.3653 - 0.3143 - 0.5933 - 0.5941 Very serious (0.1073) (0.0770) (0.1008) (0.1013) (0.1109) (0.1053) Variance components Country - level variation 0.3351 0.3160 0.3304 0.3371 0.3361 0.3308 (0.0781 ) (0.0770 ) (0.0771 ) (0.0801 ) (0.082 1 ) (0.0774 ) N umber of respondents 56842 56200 58402 55866 53851 58402 Number of countries 40 39 41 40 39 41 Notes: Five MI datasets were used, created by Amelia II. Cell entries are ordered logistic regression coefficients with standard error in parenth eses . Coefficients estimates that can be distinguished from zero at the conventional level statistical significance (two - tailed p < 0.05) are in bold . Four countries were excluded from Model 1 because survey questions were not asked in these countries. In China, the self - placement political ideology question was not asked. In Malaysia, the self - placement political ideology and religious attendance questions were not asked. Also, in Morocco, the rel igious attendance question was not asked. Finally, in Mexico , the question about socio - economic class was not asked. In Model 2, five countries were excluded: China, Malaysia, Morocco, Jordan, and Argentina. In Jordan and Argentina, the income decile question was not asked. In Model 3, China and Malaysia were both excluded because the self - placement political ideology question was not asked in either country. In Morocco, the religious attendance question was not asked. In Model 4, in addition to excluding to Morocco, Peru, Egypt and Malaysia are excluded because the social issues questions were not asked , so respondents from four countries were excluded . In Model 5, in addition to excluding Morocco, Peru, Egypt, and Malaysia, China was excluded because self - described political ideology was re - introduced into the mode l , so respondents from five countries were excluded . In Model 6, China, Malaysia, and Morocco were excluded. 227 Table A2.7: Random - intercept multilevel ordered logistic regression models with country - level predictors predicting perceived seriousness of cli mate change on multiple imputation datasets (World Values Survey 2005 - 2009) (1) (2) (3) (4) Political ideology - 0.0467 - 0.0467 - 0.0467 - 0.0467 (from left to right) (0.0093 ) (0.0093 ) (0.0093 ) (0.0093 ) Post - materialism index 0.0603 0.0604 0.0603 0. 0603 (0.0182 ) (0.0182 ) (0.0182 ) (0.0182 ) Interest in politics 0.0611 0.0611 0.0611 0.0611 (0.0176 ) (0.0176 ) (0.0177 ) (0.0177 ) Education level 0.0907 0.0907 0.0907 0.0907 (0.0123 ) (0.0140) (0.0123 ) (0.0123 ) Attendance of religious services - 0.0238 - 0.0238 - 0.0238 - 0.0238 (0.0109 ) (0.0109 ) (0.0109 ) (0.0109 ) Importance of God 0.0492 0.0492 0.0491 0.0491 (0.0074 ) (0.0074 ) (0.0074 ) (0.0074 ) Protestant - 0.1218 - 0.1216 - 0.1220 - 0.1217 (0.0589 ) (0.0589 ) (0.0587 ) (0.0588 ) Female 0.0592 0.0593 0.059 3 0.0593 (0.0309 ) (0.0309 ) (0.0309 ) (0.0309 ) Age under 30 - 0.0958 - 0.0959 - 0.0959 - 0.0959 (0.0364 ) (0.0364 ) (0.0364 ) (0.0364 ) Age 30 - 39 - 0.0114 - 0.0114 - 0.0114 - 0.0114 (0.0306 ) (0.0306 ) (0.0306 ) (0.0306 ) Age 50 - 59 0.0143 0.0143 0.0143 0.0143 (0. 0331) (0.0331 ) (0.0330 ) (0.0330 ) Age 60 - 69 0.0046 0.0046 0.0047 0.0047 (0.0471 ) (0.0470 ) (0.0471 ) (0.0470 ) Age 70 and older - 0.0707 - 0.0707 - 0.0706 - 0.0706 (0.0565 ) (0.0416 ) (0.0417 ) (0.0417 ) Country - level variables Natural logarithm of mean GDP per capita 0.0 608 0.0859 (0.0565 ) (0.0979 ) Mean HDI 0.4805 0.6453 (0.5901 ) (1.1687 ) Mean polity score 0.0009 0.0055 (0.0246 ) (0.0236 ) Mean GINI coefficient 0.0062 0.0061 (0.0116 ) (0.0105 ) Natural logarithm of CO 2 emissions per ca pita - 0.001 16 - 0.0111 (0.1114 ) (0.1345 ) Natural logarithm of oil and gas rents per capita - 0.0020 - 0.0025 (0.0267 ) (0.0280 ) Not serious at all| Not very serious - 3.486 - 3.6731 - 3.0523 - 3.3117 (0.5411 ) (0.4826 ) (0.9008 ) (0.8751 ) Not very se rious| Somewhat serious - 1.8627 - 2.0500 - 1.4290 - 1.6884 (0.5517 ) (0.4988 ) (0.8946 ) (0.8678 ) Somewhat serious| Very serious - 0.0600 - 0.2471 0.3737 0.1143 (0.5312 ) (0.4707 ) (0.8867 ) (0.8579 ) Variance component Country - level variance 0.3217 0.3236 0.3193 0.3209 (0.00732) (0.0744) (0.0728) (0.0738) 228 N umber of respondents 58402 58402 58402 58402 Number of countries 41 41 41 41 AIC 71555 71555 71563 71563 Note s : Five MI datasets were used, created by Amelia II. Cell entries ar e ordered logistic regression coefficients with standard error in parentheses. Coefficients estimates that can be distinguished from zero at the conventional level statistical significance (two - tailed p < 0.05) are in bold . China, Malaysia, and Morocco wer e excluded due to missing variables at the individual level. Table A2.8: Random - intercept multilevel linear models predicting perceived dangerousness of climate change (International Social Survey Programme 2009 - 2011) (1) (2) (3) (4) (5) (6) Lef t party identification 0.0127 0.0112 0.0129 0.0128 0.0128 0.0127 (0.0047) (0.0047) (0.0047) (0.0047) (0.0047) (0.0047) Right party identification - 0.0264 - 0.0250 - 0.0356 - 0.0356 - 0.0356 - 0.0357 (0.0086) (0.0089) (0.0093) (0.0093) (0.0093) (0.0093) Le vel of education 0.0083 0.0063 0.0068 0.0068 0.0068 0.0068 (0.0026) (0.0024) (0.0025) (0.0025) (0.0025) (0.0025) Self - described 0.0214 0.0224 0.0215 0.0215 0.0215 0.0215 environmental knowledge (0.0027) (0.0028) (0.0028) (0.0028) (0.0028) (0.0028) Income gap - 0.0197 - 0.0190 (0.0024) (0.0025) Private enterprise - 0.0072 - 0.0070 (0.0023) (0.0024) Post - materialism index 0.0015 0.0013 - 0.0053 - 0.0053 - 0.0054 - 0.0054 (0.0027) (0.0029) (0.0030) (0.0030) (0.0030) (0.0030) Atten dance of religious services 0.0019 0.0013 0.0014 0.0014 0.0014 0.0014 (0.0010) (0.0011) (0.0010) (0.0010) (0.0010) (0.0010) Catholic 0.0002 0.0004 (0.0074) (0.0077) P rotestant - 0.0060 - 0.0018 - 0.0070 - 0.0072 - 0.0072 - 0.0074 (0.0109) (0.0122 ) (0.0100) (0.0099) (0.0099) (0.0098) M uslim 0.0230 0.0177 (0.0189) (0.0204) F emale 0.0282 0.0254 0.0300 0.0300 0.0300 0.0300 (0.0052) (0.0053) (0.0054) (0.0054) (0.0054) (0.0054) Age younger than 30 - 0.0050 - 0.0059 - 0.0033 - 0.0033 - 0.0034 - 0.0034 (0.0065) (0.0069) (0.0066) (0.0066) (0.0066) (0.0066) Age 30 - 39 0.0085 0.0065 0.0104 0.0104 0.0104 0.0104 (0.0056) (0.0060) (0.0055) (0.0055) (0.0055) (0.0055) Age 50 - 59 - 0.0160 - 0.0176 - 0.0132 - 0.0132 - 0.0132 - 0.0132 (0.0045) (0.0047) (0.00 44) (0.0044) (0.0044) (0.0044) Age 60 - 69 - 0.0228 - 0.0219 - 0.0201 - 0.0200 - 0.0200 - 0.0200 (0.0049) (0.0055) (0.0048) (0.0048) (0.0048) (0.0048) Age 70 and older - 0.0320 - 0.0300 - 0.0278 - 0.0278 - 0.0278 - 0.0278 (0.0062) (0.0066) (0.0064) (0.0064) (0.006 4) (0.0064) Self - described income decile 0.0004 (0.0019) Country - level Natural logarithm of mean GDP - 0.0323 - 0.0375 per capita (0.0101) (0.0142) 229 Mean HDI - 0.3748 - 0.3241 (0.1234) (0.2706 ) Mean polity score 0.0183 0.0129 (0.0123) (0.0128) Mean GINI coefficient 0.0024 0.0025 (0.0020) (0.0023) Natural logarithm of mean CO 2 0.0118 0.0064 emissions per capita (in tons) (0.0186) (0.0247) Natural logari thm of oil and gas 0.0003 - 0.0003 rents per capita (0.0043) (0.0046) Constant 0.6745 0.6787 1.0079 1.0053 0.7880 0.7496 (0.0138) (0.0199) (0.1036) (0.1071) (0.1595) (0.2060) Variance components Country level standard deviation 0 .0657 0.0657 0.0595 0.0602 0.0558 0.0575 (0.0083 ) (0.0089 ) (0.0088 ) (0.0087 ) (0.0066 ) (0.0071 ) Individual - level standard deviation 0.2322 0.2312 0.2328 0.2328 0.2328 0.2328 (0.0044 ) (0.0046 ) (0.0044 ) (0.0044 ) (0.0044 ) (0.0044 ) N umber of respondents 2 8017 26257 29526 29526 29526 29526 Number of countries 29 27 29 29 29 29 AIC - 1506 - 1546 - 1498 - 1497 - 1492 - 1491 Note s : Cell entries linear regression coefficients with the standard errors in parentheses, clustered by country in parentheses. The depende nt variable is perceived dangerousness of climate change, ranging from 0 to 1. Coefficients estimates that can be distinguished from zero at the conventional level statistical significance (two - tailed p < 0.05) are in bold . Listwise deletion of missing val ues. In Model 1, Israel, Taiwan, and Turkey were all excluded from the model because relevant survey questions were not asked in these countries. In Taiwan and Israel, respondents were not asked about their political party identification. In Turkey, respon dents were not asked to self - evaluate their knowledge of the causes and solutions of environmental problems. In Model 2, in addition to excluding Israel, Taiwan, and Turkey, New Zealand and Great Britain were excluded, as respondents were not asked about t heir income. For Models 3 - 6, Israel, Taiwan, and Turkey were excluded. 230 Table A2.9: Random - intercept multilevel linear models with polynomial terms for age predicting perceived dangerousness of climate change (International Social Survey Programme 2009 - 2 011) (1) (2) L eft party identification 0.0130 0.0127 (0.0047) (0.0047) Right party identification - 0.0261 - 0.0263 (0.0085) (0.0085) Level of education 0.0091 0.0086 (0.0026) (0.0026) Self - described environmental knowledge 0.0212 0.0212 (0. 0026) (0.0026) Income gap - 0.0196 - 0.0196 (0.0024) (0.0024) Private enterprise - 0.0073 - 0.0072 (0.0023) (0.0023) Post - materialism index 0.0013 0.0013 (0.0027) (0.0027) Attendance of religious services 0.0018 0.0019 (0.0010) (0.0010) C atholic 0 .0006 0.0005 (0.0074) (0.0073) P rotestant - 0.0061 - 0.0060 (0.0108) (0.0109) M uslim 0.0243 0.0236 (0.0189) (0.0188) F emale 0.0284 0.0284 (0.0052) (0.0052) A ge - 0.0006 0.0008 (0.0002) (0.0006) Age 2 - 0.0000 1 (0.0000 05 ) Constant 0.6919 0.66 41 (0.0126) (0.0171) Variance components Country - level standard deviation 0.0659 0.0659 (0.0083 ) (0 .0083 ) Individual - level standard deviation 0.2324 0.2323 (0.0044 ) (0.0044 ) N umber of individuals 28017 28017 Number of countries 29 29 AIC - 149 2 - 1497 Note s : Cell entries are linear regression coefficients with the standard errors in parentheses, clustered by country in parentheses. The dependent variable is perceived dangerousness of climate change, ranging from 0 to 1. Coefficients estimates t hat can be distinguished from zero at the conventional level statistical significance (two - tailed p < 0.05) are in bold . Listwise deletion used for missing values. Israel, Taiwan, and Turkey were all excluded from the model s because relevant survey questio ns were not asked in these countries. 231 Table A2.10: Random - intercept multilevel linear models with individual - level interactions predicting perceived dangerousness of climate change (International Social Science Programme 2009 - 2011) (1) (2) (3) (4) Left party identification 0.0128 0.0130 (0.0049) (0.0046) Right party identification - 0.0350 - 0.0351 (0.0093) (0.0092) P arty 0.0482 0.0482 (left=1; right=0) (0.0092) (0.0089) Level of education 0.0044 0.0068 0.0044 0.0090 (0.0 029) (0.0025) (0.0046) (0.0027) Self - described environmental knowledge 0.0215 0.0198 0.0222 0.0097 (0.0028) (0.0034) (0.0037) (0.0058) Left party identification X Education 0.0072 (0.0032) Right party identification X Education - 0.0010 (0 .0051) Left party identification X Environmental 0.0096 Knowledge (0.0055) Right party identification X Environmental - 0.0095 Knowledge (0.0065) Party X Education 0.0072 (0.0045) Party X Environmental Knowledge 0 .0185 (0.0075) Post - materialism index - 0.0054 - 0.0054 - 0.0062 - 0.0063 (0.0030) (0.0030) (0.0034) (0.0034) A ttendance of religious service 0.0014 0.0014 0.0030 0.0030 (0.0011) (0.0011) (0.0017) (0.0017) Catholic 0.0016 0.0018 0.0000 0.0005 (0. 0074) (0.0074) (0.0088) (0.0088) Protestant - 0.0063 - 0.0061 - 0.0103 - 0.0100 (0.0113) (0.0111) (0.0128) (0.0127) Muslim 0.0181 0.0189 - 0.0108 - 0.0098 (0.0175) (0.0172) (0.0274) (0.0267) Female 0.0300 0.0300 0.0323 0.0322 (0.0054) (0.0054) (0.0058) (0.0057) Age younger than 30 - 0.0033 - 0.0034 0.0054 0.0053 (0.0065) (0.0065) (0.0073) (0.0073) Age 30 - 39 0.0104 0.0103 0.0136 0.0132 (0.0055) (0.0055) (0.0074) (0.0073) Age 50 - 59 - 0.0128 - 0.0130 - 0.0170 - 0.0171 (0.0044) (0.0044) (0.0063) (0.0062) Age 60 - 69 - 0.0196 - 0.0197 - 0.0217 - 0.0217 (0.0047) (0.0048) (0.0059) (0.0059) Age 70 and older - 0.0274 - 0.0277 - 0.0288 - 0.0291 (0.0064) (0.0064) (0.0080) (0.0080) 232 Variance components Country - level 0.0663 0.0661 0.0689 0. 0687 (0.008 ) (0.008 ) (0.0088 ) (0.0088 ) Individual - level 0.2327 0.2327 0.2341 0.2340 (0.0044 ) (0.0044 ) (0.0042 ) (0.0042 ) N umber of respondents 29526 29526 16220 16220 Number of countries 29 29 29 29 A IC - 1497 - 1502 - 626 - 635 Note s : Cell entries lin ear regression coefficients with the standard errors in parentheses, clustered by country in parentheses. The dependent variable is perceived dangerousness of climate change, ranging from 0 to 1. Coefficients estimates that can be distinguished from zero a t the conventional level statistical significance (two - tailed p < 0.05) are in bold . Listwise deletion used for missing values. Israel, Taiwan, and Turkey were all excluded from the model s because relevant survey questions were not asked in these countries . 233 Table A2.11: Random - intercept multilevel linear models predicting perceived dangerousness of climate change on multiple - imputation datasets (International Social Survey Programme 2009 - 2011) (1) (2) (3) (4) (5) (6) Left party identification 0.0146 0.0132 0.0148 0.0147 0.0147 0.0147 (0.0495) (0.0052) (0.0050) (0.0050) (0.0050) (0.0050) Right party identification - 0.0339 - 0.0317 - 0.0336 - 0.0336 - 0.0336 - 0.0336 (0.0084) (0.0094) (0.0084) (0.0084) (0.0084) (0.0084) Level of education 0.0081 0.0061 0.0080 0.0081 0.0080 0.0081 (0.0024) (0.0022) (0.0024) (0.0024) (0.0024) (0.0024) Environmental knowledge 0.0206 0.0210 0.0206 0.0206 0.0206 0.0206 (0.0027) (0.0028) (0.0027) (0.0027) (0.0027) (0.0027) Income gap - 0.0198 - 0.0191 - 0.0197 - 0.01 97 - 0.0198 - 0.0198 (0.0026) (0.0028) (0.0026) (0.0026) (0.0026) (0.0026) Private enterprise - 0.0105 - 0.0102 - 0.0105 - 0.0105 - 0.0105 - 0.0105 (0.0029) (0.0031) (0.0029) (0.0029) (0.0029) (0.0029) Post - materialism index 0.0019 0.0014 0.0019 0.0019 0.001 9 0.0019 (0.0024) (0.0025) (0.0024) (0.0024) (0.0025) (0.0025) Attendance of religious services 0.0011 0.0007 0.0011 0.0011 0.0011 0.0011 (0.0010) (0.0010) (0.0010) (0.0010) (0.0010) (0.0010) Catholic 0.0011 0.0006 0.0010 0.0010 0.0010 0.0010 (0.00 64) (0.0068) (0.0063) (0.0064) (0.0064) (0.0064) Protestant - 0.0056 - 0.0025 - 0.0053 - 0.0054 - 0.0054 - 0.0055 (0.0081) (0.0091) (0.0080) (0.0080) (0.0080) (0.0080) Muslim 0.0153 0.0093 0.0152 0.0152 0.0152 0.0153 (0.0164) (0.0165) (0.0165) (0.0165) (0. 0165) (0.0165) Self - described income decile 0.0008 (0.0016) Female 0.0319 0.0300 0.0320 0.0320 0.0320 0.0320 (0.0048) (0.0049) (0.0048) (0.0048) (0.0048) (0.0048) Age younger than 30 - 0.0067 - 0.0068 - 0.0067 - 0.0067 - 0.0067 - 0.0067 (0.006 1) (0.0064) (0.0061) (0.0061) (0.0061) (0.0061) Age 30 - 39 0.0067 0.0050 0.0067 0.0067 0.0067 0.0067 (0.0051) (0.0054) (0.0051) (0.0051) (0.0051) (0.0051) Age 50 - 59 - 0.0164 - 0.0163 - 0.0164 - 0.0164 - 0.0164 - 0.0164 (0.0059) (0.0061) (0.0059) (0.0059) (0 .0059) (0.0059) Age 60 - 69 - 0.0280 - 0.0275 - 0.0280 - 0.0280 - 0.0279 - 0.0279 (0.0058) (0.0063) (0.0058) (0.0058) (0.0058) (0.0058) Age 70 and older - 0.0364 - 0.0335 - 0.0364 - 0.0363 - 0.0363 - 0.0363 (0.0071) (0.0065) (0.0071) (0.0071) (0.0071) (0.0071) Co untry - level Natural logarithm of mean GDP - 0.0342 - 0.0348 per capita (0.0098) (0.0149) Mean HDI - 0.4000 - 0.2998 (0.1202) (0.2699) Mean polity score 0.0155 0.0105 (0.0128) (0.0130) Mean GINI coefficient 0.0 024 0.0025 (0.0019) (0.0024) Natural logarithm of mean CO 2 0.0054 0.0004 emissions per capita (in tons) (0.0183) (0.0231) 234 Natural logarithm of oil and gas - 0.0002 - 0.0008 rents per capita (0.0043) (0.0 045) Constant 0.6719 0.6731 1.0098 1.0097 0.7861 0.7493 (0.0135) (0.0174) (0.0996) (0.1037) (0.1572) (0.2136) Variance components Country - level 0.0666 0.0677 0.0589 0.0597 0.0559 0.0575 (0.0080) (0.0086 ) (0.0086 ) (0.0086 ) (0.0066 ) (0.0 071 ) Individual - level 0.2356 0.2343 0.2356 0.2356 0.2356 0.2356 (0.0043) (0.0044 ) (0.0043 ) (0.0043 ) (0.0043 ) (0.0043 ) N umber of respondents 40109 38009 40109 40109 40109 40109 Number of countries 29 27 29 29 29 29 Notes: Five MI datasets were used, c reated by Amelia II. Notes : Cell entries are linear regression coefficients with the standard errors in parentheses, clustered by country in parentheses. The dependent variable is perceived dangerousness of climate change, ranging from 0 to 1. Coefficients estimates that can be distinguished from zero at the conventional level statistical significance (two - tailed p < 0.05) are in bold . Israel, Taiwan, and Turkey were all excluded from the models because relevant survey questions were not asked in these coun tries. 235 Appendix C : Additional models, a nalysis and robustness c hecks for Chapter 5 Table A3.1: Random - intercept , random - coefficient multilevel linear models predicting concern about environmental issues (International Social Survey Programme 2009 - 2011) (1) (2) (3) (4) (5) (6) L e ft party identification 0.0207 0.0165 - 0.0343 - 0.0045 0.0165 0.0112 (0.0058) (0.0058) (0.0586) (0.0594) (0.0121) (0.0066) Right party identification - 0.0189 - 0.0148 0.3297 0.2890 0.0139 - 0.0292 (0.0089) (0.0098) (0.0761) (0.0995) (0.0212) (0.0092) Le vel of education 0.0080 0.0078 0.0077 0.0078 0.0078 0.0078 (0.0018) (0.0017) (0.0017) (0.0017) (0.0017) (0.0017) Environmental knowledge 0.0906 0.0903 0.0903 0.0903 0.0903 0.0903 (0.0068) (0.0068) (0.0067) (0.0067) (0.0068) (0.0067) Post - materialism index 0.0119 0.0114 0.0114 0.0114 0.0114 0.0113 (0.0038) (0.0038) (0.0038) (0.0038) (0.0038) (0.0038) Attendance of religious services 0.0035 0.0034 0.0034 0.0035 0.0035 0.0034 (0.0015) (0.0015) (0.0015) (0.0015) (0.0015) (0.0015) Protestant - 0.0010 - 0.0001 - 0.0001 - 0.0002 - 0.0001 - 0.0004 (0.0091) (0.0089) (0.0089) (0.0089) (0.0089) (0.0089) Female 0.0454 0.0447 0.0448 0.0448 0.0447 0.0447 (0.0041) (0.0040) (0.0040) (0.0041) (0.0040) (0.0040) Age younger than 30 - 0.0306 - 0.0307 - 0.0306 - 0.0306 - 0.0307 - 0.0307 (0.0063) (0.0063) (0.0063) (0.0063) (0.0063) (0.0063) Age 30 - 39 - 0.0133 - 0.0130 - 0.0129 - 0.0129 - 0.0130 - 0.0130 (0.0050) (0.0050) (0.0050) (0.0050) (0.0050) (0.0051) Age 50 - 59 0.0118 0.0116 0.0115 0.0116 0.0116 0.0116 (0.0050) (0.004 9) (0.0049) (0.0049) (0.0049) (0.0049) Age 60 - 69 0.0074 0.0073 0.0074 0.0073 0.0073 0.0074 (0.0069) (0.0069) (0.0068) (0.0069) (0.0068) (0.0068) Age 70 and older 0.0066 0.0069 0.0072 0.0071 0.0070 0.0070 (0.0098) (0.0097) (0.0098) (0.0098) (0.0098) ( 0.0097) Country - level variables Natural logarithm of GDP per capita 0.0022 (0.0136) HDI 0.1543 (0.1905) OECD country 0.0333 (0.0310) Post - communist country - 0.0725 (0.0315) Cross - level interactio n terms Left X Natural log of GDP per capita 0.0050 (0.0058) Right X Natural log of GDP per capita - 0.0342 (0.0075) Left X HDI 0.0240 (0.0694) Right X HDI - 0.3544 (0.1140) 236 L eft X OECD country - 0.0006 (0.0139) Right X OECD country - 0.0393 (0.0235) Left X post - communist country 0.0150 (0.0126) Right X post - communist country 0.0498 (0.0219) Constant 0.6506 0.6530 0.6309 0.5224 0.6308 0.6763 (0.0156) (0.0158) (0.1329) (0.1572) (0.0220) (0.0166) Variance components Country - level standard deviation 0.0717 (0.0082 ) 0.0788 (0.0105) 0.0787 (0.0105 ) 0.0780 (0.0107 ) 0.0772 (0.0118) 0.0709 (0.0088) Individual - level sta ndard deviation 0.2480 (0.0042 ) 0.2474 (0.0042 ) 0.2474 (0.0042 ) 0.2474 (0.0042 ) 0.2474 (0.0042) 0.2474 (0.0042) Left party identification standard deviation 0.0222 (0.0046 ) 0 .0229 ( 0.0045 ) 0.0228 (0.0044 ) 0.0227 (0.0045) 0.0208 (0.0044) Right - party identification standard deviation 0.0419 (0.0071 ) 0.0313 (0.0081 ) 0.0342 (0.0073 ) 0.0389 (0.0077) 0.0353 (0.0088) N umber of respondents 30471 30471 30471 30471 30471 30471 Number of countries 29 29 29 29 29 29 AIC 1189 1144 1138 114 2 1146 1142 Notes : Data is weighted. Cell entries are linear regression coefficients with standard error clustered by country in parentheses with list wise deletion . Bold indicates two - tailed p <0.05. Israel, Taiwan, and Turkey not included because of omit ted questions in these countries. 237 Table A3 .2: Random - intercept , random - coefficient m ultilevel linear models predicting perceived dangerousness of climate change (International Social Survey Programme 2009 - 2011) (1) (2) (3) (4) (5) (6) Left party identification 0.0127 0.0111 - 0.04 90 - 0.0334 0.0093 0.0161 (0.0047) (0.0046) (0.0645) (0.0585) (0.0078) (0.0055) Right party identification - 0.0359 - 0.0319 0.0806 0.1096 - 0.0196 - 0.0416 (0.0094) (0.0098) (0.1203) (0.1310) (0.0195) (0.0109) Level of education 0.0068 0.0068 0.0067 0.00 67 0.0067 0.0067 (0.0025) (0.0025) (0.0025) (0.0025) (0.0025) (0.0025) E nvironmental knowledge 0.0215 0.0214 0.0214 0.0214 0.0214 0.0214 (0.0028) (0.0028) (0.0028) (0.0028) (0.0028) (0.0028) Post - materialism index - 0.0054 - 0.0055 - 0.0055 - 0.0055 - 0.0 055 - 0.0056 (0.0030) (0.0030) (0.0031) (0.0030) (0.0030) (0.0030) Attendance of religious services 0.0015 0.0014 0.0013 0.0013 0.0014 0.0014 (0.0010) (0.0010) (0.0010) (0.0010) (0.0010) (0.0010) Protestant - 0.0074 - 0.0049 - 0.0047 - 0.0048 - 0.0049 - 0.0 055 (0.0100) (0.0095) (0.0094) (0.0094) (0.0094) (0.0094) Female 0.0300 0.0297 0.0297 0.0297 0.0297 0.0297 (0.0054) (0.0054) (0.0054) (0.0054) (0.0054) (0.0054) Age younger than 30 - 0.0033 - 0.0043 - 0.0044 - 0.0044 - 0.0043 - 0.0043 (0.0066) (0.0065) ( 0.0065) (0.0065) (0.0065) (0.0065) Age 30 - 39 0.0104 0.0099 0.0099 0.0099 0.0099 0.0101 (0.0055) (0.0056) (0.0056) (0.0056) (0.0056) (0.0056) Age 50 - 59 - 0.0132 - 0.0136 - 0.0136 - 0.0136 - 0.0136 - 0.0136 (0.0045) (0.0045) (0.0045) (0.0045) (0.0045) (0.004 5) Age 60 - 69 - 0.0201 - 0.0199 - 0.0199 - 0.0199 - 0.0199 - 0.0196 (0.0048) (0.0048) (0.0048) (0.0048) (0.0048) (0.0048) Age 70 and older - 0.0278 - 0.0279 - 0.0278 - 0.0278 - 0.0279 - 0.0273 (0.0064) (0.0065) (0.0065) (0.0065) (0.0065) (0.0065) Country - level Natural logarithm of mean GDP per capita - 0.0325 (0.0106) Mean HDI - 0.3707 (0.1323) OECD country - 0.0050 (0.0269) Post - communist country - 0.0125 (0.0253) Cross - level interaction terms Left X natural logarithm of 0.0061 mean GDP per capita (0.0065) Right X natural logarithm of - 0.0112 mean GDP per capita (0.0119) Left X mean HDI 0.0528 (0.0691) Right X mean HDI - 0.1655 (0.1516) L eft X OECD country 0.0026 (0.0099) Right X OECD country - 0.0168 (0.0228) 238 Left X Post - communist country - 0.0191 (0.0085) Right X Post - communist country 0.0410 (0.0210) Constant 0.688 1 0.6889 1.0108 1.0024 0.6922 0.6927 (0.0139) (0.0142) (0.1084) (0.1139) (0.0232) (0.0178) Variance components Country - level standard deviation 0.0662 (0.0080 ) 0.065 (0.0082 ) 0.0581 (0.0081 ) 0.0591 (0.0081 ) 0.0650 (0.0083 ) 0.0650 (0.0080 ) Individual - level standard deviation 0.2328 (0.0044 ) 0.2322 (0.0044 ) 0.2322 (0.0044 ) 0.2322 (0.0044 ) 0.2322 (0.0044 ) 0.02322 (0.0044 ) Left party identification standard deviation 0.0135 (0.004 ) 0.0123 (0.0044 ) 0.0131 (0.0040 ) 0.0135 0.0095 (0.0039 ) (0.0058 ) Right party identification standard deviation 0.0445 (0.0065 ) 0.0427 (0.006 ) 0.0424 (0.006 ) 0.0434 (0.0066 ) 0.0409 (0.0072 ) Number of respondents 29526 29526 29526 29526 29526 29526 Number of countries 29 29 29 29 29 29 AIC - 1495 - 1552 - 1551 - 1550 - 1546 - 1558 Notes : Data are weighted. Cell entries are linear regression coefficients with standard error clustered by country in parentheses with list wise deletion . Bold indicates two - tailed p <0.05. Israel, Taiwan, and Turkey not included because of omitted questions in these countries. 239 Appendix D : Additional models, a nalysis and robustness c hecks for Chapter 6 Table A4.1 : Ordered logistic regression models predicting perceived seriousness of environmental problems in the Unite d States (World Values Survey 2005 - 2009) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political ideology - 0.4226 - 0.2128 - 0.2513 (from left to right) (0.0543) (0 .0575) (0.0475) Post - materialism index 0.0914 0.0842 0.2048 (0.0531) (0.0629) (0.0574) Interest in politics - 0.0630 0.1971 0.0283 (0.0894) (0.0928) (0.0854) Level of education 0.0275 - 0.0409 - 0.0031 (0.0583) (0.0649) (0.0581) Attendance of r eligious services - 0.1114 - 0.0442 - 0.1628 (0.0401) (0.0426) (0.0420) Importance of God 0.1230 0.1178 0.1438 (0.0330) (0.0337) (0.0333) Protestant - 0.2440 - 0.1101 - 0.1078 (0.1598) (0.1695) (0.1594) Female 0.2785 0.3983 0.4595 (0.1424) (0.1543) (0 .1426) Age younger than 30 - 0.0740 - 0.1483 - 0.2895 (0.2418) (0.2481) (0.2425) Age 30 - 39 - 0.0562 0.1253 - 0.1839 (0.2171) (0.2479) (0.2219) Age 50 - 59 0.0327 - 0.0333 - 0.3521 (0.2217) (0.2242) (0.2226) Age 60 - 69 0.1159 - 0.1047 - 0.4117 (0.2752) (0.2 824) (0.2772) Age 70 and older 0.1839 0.0488 - 0.3496 (0.2748) (0.2623) (0.2588) Not serious at all| Not very serious - 5.0108 - 5.6436 - 4.6128 (0.4409) (0.4789) (0.4122) Not very serious | Somewhat serious - 3.5811 - 3.6840 - 2.9518 (0.4199) (0.4095) ( 0.3595) Somewhat serious | Very serious - 1.8762 - 1.5047 - 0.9767 (0.3950) (0.3968) (0.3448) Number of respondents 1154 1149 1152 AIC 2504 1860 2436 Standard errors in parentheses. p < 0.05 in bold. Listwise deletion and weighted. Respondents were aske d , Now following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 240 Table A4 .2 : Ordered logistic regression models predicting perceived seriousness of envir onmental problems in Australia (World Values Survey 2005 - 2009) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal spec ies or biodiversity Political ideology - 0.1042 - 0.0397 - 0.0306 (from left to right) (0.0363) (0.0401) (0.0351) Post - materialism index 0.0893 0.1162 0.2134 (0.0530) (0.0633) (0.0506) Interest in politics 0.0994 0.1206 0.0798 (0.0753) (0.08 90) (0.0762) Level of education - 0.0153 0.0233 - 0.0128 (0.0390) (0.0437) (0.0379) Attendance of religious services - 0.0599 - 0.0912 - 0.0845 (0.0403) (0.0467) (0.0383) Importance of God 0.0300 0.0096 - 0.0131 (0.0243) (0.0284) (0.0235) Protestant - 0 .0544 - 0.0351 - 0.0140 (0.1406) (0.1641) (0.1401) Female 0.4566 0.4157 0.3660 (0.1261) (0.1550) (0.1271) Age younger than 30 - 0.0469 - 0.1465 0.1968 (0.2190) (0.2850) (0.2266) Age 30 - 39 0.0804 - 0.2918 0.2753 (0.2147) (0.2646) (0.2115) Age 50 - 59 0 .2020 0.0524 0.1929 (0.1979) (0.2465) (0.1860) Age 60 - 69 - 0.2279 - 0.2422 0.0120 (0.2012) (0.2421) (0.1951) Age 70 and older - 0.3801 - 0.3800 - 0.3521 (0.2290) (0.2658) (0.2221) Not serious at all| Not very serious - 4.7491 - 6.9913 - 5.4040 (0.3503) (1.0411) (0.5128) Not very serious | Somewhat serious - 3.0182 - 4.2585 - 2.5804 (0.2927) (0.3787) (0.2786) Somewhat serious | Very serious - 0.9110 - 1.4016 - 0.3428 (0.2726) (0.3217) (0.2676) Number of respondents 1245 1246 1245 AIC 2118 1382 2105 Sta ndard errors in parentheses. p < 0.05 in bold. Listwise deletion and weighted. Respondents were asked , Now oblems in the world as a whole. Please, tell me how serious you consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 241 Table A4.3 : Ordered logistic regression models predicting perceived seriousness of environmental problems in Canada (World Values Survey 2005 - 2009) Global warm ing or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political ideology - 0.1309 - 0.1155 - 0.1288 (from left to right) (0.0459) (0.0511) (0.0375) Post - materialism index 0.0630 0.1073 0.1135 (0.0622) (0.0651) (0.0621) Interest in politics 0.1053 - 0.0139 0.0398 (0.0805) (0.0889) (0.0724) Level of education - 0.0411 - 0.0429 0.0208 (0.0459) (0.0571) (0.0461) Attendance of religious services - 0.0565 - 0.0522 - 0.0461 (0.0415) (0.05 70) (0.0444) Importance of God 0.0364 0.0680 0.0470 (0.0294) (0.0370) (0.0300) Protestant - 0.1609 0.0371 - 0.1063 (0.1967) (0.2325) (0.1919) Female 0.2389 0.1583 0.1759 (0.1426) (0.1737) (0.1414) Age younger than 30 - 0.1157 0.0056 0.3442 (0.2133 ) (0.2560) (0.2075) Age 30 - 39 - 0.1037 - 0.0210 0.0512 (0.2363) (0.2715) (0.2286) Age 50 - 59 0.0217 - 0.0293 0.1409 (0.2108) (0.2516) (0.2140) Age 60 - 69 - 0.2650 - 0.2883 - 0.3138 (0.2411) (0.3063) (0.2533) Age 70 and older - 0.2166 0.2283 - 0.2132 (0.2 698) (0.3258) (0.2607) Not serious at all| Not very serious - 5.1451 - 5.6962 - 5.2302 (0.4122) (0.6182) (0.4713) Not very serious | Somewhat serious - 3.5111 - 4.5698 - 3.1161 (0.3577) (0.4346) (0.3000) Somewhat serious | Very serious - 1.3911 - 1.8953 - 1. 0434 (0.3262) (0.3855) (0.2808) Number of respondents 1526 1554 1538 AIC 2641 1892 2761 Standard errors in parentheses. p < 0.05 in bold. Listwise deletion and weighted. Respondents were asked , Now oblems in the world a s a whole. Please, tell me how serious you consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 242 Table A4.4 : Ordered logistic regression models predicting perceived s eriousness of environmental problems in Chile (World Values Survey 2005 - 2009) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political ideology 0.0225 0.0086 - 0.031 0 (from left to right) (0.0536) (0.0597) (0.0563) Post - materialism index 0.1548 0.0423 0.0942 (0.0881) (0.0931) (0.0820) Interest in politics 0.1407 0.1115 0.1958 (0.1331) (0.1466) (0.1267) Level of education 0.2128 0.1830 * 0.0816 (0.0686) ( 0.0751) (0.0602) Attendance of religious services - 0.0287 - 0.0939 - 0.1334 (0.0639) (0.0728) (0.0626) Importance of God 0.1293 0.2411 0.2076 * (0.0526) (0.0526) (0.0536) Protestant - 0.0737 - 0.4428 - 0.3558 (0.3227) (0.3545) (0.3403) Female 0.2182 0. 2225 0.4788 (0.2402) (0.2822) (0.2498) Age younger than 30 - 0.5585 - 0.4992 - 0.2242 (0.3481) (0.3971) (0.3315) Age 30 - 39 - 0.1334 - 0.1415 0.0255 (0.3370) (0.3654) (0.3368) Age 50 - 59 - 0.5047 - 0.1146 - 0.6291 (0.3444) (0.4416) (0.3672) Age 60 - 69 0.3 523 0.0884 0.6034 (0.4809) (0.4682) (0.5848) Age 70 and older 0.0136 - 0.1022 - 0.1135 (0.4655) (0.5399) (0.4925) Not serious at all| Not very serious - 4.4610 (0.5898) Not very serious | Somewhat serious - 3.0441 - 3.2502 - 3.0174 (0.4387) (0.571 9) (0.4515) Somewhat serious | Very serious - 0.6554 - 0.8130 - 0.7260 (0.4088) (0.4815) (0.4277) Number of respondents 625 635 632 AIC 861 686 769 Standard errors in parentheses. p < 0.05 in bold. Listwise deletion and weighted. Respondents were asked , Now following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 243 Table A4.5 : Ordered logistic regression models predicting perceived seriousness of environmental problems in Italy (World Values Survey 2005 - 2009) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biod iversity Political ideology - 0.0776 - 0.0214 - 0.0770 (from left to right) (0.0440) (0.0493) (0.0407) Post - materialism index - 0.0033 0.2298 0.0108 (0.0749) (0.0827) (0.0673) Interest in politics - 0.0079 0.0055 0.0750 (0.1172) (0.1281) (0.10 29) Level of education - 0.0387 - 0.0095 - 0.0266 (0.0563) (0.0600) (0.0512) Attendance of religious services - 0.1619 - 0.2323 - 0.1826 (0.0753) (0.0770) (0.0625) Importance of God 0.1277 0.1692 0.1315 (0.0514) (0.0527) (0.0465) Protestant -- -- -- Female 0.2605 0.3347 0.1105 (0.1891) (0.2075) (0.1710) Age younger than 30 - 0.5040 - 0.6202 - 0.4326 (0.3043) (0.3433) (0.2614) Age 30 - 39 - 0.2377 - 0.3035 0.0954 (0.2932) (0.3428) (0.2588) Age 50 - 59 - 0.0597 - 0.2358 - 0.4000 (0.3209) (0.3692) (0. 2736) Age 60 - 69 - 0.6860 - 0.7442 - 0.7536 (0.3150) (0.3462) (0.2793) Age 70 and older - 0.6323 - 0.7071 - 0.8440 (0.4301) (0.4427) (0.4365) Not serious at all| Not very serious - 5.2388 - 6.7767 - 5.4926 (0.4824) (1.0768) (0.5339) Not very serious | Some what serious - 3.4798 - 4.9723 * - 3.3383 (0.3715) (0.5475) (0.3346) Somewhat serious | Very serious - 1.4445 - 1.5183 - 0.9603 (0.3491) (0.3921) (0.3048) N umber of respondents 674 681 674 AIC 1004 743 1148 Standard errors in parentheses. p < 0.05 in bold . Listwise deletion and weighted. Respondents were asked , Now following to be for the world as a whole. Is it very serious, somewhat seriou s, not very serious or not serious at all? There were no Protestants in the sample in Italy, so this variable was not included in the model. 244 Tab le A4.6 : Ordered logistic regression models predicting perceived seriousness of environmental problem s in Spain (World Values Survey 2005 - 2009 ) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political ideology - 0.2294 - 0.2640 - 0.2216 (from left to right) (0.0415) (0.0453) (0.0431) Post - materialism index - 0.1852 - 0.1630 - 0.2518 (0.0613) (0.0647) (0.0620) Interest in politics - 0.2502 - 0.3077 - 0.3000 (0.0909) (0.1076) (0.0961) Level of education 0.1039 0.0346 0.0582 (0.0455) (0.0469) (0.0439) Attendance of religious services - 0.0164 - 0.0082 - 0.0764 (0.0528) (0.0628) (0.0577) Importance of God 0.0482 0.0457 0.0641 (0.0313) (0.0371) (0.0340) Protestant 12.9063 11.5962 14.0554 (0.5531) (0.5943) (0.5675) Female 0.0554 0.0217 0.1504 (0.1606) (0.182 0) (0.1621) Age younger than 30 0.0793 0.2067 0.2039 (0.2603) (0.2959) (0.2650) Age 30 - 39 - 0.0257 - 0.0791 0.0446 (0.2550) (0.2808) (0.2556) Age 50 - 59 - 0.2414 - 0.2442 - 0.0668 (0.2635) (0.3029) (0.2744) Age 60 - 69 0.3689 0.1055 0.2907 (0.3091) (0. 3290) (0.3001) Age 70 and older - 0.4583 - 0.3458 - 0.3521 (0.2798) (0.3359) (0.2851) Not serious at all| Not very serious - 8.0807 - 8.2705 - 7.6428 (1.0217) (1.0509) (1.0695) Not very serious | Somewhat serious - 4.6708 - 5.0685 - 4.3043 (0.3675) (0.3992 ) (0.3743) Somewhat serious | Very serious - 2.0774 - 2.7065 - 1.8880 (0.3154) (0.3601) (0.3340) N umber of respondents 952 954 954 AIC 1211 1017 1199 Standard errors in parentheses. p < 0.05 in bold. Listwise deletion and weighted. Respondents were aske d , Now following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 245 Table A4. 7 : Ordered logistic regression models predicting perceived seriousness of environmental problems in Switzerland (World Values Survey 2005 - 2009) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal spe cies or biodiversity Political ideology - 0.2014 - 0.1518 - 0.0879 (from left to right) (0.0404) (0.0422) (0.0363) Post - materialism index 0.0857 0.1147 0.1431 (0.0602) (0.0639) (0.0587) Interest in politics 0.0658 0.0795 - 0.0758 (0.0928) (0. 0932) (0.0845) Level of education 0.0566 - 0.0315 - 0.1274 (0.0438) (0.0462) (0.0421) Attendance of religious services - 0.0484 - 0.0630 0.0067 (0.0468) (0.0467) (0.0424) Importance of God 0.0542 0.0089 - 0.0153 (0.0288) (0.0284) (0.0266) Protestant - 0.2553 - 0.0421 - 0.1623 (0.1429) (0.1527) (0.1416) Female 0.4629 0.1634 0.4032 (0.1404) (0.1500) (0.1388) Age younger than 30 0.2591 0.2075 0.4911 (0.2996) (0.3209) (0.3199) Age 30 - 39 0.4852 0.0590 0.0528 (0.2361) (0.2419) (0.2014) Age 50 - 59 - 0. 1564 0.2492 0.1478 (0.1999) (0.2301) (0.2151) Age 60 - 69 0.1581 - 0.2035 0.0214 (0.2142) (0.2190) (0.1993) Age 70 and older - 0.1249 0.1289 0.4880 (0.2355) (0.2465) (0.2451) Not serious at all| Not very serious - 4.7302 - 7.9762 - 4.8752 (0.3895) (1.0 611) (0.4176) Not very serious | Somewhat serious - 2.6874 - 3.8949 - 2.4483 (0.3118) (0.3485) (0.2954) Somewhat serious | Very serious - 0.5347 - 1.1549 - 0.0037 (0.2974) (0.3214) (0.2854) N umber of respondents 1015 1016 1016 AIC 2079 1626 2097 Standar d errors in parentheses. p < 0.05 in bold. Listwise deletion and weighted. Respondents were asked , Now oblems in the world as a whole. Please, tell me how serious you consider each of the following to be for the world as a w hole. Is it very serious, somewhat serious, not very serious or not serious at all? 246 Table A4.8 : Ordered logistic regression models predicting perceived seriousness of environmental problems in South Korea (World Values Survey 2005 - 2009) Global war ming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political ideology - 0.1035 * - 0.0375 - 0.0003 (from left to right) (0.0430) (0.0435) (0.0393) Post - materialism index 0.0192 0.05 68 - 0.0171 (0.0845) (0.0801) (0.0785) Interest in politics 0.1162 - 0.0021 0.0449 (0.1083) (0.1128) (0.1059) Level of education 0.0464 0.0664 0.2228 (0.0607) (0.0682) (0.0612) Attendance of religious services - 0.0025 0.0467 - 0.0341 (0.0470) (0.04 66) (0.0498) Importance of God 0.0011 0.0120 0.0549 (0.0375) (0.0379) (0.0373) Protestant - 0.0713 - 0.3167 - 0.2279 (0.2262) (0.2141) (0.1998) Female - 0.0225 0.0588 0.0368 (0.1522) (0.1511) (0.1609) Age younger than 30 0.0357 - 0.0560 - 0.2608 (0.2 139) (0.2203) (0.2255) Age 30 - 39 0.1133 0.0942 - 0.2131 (0.2199) (0.2266) (0.2132) Age 50 - 59 0.2268 0.1844 0.0576 (0.2660) (0.2577) (0.2813) Age 60 - 69 - 0.7963 - 0.0997 - 0.4882 (0.2968) (0.2964) (0.3098) Age 70 and older - 1.0156 0.3094 0.8924 (0.7 308) (0.6311) (0.6288) Not serious at all| Not very serious - 7.5663 - 6.7865 (1.0668) (0.7992) Not very serious | Somewhat serious - 3.8637 - 3.2948 - 2.7183 (0.3985) (0.4276) (0.3930) Somewhat serious | Very serious - 0.6216 0.0893 0.7251 (0.3581) ( 0.3779) (0.3796) Number of respondents 1187 1187 1187 AIC 1957 2014 2013 Standard errors in parentheses. p < 0.05 in bold. Listwise deletion and weighted. Respondents were asked , Now oblems in the world as a whole. Please , tell me how serious you consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 247 Table A4.9 : Ordered logistic regression models predicting perceived seriousness of env ironmental problems in Norway (World Values Survey 2005 - 2009) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political ideology - 0.0935 - 0.0886 - 0.1010 (from left to right) (0.0355) (0.0419) (0.0375) Post - materialism index 0.3507 0.2002 0.3681 (0.0703) (0.0692) (0.0695) Interest in politics 0.0474 0.1212 0.0811 (0.1007) (0.1043) (0.1051) Level of education 0.1300 0.0861 0.0038 (0.0400) (0.0437) (0.0424 ) Attendance of religious services 0.0522 - 0.0596 - 0.0002 (0.0521) (0.0561) (0.0540) Importance of God 0.0238 0.0729 0.0327 (0.0293) (0.0313) (0.0307) Protestant - 0.3185 - 0.4286 - 0.5913 (0.1467) (0.1619) (0.1549) Female 0.2668 0.3591 0.2259 (0. 1363) (0.1510) (0.1400) Age younger than 30 - 0.2417 0.0127 0.2186 (0.2147) (0.2311) (0.2217) Age 30 - 39 - 0.3578 - 0.0819 - 0.1336 (0.2052) (0.2227) (0.2108) Age 50 - 59 - 0.2140 0.3836 - 0.1548 (0.2208) (0.2513) (0.2103) Age 60 - 69 - 0.2985 0.2011 - 0.0649 (0.2312) (0.2438) (0.2310) Age 70 and older 0.1679 0.9416 0.3399 (0.2841) (0.3544) (0.3178) Not serious at all| Not very serious - 4.8097 - 5.8127 - 6.1624 (0.4120) (0.6087) (0.5449) Not very serious | Somewhat serious - 3.1522 - 3.7963 - 3.7534 (0.3 293) (0.3850) (0.3541) Somewhat serious | Very serious - 0.9235 - 1.1508 - 1.2976 (0.3131) (0.3471) (0.3253) Number of respondents 966 969 969 AIC 1717 1341 1564 Standard errors in parentheses. p < 0.05 in bold. Listwise deletion and weighted. Responden ts were asked , Now oblems in the world as a whole. Please, tell me how serious you consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 248 Table A4.10 : Ordered logistic regression models predicting perceived seriousness of environmental problems in Finland (World Values Survey 2005 - 2009) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or a nimal species or biodiversity Political ideology - 0.0821 - 0.0948 - 0.0720 (from left to right) (0.0370) (0.0386) (0.0364) Post - materialism index 0.0963 0.0029 0.0013 (0.0656) (0.0657) (0.0640) Interest in politics 0.0578 0.1991 0.0656 (0.0 927) (0.0930) (0.0881) Level of education 0.0964 0.0416 0.0221 (0.0387) (0.0404) (0.0392) Attendance of religious services 0.0042 - 0.0372 0.0642 (0.0521) (0.0563) (0.0524) Importance of God 0.0032 0.0481 - 0.0011 (0.0304) (0.0298) (0.0286) Protest ant -- -- -- Female 0.4748 0.6508 0.5071 (0.1433) (0.1533) (0.1425) Age younger than 30 - 0.1147 - 0.5244 - 0.0248 (0.2296) (0.2510) (0.2214) Age 30 - 39 0.0406 - 0.3623 - 0.2701 (0.2134) (0.2410) (0.2152) Age 50 - 59 - 0.0052 - 0.5397 - 0.3333 (0.220 2) (0.2457) (0.2267) Age 60 - 69 0.0769 - 0.5522 - 0.3385 (0.2506) (0.2773) (0.2467) Age 70 and older 0.1404 - 0.5337 - 0.0896 (0.2548) (0.2909) (0.2765) Not serious at all| Not very serious - 5.2717 - 6.1876 - 4.7104 (0.5007) (0.5917) (0.3959) Not very s erious | Somewhat serious - 2.4810 - 3.9280 - 2.2844 (0.2890) (0.3561) (0.2888) Somewhat serious | Very serious - 0.1747 - 1.0424 0.0948 (0.2782) (0.2979) (0.2767) Number of respondents 847 853 847 AIC 1609 1340 1748 Standard errors in parentheses. p < 0.05 in bold. Listwise deletion and weighted. Respondents were asked , Now following to be for the world as a whole. Is it very serious, som ewhat serious, not very serious or not serious at all? There were no Pr otestants in the sample in Finland , so this variable was not included in the model. 249 Table A4.11 : Ordered logistic regression models predicting perceived seriousness of environmental problems in Germany (World Values Survey 2005 - 2009) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political ideology - 0.0854 - 0.1018 - 0.1155 (from left to right) (0.0332) (0.0315) (0.0305) Post - materialism index 0.0861 0.0602 0.1245 (0.0541) (0.0500) (0.0487) Interest in politics 0.0964 0.2710 0.1063 (0.0730) (0.0791) (0.0652) Level of education 0.0853 0.0433 0.0737 (0.0325) (0.0337) (0.0279) Attend ance of religious services - 0.0115 - 0.1400 - 0.0792 (0.0463) (0.0478) (0.0418) Importance of God 0.0287 0.0661 * 0.0257 (0.0256) (0.0265) (0.0231) Protestant -- -- -- Female 0.3060 0.4967 0.3994 (0.1124) (0.1134) (0.1093) Age younger than 30 - 0.1810 0.1070 - 0.0541 (0.1908) (0.1935) (0.1798) Age 30 - 39 - 0.0395 - 0.0964 - 0.1562 (0.1807) (0.1794) (0.1836) Age 50 - 59 0.0298 - 0.1761 - 0.1434 (0.1954) (0.2045) (0.1749) Age 60 - 69 0.0892 0.0206 - 0.0659 (0.1801) (0.1778) (0.1688) Age 70 and olde r 0.0332 - 0.3311 - 0.1267 (0.1932) (0.1750) (0.1825) Not serious at all| Not very serious - 5.4579 - 4.9766 - 4.9608 (0.3885) (0.3914) (0.3141) Not very serious | Somewhat serious - 2.5777 - 1.9760 - 1.8627 (0.2591) (0.2399) (0.2138) Somewhat serious | V ery serious - 0.2010 0.1624 0.2573 (0.2385) (0.2256) (0.2088) Number of respondents 1648 1664 1656 AIC 3093 3348 3478 Standard errors in parentheses. p < 0.05 in bold. Listwise deletion and weighted. Respondents were asked , Now ental problems in the world as a whole. Please, tell me how serious you consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? Germany were not asked about their religio us identity, so the Protestant variable was not included in the model. 250 Table A4.12 : Ordered logistic regression models predicting perceived seriousness of environmental problems in Sweden (World Values Survey 2005 - 2009) Global warming or the green house effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political ideology - 0.1030 - 0.0912 - 0.1327 (from left to right) (0.0367) (0.0405) (0.0354) Post - materialism index 0.0556 0.1009 0.0070 (0.069 4) (0.0713) (0.0675) Interest in politics 0.0857 0.1591 0.1341 (0.0959) (0.1100) (0.0976) Level of education 0.0612 0.0319 - 0.0825 (0.0463) (0.0504) (0.0486) Attendance of religious services - 0.0284 - 0.0898 - 0.1082 (0.0640) (0.0666) (0.0635) Impo rtance of God 0.0211 - 0.0065 0.0138 (0.0302) (0.0308) (0.0285) Protestant - 2.0272 - 1.1979 - 0.4769 (1.5016) (2.3774) (2.0109) Female 0.5473 0.8949 0.5475 (0.1578) (0.1649) (0.1460) Age younger than 30 - 0.3975 - 0.3517 0.1156 (0.2438) (0.2585) (0.2 282) Age 30 - 39 - 0.0428 - 0.3259 - 0.1109 (0.2539) (0.2632) (0.2356) Age 50 - 59 - 0.0493 0.2488 0.1205 (0.2447) (0.2714) (0.2209) Age 60 - 69 - 0.4463 - 0.3830 - 0.2274 (0.2605) (0.2756) (0.2453) Age 70 and older 0.0983 0.1850 - 0.2113 (0.3115) (0.3303) ( 0.3064) Not serious at all| Not very serious - 6.1300 - 6.8355 - 5.9905 (0.6541) (1.0764) (0.5837) Not very serious | Somewhat serious - 3.2982 - 3.9998 - 3.1033 (0.3595) (0.4175) (0.3241) Somewhat serious | Very serious - 0.8670 - 0.9926 - 0.9135 (0.3290) (0.3440) (0.3014) Number of respondents 915 925 923 AIC 1473 1164 1600 Standard errors in parentheses. p < 0.05 in bold. Listwise deletion and weighted. Respondents were asked , Now oblems in the world as a whole. Please, tell me how serious you consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 251 Table A4.13 : Ordered logistic regression models predicting perceived seriousness of env ironmental problems in Japan (World Values Survey 2005 - 2009) (1) (2) (3) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political ideology - 0.1003 - 0.0244 - 0.1074 (fr om left to right) (0.0499) (0.0468) (0.0472) Post - materialism index 0.1049 0.1074 0.1908 (0.0793) (0.0710) (0.0695) Interest in politics 0.2197 0.1787 0.1959 (0.1324) (0.1202) (0.1238) Level of education 0.0573 0.0018 0.0503 (0.0648) (0.0575) (0.0597) Attendance of religious services 0.0013 0.0794 0.0069 (0.0797) (0.0701) (0.0681) Importance of God 0.0733 0.0574 0.0554 (0.0363) (0.0333) (0.0321) Protestant - 0.4212 - 0.6279 0.7375 (0.9223) (0.7843) (0.8892) Female 0.5198 * 0.4471 0.4008 (0.1791) (0.1611) (0.1655) Age younger than 30 - 0.1482 0.1900 0.3891 (0.2974) (0.2676) (0.2611) Age 30 - 39 0.3368 0.1285 0.2176 (0.3246) (0.2799) (0.2484) Age 50 - 59 - 0.3107 0.0421 - 0.0647 (0.2662) (0.2498) (0.2223) Age 60 - 69 - 0.2094 - 0.1385 - 0. 2470 (0.3002) (0.2622) (0.2395) Age 70 and older - 0.4064 - 0.4166 0.2709 (0.3443) (0.3140) (0.3189) Not serious at all| Not very serious - 6.1560 (0.8327) Not very serious | Somewhat serious - 4.3250 - 3.5287 - 2.8565 (0.4809) (0.4026) (0.3629) Somewhat serious | Very serious - 1.2697 - 0.3105 - 0.1601 (0.3929) (0.3527) (0.3515) Number of respondents 673 671 667 AIC 901 1067 1223 Standard errors in parentheses. p < 0.05 in bold. Listwise deletion and weighted. Respondents were asked , Now consider environmental pr oblems in the world as a whole. Please, tell me how serious you consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 252 Table A4.14 : Ordered l ogistic regression models predicting perceived seriousness of environmental problems in Mexico (World Values Survey 2005 - 2009) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodivers ity Political ideology - 0.0274 - 0.0194 - 0.0202 (from left to right) (0.0229) (0.0286) (0.0251) Post - materialism index - 0.1346 - 0.0244 - 0.0779 (0.0593) (0.0749) (0.0629) Interest in politics 0.1313 0.2370 * 0.0616 (0.0670) (0.0859) (0.0735) Level of education 0.2307 0.2764 0.2247 (0.0344) (0.0450) (0.0395) Attendance of religious services - 0.1115 - 0.0465 - 0.0653 (0.0376) (0.0494) (0.0436) Importance of God 0.0561 0.1224 0.1095 (0.0477) (0.0546) (0.0461) Protestant 1.2042 14.3232 1. 5021 (0.8065) (0.3402) (1.0492) Female - 0.1446 - 0.0815 - 0.1087 (0.1307) (0.1740) (0.1455) Age younger than 30 - 0.6431 - 0.3083 - 0.2380 (0.1941) (0.2809) (0.2185) Age 30 - 39 0.1572 - 0.4564 0.1274 (0.2115) (0.2778) (0.2344) Age 50 - 59 0.1717 - 0.0139 0.1212 (0.2507) (0.3342) (0.2721) Age 60 - 69 0.4956 0.1158 0.5704 (0.3088) (0.3722) (0.3409) Age 70 and older 0.1945 0.9930 0.0552 (0.3532) (0.6168) (0.4113) Not serious at all| Not very serious - 4.2540 - 4.3838 - 3.7757 (0.3851) (0.4722) (0.3699) Not very serious | Somewhat serious - 2.5331 - 3.2447 - 2.5007 (0.3125) (0.4193) (0.3283) Somewhat serious | Very serious - 0.9491 - 1.6601 - 1.0032 (0.2962) (0.3817) (0.3059) Number of respondents 1220 1240 1240 AIC 1994 1259 1658 Standard errors in p arentheses. p < 0.05 in bold. Listwise deletion and weighted. Respondents were asked , Now oblems in the world as a whole. Please, tell me how serious you consider each of the following to be for the world as a whole. Is it v ery serious, somewhat serious, not very serious or not serious at all? 253 Table A4.15 : Ordered logistic regression models predicting perceived seriousness of environmental problems in Slovenia (World Values Survey 2005 - 2009) Global warming or the gre enhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political ideology - 0.0499 - 0.0469 - 0.0697 (from left to right) (0.0428) (0.0441) (0.0424) Post - materialism index 0.1026 - 0.0999 - 0.0876 (0 .0866) (0.0884) (0.0858) Interest in politics 0.2572 0.1297 0.1476 (0.1095) (0.1146) (0.1058) Level of education - 0.0928 - 0.0203 - 0.0689 (0.0525) (0.0536) (0.0511) Attendance of religious services - 0.0472 - 0.0742 - 0.0377 (0.0590) (0.0585) (0.0540) Importance of God 0.0411 0.0348 0.0259 (0.0351) (0.0347) (0.0332) Protestant - 0.7881 - 0.5693 0.0160 (0.8189) (0.7679) (0.7381) Female - 0.0709 0.1339 0.2106 (0.1733) (0.1800) (0.1664) Age younger than 30 0.0638 0.2365 0.1667 (0.2630) (0.2836) ( 0.2600) Age 30 - 39 - 0.3373 0.0299 - 0.1309 (0.2727) (0.2867) (0.2735) Age 50 - 59 0.1219 0.1617 0.0631 (0.2787) (0.2881) (0.2666) Age 60 - 69 0.0718 0.1400 0.0965 (0.3168) (0.3103) (0.3232) Age 70 and older - 0.1644 - 0.1624 - 0.3134 (0.3338) (0.3610) ( 0.3217) Not serious at all| Not very serious - 5.4597 - 5.0939 - 4.7657 (0.5819) (0.5569) (0.4746) Not very serious | Somewhat serious - 3.1351 - 3.7516 - 3.1075 (0.3451) (0.4071) (0.3763) Somewhat serious | Very serious - 0.9333 - 0.8306 - 0.5702 (0.3246) (0.3390) (0.3273) Number of respondents 571 589 585 AIC 1000 887 1063 Standard errors in parentheses. p < 0.05 in bold. Listwise deletion and weighted. Respondents were asked , Now tell me how serious you consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 254 Table A4.16 : Ordered logistic regression models predicting perceived seriousness of env ironmental problems in Turkey (World Values Survey 2005 - 2009) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political ideology - 0.1548 - 0.0598 - 0.0656 (from left to right) (0.0473) (0.0395) (0.0355) Post - materialism index - 0.2126 - 0.3392 - 0.1883 (0.0787) (0.0724) (0.0671) Interest in politics - 0.0244 0.0611 0.0337 (0.1018) (0.0926) (0.0912) Level of education 0.1205 0.1541 0.0943 (0.0506) (0.0484) (0. 0417) Attendance of religious services 0.0654 0.0576 0.0722 (0.0556) (0.0546) (0.0501) Importance of God 0.0754 - 0.0198 0.0069 (0.0654) (0.0603) (0.0595) Protestant -- -- -- Female 0.1735 - 0.0732 - 0.2684 (0.2413) (0.2282) (0.2108) Age young er than 30 - 0.5749 - 0.2592 - 0.4243 (0.2835) (0.2736) (0.2361) Age 30 - 39 - 0.3466 - 0.0029 - 0.2845 (0.3326) (0.3172) (0.2755) Age 50 - 59 - 0.4500 0.3092 - 0.3558 (0.3632) (0.3800) (0.3254) Age 60 - 69 0.2059 - 0.8371 - 0.6155 (0.6036) (0.4390) (0.4746) A ge 70 and older 0.0618 0.2590 - 0.3133 (0.6864) (0.7149) (0.5335) Not serious at all| Not very serious - 6.0114 - 5.9415 - 5.2802 (0.5188) (0.6917) (0.4851) Not very serious | Somewhat serious - 4.3253 - 4.4315 - 3.9957 (0.4214) (0.4284) (0.3775) Somewha t serious | Very serious - 2.9067 - 2.3967 - 2.2397 (0.4065) (0.3837) (0.3490) Number of respondents 1028 1044 1038 AIC 931 959 1251 Ordered logistic coefficient. Standard errors in parentheses. p < 0.05 in bold. Listwise deletion. Data weighted. Respond ents were asked , serious you consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all ? There were n o Protestants in the sample in Turkey , so this variable was not included in the model. 255 Table A4.17 : Ordered logistic regression models predicting perceived seriousness of environmental problems in Poland (World Values Survey 2005 - 2009) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political ideology - 0.0107 - 0.0178 - 0.0121 (from left to right) (0.0386) (0.0443) (0.0409) Post - materialism inde x - 0.0092 0.0851 0.0181 (0.0900) (0.0984) (0.0909) Interest in politics 0.0465 0.3320 * 0.1514 (0.1063) (0.1161) (0.1074) Level of education - 0.0303 - 0.0488 - 0.0662 (0.0513) (0.0574) (0.0516) Attendance of religious services - 0.0258 - 0.0756 - 0.0529 (0.0678) (0.0754) (0.0735) Importance of God 0.1048 0.1096 0.1487 (0.0490) (0.0496) (0.0544) Protestant 0.8636 14.9454 0.3045 (1.1710) (0.5856) (1.0543) Female - 0.3500 - 0.0817 0.0599 (0.1663) (0.1904) (0.1694) Age younger than 30 0.3730 0.2453 0.1709 (0.2513) (0.2629) (0.2506) Age 30 - 39 0.5541 0.7450 0.4485 (0.2830) (0.3276) (0.2968) Age 50 - 59 0.2800 0.3581 0.3616 (0.2418) (0.2615) (0.2440) Age 60 - 69 0.5392 0.4462 0.3400 (0.3120) (0.3565) (0.3066) Age 70 and older 0.1271 0.1344 0.09 46 (0.3316) (0.3681) (0.3323) Not serious at all| Not very serious - 4.3125 - 4.9057 - 4.4129 (0.5575) (0.6533) (0.5972) Not very serious | Somewhat serious - 2.5205 - 3.5058 - 2.4066 (0.3520) (0.4225) (0.3589) Somewhat serious | Very serious 0.1022 - 0. 6645 0.1159 (0.3084) (0.3272) (0.3163) Number of respondents 671 675 674 AIC 1177 920 1147 Standard errors in parentheses. p < 0.05 in bold. Listwise deletion and weighted. Respondents were asked , Now d as a whole. Please, tell me how serious you consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 256 Table A4.18 : Linear regression models predicting perceived seriou sness of envir onmental problems in the United States (World Values Survey 2005 - 2009) (1) (2) (3) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political ideology - 0.0613 - 0.0227 - 0.0353 (from left to right) (0.0076) (0.0064) (0.0069) Post - materialism index 0.0143 0.0067 0.0251 (0.0074) (0.0062) (0.0075) Interest in politics - 0.0134 0.0133 0.0002 (0.0126) (0.0091) (0.0116) Level of education 0.0028 - 0.0020 0.001 9 (0.0086) (0.0066) (0.0080) Attendance of religious services - 0.0146 - 0.0025 - 0.0207 (0.0055) (0.0044) (0.0055) Importance of God 0.0162 0.0103 0.0192 (0.0046) (0.0036) (0.0047) Protestant - 0.0449 - 0.0083 - 0.0160 (0.0239) (0.0162) (0.0220) Fem ale 0.0418 0.0357 0.0611 (0.0204) (0.0153) (0.0194) Age younger than 30 - 0.0176 - 0.0273 - 0.0424 (0.0359) (0.0254) (0.0329) Age 30 - 39 - 0.0090 - 0.0008 - 0.0198 (0.0303) (0.0227) (0.0278) Age 50 - 59 0.0161 - 0.0027 - 0.0447 (0.0302) (0.0200) (0.0277) Age 60 - 69 0.0233 - 0.0179 - 0.0579 (0.0391) (0.0277) (0.0376) Age 70 and older 0.0265 0.0119 - 0.0424 (0.0388) (0.0219) (0.0352) Intercept 1.0386 0.9602 0.9148 (0.0572) (0.0426) (0.0503) N umber of respondents 1154 1149 1152 R 2 0.1732 0.0634 0.1244 adj. R 2 0.1637 0.0527 0.1144 OLS coefficients. Standard errors in parentheses. p < 0.05 in bold. Listwise deletion. Data weighted. The dependent variable ranges from 0 to 1 (0 = not serious at all; 0.33 = not very serious; somewhat serious = 0.67; 1.0 = v ery serious). Respondents were asked , oblems in the world as a whole. Please, tell me how serious you consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 257 Table A6.19 : Linear regression models predicting perceived seriousness of envir onmental problems in Italy (World Values Survey 2005 - 2009) (1) (2) (3) Global warming or the greenhouse effect Pollution of rivers, lakes, and ocean s Loss of plant or animal species or biodiversity Political ideology - 0.0094 - 0.0012 - 0.0072 (from left to right) (0.0041) (0.0026) (0.0042) Post - materialism index - 0.0024 0.0126 0.0030 (0.0070) (0.0048) (0.0071) Interest in politics - 0.000 3 0.0017 0.0072 (0.0103) (0.0071) (0.0103) Level of education - 0.0047 - 0.0005 - 0.0033 (0.0050) (0.0036) (0.0051) Attendance of religious services - 0.0104 - 0.0105 - 0.0149 (0.0073) (0.0039) (0.0059) Importance of God 0.0099 0.0098 0.0128 (0.0048) (0.0030) (0.0046) Protestant -- -- -- Female 0.0245 0.0187 0.0128 (0.0170) (0.0114) (0.0170) Age younger than 30 - 0.0363 - 0.0340 - 0.0503 (0.0246) (0.0176) (0.0253) Age 30 - 39 - 0.0102 - 0.0188 0.0046 (0.0219) (0.0163) (0.0215) Age 50 - 59 - 0.001 3 - 0.0108 - 0.0450 (0.0230) (0.0163) (0.0252) Age 60 - 69 - 0.0588 - 0.0425 - 0.0837 (0.0281) (0.0192) (0.0286) Age 70 and older - 0.0630 - 0.0440 - 0.1004 (0.0455) (0.0272) (0.0441) Intercept 0.9398 0.9312 0.9008 (0.0298) (0.0197) (0.0298) N umber of re spondents 674 681 674 R 2 0.0371 0.0520 0.0544 adj. R 2 0.0196 0.0350 0.0372 OLS coefficients. Standard errors in parentheses. p < 0.05 in bold. Listwise deletion. Data weighted. The dependent variable ranges from 0 to 1 (0 = not serious at all; 0.33 = no t very serious; somewhat serious = 0.67; 1.0 = very serious). Respondents were asked , oblems in the world as a whole. Please, tell me how serious you consider each of the following to be for the world as a whole. Is it v ery serious, somewhat serious, not very serious or not serious at all? 258 Table A4.20 : Linear regression models predicting perceived seriousness of envir onmental problems in Spain (World Values Survey 2005 - 2009) (1) (2) (3) Global warming or the greenho use effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political ideology - 0.0163 - 0.0158 - 0.0155 (from left to right) (0.0031) (0.0028) (0.0031) Post - materialism index - 0.0122 - 0.0089 - 0.0174 (0.00 44) (0.0035) (0.0042) Interest in politics - 0.0195 - 0.0213 - 0.0237 (0.0065) (0.0065) (0.0068) Level of education 0.0062 0.0026 0.0048 (0.0031) (0.0025) (0.0029) Attendance of religious services 0.0005 0.0020 - 0.0025 (0.0037) (0.0036) (0.0041) Imp ortance of God 0.0020 0.0011 0.0021 (0.0020) (0.0019) (0.0021) Protestant 0.0724 0.0509 0.0862 (0.0171) (0.0188) (0.0197) Female 0.0024 0.0026 0.0123 (0.0109) (0.0099) (0.0109) Age younger than 30 0.0030 0.0061 0.0087 (0.0171) (0.0147) (0.0172) Age 30 - 39 - 0.0014 - 0.0102 - 0.0013 (0.0172) (0.0157) (0.0177) Age 50 - 59 - 0.0208 - 0.0157 - 0.0064 (0.0204) (0.0181) (0.0203) Age 60 - 69 0.0174 0.0056 0.0191 (0.0199) (0.0169) (0.0205) Age 70 and older - 0.0383 - 0.0297 - 0.0279 (0.0227) (0.0221) (0.02 30) Intercept 0.9768 1.0025 0.9592 (0.0214) (0.0196) (0.0238) N umber of respondents 952 954 954 R 2 0.0656 0.0650 0.0739 adj. R 2 0.0527 0.0520 0.0611 OLS coefficients. Standard errors in parentheses. p < 0.05 in bold. Listwise deletion. Data weighted . The dependent variable ranges from 0 to 1 (0 = not serious at all; 0.33 = not very serious; somewhat serious = 0.67; 1.0 = very serious). Respondents were asked , oblems in the world as a whole. Please, tell me how seri ous you consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 259 Table A4.21: Linear regression models predicting perceived seriousness of environmental problems in Swit zerland (World Values Survey 2005 - 2009) (1) (2) (3) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political ideology - 0.0249 - 0.0138 - 0.0119 (from left to right) (0. 0047) (0.0037) (0.0044) Post - materialism index 0.0114 0.0109 0.0161 (0.0071) (0.0056) (0.0068) Interest in politics 0.0108 0.0070 - 0.0066 (0.0113) (0.0082) (0.0098) Level of education 0.0057 - 0.0031 - 0.0148 (0.0052) (0.0040) (0.0049) Attenda nce of religious services - 0.0044 - 0.0054 0.0023 (0.0054) (0.0040) (0.0050) Importance of God 0.0051 - 0.0001 - 0.0030 (0.0034) (0.0024) (0.0032) Protestant - 0.0349 - 0.0048 - 0.0191 (0.0173) (0.0134) (0.0167) Female 0.0669 0.0199 0.0535 (0.0171) (0 .0132) (0.0166) Age younger than 30 0.0227 0.0079 0.0308 (0.0347) (0.0270) (0.0386) Age 30 - 39 0.0480 0.0013 0.0087 (0.0263) (0.0210) (0.0241) Age 50 - 59 - 0.0122 0.0206 0.0132 (0.0243) (0.0189) (0.0255) Age 60 - 69 0.0182 - 0.0170 0.0039 (0.0254) (0 .0195) (0.0237) Age 70 and older - 0.0236 0.0037 0.0475 (0.0289) (0.0221) (0.0276) Intercept 0.8602 0.9219 0.8158 (0.0345) (0.0272) (0.0332) N umber of respondents 1015 1016 1016 R 2 0.1062 0.0480 0.0564 adj. R 2 0.0946 0.0356 0.0442 OLS coefficients . Standard errors in parentheses. p < 0.05 in bold. Listwise deletion. Data weighted. The dependent variable ranges from 0 to 1 (0 = not serious at all; 0.33 = not very serious; somewhat serious = 0.67; 1.0 = very serious). Respondents were asked , Now let oblems in the world as a whole. Please, tell me how serious you consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 260 Table A4.22 : Linear regression models predicting perceived seriousness of envir onmental problems in South Korea (World Values Survey 2005 - 2009) (1) (2) (3) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or bio diversity Political ideology - 0.0088 - 0.0029 0.0004 (from left to right) (0.0039) (0.0041) (0.0036) Post - materialism index 0.0013 0.0036 - 0.0019 (0.0076) (0.0074) (0.0072) Interest in politics 0.0116 0.0022 0.0055 (0.0096) (0.0105) (0.009 8) Level of education 0.0052 0.0076 0.0204 (0.0054) (0.0062) (0.0052) Attendance of religious services - 0.0002 0.0046 - 0.0023 (0.0043) (0.0044) (0.0046) Importance of God - 0.0003 0.0002 0.0041 (0.0034) (0.0036) (0.0034) Protestant - 0.0005 - 0.0229 - 0.0168 (0.0204) (0.0198) (0.0182) Female - 0.0018 0.0035 0.0049 (0.0139) (0.0142) (0.0148) Age younger than 30 0.0020 - 0.0057 - 0.0218 (0.0199) (0.0211) (0.0215) Age 30 - 39 0.0118 0.0114 - 0.0138 (0.0200) (0.0213) (0.0200) Age 50 - 59 0.0224 0.0203 0.0101 (0.0242) (0.0241) (0.0259) Age 60 - 69 - 0.0809 - 0.0166 - 0.0442 (0.0276) (0.0282) (0.0283) Age 70 and older - 0.1015 0.0387 0.0893 (0.0694) (0.0584) (0.0572) Intercept 0.8737 0.8080 0.7512 (0.0322) (0.0347) (0.0342) N umber of respondents 11 87 1187 1187 R 2 0.0703 0.0175 0.0583 adj. R 2 0.0600 0.0066 0.0479 OLS coefficients. Standard errors in parentheses. p < 0.05 in bold. Listwise deletion. Data weighted. The dependent variable ranges from 0 to 1 (0 = not serious at all; 0.33 = not very se rious; somewhat serious = 0.67; 1.0 = very serious). Respondents were asked , oblems in the world as a whole. Please, tell me how serious you consider each of the following to be for the world as a whole. Is it very serio us, somewhat serious, not very serious or not serious at all? 261 Table A4.23 : Linear regression models predicting perceived seriousness of envir onmental problems in Norway (World Values Survey 2005 - 2009) (1) (2) (3) Global warming or the greenhouse effe ct Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political ideology - 0.0102 - 0.0059 - 0.0091 (from left to right) (0.0037) (0.0032) (0.0034) Post - materialism index 0.0317 0.0158 0.0306 (0.0071) (0.0055 ) (0.0061) Interest in politics 0.0065 0.0105 0.0098 (0.0111) (0.0081) (0.0100) Level of education 0.0153 0.0048 0.0010 (0.0045) (0.0035) (0.0041) Attendance of religious services 0.0054 - 0.0044 0.0013 (0.0053) (0.0042) (0.0049) Importance of God 0.0019 0.0054 0.0015 (0.0030) (0.0023) (0.0028) Protestant - 0.0294 - 0.0357 - 0.0506 (0.0154) (0.0119) (0.0134) Female 0.0353 0.0299 0.0238 (0.0144) (0.0117) (0.0129) Age younger than 30 - 0.0317 - 0.0034 0.0099 (0.0221) (0.0191) (0.0196) Age 30 - 3 9 - 0.0425 - 0.0040 - 0.0180 (0.0216) (0.0179) (0.0198) Age 50 - 59 - 0.0280 0.0197 - 0.0182 (0.0225) (0.0188) (0.0199) Age 60 - 69 - 0.0320 0.0122 - 0.0084 (0.0240) (0.0194) (0.0217) Age 70 and older 0.0158 0.0555 0.0249 (0.0306) (0.0236) (0.0285) Interc ept 0.8847 0.9093 0.9224 (0.0330) (0.0268) (0.0289) N umber of respondents 966 969 969 R 2 0.0820 0.0541 0.0704 adj. R 2 0.0695 0.0413 0.0578 OLS coefficients. Standard errors in parentheses. p < 0.05 in bold. Listwise deletion. Data weighted. The depen dent variable ranges from 0 to 1 (0 = not serious at all; 0.33 = not very serious; somewhat serious = 0.67; 1.0 = very serious). Respondents were asked , oblems in the world as a whole. Please, tell me how serious you con sider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 262 Table A4.24 : Linear regression models predicting perceived seriousness of envir onmental problems in Finland (World Values Survey 2005 - 2009) (1) (2) (3) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political ideology - 0.0096 - 0.0079 - 0.0098 (from left to right) (0.0041) (0.0033) (0.0043) Post - materialism index 0.0114 0.0023 0.0009 (0.0072) (0.0056) (0.0076) Interest in politics 0.0046 0.0189 0.0085 (0.0102) (0.0081) (0.0106) Level of education 0.0112 0.0048 0.0035 (0.0043) (0.0036) (0.0048) Attendance of religious s ervices - 0.0002 - 0.0041 0.0073 (0.0057) (0.0048) (0.0063) Importance of God 0.0010 0.0044 0.0001 (0.0033) (0.0027) (0.0034) Protestant -- -- -- Female 0.0559 0.0563 0.0666 (0.0161) (0.0137) (0.0174) Age younger than 30 - 0.0199 - 0.0463 - 0.003 6 (0.0254) (0.0218) (0.0259) Age 30 - 39 0.0057 - 0.0212 - 0.0322 (0.0232) (0.0190) (0.0263) Age 50 - 59 - 0.0063 - 0.0403 - 0.0421 (0.0246) (0.0204) (0.0267) Age 60 - 69 - 0.0018 - 0.0430 - 0.0411 (0.0282) (0.0238) (0.0296) Age 70 and older 0.0152 - 0.0422 - 0.0156 (0.0285) (0.0258) (0.0334) Intercept 0.8236 0.8993 0.7970 (0.0306) (0.0246) (0.0325) N umber of respondents 847 853 847 R 2 0.0413 0.0544 0.0393 adj. R 2 0.0275 0.0409 0.0255 OLS coefficients. Standard errors in parentheses. p < 0.05 in bold. Listwise deletion. Data weighted. The dependent variable ranges from 0 to 1 (0 = not serious at all; 0.33 = not very serious; somewhat serious = 0.67; 1.0 = very serious). Respondents were asked , oblems in the world as a whole. Please, tell me how serious you consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 263 Table A4.25 : Linear regression models predicting perceived seriousness o f envir onmental problems in Germany (World Values Survey 2005 - 2009) (1) (2) (3) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political ideology - 0.0099 - 0.0123 - 0.01 51 (from left to right) (0.0037) (0.0038) (0.0037) Post - materialism index 0.0105 0.0057 0.0153 (0.0060) (0.0061) (0.0061) Interest in politics 0.0103 0.0339 0.0142 (0.0079) (0.0097) (0.0080) Level of education 0.0095 0.0058 0.0098 (0.0035) ( 0.0040) (0.0035) Attendance of religious services - 0.0001 - 0.0169 - 0.0113 (0.0051) (0.0059) (0.0052) Importance of God 0.0024 0.0072 0.0039 (0.0028) (0.0031) (0.0028) Protestant -- -- -- Female 0.0347 0.0602 0.0524 (0.0125) (0.0139) (0.0137) Age younger than 30 - 0.0181 0.0070 - 0.0051 (0.0215) (0.0236) (0.0222) Age 30 - 39 0.0023 - 0.0126 - 0.0197 (0.0196) (0.0218) (0.0226) Age 50 - 59 0.0096 - 0.0230 - 0.0176 (0.0210) (0.0254) (0.0222) Age 60 - 69 0.0092 - 0.0034 - 0.0103 (0.0200) (0.0211) (0 .0210) Age 70 and older 0.0063 - 0.0404 - 0.0162 (0.0217) (0.0216) (0.0229) Intercept 0.8249 0.7774 0.7662 (0.0276) (0.0285) (0.0259) N umber of respondents 1648 1664 1656 R 2 0.0361 0.0582 0.0524 adj. R 2 0.0290 0.0513 0.0455 OLS coefficients. Standa rd errors in parentheses. p < 0.05 in bold. Listwise deletion. Data weighted. The dependent variable ranges from 0 to 1 (0 = not serious at all; 0.33 = not very serious; somewhat serious = 0.67; 1.0 = very serious). Respondents were asked , der environmental pr oblems in the world as a whole. Please, tell me how serious you consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 264 Table A4.26 : Linear regressi on models predicting perceived seriousness of envir onmental problems in Canada (World Values Survey 2005 - 2009) (1) (2) (3) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political ideology - 0.0128 - 0.0074 - 0.0136 (from left to right) (0.0046) (0.0031) (0.0038) Post - materialism index 0.0057 0.0061 0.0075 (0.0061) (0.0039) (0.0062) Interest in politics 0.0076 - 0.0028 0.0001 (0.0079) (0.0053) (0.0071) Level of education - 0.0051 - 0.0034 0.0017 (0.0045) (0.0044) (0.0053) Attendance of religious services - 0.0047 0.0000 - 0.0047 (0.0039) (0.0044) (0.0050) Importance of God 0.0037 0.0029 0.0041 (0.0028) (0.0029) (0.0033) Protestant - 0.0111 0.0062 - 0.0091 ( 0.0194) (0.0140) (0.0202) Female 0.0333 0.0129 0.0277 (0.0138) (0.0121) (0.0150) Age younger than 30 - 0.0171 - 0.0060 0.0314 (0.0203) (0.0158) (0.0192) Age 30 - 39 - 0.0145 - 0.0030 0.0032 (0.0221) (0.0157) (0.0224) Age 50 - 59 - 0.0061 - 0.0048 0.0146 (0.0189) (0.0144) (0.0206) Age 60 - 69 - 0.0267 - 0.0381 - 0.0368 (0.0236) (0.0292) (0.0322) Age 70 and older - 0.0351 0.0074 - 0.0341 (0.0272) (0.0178) (0.0289) Intercept 0.9348 0.9654 0.9041 (0.0318) (0.0248) (0.0289) N umber of respondents 1526 1554 1 538 R 2 0.0296 0.0223 0.0413 adj. R 2 0.0212 0.0141 0.0332 OLS coefficients. Standard errors in parentheses. p < 0.05 in bold. Listwise deletion. Data weighted. The dependent variable ranges from 0 to 1 (0 = not serious at all; 0.33 = not very serious; so mewhat serious = 0.67; 1.0 = very serious). Respondents were asked , oblems in the world as a whole. Please, tell me how serious you consider each of the following to be for the world as a whole. Is it very serious, somew hat serious, not very serious or not serious at all? 26 5 Table A4.27 : Linear regression models predicting perceived seriousness of envir onmental problems in Australia (World Values Survey 2005 - 2009) (1) (2) (3) Global warming or the greenhouse effect Pol lution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political ideology - 0.0125 - 0.0022 - 0.0035 (from left to right) (0.0040) (0.0023) (0.0036) Post - materialism index 0.0070 0.0056 0.0197 (0.0055) (0.0037) (0.0 049) Interest in politics 0.0071 0.0080 0.0068 (0.0080) (0.0052) (0.0077) Level of education - 0.0008 0.0008 - 0.0017 (0.0041) (0.0027) (0.0040) Attendance of religious services - 0.0049 - 0.0037 - 0.0071 (0.0045) (0.0029) (0.0040) Importance of God 0 .0017 - 0.0002 - 0.0013 (0.0026) (0.0018) (0.0023) Protestant 0.0028 - 0.0011 - 0.0026 (0.0149) (0.0096) (0.0146) Female 0.0439 0.0222 0.0319 (0.0131) (0.0090) (0.0128) Age younger than 30 - 0.0025 - 0.0134 0.0147 (0.0210) (0.0156) (0.0212) Age 30 - 39 0.0048 - 0.0239 0.0266 (0.0220) (0.0166) (0.0206) Age 50 - 59 0.0144 0.0004 0.0170 (0.0194) (0.0124) (0.0183) Age 60 - 69 - 0.0210 - 0.0139 - 0.0023 (0.0217) (0.0135) (0.0206) Age 70 and older - 0.0439 - 0.0271 - 0.0416 (0.0256) (0.0162) (0.0244) Interce pt 0.8977 0.9334 0.8444 (0.0288) (0.0183) (0.0272) N umber of respondents 1245 1246 1245 R 2 0.0402 0.0208 0.0448 adj. R 2 0.0300 0.0104 0.0347 OLS coefficients. Standard errors in parentheses. p < 0.05 in bold. Listwise deletion. Data weighted. The dep endent variable ranges from 0 to 1 (0 = not serious at all; 0.33 = not very serious; somewhat serious = 0.67; 1.0 = very serious). Respondents were asked , oblems in the world as a whole. Please, tell me how serious you c onsider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 266 Table A4.28 : Linear regression models predicting perceived seriousness of envir onmental problems in Japan (World Values Survey 2005 - 2009) (1) (2) (3) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political ideology - 0.0072 - 0.0024 - 0.0103 (from left to right) (0.0037) (0.0042) (0.0049) Post - materialism index 0.0045 0.0092 0.0187 (0.0057) (0.0064) (0.0072) Interest in politics 0.0153 0.0147 0.0219 (0.0098) (0.0107) (0.0135) Level of education 0.0033 - 0.0009 0.0042 (0.0050) (0.0051) (0.0065) Attendance of religious services 0.0015 0.0080 0.0010 (0.0058) (0.0062) (0.0072) Importance of God 0.0040 0.0049 0.0048 (0.0027) (0.0030) (0.0035) Protestant - 0.0243 - 0.0508 0.0681 (0.0668) (0.0691) (0.0660) Female 0.0421 0.0420 0.0397 (0.0130) (0.0145) (0.0174) Age y ounger than 30 - 0.0082 0.0183 0.0200 (0.0211) (0.0236) (0.0275) Age 30 - 39 0.0168 0.0108 0.0130 (0.0220) (0.0250) (0.0257) Age 50 - 59 - 0.0224 - 0.0000 - 0.0159 (0.0192) (0.0223) (0.0232) Age 60 - 69 - 0.0200 - 0.0127 - 0.0376 (0.0215) (0.0235) (0.0255) Age 70 and older - 0.0297 - 0.0371 0.0047 (0.0256) (0.0292) (0.0327) Intercept 0.9217 0.8535 0.8351 (0.0293) (0.0318) (0.0375) N umber of respondents 673 671 667 R 2 0.0345 0.0326 0.0446 adj. R 2 0.0154 0.0134 0.0256 OLS coefficients. Standard errors i n parentheses. p < 0.05 in bold. Listwise deletion. Data weighted. The dependent variable ranges from 0 to 1 (0 = not serious at all; 0.33 = not very serious; somewhat serious = 0.67; 1.0 = very serious). Respondents were asked , mental pr oblems in the world as a whole. Please, tell me how serious you consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 267 Table A4.29 : Linear regression models p redicting perceived seriousness of envir onmental problems in Mexico (World Values Survey 2005 - 2009) (1) (2) (3) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Politica l ideology - 0.0019 - 0.0009 - 0.0009 (from left to right) (0.0024) (0.0019) (0.0022) Post - materialism index - 0.0136 - 0.0019 - 0.0077 (0.0061) (0.0046) (0.0053) Interest in politics 0.0110 0.0119 * 0.0057 (0.0066) (0.0048) (0.0058) Level of educati on 0.0238 0.0168 * 0.0201 * (0.0034) (0.0028) (0.0033) Attendance of religious services - 0.0113 - 0.0037 - 0.0053 (0.0036) (0.0027) (0.0033) Importance of God 0.0066 0.0078 0.0092 * (0.0050) (0.0040) (0.0046) Protestant 0.0964 0.0711 * 0.0787 * (0.0350 ) (0.0139) (0.0260) Female - 0.0221 - 0.0088 - 0.0124 (0.0137) (0.0106) (0.0121) Age younger than 30 - 0.0478 - 0.0037 - 0.0087 (0.0204) (0.0159) (0.0182) Age 30 - 39 0.0224 - 0.0065 0.0098 (0.0204) (0.0162) (0.0190) Age 50 - 59 0.0315 0.0122 0.0183 (0.02 52) (0.0210) (0.0239) Age 60 - 69 0.0491 0.0190 0.0473 (0.0294) (0.0231) (0.0253) Age 70 and older 0.0505 0.0743 0.0205 (0.0370) (0.0225) (0.0389) Intercept 0.8512 0.9046 0.8624 (0.0317) (0.0268) (0.0296) N umber of respondents 1220 1240 1240 R 2 0. 0789 0.0584 0.0525 adj. R 2 0.0689 0.0484 0.0424 OLS coefficients. Standard errors in parentheses. p < 0.05 in bold. Listwise deletion. Data weighted. The dependent variable ranges from 0 to 1 (0 = not serious at all; 0.33 = not very serious; somewhat ser ious = 0.67; 1.0 = very serious). Respondents were asked , oblems in the world as a whole. Please, tell me how serious you consider each of the following to be for the world as a whole. Is it very serious, somewhat seriou s, not very serious or not serious at all? 268 Table A4.30 : Linear regression models predicting perceived seriousness of envir onmental problems in Slovenia (World Values Survey 2005 - 2009) (1) (2) (3) Global warming or the greenhouse effect Pollution of r ivers, lakes, and oceans Loss of plant or animal species or biodiversity Political ideology - 0.0062 - 0.0049 - 0.0064 (from left to right) (0.0044) (0.0040) (0.0045) Post - materialism index 0.0079 - 0.0062 - 0.0064 (0.0090) (0.0079) (0.0094) Int erest in politics 0.0279 0.0137 0.0163 (0.0114) (0.0105) (0.0117) Level of education - 0.0062 - 0.0023 - 0.0067 (0.0053) (0.0046) (0.0054) Attendance of religious services - 0.0030 - 0.0050 - 0.0010 (0.0060) (0.0050) (0.0057) Importance of God 0.0032 0. 0020 0.0004 (0.0035) (0.0029) (0.0035) Protestant - 0.0905 - 0.0630 - 0.0048 (0.0949) (0.0784) (0.0803) Female - 0.0044 0.0140 0.0273 (0.0179) (0.0153) (0.0178) Age younger than 30 - 0.0047 0.0108 0.0157 (0.0274) (0.0238) (0.0268) Age 30 - 39 - 0.0415 0.0053 - 0.0142 (0.0301) (0.0237) (0.0295) Age 50 - 59 0.0007 0.0051 0.0017 (0.0277) (0.0245) (0.0284) Age 60 - 69 - 0.0019 0.0145 - 0.0021 (0.0308) (0.0242) (0.0337) Age 70 and older - 0.0118 - 0.0329 - 0.0376 (0.0336) (0.0368) (0.0390) Intercept 0.9019 0.8992 0.8613 (0.0338) (0.0293) (0.0343) N umber of respondents 571 589 585 R 2 0.0296 0.0165 0.0156 adj. R 2 0.0070 - 0.0058 - 0.0068 OLS coefficients. Standard errors in parentheses. p < 0.05 in bold. Listwise deletion. Data weighted. The dependent var iable ranges from 0 to 1 (0 = not serious at all; 0.33 = not very serious; somewhat serious = 0.67; 1.0 = very serious). Respondents were asked , oblems in the world as a whole. Please, tell me how serious you consider ea ch of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 269 Table A4.31 : Linear regression models predicting perceived seriousness of envir onmental problems in Turkey (World Values Su rvey 2005 - 2009) (1) (2) (3) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political ideology - 0.0076 - 0.0023 - 0.0046 (from left to right) (0.0022) (0.0021) (0.0027) Post - materialism index - 0.0097 - 0.0133 - 0.0137 (0.0040) (0.0038) (0.0051) Interest in politics - 0.0010 0.0048 0.0089 (0.0050) (0.0048) (0.0067) Level of education 0.0052 0.0079 0.0078 (0.0025) (0.0024) (0.0030) Attendance of religious servic es 0.0020 0.0032 0.0041 (0.0027) (0.0028) (0.0037) Importance of God 0.0037 - 0.0002 0.0018 (0.0028) (0.0026) (0.0036) Protestant -- -- -- Female 0.0118 0.0012 - 0.0111 (0.0118) (0.0104) (0.0149) Age younger than 30 - 0.0311 - 0.0167 - 0.0263 ( 0.0118) (0.0116) (0.0138) Age 30 - 39 - 0.0183 - 0.0064 - 0.0210 (0.0143) (0.0147) (0.0178) Age 50 - 59 - 0.0268 0.0040 - 0.0324 (0.0183) (0.0171) (0.0234) Age 60 - 69 0.0034 - 0.0675 - 0.0667 (0.0254) (0.0364) (0.0440) Age 70 and older 0.0037 0.0038 - 0.0224 (0.0258) (0.0305) (0.0362) Intercept 0.9872 0.9644 0.9634 (0.0177) (0.0191) (0.0233) N umber of respondents 1028 1044 1038 R 2 0.0324 0.0425 0.0376 adj. R 2 0.0210 0.0314 0.0263 OLS coefficients. Standard errors in parentheses. p < 0.05 in bold. List wise deletion. Data weighted. The dependent variable ranges from 0 to 1 (0 = not serious at all; 0.33 = not very serious; somewhat serious = 0.67; 1.0 = very serious). Respondents were asked , oblems in the world as a who le. Please, tell me how serious you consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 270 Table A4.32 : Linear regression models predicting perceived seriousness of en vir onmental problems in Chile (World Values Survey 2005 - 2009) (1) (2) (3) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political ideology 0.0002 - 0.0004 - 0.0033 (from le ft to right) (0.0043) (0.0033) (0.0040) Post - materialism index 0.0090 0.0039 0.0050 (0.0068) (0.0054) (0.0056) Interest in politics 0.0133 0.0026 0.0134 (0.0107) (0.0080) (0.0080) Level of education 0.0129 0.0101 0.0051 (0.0055) (0.0044) (0.0 039) Attendance of religious services - 0.0006 - 0.0048 - 0.0072 (0.0058) (0.0043) (0.0042) Importance of God 0.0095 0.0145 0.0147 (0.0049) (0.0040) (0.0045) Protestant - 0.0063 - 0.0290 - 0.0315 (0.0300) (0.0257) (0.0280) Female 0.0113 0.0164 0.0364 * (0.0192) (0.0173) (0.0178) Age younger than 30 - 0.0462 - 0.0304 - 0.0105 (0.0274) (0.0247) (0.0208) Age 30 - 39 - 0.0206 - 0.0056 0.0048 (0.0256) (0.0182) (0.0202) Age 50 - 59 - 0.0566 - 0.0124 - 0.0507 (0.0296) (0.0264) (0.0287) Age 60 - 69 0.0144 0.0113 0 .0192 (0.0352) (0.0259) (0.0352) Age 70 and older - 0.0036 - 0.0142 - 0.0135 (0.0358) (0.0392) (0.0404) Intercept 0.8853 0.8866 0.8808 (0.0325) (0.0295) (0.0296) N umber of respondents 625 635 632 R 2 0.0509 0.0621 0.0725 adj. R 2 0.0307 0.0425 0.0530 OLS coefficients. Standard errors in parentheses. p < 0.05 in bold. Listwise deletion. Data weighted. The dependent variable ranges from 0 to 1 (0 = not serious at all; 0.33 = not very serious; somewhat serious = 0.67; 1.0 = very serious). Respondents we re asked , oblems in the world as a whole. Please, tell me how serious you consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 271 Ta ble A4.33 : Linear regression models predicting perceived seriousness of envir onmental problems in Poland (World Values Survey 2005 - 2009) (1) (2) (3) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal s pecies or biodiversity Political ideology - 0.0010 - 0.0015 - 0.0022 (from left to right) (0.0040) (0.0034) (0.0041) Post - materialism index 0.0010 0.0071 0.0031 (0.0096) (0.0079) (0.0092) Interest in politics 0.0111 0.0255 0.0135 (0.0113) (0.0095 ) (0.0112) Level of education - 0.0052 - 0.0034 - 0.0049 (0.0054) (0.0046) (0.0052) Attendance of religious services - 0.0041 - 0.0064 - 0.0037 (0.0072) (0.0053) (0.0072) Importance of God 0.0133 0.0092 0.0163 (0.0060) (0.0041) (0.0057) Protestant 0.07 76 0.1126 0.0342 (0.0732) (0.0199) (0.0840) Female - 0.0285 - 0.0016 0.0089 (0.0171) (0.0148) (0.0168) Age younger than 30 0.0486 0.0227 0.0276 (0.0284) (0.0238) (0.0281) Age 30 - 39 0.0636 0.0565 0.0491 (0.0300) (0.0245) (0.0302) Age 50 - 59 0.0357 0.0332 0.0509 * (0.0264) (0.0218) (0.0255) Age 60 - 69 0.0581 0.0345 0.0499 (0.0315) (0.0268) (0.0296) Age 70 and older 0.0229 0.0174 0.0264 (0.0353) (0.0295) (0.0336) Intercept 0.7840 0.8678 0.7830 (0.0341) (0.0277) (0.0345) N umber of respondents 671 675 674 R 2 0.0329 0.0369 0.0400 adj. R 2 0.0138 0.0179 0.0211 OLS coefficients. Standard errors in parentheses. p < 0.05 in bold. Listwise deletion. Data weighted. The dependent variable ranges from 0 to 1 (0 = not serious at all; 0.33 = not very se rious; somewhat serious = 0.67; 1.0 = very serious). Respondents were asked , oblems in the world as a whole. Please, tell me how serious you consider each of the following to be for the world as a whole. Is it very serio us, somewhat serious, not very serious or not serious at all? 272 Table A6.34 : Linear regression models predicting perceived seriousness of envir onmental problems in Sweden (World Values Survey 2005 - 2009) (1) (2) (3) Global warming or the greenhouse effe ct Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political ideology - 0.0093 - 0.0045 - 0.0130 (from left to right) (0.0034) (0.0028) (0.0035) Post - materialism index 0.0068 0.0068 - 0.0007 (0.0064) (0.004 8) (0.0068) Interest in politics 0.0056 0.0121 0.0097 (0.0089) (0.0074) (0.0097) Level of education 0.0052 0.0015 - 0.0070 (0.0044) (0.0034) (0.0048) Attendance of religious services - 0.0007 - 0.0050 - 0.0110 (0.0058) (0.0047) (0.0065) Importance of God 0.0013 - 0.0008 0.0020 (0.0027) (0.0021) (0.0028) Protestant - 0.2259 - 0.1451 - 0.0956 (0.1586) (0.1910) (0.1872) Female 0.0491 0.0561 0.0629 (0.0148) (0.0108) (0.0145) Age younger than 30 - 0.0297 - 0.0283 0.0097 (0.0239) (0.0177) (0.0221) Age 30 - 39 0.0055 - 0.0255 - 0.0183 (0.0226) (0.0190) (0.0240) Age 50 - 59 0.0017 0.0132 0.0085 (0.0214) (0.0152) (0.0211) Age 60 - 69 - 0.0341 - 0.0317 - 0.0312 (0.0253) (0.0201) (0.0262) Age 70 and older 0.0098 0.0096 - 0.0281 (0.0280) (0.0205) (0.0310) In tercept 0.8852 0.8963 0.8879 (0.0295) (0.0227) (0.0295) N umber of respondents 915 925 923 R 2 0.0462 0.0596 0.0570 adj. R 2 0.0324 0.0462 0.0435 OLS coefficients. Standard errors in parentheses. p < 0.05 in bold. Listwise deletion. Data weighted. The d ependent variable ranges from 0 to 1 (0 = not serious at all; 0.33 = not very serious; somewhat serious = 0.67; 1.0 = very serious). Respondents were asked , oblems in the world as a whole. Please, tell me how serious you consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 273 Table A 4 .35: Regression models predicting climate change attitudes in United States (interaction between politi cal i deology and political knowledge (2008 and 2012 American National Elections Studies) Concern about climate change Belief in climate change scale Cause of climate change human activity Cause of climate change human activity Linear (2008 ANES) Linear (2008 ANES) Binary logit (2008 ANES) Binary logit (2012 ANES) Political ideology - 0.0285 - 0.1133 0.1011 0.0055 (liberal to conservative) (0.0180) (0.1811) (0.1242) (0.0893) Political knowledge 0.0374 0.5075 0.6032 0.8256 (0.0165) (0.1694) (0.1384) (0.1384) Political ideology X - 0.0101 - 0.1662 - 0.1250 - 0.1847 Political knowledge (0.0039) (0.0418) (0.0307) (0.0337) Biblical literalism 0.0451 0.3937 0.2845 0.1409 (0.0188) (0.2104) (0.1322) (0.0413) Level of educatio n 0.0152 0.0664 - 0.1004 0.1688 (0.0094) (0.1118) (0.0789) (0.0626) Age younger than 30 0.0086 0.2539 - 0.1865 0.1458 (0.0420) (0.4974) (0.2870) (0.1468) Age 30 - 39 0.0438 - 0.3858 0.0336 0.0890 (0.0350) (0.3727) (0.2669) (0.1478) Age 50 - 59 - 0. 0119 - 0.0648 - 0.2262 - 0.0176 (0.0328) (0.3503) (0.2252) (0.1355) Age 60 - 69 0.0011 0.0028 - 0.5142 - 0.4321 (0.0331) (0.3718) (0.2641) (0.1422) Age 70 and older - 0.0044 - 0.4486 - 0.2085 - 0.3588 (0.0370) (0.3941) (0.2914) (0.1641) Female 0.0288 0 .4447 - 0.0255 - 0.0140 (0.0239) (0.2433) (0.1750) (0.0879) Constant 0.5817 6.3330 - 1.2585 - 1.5936 (0.0944) (0.8406) (0.6651) (0.4103) N umber of respondents 1065 759 1386 4919 AIC 144 3488 1671 5682 OLS or logit coefficient estimate. Standard e rrors in parentheses. Two tailed p - value<0.05 in bold . Data are weighted. Listwise deletion. Political knowledge ranges from 0 to 6 for the 2008 American National Elections Study and 0 to 5 for the 2012 American National Elections Study. 274 Table A4.36: Reg ression models predicting climate change attitudes in United States for liberals and conservatives (2008 and 2012 American National Elections Stu dies) Concern about climate change Linear 2008 ANES Belief in climate change scale Linear 2008 ANES Cause of climate change human activity Binary logit 2008 ANES Liberals Conservatives Liberals Conservatives Liberals Conservatives Political knowledge 0.0169 - 0.0242 0.2224 - 0.5228 0.2536 - 0.0666 (0.0105) (0.0124) (0.1093) (0.1572) (0.0957) (0.0886) Biblical literalism 0.0482 0.0538 0.6003 0.4021 0.5647 0.1613 (0.0291) (0.0253) (0.2817) (0.3116) (0.2154) (0.1761) Level of education 0.0149 0.0082 0.0838 0.0083 - 0.0595 - 0.1783 (0.0165) (0.0127) (0.1506) (0.1625) (0.1243) (0.1084) Age younger tha n 30 0.0342 - 0.0637 0.5883 0.0092 0.0990 - 0.6298 (0.0494) (0.0702) (0.5716) (0.9534) (0.4310) (0.4442) Age 30 - 39 0.1184 0.0223 - 0.3428 - 0.0527 - 0.4055 0.3245 (0.0496) (0.0478) (0.5670) (0.5164) (0.3994) (0.3367) Age 50 - 59 0.0515 - 0.0299 0.6443 - 0.174 5 - 0.1497 - 0.1230 (0.0569) (0.0431) (0.3831) (0.5260) (0.3693) (0.2978) Age 60 - 69 0.0404 - 0.0181 0.1558 - 0.0015 - 0.3332 - 0.7260 (0.0507) (0.0485) (0.4607) (0.5874) (0.4389) (0.3516) Age 70 and older 0.0911 - 0.0411 0.0076 - 0.7420 0.4617 - 0.5845 (0.0 415) (0.0519) (0.4910) (0.5721) (0.4833) (0.4547) Female 0.0093 0.0940 0.5762 0.7896 - 0.3825 0.3260 (0.0362) (0.0352) (0.3168) (0.4018) (0.2893) (0.2462) Constant 0.4723 0.4274 4.9603 5.9284 - 1.4577 - 0.4869 (0.0950) (0.0841) (0.6784) (0.9259) (0.6610 ) (0.5417) N umber of respondents 358 578 287 386 475 756 AIC - 43 175 1154 1892 582 918 275 Cause of climate change human activity Binary logit 2012 ANES Liberals Conservatives Political knowledge 0.3033 - 0.1377 (0.0882) (0.0924) Biblical literalism 0.2803 - 0.0172 (0.0674) (0.0785) Level of education 0.2405 0.0781 (0.1085) (0.1208) Age younger than 30 0.4239 0.0668 (0.2521) (0.2880) Age 30 - 39 0.2086 0.2935 (0.2470) (0.2697) Age 50 - 59 - 0.0803 0.1904 (0.2325) (0.2343) Age 60 - 69 - 0.4787 - 0.3958 (0.2417) (0.2535) Age 70 and older - 0.1071 - 0.5535 (0.2968) (0.3000) Female 0.0396 0.3071 (0.1531) (0.1564) Constant - 2.0881 - 1.3002 (0.3620) (0.3993) N umber of respondents 1417 1794 AIC 1754 1879 OLS or lo git coefficient estimate. Standard errors in parentheses. Two tailed p - value<0.1 in bold . Listwise deletion and 0 to 5 for the 2012 American National Elections Study. 276 Table A 4 .37: Regression models predicting climate change attitudes in United States for liberals and conservatives interest in politics instead of political knowledge (2008 Ame rican National Elections Study) Concern about climate change Belief in clim ate change scale Cause of climate change human activity Linear (2008 ANES) Linear (2008 ANES) Binary logit (2008 ANES) Political ideology - 0.0139 - 0.0937 0.0682 (liberal to conservative) (0.0249) (0.2625) (0.1674) Interest in politics 0 .0896 0.7090 0.6093 (0.0291) (0.3037) (0.2145) Political ideology X - 0.0145 - 0.1809 - 0.1122 Interest in politics (0.0063) (0.0655) (0.0441) Biblical literalism 0.0412 0.3576 0.3059 (0.0184) (0.2087) (0.1309) Level of education 0.0088 - 0.0187 - 0.0864 (0.0099) (0.1086) (0.0764) Age younger than 30 0.0068 0.0663 - 0.2497 (0.0412) (0.5067) (0.2929) Age 30 - 39 0.0401 - 0.5685 - 0.0319 (0.0354) (0.3667) (0.2765) Age 50 - 59 - 0.0294 - 0.1664 - 0.2979 (0.0339) (0.3820) (0.2300) Age 60 - 69 - 0.0213 - 0.1425 - 0.5655 (0.0339) (0.4090) (0.2616) Age 70 and older - 0.0409 - 0.7851 - 0.3267 (0.0373) (0.4365) (0.2860) Female 0.0320 0.5057 - 0.0806 (0.0234) (0.2473) (0.1695) Constant 0.4342 6.1286 - 1.3419 (0.1297) (1.2857) (0.8908) Number of responden ts 1064 758 1385 AIC 137 3510 1690 Standard errors in parentheses. Two tailed p - value<0.05 in bold . Data are weighted. 277 Table A 4 .38: Binary logistic model predicting cause of climate change in United States for liberals and conservatives attention to politics instead of political knowledge (2012 Ame rican National Elections Study) Cause of climate change human activity Binary logit (2012 ANES) Political ideology (from left to right) - 0.1580 (0.1043) Attention to politics 0.3253 (0.1181) Political ideology X Attention to politics - 0.0823 (0.0284) Level of education 0.1691 (0.0401) Biblical literalism 0.1939 (0.0619) Age younger than 30 0.1480 (0.1472) Age 30 - 39 0.1070 (0.1465) Age 50 - 59 0.0038 (0.1353) Age 60 - 69 - 0.3703 (0.1395) Age 70 and older - 0.2886 (0.1632) Female - 0.0368 (0.0881) Constant - 0.8629 (0.4749) N umber of respondents 4919 AIC 5730 Standard errors in parentheses. Two tailed p - value<0.05 in bold . Data are weighted. 278 Table A 4 .39: Regression mod els predicting climate change attitudes in United States for liberals and conservatives interest in politics instead of political knowledge (2008 Ameri can National Elections Studies) Concern about climate change Linear 2008 ANES Belief in climate chan ge scale Linear 2008 ANES Cause of climate change Binary logit 2008 ANES Liberal Conservative Liberal Conservative Liberal Conservative Interest in politics 0.0632 - 0.0101 0.3914 - 0.3872 0.2654 - 0.0639 (0.0200) (0.0160) (0.2035) (0.2161) ( 0.1625) (0.1107) Biblical literalism 0.0550 0.0496 0.6923 0.3498 0.6484 0.1425 (0.0288) (0.0249) (0.2876) (0.3008) (0.2109) (0.1756) Level of education 0.0067 0.0004 0.0793 - 0.1608 - 0.0150 - 0.1942 (0.0167) (0.0135) (0.1322) (0.1678) (0.1221) (0.1057) Age younger than 30 0.0515 - 0.0864 0.6699 - 0.5295 0.0815 - 0.7073 (0.0522) (0.0642) (0.6009) (0.9334) (0.4360) (0.4456) Age 30 - 39 0.1065 0.0175 - 0.5214 - 0.3345 - 0.5143 0.2941 (0.0482) (0.0490) (0.5232) (0.5267) (0.4149) (0.3373) Age 50 - 59 0.0321 - 0. 0388 0.4550 - 0.1950 - 0.2039 - 0.1433 (0.0530) (0.0456) (0.3884) (0.5727) (0.3621) (0.3058) Age 60 - 69 0.0092 - 0.0336 0.0281 - 0.1356 - 0.3661 - 0.7756 (0.0504) (0.0489) (0.4846) (0.6396) (0.4284) (0.3593) Age 70 and older 0.0433 - 0.0603 - 0.2909 - 1.1754 0. 3523 - 0.6381 (0.0428) (0.0516) (0.5545) (0.6378) (0.4692) (0.4512) Female - 0.0041 0.1063 0.3526 1.1119 - 0.5658 0.3440 (0.0339) (0.0353) (0.3024) (0.4128) (0.2801) (0.2405) Constant 0.3354 0.4080 4.3931 5.9543 - 1.6643 - 0.4046 (0.1064) (0.0829) (0.88 97) (1.0382) (0.7647) (0.5937) N umber of respondents 358 577 287 385 475 755 AIC - 63 182 1150 1906 588 918 Standard errors in parentheses. Two tailed p - value<0.1 in bold . Data are weighted. 279 Table A 4 .40: Binary logistic models predictin g cause of climate change in United States for liberals and conservatives attention to politics instead of political knowledge (2012 Ameri can National Elections Studies) Cause of climate change human activity Binary logit (2012 ANES) Liberals con servatives Attention to politics 0.0970 - 0.1971 (0.0702) (0.0748) Level of education 0.3361 - 0.0158 (0.0653) (0.0765) Biblical literalism 0.2899 0.0608 (0.1075) (0.1207) Age younger than 30 0.3808 0.0815 (0.2485) (0.2867) Age 30 - 39 0.2299 0.3081 (0.2474) (0.2706) Age 50 - 59 - 0.0461 0.2227 (0.2310) (0.2366) Age 60 - 69 - 0.4071 - 0.3305 (0.2364) (0.2564) Age 70 and older - 0.0103 - 0.4693 (0.2972) (0.3049) Female - 0.0082 0.2642 (0.1513) (0.1606) Constant - 1.9361 - 0.9502 (0.4043) (0 .4456) N umber of respondents 1417 1794 AIC 1772 1871 Standard errors in parentheses. Two tailed p - value<0.05 in bold . Data are weighted. 280 Table A4.41 : Ordered logistic regression models predicting perceived seriousness of environmental problems in the United States among those on the left (World Values Survey 2005 - 2009) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political interest 0.6824 0.3766 0.5879 (low; medium; high) (0.2365) (0.2925) (0.2068) Level of education 0.0408 0.2682 0.2275 (0.1821) (0.2018) (0.1719) Lower class - 0.8200 - 0.4764 - 0.6289 (1.4350) (1.4862) (1.2696) Working class 0.1747 - 0.3950 - 0.0592 (0.5252) (0.5795) (0.5288) Upper class - 0.4602 - 1.1160 * - 0.6918 (0.4283) (0.4678) (0.4146) Female 0.4630 1.1068 * 0.5588 (0.3917) (0.4448) (0.3763) Age younger than 30 - 0.3977 - 0.2264 - 1.1775 (0.7623) (0.7118) (0.7818) Age 30 - 39 - 0.0527 1.0746 - 0.1564 (0.5854) (0.6204) (0.5102) Age 50 - 59 0.5823 0.7634 - 0.1133 (0.5532) (0.5081) (0.4787) Age 60 - 69 - 0.3569 - 0.1589 - 1.2926 (0.6419) (0.8144) (0.7099) Age 70 and older 2.5007 2.4563 0.4559 (1.1475) (1.0874) (0.6653) Not serious at all | Not very serious - 3.8007 - 4.3308 (0. 7272) (0.8213) Not very serious | Somewhat serious - 1.9782 - 2.3730 - 2.3052 (0.7098) (0.6403) (0.5960) Somewhat serious | Very serious - 0.1914 - 0.1249 - 0.0949 (0.6896) (0.6831) (0.6230) Number of respondents 205 206 206 AIC 387 273 391 Standard er rors in parentheses. p < 0.10 in bold. Listwise deletion and weighted. Left was defined as respondents who chose 1, 2, 3, or 4 to describe their political ideology on the left - right continuum from 1 to 10. Respondents were asked , nmental problems in the world as a whole. Please, tell me how serious you consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 281 Table A4.42 : Ordered logistic regress ion models predicting perceived seriousness of environmental problems in the United States among those on the right (World Values Survey 2005 - 2009) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political interest - 0.3708 0.0782 - 0.0901 (low; medium; high) (0.2238) (0.2227) (0.2164) Level of education - 0.1071 - 0.0963 - 0.1393 (0.1123) (0.1304) (0.1154) Lower class 2.3399 1.7934 15.8900 (1.0146) (1.1637) (0 .5748) Working class - 0.0386 0.4168 0.1164 (0.3544) (0.3859) (0.3833) Upper class 0.1763 0.1328 0.4464 (0.3017) (0.3158) (0.2902) Female 0.2840 0.1856 0.4969 (0.2793) (0.2853) (0.2992) Age younger than 30 - 0.5962 - 0.3183 - 0.6645 (0.4234) (0.470 0) (0.4578) Age 30 - 39 - 0.6028 - 0.0914 - 0.7480 (0.3246) (0.4168) (0.4089) Age 50 - 59 - 0.7510 - 0.5763 - 1.2833 (0.5565) (0.4516) (0.4963) Age 60 - 69 - 0.1599 0.0342 - 0.3889 (0.4084) (0.5106) (0.4960) Age 70 and older - 0.3168 - 0.1933 - 0.8308 (0.4513) (0.4252) (0.5096) Not serious at all | Not very serious - 1.9191 - 3.3955 - 2.8393 (0.4647) (0.5756) (0.5183) Not very serious | Somewhat serious - 0.6830 - 2.1362 - 1.3871 (0.4374) (0.4739) (0.4715) Somewhat serious | Very serious 0.8312 - 0.3141 0.4828 (0.4360) (0.4350) (0.4730) Number of respondents 332 330 331 AIC 888 675 807 Standard errors in parentheses. p < 0.10 in bold. Listwise deletion and weighted. Right was defined as respondents who chose 7, 8, 9, or 10 to describe their political ideolog y on the left - right continuum from 1 to 10. Respondents were asked , consider each of the following to be for the world as a whole. Is it very serious, somew hat serious, not very serious or not serious at all? 282 Table A4.43 : Ordered logistic regression models predicting perceived seriousness of environmental problems in the Switzerland among those on the left (World Values Survey 2005 - 2009) Global warm ing or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political interest 0.1859 0.0859 - 0.1043 (low; medium; high) (0.1705) (0.1868) (0.1588) Level of education 0.1293 - 0.0517 - 0.11 55 (0.0743) (0.0885) (0.0728) Lower class 0.2897 0.1165 1.4882 (0.7313) (1.0739) (1.0990) Working class - 0.2428 - 0.0404 0.1034 (0.4174) (0.4582) (0.4628) Upper class - 0.1243 - 0.0789 0.0708 (0.2507) (0.2698) (0.2385) Female 0.4248 0.3718 0.3513 (0.2535) (0.2578) (0.2363) Age younger than 30 0.3685 0.3315 0.1680 (0.5145) (0.4890) (0.4206) Age 30 - 39 0.7262 0.2376 - 0.0186 (0.3545) (0.3978) (0.3173) Age 50 - 59 0.2294 0.5193 0.0380 (0.3234) (0.3874) (0.3146) Age 60 - 69 0.6300 - 0.5317 0.1530 (0.3814) (0.3746) (0.3639) Age 70 and older - 0.1468 0.1910 0.4094 (0.4399) (0.4957) (0.5501) Not serious at all | Not very serious - 5.4999 (1.0443) Not very serious | Somewhat serious - 2.0281 - 3.8010 - 2.4199 (0.4566) (0.5502) (0.3903) Somew hat serious | Very serious 0.4688 - 0.6397 0.3623 (0.3686) (0.3667) (0.3441) Number of respondents 381 381 380 AIC 645 527 705 Standard errors in parentheses. p < 0.10 in bold. Listwise deletion and weighted. Left was defined as respondents who chose 1 , 2, 3, or 4 to describe their political ideology on the left - right continuum from 1 to 10. Respondents were asked , consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 283 Table A4.44 : Ordered logistic regression models predicting perceived seriousness of environmental problems in Switzerland among those on the right (Wor ld Values Survey 2005 - 2009) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political interest - 0.0341 0.0039 - 0.0925 (low; medium; high) (0.1752) (0.1819) (0.1684) Level of education 0.0818 - 0.0541 - 0.1736 * (0.0891) (0.0914) (0.0881) Lower class -- -- -- Working class - 0.0600 - 0.3410 0.4387 (0.6072) (0.5034) (0.6179) Upper class 0.5742 - 0.1681 0.2942 (0.2696) (0.2959) (0.2754) Female 0.7059 - 0.0 618 0.6846 (0.2572) (0.2580) (0.2640) Age younger than 30 0.3424 - 0.2526 0.6867 (0.4805) (0.5667) (0.6792) Age 30 - 39 0.6240 - 0.1408 0.7197 (0.4307) (0.4716) (0.4152) Age 50 - 59 0.0525 0.1419 0.6746 (0.3826) (0.4053) (0.3845) Age 60 - 69 0.5282 - 0. 0877 0.2691 (0.3926) (0.3841) (0.3904) Age 70 and older - 0.3777 - 0.0005 0.6182 (0.4055) (0.4006) (0.4319) Not serious at all | Not very serious - 2.3396 - 2.8700 (0.4648) (0.5437) Not very serious | Somewhat serious - 0.3077 - 2.7156 - 0.9658 (0.34 81) (0.4468) (0.4100) Somewhat serious | Very serious 1.4623 - 0.4493 1.4198 (0.3572) (0.3856) (0.4121) Number of respondents 289 291 290 AIC 725 574 677 Standard errors in parentheses. p < 0.10 in bold. Listwise deletion and weighted. Right was defin ed as respondents who chose 7, 8, 9, or 10 to describe their political ideology on the left - right continuum from 1 to 10. Respondents were asked , consider e ach of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 284 Table A4.45 : Ordered logistic regression models predicting perceived seriousness of environmental problems in Canada amon g those on the left (World Values Survey 2005 - 2009) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political interest 0.9458 0.5265 0.6576 (low; medium; high) (0.2 212) (0.2548) (0.2152) Level of education - 0.1663 - 0.0599 - 0.0688 (0.1067) (0.1222) (0.1045) Lower class - 1.6411 - 1.3043 0.9572 (1.3614) (0.9379) (0.8273) Working class - 0.4388 1.0186 0.1292 (0.3617) (0.6442) (0.3758) Upper class 0.2074 - 0.0 740 0.2737 (0.3455) (0.3995) (0.3227) Female 0.4102 0.4758 0.2784 (0.2938) (0.3549) (0.2839) Age younger than 30 - 0.0238 0.3247 0.5826 (0.3949) (0.5058) (0.4338) Age 30 - 39 0.3625 0.8371 0.6100 (0.4494) (0.5303) (0.4976) Age 50 - 59 1.4033 - 0.2840 0.1284 (0.5094) (0.5396) (0.4275) Age 60 - 69 0.5223 0.9273 - 0.1597 (0.5978) (0.7180) (0.5091) Age 70 and older - 0.9550 0.4111 - 1.3946 (0.5855) (0.7950) (0.5330) Not serious at all | Not very serious - 5.3339 - 6.0138 (0.7394) (0.9931) Not very serious | Somewhat serious - 3.5196 - 5.6115 - 3.6451 (0.5154) (0.9663) (0.5124) Somewhat serious | Very serious - 1.4058 - 1.4864 - 0.9309 (0.4289) (0.5306) (0.4330) Number of respondents 378 383 378 AIC 523 350 526 Standard errors in parentheses. p < 0 .10 in bold. Listwise deletion and weighted. Left was defined as respondents who chose 1, 2, 3, or 4 to describe their political ideology on the left - right continuum from 1 to 10. Respondents were asked , rld as a whole. Please, tell me how serious you consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 285 Table A4.46 : Ordered logistic regression models predicting perc eived seriousness of environmental problems in Canada among those on the right (World Values Survey 2005 - 2009) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Politi cal interest - 0.2141 0.1006 - 0.1128 (low; medium; high) (0.1729) (0.1925) (0.1823) Level of education - 0.0465 - 0.0981 0.1002 (0.0769) (0.0840) (0.0775) Lower class 1.2911 1.2251 0.5652 (0.7281) (0.7165) (0.5744) Working class 0.4672 0.1943 0.3 184 (0.3208) (0.3866) (0.3492) Upper class 0.7176 - 0.1321 - 0.1508 (0.3125) (0.3228) (0.2785) Female 0.1842 0.2946 0.2737 (0.2517) (0.2985) (0.2532) Age younger than 30 0.2988 - 0.0498 0.5929 (0.4320) (0.4840) (0.4309) Age 30 - 39 0.2292 0.2094 0.0 379 (0.4453) (0.4313) (0.3998) Age 50 - 59 0.4384 0.6467 0.4841 (0.3761) (0.4632) (0.3960) Age 60 - 69 0.1294 0.4512 0.0947 (0.4018) (0.4847) (0.4382) Age 70 and older 0.7718 1.0808 0.8126 (0.4298) (0.4781) (0.4489) Not serious at all | Not very se rious - 3.2633 - 5.4267 - 4.1127 (0.5718) (1.0519) (0.7279) Not very serious | Somewhat serious - 1.7126 - 4.0516 - 1.4660 (0.3975) (0.5815) (0.4256) Somewhat serious | Very serious 0.4370 - 0.7735 0.4129 (0.3881) (0.4200) (0.3854) Number of respondents 409 425 423 AIC 853 598 904 Standard errors in parentheses. p < 0.10 in bold. Listwise deletion and weighted. Right was defined as respondents who chose 7, 8, 9, or 10 to describe their political ideology on the left - right continuum from 1 to 10. Respond ents were asked , consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all ? 286 Table A4.47 : Ordered logistic regression models predicting perceived seriousness of environmental problems in Norway among those on the left (World Values Survey 2005 - 2009) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political interest 0.0348 0.0905 0.3809 (low; medium; high) (0.2184) (0.2673) (0.2341) Level of education 0.1848 0.2850 0.0535 (0.0787) (0.0912) (0.0920) Lower class 0.4355 0.7929 0.89 98 (0.7316) (0.9408) (0.9650) Working class - 0.1909 - 0.1172 - 0.1566 (0.3469) (0.3988) (0.3482) Upper class - 0.4808 - 0.7570 - 0.1743 (0.3118) (0.3726) (0.3325) Female 0.2318 0.5386 0.2753 (0.2720) (0.3110) (0.2814) Age younger than 30 - 0.1627 0.3 702 0.1999 (0.4531) (0.4793) (0.4883) Age 30 - 39 - 0.5240 - 0.2529 - 0.7163 (0.4404) (0.4842) (0.4403) Age 50 - 59 - 0.3684 0.8071 - 0.4981 (0.4150) (0.5066) (0.4337) Age 60 - 69 - 0.7080 0.5946 - 0.3837 (0.4943) (0.5752) (0.5205) Age 70 and older 0.3655 0 .5076 - 0.4898 (0.6602) (0.6773) (0.7129) Not serious at all | Not very serious - 4.2345 - 4.8039 - 5.7777 (0.6699) (1.0419) (1.0864) Not very serious | Somewhat serious - 3.2922 - 2.8263 - 3.4263 (0.5510) (0.5558) (0.5186) Somewhat serious | Very seriou s - 0.7415 - 0.4174 - 1.0926 (0.4386) (0.4729) (0.4462) Number of respondents 284 284 284 AIC 441.8809 342.2089 409.5932 Standard errors in parentheses. p < 0.10 in bold. Listwise deletion and weighted. Left was defined as respondents who chose 1, 2, 3, or 4 to describe their political ideology on the left - right continuum from 1 to 10. Respondents were asked , consider each of the following to be for the wor ld as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 287 Table A4.48 : Ordered logistic regression models predicting perceived seriousness of environmental problems in Norway among those on the right (World Values Sur vey 2005 - 2009) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political interest - 0.0462 0.0730 - 0.1061 (low; medium; high) (0.2017) (0.2166) (0.2086) Level of education 0.2489 0.0594 0.0876 (0.0759) (0.0725) (0.0727) Lower class 0.1142 - 0.2255 - 0.1665 (0.6974) (0.8916) (0.7519) Working class 0.0268 0.2025 - 0.0597 (0.3362) (0.3680) (0.3433) Upper class - 0.2997 - 0.2973 - 0.5031 (0.2423) (0.2530) (0.24 15) Female 0.3303 0.3238 0.0107 (0.2281) (0.2487) (0.2281) Age younger than 30 - 0.0692 - 0.0405 0.8466 (0.3346) (0.3704) (0.3817) Age 30 - 39 - 0.3322 0.1026 0.1252 (0.3278) (0.3513) (0.3271) Age 50 - 59 - 0.0325 - 0.1057 - 0.0506 (0.3631) (0.4038) (0.3 512) Age 60 - 69 - 0.1037 0.0076 0.4052 (0.3773) (0.3852) (0.3540) Age 70 and older 0.0310 0.7598 1.2839 (0.4196) (0.5071) (0.4936) Not serious at all | Not very serious - 3.5459 - 5.6552 - 4.8991 (0.4979) (1.0542) (0.7352) Not very serious | Somewhat serious - 1.7847 - 3.3204 - 2.2083 (0.3639) (0.4259) (0.3710) Somewhat serious | Very serious 0.2514 - 0.4714 0.1115 (0.3466) (0.3345) (0.3428) Number of respondents 337 338 338 AIC 679.9312 526.0186 631.3425 Standard errors in parentheses. p < 0.10 in bold. Listwise deletion and weighted. Right was defined as respondents who chose 7, 8, 9, or 10 to describe their political ideology on the left - right continuum from 1 to 10. Respondents were asked , as a whole. Please, tell me how serious you consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 288 Table A4.49: Ordered logistic regression models predicting perceive d seriousness of environmental problems in Germany among those on the left (World Values Survey 2005 - 2009) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political interest 0.0386 0.4165 0.0939 (low; medium; high) (0.1277) (0.1387) (0.1138) Level of education 0.1001 0.0313 0.0805 (0.0512) (0.0537) (0.0433) Lower class 0.1275 0.3196 0.5748 (0.6097) (0.5297) (0.5100) Working class - 0.0468 0.0009 - 0.0508 (0.2208) (0.2233) (0.2253) Upper class 0.2809 0.0591 - 0.0606 (0.2285) (0.2238) (0.2076) Female 0.2505 0.3110 0.3922 (0.1762) (0.1865) (0.1762) Age younger than 30 - 0.1535 0.4150 0.2257 (0.3201) (0.3329) (0.3262) Age 30 - 39 0.2520 0.3286 0.3165 ( 0.2875) (0.2807) (0.2881) Age 50 - 59 - 0.3483 - 0.4021 - 0.3892 (0.2980) (0.3392) (0.2431) Age 60 - 69 0.1609 0.2944 0.1070 (0.2764) (0.2724) (0.2514) Age 70 and older 0.2426 - 0.2665 - 0.0418 (0.3051) (0.2749) (0.2709) Not serious at all | Not very seri ous - 5.1552 - 4.7454 - 5.4177 (0.8069) (0.7455) (0.6664) Not very serious | Somewhat serious - 2.2642 - 1.8871 - 1.7602 (0.3113) (0.3024) (0.2433) Somewhat serious | Very serious 0.1285 0.2811 0.4999 (0.2549) (0.2618) (0.2309) Number of respondents 686 690 684 AIC 1273 1370 1387 Standard errors in parentheses. p < 0.10 in bold. Listwise deletion and weighted. Left was defined as respondents who chose 1, 2, 3, or 4 to describe their political ideology on the left - right continuum from 1 to 10. Responden ts were asked , consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 289 Table A4.50 : Ordered logistic regression models predicting perceived seriousness of environmental problems in Germany among those on the right (World Values Survey 2005 - 2009) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political interest 0.4317 0.5107 0.0954 (low; medium; high) (0.1927) (0.2028) (0.1893) Level of education 0.0046 0.0870 0.1021 (0.0752) (0.0841) (0.0829) Lower class - 1.2211 - 0.6647 - 0 .2287 (0.9378) (0.5121) (0.5916) Working class 0.1301 0.1337 - 0.1874 (0.3292) (0.3308) (0.3510) Upper class 0.7280 0.2960 0.3872 (0.3495) (0.3188) (0.3066) Female 0.6046 0.6072 0.2950 (0.2609) (0.2710) (0.2515) Age younger than 30 - 0.5169 - 0.72 93 - 0.5814 (0.4662) (0.4495) (0.4680) Age 30 - 39 0.1258 - 0.3269 - 0.2100 (0.4378) (0.4547) (0.4993) Age 50 - 59 0.5745 0.3632 0.0205 (0.4504) (0.4680) (0.4730) Age 60 - 69 - 0.3731 - 0.4014 - 0.7696 (0.4420) (0.4989) (0.4580) Age 70 and older - 0.1774 - 0 .8642 0.0564 (0.4400) (0.4412) (0.4519) Not serious at all | Not very serious - 4.0165 - 5.7998 - 4.2198 (0.6147) (1.0992) (0.5784) Not very serious | Somewhat serious - 1.5800 - 1.6349 - 1.2332 (0.4114) (0.4359) (0.4475) Somewhat serious | Very serious 0.6079 0.6476 0.7674 (0.3792) (0.3931) (0.4242) Number of respondents 285 284 284 AIC 597 592 652 Standard errors in parentheses. p < 0.10 in bold. Listwise deletion and weighted. Right was defined as respondents who chose 7, 8, 9, or 10 to describe their political ideology on the left - right continuum from 1 to 10. Respondents were asked , consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 290 Table A4.51 : Ordered logistic regression models predicting perceived seriousness of environmental problems in Australia among those on the left (World Values Survey 2005 - 2009) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political interest 0.4827 0.4108 0.4209 (low; medium; high) (0.2282) (0.3170) (0.2455) Level of education 0.13 22 0.2678 0.1919 (0.1271) (0.1480) (0.1235) Lower class - 0.2834 13.6589 - 0.6864 (1.1151) (0.6421) (1.7674) Working class 0.0341 - 0.1541 - 0.2229 (0.4707) (0.5424) (0.4314) Upper class 0.3713 0.7371 0.0106 (0.3748) (0.5732) (0.3990) Female 0.1075 0.0766 0.3690 (0.3052) (0.4320) (0.3175) Age younger than 30 0.3326 - 0.8354 0.2728 (0.4891) (0.6423) (0.5562) Age 30 - 39 0.2090 - 0.5782 - 0.1385 (0.6019) (0.7457) (0.6197) Age 50 - 59 0.1478 - 0.2659 - 0.3557 (0.4315) (0.6035) (0.4140) Age 60 - 69 0.3 360 0.9723 0.7759 (0.5180) (0.9326) (0.5966) Age 70 and older - 0.2632 0.0848 - 0.7616 (0.5955) (0.7217) (0.5324) Not serious at all | Not very serious - 4.1578 - 4.9675 - 4.4326 (0.8419) (1.0144) (0.7010) Not very serious | Somewhat serious - 2.6940 - 3 .5648 - 2.8540 (0.5273) (0.6753) (0.5477) Somewhat serious | Very serious - 0.6662 - 1.4760 - 0.6540 (0.5143) (0.6498) (0.4969) Number of respondents 281 280 281 AIC 353 230 358 Standard errors in parentheses. p < 0.10 in bold. Listwise deletion and we ighted. Left was defined as respondents who chose 1, 2, 3, or 4 to describe their political ideology on the left - right continuum from 1 to 10. Respondents were asked , erious you consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 291 Table A4.52 : Ordered logistic regression models predicting perceived seriousness of environmental pr oblems in Australia among those on the right (World Values Survey 2005 - 2009) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political interest 0.0368 0.4056 0.2075 (low; medium; high) (0.1844) (0.1937) (0.1838) Level of education - 0.1669 - 0.0856 - 0.1443 (0.0719) (0.0860) (0.0727) Lower class 1.3051 0.1341 - 0.0017 (1.1716) (0.8733) (0.6380) Working class - 0.1000 0.4023 - 0.2296 (0.3020) (0.3695) (0.3130) Upper class 0.1912 0.6201 0.0394 (0.2528) (0.3247) (0.2606) Female 0.6973 * 0.5813 0.3858 (0.2316) (0.2892) (0.2326) Age younger than 30 - 0.6404 - 0.0974 0.1438 (0.4748) (0.6714) (0.6225) Age 30 - 39 - 0.0262 - 0.1533 0.6420 (0.4734) (0.6485) (0.563 7) Age 50 - 59 0.3340 0.2036 - 0.1731 (0.3962) (0.5046) (0.3910) Age 60 - 69 - 0.3190 - 0.5657 - 0.9448 (0.3755) (0.4764) (0.3826) Age 70 and older - 0.4982 - 0.7777 - 0.9034 (0.3724) (0.4731) (0.3919) Not serious at all | Not very serious - 3.6843 - 6.4250 (0.4705) (1.1780) Not very serious | Somewhat serious - 2.2176 - 4.4246 - 2.9644 (0.4124) (0.6228) (0.4289) Somewhat serious | Very serious - 0.4075 - 1.1668 - 0.9042 (0.3920) (0.5051) (0.4013) Number of respondents 343 342 341 AIC 693 416 637 Standa rd errors in parentheses. p < 0.10 in bold. Listwise deletion and weighted. Right was defined as respondents who chose 7, 8, 9, or 10 to describe their political ideology on the left - right continuum from 1 to 10. Respondents were asked , environmental problems in the world as a whole. Please, tell me how serious you consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 292 Table A4.53 : Ordered logistic r egression models predicting perceived seriousness of environmental problems in Sweden among those on the left (World Values Survey 2005 - 2009) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal speci es or biodiversity Political interest 0.5121 0.3220 0.5121 (low; medium; high) (0.2039) (0.2361) (0.2022) Level of education 0.0479 0.0116 - 0.2424 (0.0906) (0.0849) (0.0912) Lower class 1.1557 1.8870 0.1638 (0.7442) (0.8964) (0.6684) Wor king class - 0.3798 - 0.0602 - 0.0599 (0.3330) (0.3657) (0.3418) Upper class - 0.2109 0.7357 - 0.1667 (0.3275) (0.3830) (0.3182) Female 0.7386 0.8708 0.5832 (0.2659) (0.3085) (0.2532) Age younger than 30 - 0.2190 - 0.9125 - 0.0715 (0.4013) (0.4370) (0.3 944) Age 30 - 39 0.1274 - 0.1399 0.6979 (0.4379) (0.5258) (0.4505) Age 50 - 59 - 0.0278 0.0749 0.0133 (0.4235) (0.4695) (0.3739) Age 60 - 69 - 0.2340 - 0.5925 - 0.5030 (0.5289) (0.5138) (0.4637) Age 70 and older 0.0946 - 0.5189 - 1.2431 (0.6774) (0.6522) (0 .5897) Not serious at all | Not very serious - 5.6436 - 6.3362 (1.1106) (1.1165) Not very serious | Somewhat serious - 3.2303 - 3.6883 - 3.4663 (0.5432) (0.5088) (0.5010) Somewhat serious | Very serious - 0.6595 - 1.1043 * - 1.2728 (0.4835) (0.4797) (0.4 477) Number of respondents 324 332 331 AIC 456 384 507 Standard errors in parentheses. p < 0.10 in bold. Listwise deletion and weighted. Left was defined as respondents who chose 1, 2, 3, or 4 to describe their political ideology on the left - right conti nuum from 1 to 10. Respondents were asked , consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very ser ious or not serious at all? 293 Table A4.54 : Ordered logistic regression models predicting perceived seriousness of environmental problems in Sweden among those on the right (World Values Survey 2005 - 2009) Global warming or the greenhouse effect Poll ution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political interest - 0.1162 - 0.2022 - 0.0171 (low; medium; high) (0.1691) (0.1761) (0.1702) Level of education 0.0487 0.0943 0.0720 (0.0771) (0.0937) (0.0799) Lower class - 0.1436 1.4299 1.3909 (0.7526) (1.0416) (0.8298) Working class 0.4169 1.2350 0.5547 (0.5353) (0.6960) (0.5065) Upper class 0.0178 0.0253 0.0588 (0.2242) (0.2523) (0.2283) Female 0.4059 0.9116 0.6279 (0.2232) (0.2501) (0.2142) Age yo unger than 30 - 0.3281 - 0.1266 0.3140 (0.3776) (0.4100) (0.3479) Age 30 - 39 - 0.2492 - 0.3675 - 0.3751 (0.3870) (0.3771) (0.3629) Age 50 - 59 0.0285 - 0.0195 0.0237 (0.3781) (0.4005) (0.3263) Age 60 - 69 - 0.3638 - 0.1152 - 0.0003 (0.3789) (0.4042) (0.3446) Age 70 and older 0.2351 0.3939 0.1429 (0.4615) (0.4959) (0.4236) Not serious at all | Not very serious - 4.6685 - 4.7713 (0.7476) (0.7750) Not very serious | Somewhat serious - 2.6694 - 3.7905 - 1.6940 (0.4449) (0.5863) (0.3692) Somewhat serious | V ery serious - 0.1044 - 0.1231 0.4990 (0.3769) (0.4279) (0.3518) Number of respondents 376 379 377 AIC 668 508 735 Standard errors in parentheses. p < 0.10 in bold. Listwise deletion and weighted. Right was defined as respondents who chose 7, 8, 9, or 10 to describe their political ideology on the left - right continuum from 1 to 10. Respondents were asked , consider each of the following to be for the world a s a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 294 Table A4.55 : Ordered logistic regression models predicting perceived seriousness of environmental problems in Japan among those on the left (World Values Survey 20 05 - 2009) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political interest 0.3983 0.4265 0.2978 (low; medium; high) (0.2650) (0.2428) (0.2537) Level of educa tion 0.0262 - 0.0440 0.0712 (0.1542) (0.1302) (0.1452) Lower class - 0.3526 - 0.3456 - 0.6211 (0.6433) (0.5312) (0.6826) Working class 0.2194 0.6650 0.3718 (0.4247) (0.3951) (0.3701) Upper class - 0.0092 0.0383 0.1126 (0.5307) (0.4175) (0.4135) Fema le 1.1559 1.0356 0.8747 (0.3965) (0.3302) (0.3237) Age younger than 30 0.8618 1.1596 0.2947 (0.5087) (0.4418) (0.4361) Age 30 - 39 2.0291 1.2885 0.4388 (0.9396) (0.6326) (0.5570) Age 50 - 59 1.1635 0.3900 0.0951 (0.5807) (0.4863) (0.4157) Age 60 - 69 0.7456 0.8792 0.9244 (0.5639) (0.4722) (0.4520) Age 70 and older - 0.0605 0.2975 0.3107 (0.6324) (0.6103) (0.6024) Not serious at all | Not very serious - 3.3455 (0.7917) Not very serious | Somewhat serious - 3.0591 - 2.4612 - 2.0758 (0.7992) (0 .6443) (0.6400) Somewhat serious | Very serious 0.0395 0.8064 0.8938 (0.6251) (0.5584) (0.5906) Number of respondents 195 193 193 AIC 238 305 355 Standard errors in parentheses. p < 0.10 in bold. Listwise deletion and weighted. Left was defined as re spondents who chose 1, 2, 3, or 4 to describe their political ideology on the left - right continuum from 1 to 10. Respondents were asked , consider each of th e following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 295 Table A4.56 : Ordered logistic regression models predicting perceived seriousness of environmental problems in Japan among those on the right (World Values Survey 2005 - 2009) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political interest 0.0108 0.1536 - 0.0057 (low; medium; high) (0.2418) (0. 2312) (0.2331) Level of education 0.0545 - 0.0050 0.0335 (0.1140) (0.1189) (0.1273) Lower class 1.0954 1.5502 1.4940 (0.7930) (0.7980) (0.6706) Working class 0.1833 0.4606 0.0751 (0.3742) (0.3587) (0.3509) Upper class 0.3985 - 0.0026 - 0.1558 (0.5353) (0.4125) (0.4543) Female 0.4859 0.6295 - 0.0372 (0.3212) (0.3074) (0.3025) Age younger than 30 0.1154 - 0.3389 1.0381 (0.6887) (0.5530) (0.5977) Age 30 - 39 - 0.0512 - 0.2108 - 0.1381 (0.6231) (0.5552) (0.4914) Age 50 - 59 - 0.1326 0.3495 0.3896 (0.5419) (0.5124) (0.4420) Age 60 - 69 - 0.6243 - 0.5891 - 0.8047 (0.5675) (0.5134) (0.4934) Age 70 and older - 0.4583 - 0.6786 - 0.2065 (0.5934) (0.5510) (0.5382) Not very serious | Somewhat serious - 3.7086 - 3.5586 - 2.4071 (0.6468) (0.7117) (0.5803) S omewhat serious | Very serious - 0.5585 - 0.0865 0.4380 (0.5664) (0.5525) (0.5743) Number of respondents 199 200 195 AIC 294 322 368 Standard errors in parentheses. p < 0.10 in bold. Listwise deletion and weighted. Right was defined as respondents who c hose 7, 8, 9, or 10 to describe their political ideology on the left - right continuum from 1 to 10. Respondents were asked , consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 296 Table A4.57 : Ordered logistic regression models predicting perceived seriousness of environmental problems in Turkey among those on the left (Wo rld Values Survey 2005 - 2009) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Political interest - 0.2857 0.1521 0.1815 (low; medium; high) (0.3289) (0.2953) (0.2804) Level of education 0.2071 0.1432 0.1368 (0.0893) (0.0906) (0.0786) Lower class - 0.1282 - 0.8227 - 0.1794 (0.8956) (0.7555) (0.6186) Working class 0.8216 0.2610 0.2746 (0.5941) (0.5843) (0.4568) Upper class 0.0687 - 0.1014 0.0269 (0.5198) (0. 4930) (0.4340) Female 0.1863 0.0118 - 0.4123 (0.4756) (0.4306) (0.3826) Age younger than 30 - 2.5140 - 0.9927 - 0.9513 (1.1047) (0.5846) (0.5003) Age 30 - 39 - 1.4871 - 0.9650 - 0.8677 (1.2707) (0.7180) (0.6181) Age 50 - 59 - 2.0200 * 0.2281 - 1.1309 (1.2145 ) (0.8830) (0.6466) Age 60 - 69 - 1.8471 - 1.0960 - 0.6654 (1.6311) (1.6197) (1.5332) Age 70 and older 12.0359 13.0971 12.6362 (1.0383) (0.7370) (0.6075) Not serious at all | Not very serious - 6.5834 - 5.3089 - 5.5876 (1.2845) (1.0352) (0.9534) Not very serious | Somewhat serious - 5.0873 - 4.9959 - 4.4517 (1.1534) (0.9592) (0.6862) Somewhat serious | Very serious - 3.6028 - 2.6728 - 2.5056 (1.0842) (0.7039) (0.5513) Number of respondents 276 281 280 AIC 218 244 312 Standard errors in parentheses. p < 0.10 in bold. Listwise deletion and weighted. Left was defined as respondents who chose 1, 2, 3, or 4 to describe their political ideology on the left - right continuum from 1 to 10. Respondents were asked , orld as a whole. Please, tell me how serious you consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 297 Table A4.58 : Ordered logistic regression models predicting per ceived seriousness of environmental problems in Turkey among those on the right (World Values Survey 2005 - 2009) Global warming or the greenhouse effect Pollution of rivers, lakes, and oceans Loss of plant or animal species or biodiversity Polit ical interest 0.2897 - 0.0537 0.1154 (low; medium; high) (0.1997) (0.1816) (0.1867) Level of education 0.1103 0.1535 0.0773 (0.0765) (0.0761) (0.0652) Lower class 0.8549 0.6248 0.1781 (0.7152) (0.5963) (0.5642) Working class 0.4947 1.0017 0.646 9 (0.3560) (0.3905) (0.3382) Upper class 0.2077 0.5312 0.3367 (0.3150) (0.3275) (0.2854) Female 0.3127 - 0.1944 - 0.4659 (0.2929) (0.2867) (0.2635) Age younger than 30 - 0.1828 - 0.0454 - 0.3466 (0.3417) (0.3678) (0.3220) Age 30 - 39 0.3444 0.1859 - 0. 1183 (0.4259) (0.4215) (0.3817) Age 50 - 59 - 0.1532 0.4445 - 0.0685 (0.4308) (0.5121) (0.4549) Age 60 - 69 0.8498 0.0978 - 0.1681 (0.6802) (0.5379) (0.5760) Age 70 and older 0.4401 0.2021 - 0.6058 (0.7413) (0.7951) (0.6520) Not serious at all | Not ve ry serious - 4.3290 - 5.9357 - 4.7076 (0.5132) (1.0269) (0.5244) Not very serious | Somewhat serious - 2.7780 - 3.4880 - 3.1669 (0.3640) (0.3993) (0.3279) Somewhat serious | Very serious - 1.4167 - 1.3769 - 1.5838 (0.3389) (0.3438) (0.3155) Number of respo ndents 455 465 461 AIC 544.0863 485.8105 630.5880 Standard errors in parentheses. p < 0.10 in bold. Listwise deletion and weighted. Right was defined as respondents who chose 7, 8, 9, or 10 to describe their political ideology on the left - right continuum from 1 to 10. Respondents were asked , consider each of the following to be for the world as a whole. Is it very serious, somewhat serious, not very serious or not serious at all? 298 Figure A4.1 : Percentage of respondents to say climate change is very serious by political ideology and interest in politics , by OECD country (World Values Survey 2005 - 2009) Notes: Column percenta ges are reported in the graph. Left are considered those from one to four on the political ideology left - right continuum, while right are from seven to 10. Political interest has been collapsed into three different categories: low, medium, and high. Low ca Political interest was collapsed this way in order to overcome small sample sizes in each cell and also show patterns across countries. I only include countries which had at least 25 respondents in each cell, so only 9 countries were included and eight OECD countries were dropped. 299 App endix E : Additional models, a nalysis and robust ness c hecks for Chapter 7 Table A5.1: Random - intercept multilevel linear models predicting perceived dangerousness of environmental problems (International Social Science Survey 2009 - 2011) Pollution from cars Pollution from industry Pollution of rivers, lakes, and streams Pesticides and chemicals in farming Genetically modified crops Nuclear power plants Dangerousness 0.0850 0.0805 0.0921 0.0806 0.0957 0.0824 of climate change (0.0022) (0.0026) (0.0041) (0.0028) (0.0046) (0.0036) Level of education 0.0008 0.0004 0.0032 0.0056 - 0.0032 - 0.0183 (0.0015) (0.0011) (0.0016) (0.0012) (0.0020) (0.0030) Female 0.0175 0.0126 0.0115 0.0200 0.0429 0.0698 (0.0027) (0.0029) (0.0031) (0.0040) (0.0057) (0.0088) Age under 30 - 0.0002 0.0111 - 0.0050 - 0.0370 * - 0.0484 0.0042 (0.0042) (0.0037) (0.0043) (0.0051) (0.0075) (0.0063) Age 30 - 39 0.0007 0.0088 - 0.0037 - 0.0117 - 0.0163 0.0037 (0.0040) (0.0044) (0.0036) (0.0043) (0.0056) (0.0069) Age 50 - 59 - 0.0032 - 0.0089 - 0.0064 0.0048 0.0013 - 0.010 8 (0.0046) (0.0030) (0.0045) (0.0029) (0.0053) (0.0056) Age 60 - 69 0.0061 - 0.0127 0.0004 0.0068 - 0.0019 - 0.0309 (0.0045) (0.0037) (0.0055) (0.0043) (0.0067) (0.0061) Age 70 and older 0.0183 - 0.0176 - 0.0067 0.0013 - 0.0055 - 0.0299 (0.0052) (0.0046) (0 .0060) (0.0045) (0.0074) (0.0085) Constant 0.4135 0.5232 0.4613 0.4725 0.3686 0.4921 (0.0129) (0.0118) (0.0131) (0.0131) (0.0164) (0.0193) Variance components Country - level standard 0.0542 0.0399 0.0553 0.0534 0.0761 0.0905 Deviation (0.0055 ) (0.0041 ) (0.0065 ) (0.0063 ) (0.0072 ) (0.0083 ) Individual - level 0.1893 0.1795 0.1991 0.2016 0.2328 0.2437 standard deviation (0.0036 ) (0.0029 ) (0.0046 ) (0.0043 ) (0.0046 ) (0.0051 ) N umber of respondents 41806 41912 41812 41596 38990 40751 Number of countries 32 32 32 32 32 32 AIC - 14086 - 17203 - 11155 - 10379 - 1957 540 300 Table A5.1 (cont d) Notes: Linear regression coefficients with standard error in parentheses. Coefficients estimates that can be distinguished from zero at the conventional level statistical significance (two - tailed p < 0.05) are in bold . Listwise deletion used for missing values. The dependent variable, perceived dangerous of each environmental problem, ranges from 0 to 1: 0=not dangerous at all; 0.25=somewhat dangerous; 0.75=very dangerous; 1=extremely dangerous Table A5.2: Linear models predicting perceived dangerousness of environmental problems, United States (International Social Science Survey 2009 - 2011) (1) (2) (3) (4) (5) (6) Pollution from cars Pollution from industry Pollution of rivers, lake s, and streams Pesticides and chemicals in farming Genetically modified crops Nuclear power stations Dangerousness of 0.0955 0.0862 0.0777 0.0664 0.0674 0.0797 climate change (0.0053) (0.0051) (0.0058) (0.0063) (0.0079) (0.0071) Level of educati on - 0.0186 - 0.0074 0.0069 - 0.0126 - 0.0143 - 0.0554 (0.0052) (0.0054) (0.0062) (0.0059) (0.0073) (0.0074) Female 0.0147 0.0304 0.0165 0.0459 0.0732 0.1278 (0.0118) (0.0122) (0.0128) (0.0131) (0.0165) (0.0163) Age under 30 - 0.0008 - 0.0149 0.0109 - 0.0505 - 0.0112 0.0678 (0.0192) (0.0193) (0.0217) (0.0208) (0.0257) (0.0244) Age 30 - 39 - 0.0263 - 0.0129 0.0016 - 0.0239 - 0.0144 0.0397 (0.0179) (0.0191) (0.0194) (0.0207) (0.0288) (0.0265) Age 50 - 59 0.0164 0.0294 0.0125 0.0247 0.0031 0.0218 (0.0206) (0.0195 ) (0.0192) (0.0205) (0.0273) (0.0265) Age 60 - 69 0.0396 0.0259 0.0344 0.0316 - 0.0565 - 0.0622 (0.0191) (0.0195) (0.0221) (0.0208) (0.0280) (0.0273) Age 70 and older - 0.0257 - 0.0540 - 0.0263 - 0.0272 - 0.0605 - 0.0461 (0.0217) (0.0222) (0.0254) (0.0257) (0. 0315) (0.0324) Constant 0.3986 0.4991 0.5293 0.5268 * 0.3805 0.5425 (0.0200) (0.0187) (0.0217) (0.0309) (0.0382) (0.0380) N umber of respondents 1304 1309 1305 1296 1134 1262 R 2 0.2727 0.2474 0.1786 0.1530 0.1338 0.2530 adjusted R 2 0.2682 0.2428 0.1735 0.1478 0.1276 0.2483 Notes: Linear regression coefficients with standard error in parentheses. Coefficients estimates that can be distinguished from zero at the conventional level statistical significance (two - tailed p < 0.05) are in bold . Listwise deletion used for missing values. 301 Table A5.3 : Correlations with perceived dangerousness of climate change a cross countries (International Social Survey Programme 2009 - 2011) Item Argentina Austria Belgium Bulgaria Canada Chile Taiwan Croatia Czech Republic Air po llution from cars 0.367 0.348 0.442 0.385 0.447 0.396 0.254 0.442 0.380 Air pollution from industry 0.419 0.299 0.374 0.402 0.472 0.501 0.379 0.460 0.415 Pollution of rivers, lakes, and streams 0.451 0.343 0.418 0.361 0.428 0.578 0.427 0.5 80 0.440 Pesticides and chemicals used in farming 0.377 0.345 0.384 0.382 0.343 0.423 0.345 0.472 0.384 Modifying the genes of certain crops 0.393 0.320 0.364 0.375 0.332 0.458 0.295 0.561 0.402 Nuclear power station s 0.419 0.211 0.316 0.2 92 0.187 0.500 0.173 0.397 0.287 302 Item Denmark Finland France Germany Israel Japan South Korea Latvia Lithuania Air pollution from cars 0.408 0.489 0.440 0.343 0.282 0.403 0.355 0.430 0.353 Air pollution from industry 0.376 0 .378 0.339 0.354 0.251 0.432 0.407 0.401 0.359 Pollution of rivers, lakes, and streams 0.346 0.315 0.356 0.310 0.316 0.426 0.404 0.436 0.475 Pesticides and chemicals used in farming 0.349 0.347 0.302 0.289 0.245 0.338 0.336 0.403 0.353 Modi fying the genes of certain crops 0.244 0.292 0.340 0.308 0.313 0.394 0.411 0.456 0.382 Nuclear power station s 0.196 0.328 0.294 0.291 0.163 0.277 0.253 0.367 0.380 303 Item Mexico New Zealand Norway Philippines Russia Slova k Republic Slovenia South Africa Spain Air pollution from cars 0.299 0.473 0.466 0.370 0.393 0.408 0.304 0.365 0.390 Air pollution from industry 0.439 0.480 0.434 0.389 0.385 0.396 0.316 0.410 0.398 Pollution of rivers, lakes, and streams 0.470 0.365 0.321 0.337 0.394 0.523 0.405 0.539 0.380 Pesticides and chemicals used in farming 0.282 0.360 0.342 0.317 0.392 0.460 0.323 0.416 0.339 Modifying the genes of certain crops 0.389 0.301 0.263 0.456 0.438 0.505 0.373 0.466 0.323 Nuclear power station s 0.421 0.344 0.307 0.395 0.386 0.259 0.363 0.385 0.310 304 Item Sweden Switzerland Turkey Great Britain United States All Countries Air pollution from cars 0.491 0.334 0.492 0.537 0.487 0.429 Air pollution from industry 0.421 0.309 0.449 0.520 0.485 0.427 Pollution of rivers, lakes, and streams 0.365 0.252 0.579 0.364 0.396 0.426 Pesticides and chemicals used in farming 0.368 0.267 0.522 0.379 0.350 0.376 Modifying the genes of certain crops 0.23 1 0.302 0.566 0.259 0.310 0.390 Nuclear power station s 0.222 0.258 0.432 0.325 0.369 0.341 305 Table A5.4: Exploratory factor analysis with promax rotation of perceived dangerousness of environmental problems (International Social Science Progra mme 2009 - 2011) Country Argentina Austria Belgium Bulgaria Canada Chile Factor 1 2 1 2 1 2 1 2 1 2 1 2 Climate Change 0.63 - 0.40 0.19 0.63 - 0.61 - 0.54 0.09 0.55 0.21 Air pollution from cars 0.58 - 0.82 - 0.24 0.65 - 0.60 - 0.80 - 0.08 - 0.07 0.72 Air pollution from industry 0.71 - 0.75 - 0.13 0.66 - 0.66 - 0.75 0.05 0.02 0.75 Pollution of rivers, lakes, and streams 0.70 - 0.40 0.32 0.66 - 0.61 - 0.41 0.30 0.48 0.24 Pesticides and chemicals used in farming 0.66 - 0.40 0.20 0.67 - 0.68 - 0.27 0.50 0.33 0.37 Modifying the genes of certain crops 0.57 - - 0.23 0.96 0.49 - 0.62 - - 0.09 0.72 0.66 0.00 Nuclear power station s 0.59 - 0.03 0.46 0.46 - 0.38 - - 0.09 0.58 0.83 - 0.10 Variance Explained 2.58 2.084 1.883 2.59 2.51 2.5 8 2.31 2.96 2.80 Correlation with Factor 1 - 0.67 - 0.71 0.79 306 Country Taiwan Croatia Czech Republic Denmark Finland Factor 1 2 1 2 1 2 1 2 1 2 Climate Change 0.73 - 0.85 - 0.10 0.63 - 0.18 0.38 0.59 - Air pollution fro m cars 0.63 - 0.18 0.50 0.62 - - 0.11 0.87 0.68 - Air pollution from industry 0.72 - - 0.15 1.07 0.71 - 0.09 0.67 0. 72 - Pollution of rivers, lakes, and streams 0.69 - 0.67 0.04 0.64 - 0.74 - 0.04 0.58 - Pesticides and chemicals used in farming 0.71 - 0 .56 0.16 0.66 - 0.82 - 0.01 0.67 - Modifying the genes of certain crops 0.66 - 0.84 - 0.17 0.56 - 0.55 - 0.06 0.45 - Nuclear power station s 0.57 - 0.39 0.22 0.39 - 0.30 0.12 0.45 - Variance Explained 3.18 2.08 3.11 2.60 2.37 2.30 2.51 Correlation with Factor 1 - 0.75 - 0.71 - 307 Country France Germany Israel Japan South Korea Factor 1 2 1 2 1 2 1 2 1 2 Climate Change 0.56 - 0.53 - 0.48 - 0.29 0.36 0.17 0.45 Air pollution from cars 0.54 - 0.57 - 0.52 - 0.79 - 0.04 0.83 - 0.10 Air pollution from industry 0.66 - 0.66 - 0.62 - 0.90 - 0.12 0.87 - 0.05 Pollution of rivers, lakes, and streams 0.65 - 0.63 - 0.55 - 0.32 0.39 0.27 0.40 Pesticides and chemicals used in farming 0.69 - 0.63 - 0.65 - 0.30 0.44 0.43 0.32 Modifying the genes of certain crops 0.55 - 0.54 - 0.55 - - 0.25 0.80 - 0.14 0.81 Nuclear power station s 0.47 - 0.48 - 0.28 - - 0.05 0.57 - 0.14 0.61 Variance Explained 2.47 2.36 1.99 2.47 2.31 2.52 2.38 Correlation with Factor 1 - - - 0.70 0.71 308 Country Latvia Lithuania Mexico New Zealand Norway Factor 1 2 1 2 1 2 1 2 1 2 Climate Change 0.61 - 0.59 - 0.63 - 0.50 0.14 0.62 - Air pollution from cars 0.72 - 0.62 - 0.50 - 0.85 - 0.11 0.69 - Air pol lution from industry 0.70 - 0.71 - 0.71 - 0.88 - 0.08 0.72 - Pollution of rivers, lakes, and streams 0.69 - 0.68 - 0.67 - 0.28 0.32 0.56 - Pesticides and chemicals used in farming 0.75 - 0.70 - 0.62 - 0.30 0.41 0.60 - Modifying the genes of certain cro ps 0.59 - 0.57 - 0.55 - - 0.18 0.83 0.46 - Nuclear power station s 0.50 - 0.49 - 0.55 - - 0.06 0.68 0.43 - Variance Explained 3.03 2.75 2.60 2.64 2.42 2.45 Correlation with Factor 1 - - - 0.73 - 309 Country Philippines Russia Slovak Republic Slovenia South Africa Factor 1 2 1 2 1 2 1 2 1 2 Climate Change 0.65 - - 0.06 0.68 0.12 0.61 0.60 - 0.70 - Air pollution from cars 0.59 - 0.60 0.13 0.72 0.01 0.61 - 0.56 - Air pollution from industry 0.63 - 0.93 - 0. 06 0.96 - 0.19 0.69 - 0.59 - Pollution of rivers, lakes, and streams 0.58 - 0.40 0.36 0.22 0.54 0.62 - 0.69 - Pesticides and chemicals used in farming 0.54 - 0.43 0.38 0.41 0.36 0.65 - 0.69 - Modifying the genes of certain crops 0.58 - - 0.05 0.73 - 0 .28 0.94 0.46 - 0.54 - Nuclear power station s 0.55 - - 0.11 0.62 0.02 0.40 0.41 - 0.53 - Variance Explained 2.44 2.85 2.81 2.60 2.54 2.41 2.78 Correlation with Factor 1 - 0.75 0.70 - - 310 Country Spain Sweden Sw itzerland Turkey Great Britain Factor 1 2 1 2 1 2 1 2 1 2 Climate Change 0.60 - 0.59 - 0.19 0.33 0.71 - 0.68 - 0.03 Air pollution from cars 0.61 - 0.71 - 0.72 - 0.05 0.67 - 0.88 - 0.12 Air pollution from industry 0.69 - 0.74 - 0.88 - 0.10 0.73 - 0. 90 - 0 .13 Pollution of rivers, lakes, and streams 0.62 - 0.65 - 0.18 0.18 0.76 - 0.34 0.21 Pesticides and chemicals used in farming 0.65 - 0.67 - 0.21 0.51 0.76 - 0.33 0.39 Modifying the genes of certain crops 0.46 - 0.41 - - 0.22 0.81 0.77 - - 0.26 0.9 7 Nuclear power station s 0.41 - 0.44 - - 0.05 0.53 0.62 - - 0.29 0.28 Variance Explained 2.41 2.64 2.06 2.05 3.61 2.73 2.13 Correlation with Factor 1 - - 0.69 - 0.70 311 Country United States Factor 1 2 Climate Chang e 0.41 0.24 Air pollution from cars 0.93 - 0.16 Air pollution from industry 0.81 0.03 Pollution of rivers, lakes, and streams 0.35 0.32 Pesticides and chemicals used in farming 0.37 0.34 Modifying the genes of certain crops - 0.22 0.81 Nuclear po wer station s - 0.03 0.61 Variance Explained 2.80 2.51 Correlation with Factor 1 0.76 312 Table A5.5: Logistic regression models predicting considering the environment most important or next most important issue, by country (International Social Survey Programme 2009 - 2011) Argentina Austria Belgium Bulgaria Canada Chile D anger ousness 0.5307 0.0302 0.5336 - 0.0251 0.5804 - 0.3742 of climate change (0.2468) (0.0949) (0.0837) (0.1726) (0.1220) (0.2940) Left party identification - 0.4336 - 0. 0696 0.3155 0.4768 0.1679 - 0.2102 (0.6319) (0.2338) (0.2887) (0.3407) (0.2470) (0.5357) Right party identification -- - 0.1704 - 0.1599 0.5658 - 0.5714 - 0.6858 (0.2385) (0.2901) (0.6006) (0.3235) (0.4952) Level of education - 0.0001 0.0279 0.0446 0.1315 0.0413 0.3425 (0.1755) (0.0832) (0.0714) (0.1270) (0.0823) (0.1697) F emale - 0.2706 0.0162 - 0.0967 - 0.2145 - 0.1591 0.1182 (0.3837) (0.1948) (0.1590) (0.3070) (0.2104) (0.3587) Age under 30 0.9336 0.2867 - 0.2321 1.8347 - 0.4210 - 0.5218 (0.6779) (0.27 06) (0.2698) (0.6824) (0.3625) (0.5322) Age 30 - 39 0.4141 0.0301 0.0493 0.9529 - 0.2605 - 0.4212 (0.7460) (0.3167) (0.2667) (0.7067) (0.4095) (0.4756) Age 50 - 59 0.9885 - 0.6265 0.3509 1.3053 - 0.0204 - 0.7236 (0.6812) (0.3337) (0.2436) (0.6653) (0.3119) (0 .5637) Age 60 - 69 0.3323 - 0.1630 0.3109 0.9753 0.1172 - 0.6220 (0.8420) (0.3447) (0.2709) (0.7043) (0.3257) (0.6316) Age 70 and older - 0.0550 - 0.1981 0.3550 1.0365 0.2114 - 2.5969 Constant - 5.5512 - 1.6626 - 2.6280 - 4.1518 - 2.3947 - 1.5339 (1.1355) (0.371 6) (0.3876) (0.8397) (0.5202) (1.0833) N umber of respondents 1066 958 1010 932 868 1371 AIC 298 837 1059 424 984 422 313 Croatia Czech Republic Denmark Finland France Germany D anger ousness 0.0011 0.2257 0.2389 0.4866 0.3638 0.1665 of climate change (0.1237) (0.0905) (0.0677) (0.0713) (0.0913) (0.0798) Left party identification 0.7431 - 0.1232 - 0.1228 - 0.3546 0.2644 0.2251 (0.3536) (0.1943) (0.1699) (0.1790) (0.2012) (0.1622) R ight party identification 0.4713 0.087 9 - 0.5937 - 0.4219 - 0.4109 - 0.2814 (0.4189) (0.2044) (0.1702) (0.2004) (0.2323) (0.2198) Level of education 0.0260 0.0809 0.0606 0.0800 0.0301 - 0.0071 (0.1013) (0.0794) (0.0771) (0.0663) (0.0640) (0.0534) F emale 0.3313 - 0.2429 0.3028 0.1837 0.3191 0.0 914 (0.2355) (0.1625) (0.1374) (0.1540) (0.1762) (0.1468) Age under 30 0.0900 0.5605 * - 0.0852 0.2045 0.5832 - 0.1446 (0.3593) (0.2556) (0.2365) (0.2448) (0.3130) (0.2443) Age 30 - 39 - 0.1568 0.1475 - 0.4542 0.2745 0.0950 - 0.0199 (0.3802) (0.2783) (0.23 97) (0.2569) (0.2445) (0.2643) Age 50 - 59 - 0.0797 0.2717 0.4025 0.0132 - 0.1863 0.0482 (0.3729) (0.2731) (0.2047) (0.2490) (0.2062) (0.2398) Age 60 - 69 0.2827 0.2473 0.1239 - 0.0880 - 0.0388 0.0887 (0.3762) (0.2867) (0.2245) (0.2603) (0.1913) (0.2430) Ag e 70 and older - 0.4523 0.1399 0.1466 0.0280 - 0.3110 0.1822 (0.5842) (0.3360) (0.2600) (0.4048) (0.2159) (0.2448) Constant - 2.8483 - 2.3595 - 1.4846 - 2.4583 - 2.4519 - 2.0771 (0.4459) (0.3182) (0.2755) (0.2968) (0.3125) (0.3176) N umber of respondents 1155 1299 1115 1092 1872 1222 AIC 609 1130 1329 1170 1923 1191 314 Japan South Korea Latvia Lithuania Mexico New Zealand D anger ousness 0.2777 0.0764 0.1525 0.1784 0.2682 0.4515 of climate change (0.1016) (0.0720) (0.1475) (0.32 04) (0.0930) (0.0783) L eft party identification 0.3553 - 0.0875 - 0.0103 - 1.0766 - 0.0677 - 0.7208 (0.2071) (0.2229) (0.3864) (0.7786) (0.1789) (0.2119) Right party identification 0.2598 - 0.2457 - 0.3711 0.8340 0.2492 - 0.8539 (0.2165) (0.2 364) (1.0562) (0.8748) (0.2371) (0.1895) Level of education - 0.1297 0.0905 - 0.1874 0.0453 - 0.1118 0.1653 (0.0644) (0.0587) (0.1454) (0.1766) (0.0599) (0.0619) Female 0.1763 - 0.1851 0.3827 0.5907 0.0073 - 0.3615 (0.1661) (0.1255) (0.337 2) (0.6010) (0.1637) (0.1698) Age under 30 0.4592 0.0163 - 0.5007 0.2380 0.0529 0.4285 (0.2840) (0.1921) (0.5385) (0.6695) (0.2363) (0.2979) Age 30 - 39 - 0.0289 - 0.1233 0.2042 0.0654 - 0.2130 0.3964 (0.2953) (0.1859) (0.5400) (0.7608) (0. 2658) (0.2935) Age 50 - 59 - 0.1449 0.5375 0.3367 - 1.6875 0.3390 0.4656 (0.2931) (0.2043) (0.4997) (1.0856) (0.2625) (0.2818) Age 60 - 69 - 0.1491 0.3635 - 0.0277 - 0.6667 - 0.2669 0.6774 (0.2876) (0.2441) (0.5858) (0.8735) (0.3516) (0.2848) Age 70 and older - 0.0548 0.0360 - 0.5129 - 0.7700 - 0.7122 0.3857 (0.3048) (0.2780) (0.7110) (0.8034) (0.3975) (0.3139) Constant - 2.8184 - 1.3576 - 3.5460 - 4.3560 - 2.8237 - 2.2271 (0.4153) (0.3217) (0.5966) (0.9681) (0.3812) (0.3282) N umber of respondents 1185 1504 902 935 1419 1030 A IC 1027 1619 354 221 1090 1027 315 Norway Philippines Russia Slovak Republic Slovenia South Africa D anger ousness 0.3294 0.0329 0.0107 0.1310 - 0.1271 - 0.1165 of climate change (0.0 673) (0.1450) (0.1002) (0.1362) (0.1232) (0.0986) L eft party identification 0.5450 - 0.0376 0.2062 0.2396 - 0.0625 - 0.3320 (0.1857) (0.4519) (0.1976) (0.2708) (0.2864) (0.2272) Right party identification - 0.5528 0.1851 - 0.6655 0.0312 0.0490 0.052 4 (0.2273) (0.4586) (0.5726) (0.6258) (0.2731) (0.6430) Level of education 0.0989 0.0630 0.0980 - 0.0235 0.1305 0.0353 (0.0526) (0.1107) (0.0792) (0.1142) (0.0888) (0.0922) F emale 0.5320 - 0.0996 - 0.0653 0.2650 - 0.0611 0.1445 (0.1321) (0.2783) (0.197 5) (0.2564) (0.2148) (0.2212) Age under 30 0.4871 0.0411 0.0823 0.0388 - 0.1775 - 0.0284 (0.2344) (0.4243) (0.2992) (0.4026) (0.3572) (0.2913) Age 30 - 39 - 0.1751 0.6251 0.0822 - 0.1670 0.1654 0.5553 (0.2200) (0.3804) (0.3057) (0.3961) (0.3446) (0.3083) Age 50 - 59 0.1829 0.0496 - 0.1354 - 0.0474 - 0.2161 - 0.1884 (0.2065) (0.5275) (0.3428) (0.3921) (0.3949) (0.4838) Age 60 - 69 0.3621 0.3793 - 0.6859 - 0.3276 0.3846 - 0.3447 (0.2115) (0.4656) (0.3692) (0.4506) (0.3832) (0.4234) Age 70 and older 0.2496 0.3199 - 0.7072 - 0.2316 0.1369 - 0.0974 (0.2644) (0.6976) (0.3800) (0.5462) (0.3935) (0.6236) Constant - 2.2249 - 2.9727 - 1.9422 - 3.0149 - 1.8010 - 2.4078 (0.2818) (0.5649) (0.4136) (0.5835) (0.4358) (0.3894) N umber of respondents 1228 1157 1476 1078 981 2816 AI C 1431 564 1079 598 666 1424 316 Spain Sweden Switzerland Turkey Great Britain United States D anger ousness 0.2634 0.4857 0.1705 - 0.3446 0.8343 0.4167 of climate change (0.0977) (0.0769) (0.0800) (0.1854) (0.2215) (0.0889) L eft party identification 0.0988 0.6745 0.4940 0.3413 - 0.1169 0.1234 (0.1856) (0.2145) (0.1631) (0.4565) (0.4165) (0.5184) R ight party identification - 0.3127 0.2670 - 0.2977 0.0475 - 0.0838 - 0.4400 (0.2571) (0.2441) (0.1721) (0.3554) (0.4366) (0.5 464) Level of education - 0.0665 0.0407 0.1518 - 0.2246 0.1867 0.1998 (0.0670) (0.0502) (0.0527) (0.1644) (0.1077) (0.0819) F emale - 0.1492 0.1534 0.1989 - 0.2039 - 0.2079 - 0.1888 (0.1462) (0.1498) (0.1369) (0.3331) (0.3044) (0.1884) Age under 30 0.0870 0.4150 0.0566 0.6996 - 0.0967 0.5135 (0.2173) (0.2455) (0.2186) (0.5488) (0.5379) (0.3293) Age 30 - 39 - 0.3684 0.5605 - 0.3599 0.2588 - 0.1431 0.6415 (0.2196) (0.2493) (0.2386) (0.5657) (0.4590) (0.3182) Age 50 - 59 - 0.1333 0.1681 0.2524 0.7468 0.0663 0.402 9 (0.2307) (0.2478) (0.2088) (0.5923) (0.4922) (0.3301) Age 60 - 69 - 0.5251 0.1192 0.2428 - 0.8235 - 0.5370 0.4258 (0.2954) (0.2421) (0.2251) (1.1243) (0.5563) (0.3289) Age 70 and older - 0.7368 0.4717 0.0223 0.5816 0.8562 0.8987 (0.3180) (0.2810) (0.24 32) (0.7379) (0.5178) (0.3625) Constant - 2.8530 - 3.1425 - 1.5247 - 3.1073 - 4.5472 - 3.5748 (0.3517) (0.3509) (0.2805) (0.8017) (0.7326) (0.6114) N umber of respondents 2351 1036 1158 1493 706 1239 A IC 1417 1137 1359 385 420 906 Note s : Cell entries are lo gistic regression coefficients with the standard error in parenthesis . Listwise deletion was used for missing values. Bold indicates two - tailed p <0.05. Models include individual survey weights. Israel and Taiwan were not included in the models because resp ondents were not asked for their partisan identification in these two countries. 317 Figure A5.1a : Effect of perceived dangerousness of climate change on considering the environment the most or next most important issue for country across natural logarithm o f mean GDP per capita by country (International Social Survey Programme 2009 - 2011) 318 Figure A5.1b : Effect of perceived dangerousness of climate change on considering the environment the most or next most important issue for country across natural logar ithm of mean HDI by country (International Social Survey Programme 2009 - 2011) Notes: The blue line is the OLS line of best fit and the blue region is its 95 percent confidence interval. The red line is the loess line of best fit. And the vertical error bar is the 95 percent confidence interval for the estimated difference in the predicted probability for ranking the environment as the most or next most important issue setting for each country , controlling for other factors . The statistical models used to produce this figure are presented in Table A5.5. 319 Table A5.6: Logistic regression models predicting considering climate change the most important environmental probl em, by country (International Social Survey Programme 2009 - 2011) Argentina Austria Belgium Bulgaria Canada Chile D anger ousness 0.2624 0.7505 0.7796 0.5530 1.1148 0.1306 of climate change (0.1721) (0.1103) (0.1019) (0.1710) (0.1478) (0.1840) L eft party identification - 0.0240 0.2143 - 0.0114 - 0.4232 0.2620 0.5216 (0.3952) (0.2077) (0.3576) (0.3130) (0.2834) (0.3829) R ight party identification 1.2040 0.0359 0.3216 0.7653 0.2197 0.3848 (0.8030) (0.2071) (0.3511) (0.4014) (0.3376) (0.457 7) Level of education 0.0098 0.0897 0.0396 0.0402 0.1767 0.1681 (0.1013) (0.0738) (0.0901) (0.1241) (0.0905) (0.1637) Female - 0.0109 0.2019 0.1982 - 0.1528 - 0.2696 0.3151 (0.2591) (0.1726) (0.1985) (0.2800) (0.2318) (0.3158) Age under 30 - 0.1673 0.23 76 0.2660 - 0.0530 0.1795 0.1306 (0.3866) (0.2498) (0.2988) (0.4645) (0.3641) (0.4215) Age 30 - 39 - 0.2655 - 0.0251 0.1776 0.0609 0.5423 0.3990 (0.4069) (0.2868) (0.3072) (0.4659) (0.3969) (0.5001) Age 50 - 59 - 0.4175 0.1408 0.0044 - 0.6243 0.0008 - 0.0935 (0.4274) (0.2691) (0.3090) (0.4535) (0.3185) (0.4580) Age 60 - 69 - 0.3445 0.1163 - 0.5675 - 0.7767 - 0.0632 - 0.6979 (0.4668) (0.3022) (0.3705) (0.4741) (0.3485) (0.6971) Age 70 and older - 0.7609 - 0.3592 - 0.8758 0.0195 0.0608 1.1192 (0.5364) (0.3679) (0.4 635) (0.4367) (0.3728) (0.6737) Constant - 3.2815 - 3.5333 - 4.0581 - 3.8963 - 4.7460 - 3.7494 (0.6410) (0.4188) (0.4847) (0.6401) (0.6329) (0.8125) N umber of respondents 1078 937 1010 903 881 1371 A IC 512 984 754 476 833 605 320 Croati a Czech Republic Denmark Finland France Germany Dangerousness 0.2130 0.4199 0.5787 0.7936 0.3420 0.6438 of climate change (0.1071) (0.1153) (0.0780) (0.0826) (0.1334) (0.0763) L eft party identification 0.3538 - 0.2171 0.2840 - 0.2874 0.262 9 0.0548 (0.3683) (0.2541) (0.1854) (0.1972) (0.3144) (0.1490) Right party identification - 0.9676 0.2684 0.2690 0.0693 - 0.1057 - 0.1112 (0.5968) (0.2466) (0.1819) (0.2083) (0.2974) (0.1847) Level of education - 0.0669 - 0.0256 - 0.0038 0.0125 0.0578 0.0 901 (0.0923) (0.1016) (0.0793) (0.0717) (0.0831) (0.0475) Female 0.2902 0.2224 - 0.0553 - 0.1999 0.1797 0.1830 (0.2014) (0.2122) (0.1470) (0.1682) (0.2412) (0.1335) Age under 30 - 0.1872 - 0.5074 - 0.0902 - 0.2096 0.0631 - 0.0975 (0.3012) (0.3133) (0.2374 ) (0.2599) (0.3866) (0.2154) Age 30 - 39 - 0.1974 - 0.5207 - 0.3798 - 0.2298 - 0.2438 0.1424 (0.3083) (0.3484) (0.2338) (0.2605) (0.3322) (0.2239) Age 50 - 59 - 0.0430 - 0.7006 - 0.0826 - 0.2152 - 0.9297 0.1979 (0.3031) (0.3582) (0.2098) (0.2504) (0.3199) (0.2070) Age 60 - 69 - 1.0814 - 0.2787 - 0.4222 - 0.2190 - 0.6328 - 0.3085 (0.4190) (0.3137) (0.2425) (0.2593) (0.2804) (0.2302) Age 70 and older 0.1200 - 0.2244 - 1.3106 - 1.2080 - 0.8275 0.1097 (0.3807) (0.3788) (0.3414) (0.5310) (0.3120) (0.2279) Constant - 2.7959 - 3 .2471 - 2.4738 - 3.1156 - 3.0295 - 3.0121 (0.4093) (0.3708) (0.3057) (0.3133) (0.4340) (0.2971) N umber of respondents 1136 1309 1103 1100 1938 1256 A IC 751 765 1189 1058 1197 1403 321 Japan South Korea Latvia Lithuania Mexico New Zeala nd D anger ousness 0.4730 0.4238 0.4834 0.7488 0.7095 0.8522 of climate change (0.0714) (0.0946) (0.1256) (0.1835) (0.1450) (0.1156) L eft party identification 0.1959 0.5792 - 0.1540 - 1.2808 - 0.0977 - 0.7232 (0.1534) (0.3403) (0.3077) (0.58 66) (0.1907) (0.2703) Right party identification - 0.0783 0.3259 - 1.0140 - 1.9656 0.0794 - 0.4836 (0.1559) (0.3587) (1.0314) (1.0381) (0.2683) (0.2437) Level of education 0.0097 0.1244 - 0.1811 - 0.0554 - 0.0097 0.1151 (0.0513) (0.0805) (0.0898) (0.1527) ( 0.0650) (0.0844) F emale 0.1026 0.1222 0.0813 - 0.3469 0.0188 0.0659 (0.1210) (0.1574) (0.2524) (0.3368) (0.1788) (0.2206) Age under 30 - 0.4001 0.6745 0.0038 - 0.3311 0.1062 0.1657 (0.2157) (0.2295) (0.4165) (0.4740) (0.2689) (0.3229) Age 30 - 39 - 0.4438 0.6516 0.1127 - 0.6321 0.4216 - 0.0397 (0.2067) (0.2180) (0.4589) (0.6551) (0.2779) (0.3060) Age 50 - 59 - 0.0347 - 0.2940 - 0.1819 - 0.5991 0.2966 - 0.4357 (0.2065) (0.3294) (0.4740) (0.5365) (0.3155) (0.3405) Age 60 - 69 - 0.1136 - 0.6337 0.2485 0.1899 - 0.1250 - 0.9682 (0.2052) (0.4395) (0.4648) (0.4640) (0.4035) (0.3711) Age 70 and older - 0.1579 - 0.1306 0.6738 - 0.9606 - 0.4723 - 0.4931 (0.2245) (0.4099) (0.4466) (0.5460) (0.4553) (0.3993) Constant - 1.2822 - 4.0209 - 3.7377 - 4.2992 - 4.6467 - 3.7954 (0.2834) ( 0.4620) (0.5065) (0.6919) (0.5871) (0.4052) N umber of respondents 1195 1498 872 912 1397 1014 A IC 1614 1125 496 384 935 719 322 Norway Philippines Russia Slovak Republic Slovenia South Africa D anger ousness 0.6953 0.1719 0.46 25 0.5249 0.4419 0.4474 of climate change (0.0730) (0.0945) (0.1132) (0.1687) (0.1376) (0.0937) L eft party identification 0.3056 - 0.3983 0.1488 0.0647 0.3631 - 0.2231 (0.1987) (0.3257) (0.2212) (0.2584) (0.3181) (0.2129) Right party identificat ion 0.0119 - 0.1219 0.5568 0.3406 0.8046 0.2991 (0.2278) (0.3605) (0.4826) (0.5302) (0.2655) (0.6742) Level of education 0.1224 0.0427 - 0.2220 0.1482 0.0064 0.1003 (0.0555) (0.0687) (0.1032) (0.0949) (0.1032) (0.0796) Female - 0.1845 - 0.3737 0.1785 - 0. 5374 - 0.1464 - 0.1916 (0.1381) (0.1810) (0.2251) (0.2348) (0.2315) (0.2060) Age under 30 - 0.1984 - 0.4373 - 0.4132 0.2478 - 0.1740 0.4920 (0.2370) (0.2702) (0.3529) (0.3556) (0.3714) (0.2741) Age 30 - 39 - 0.4146 - 0.1875 0.1807 - 0.2496 0.0112 0.7442 (0.22 72) (0.2566) (0.3244) (0.3702) (0.3754) (0.3080) Age 50 - 59 - 0.0423 - 0.0124 - 0.4660 - 0.0691 0.3346 0.6283 (0.2130) (0.2896) (0.3664) (0.3536) (0.3536) (0.3662) Age 60 - 69 0.0422 - 0.1227 - 0.8323 - 0.3227 - 0.1093 0.4170 (0.2147) (0.3365) (0.4017) (0.4369) (0.4158) (0.4177) Age 70 and older - 0.0963 - 0.4238 - 0.8794 - 1.1141 - 0.8453 0.2344 (0.2865) (0.4684) (0.4046) (0.6627) (0.4820) (0.4802) Constant - 2.6936 - 1.7207 - 3.7189 - 3.5827 - 3.7172 - 4.0532 (0.3040) (0.3657) (0.4613) (0.6248) (0.4945) (0.4199) N umber of respondents 1220 1174 1462 1054 995 2772 AIC 1345 1072 830 601 593 1711 323 Spain Sweden Switzerland Turkey Great Britain United States D anger ousness 0.4988 0.7588 0.3976 0.1786 0.6007 0.9587 of climate change (0.06 48) (0.0906) (0.0955) (0.1276) (0.1334) (0.1266) L eft party identification 0.1773 0.1411 - 0.0288 - 0.6620 0.2435 0.9479 (0.1328) (0.2146) (0.2004) (0.3966) (0.2987) (0.8554) R ight party identification 0.0237 - 0.0856 0.0959 0.0303 - 0.3481 0.6021 (0.1582) (0.2441) (0.1900) (0.2375) (0.3282) (0.8932) Level of education 0.0581 0.0662 0.1832 0.0753 0.2027 0.3306 (0.0398) (0.0558) (0.0589) (0.0880) (0.0767) (0.1173) F emale - 0.0717 - 0.2222 - 0.0457 - 0.0749 - 0.2050 - 0.5558 (0.1002) (0.1643) (0.158 0) (0.2248) (0.2213) (0.2296) Age younger than 30 0.1020 - 0.3118 - 0.0244 - 0.1427 - 0.0518 0.1518 (0.1577) (0.2599) (0.2569) (0.2997) (0.3828) (0.3697) Age 30 - 39 0.0750 0.1832 0.0524 - 0.4827 0.5186 - 0.0753 (0.1474) (0.2511) (0.2592) (0.3258) (0.3128) ( 0.3704) Age 50 - 59 - 0.0794 - 0.5030 - 0.0198 - 0.2251 0.1608 - 0.2318 (0.1632) (0.2616) (0.2454) (0.3957) (0.3440) (0.4155) Age 60 - 69 - 0.3056 - 0.4584 - 0.1331 - 0.6357 - 0.0767 - 0.2138 (0.1904) (0.2586) (0.2695) (0.5153) (0.3712) (0.3797) Age 70 and older - 0.6960 - 0.2217 - 0.4132 - 1.1214 0.1183 0.3811 (0.2079) (0.3166) (0.3019) (0.7576) (0.4112) (0.4143) Constant - 2.5529 - 3.0832 - 2.5392 - 2.9117 - 2.9869 - 5.6751 (0.2346) (0.3640) (0.3327) (0.5091) (0.4885) (0.9953) N umber of respondents 2340 1021 1169 151 0 685 1168 A IC 2529 989 1069 687 702 712 Note s : Cell entries are logistic regression coefficients with the standard error in parenthesis . Listwise deletion was used for missing values. Bold indicates two - tailed p <0.05. Models include individual survey we ights. Israel and Taiwan were not included in the models because respondents were not asked for their partisan identification in these two countries. 324 Figure A5.2a : Effect of perceived dangerousness of climate change on considering climate change the most important environmental problem facing country across natural logarithm of mean GDP per capita by country (International Social Survey Programme 2009 - 2011) 325 Figure A5.2b : Effect of perceived dangerousness of climate change on considering climate chang e the most important environmental problem facing country across mean HDI by country (International Social Survey Programme 2009 - 2011) Notes: The blue line is the OLS line of best fit and the blue region is its 95 percent confidence interval. The red li ne is the loess line of best fit. And the vertical error bar is the 95 percent confidence interval for the estimated difference in the predicted probability for ranking climate change as the most important environmental problem facing their country, settin for each country , controlling for other factors . The statistical models used to produce this figure are presented in Table A5.6. 326 REFERENCES 327 REFERENCES Abr amson, Paul R. 1997. "Postmaterialism and Environmentalism: A Comment on an Analysis and a Reappraisal." Social Science Quarterly (University of Texas Press) 78: 21 - 23. Abramson, Paul R., and Ronald Inglehart. 1995. Value Change in Global Perspective . Ann Arbor: University of Michigan Press. Aday, Sean, Steven Livingston, and Maeve Hebert. 2005. "Embedding the Truth: A Cross - Cultural Analysis of Objectivity and Television Coverage of the Iraq War." The Harvard International Journal of Press/Politic s 10: 3 - 21. Anderegg, William R. L., James W. Prall, Jacob Harold, and Stephen H. Schneider. 2010. "Expert Credibility in Climate Change." Proceedings of the National Academy of Sciences 107: 12107 - 09. Ansolabehere, Stephen, and David M. Konisky. 2014. Cheap and Clean: How Americans Think About Energy in the Age of Global Warming . Cambridge, MA: MIT Press. Bafumi, Joseph, and Andrew Gelman. 2006. "Fitting Multilevel Models When Predictors and Group Effects Correlate." In Annual Meeting of the Midwes t Political Science Association . Bafumi, Joseph, and Robert Y. Shapiro. 2009. "A New Partisan Voter." The Journal of Politics 71: 1 - 24. Barker, David C., and David H. Bearce. 2013. "End - Times Theology, the Shadow of the Future, and Public Resistance t o Addressing Global Climate Change." Political Research Quarterly 66: 267 - 79. Bartholomew, David J., Fiona Steele, Irini Moustaki, and Jane Galbraith. 2008. Analysis of Multivariate Social Science Data . Second ed. Boca Raton, FL: Chapman & Hall/CRC. B auer, Daniel J., and Sonya K. Sterba. 2011. "Fitting Multilevel Models with Ordinal Outcomes: Performance of Alternative Specifications and Methods of Estimation." Psychological Methods 16: 373 - 90. Bauer, Martin W., Nick Allum, and Steve Miller. 2007. "What Can We Learn from 25 Years of Pus Survey Research? Liberating and Expanding the Agenda." Public Understanding of Science 16: 79 - 95. Baumol, William J., and Wallace E. Oates. 1979. Economics, Environmental Policy, and the Quality of Life . Inglewo od Cliffs, New Jersey: Prentice - Hall. Begley, Sharon. 2007. "The Truth About Denial." Newsweek , (August 13, 2007). 20 - 29. Benoit, Kenneth, and Michael Laver. 2009. Party Policy in Modern Democracies . New York: Routledge. Berinsky, Adam J. 2009. In Time of War: Understanding American Public Opinion from World War Ii to Iraq . Chicago: University of Chicago Press. 328 Biel, Anders, and Andreas Nilsson. 2005. "Religious Values and Environmental Concern: Harmony and Detachment*." Social Science Quarterl y 86: 178 - 91. Bird, Deanne K., Katharine Haynes, Rob van den Honert, John McAneney, and Wouter Poortinga. 2014. "Nuclear Power in Australia: A Comparative Analysis of Public Opinion Regarding Climate Change and the Fukushima Disaster." Energy Policy 65: 644 - 53. Bohr, Jeremiah. 2014. "Public Views on the Dangers and Importance of Climate Change: Predicting Climate Change Beliefs in the United States through Income Moderated by Party Identification." Climatic Change forthcoming. Bord, Richard J., Ann Fi United State and International Perspectives." Climate Research 11: 75 - 84. Boykoff, Maxwell T. 2013. "Public Enemy No. 1?: Understanding Media Representations of Outlier Views o n Climate Change." American Behavioral Scientist 57: 796 - 817. . 2011. Who Speaks for the Climate? Making Sense of Reporting on Climate Change . New York: Cambridge University Press. Boykoff, Maxwell T., and Jules M. Boykoff. 2004. "Balance as Bias: Global Warming and the Us Prestige Press." Global Environmental Change 14: 125 - 36. Brechin, Steven R. 1999. "Objective Problems, Subjective Values, and Global Environmentalism: Evaluating the Postmaterialist Argument and Challenging a New Explanation." Social Science Quarterly 80: 790 - 809. Brechin, Steven R., and Medani Bhandari. 2011. "Perceptions of Climate Change Worldwide." Wiley Interdisciplinary Reviews: Climate Change 2: 871 - 85. Brechin, Steven R., and Willett Kempton. 1994. "Global Environ mentalism: A Challenge to the Postmaterialism Thesis?". Social Science Quarterly (University of Texas Press) 75: 245 - 69. Brownstein, Ronald. 2010. "Gop Gives Climate Science a Cold Shoulder." National Journal , (October 9, 2010). Brulle, Robert J., and David N. Pellow. 2006. "Environmental Justice: Human Health and Environmental Inequalities." Annual Review of Sociology 17: 103 - 24. Budge, Ian, Hans - Dieter Klingemann, Andrea Volkens, Judith Bara, and Eric Tanebaum. 2001. Mapping Policy Preferences: E stimates for Parties, Electors, and Governments 1945 - 1988 . New York: Oxford University Press. Buttel, Frederick M, and William L Flinn. 1978. "The Politics of Environmental Concern the Impacts of Party Identification and Political Ideology on Environmen tal Attitudes." Environment and Behavior 10: 17 - 36. Campbell, Angus, Philip E. Converse, Warren E. Miller, and Donald E. Stokes. 1960. The American Voter . New York: John Wiley & Sons. 329 Carman, Christopher Jan. 1998. "Dimensions of Environmental Policy Support in the United States: Research on the Environment." Social Science Quarterly 79: 717 - 33. Clark, Tom S. , and Drew A. Linzer. 2013. "Should I Use Fixed or Random Effects?". "The Color of News: How Different Media Have Covered the General Electio n." 2008. In Secondary The Color of News: How Different Media Have Covered the General Election , ed Secondary. Reprint, Reprint. Converse, Philip E. 1964. "The Nature of Belief Systems in Mass Publics." In Ideology and Discontent , ed. David Apter. New York: Free Press. 206 - 61. Corner, Adam, Dan Venables, Alexa Spence, Wouter Poortinga, Christina Demski, and Nick Pidgeon. 2011. "Nuclear Power, Climate Change and Energy Security: Exploring British Public Attitudes." Energy Policy 39: 4823 - 33. Dav enport, Coral. 2011. "Heads in the Sand." National Journal , (November 1, 2011). . 2015. "Senate Rejects Human Role in Climate Change." New York Times . Dickinson, Tim. 2010. "The Climate Killers." Rolling Stone . 35 - 41. Diekmann, Andreas, and Ax el Franzen. 1999. "The Wealth of Nations and Environmental Concern." Environment and Behavior 31: 540 - 49. Donner, Simon D., and Jeremy McDaniels. 2013. "The Influence of National Temperature Fluctuations on Opinons About Climate Change in the U.S. Sin ce 1990." Climatic Change . Downs, Anthony. 1957. An Economic Theory of Democracy . New York: Harper and Row. Dunlap, R. E., C. Xiao, and A. M. McCright. 2001. "Politics and Environment in America: Partisan and Ideological Cleavages in Public Support f or Environmentalism." Environmental Politics 10: 23 - 48. Dunlap, Riley E, and Richard P Gale. 1974. "Party Membership and Environmental Politics - Legislative Roll - Call Analysis." Social Science Quarterly 55: 670 - 90. Dunlap, Riley E. 2013. "Climate Chan ge Skepticism and Denial: An Introduction." American Behavioral Scientist 57: 691 - 98. Dunlap, Riley E., George H. Jr. Gallup, and Alec M. Gallup. 1993. "Of Global Concern: Results of the Health of the Planet Survey." Environment 35: 7 - 39. Dunlap, Riley E., and Peter J. Jacques. 2013. "Climate Change Denial Books and Conservative Think Tanks: Exploring the Connection." American Behavioral Scientist 57: 699 - 731. Dunlap, Riley E., and Robert Emmet Jones. 2002. "Environmental Concern: Conceptual and Me asurement Issues." In Handbook of Environmental Sociology , eds. Riley E. Dunlap and William Michelson. Westport, CT: Geenwook Press. 482 - 524. 330 Dunlap, Riley E., and Aaron M. McCright. 2011. "Organized Climate Change Denial." In Oxford Hanbook of Clim ate Change and Society , eds. John S. Dryzek, Richard B. Norgaard and David Schlosberg. New York: Oxford University Press. 144 - 59. Dunlap, Riley E., and Angela G. Mertig. 1995. "Global Concern for the Environment: Is Affluence a Prerequisite?". Journa l of Social Issues 51: 121 - 37. Dunlap, Riley E., and Richard York. 2008. "The Globalization of Environmental Concern and the Limits of the Postmaterialist Values Explanation: Evidence from Four Multinational Surveys." The Sociological Quarterly 49: 529 - 63. Eckberg, Douglas L., and T. Jean Blocker. 1989. "Variations of Religious Involvements and Environmental Concern: Testing the Lynn White Thesis." Journal of the Scientific Study of Religion 28: 509 - 17. Egan, Patrick J., and Megan Mullin. 2012. "Tur ning Personal Experience into Political Attitudes: The Effect of Local Weather on Americans' Perceptions About Global Warming." Journal of Politics 74: 796 - 809. Elsasser, Shaun W., and Riley E. Dunlap. 2013. "Leading Voices in the Denier Choir: Conserva tive American Behavioral Scientist 57: 754 - 76. Feldman, Lauren, Edward W. Maibach, Connie Roser - Renouf, and Anthony Leiserowitz. 2012. "Climate on Cable: The Nature and Impact of Global Warming Coverage on Fox News, Cnn, and Msnbc." The International Journal of Press/Politics 17: 3 - 31. Feygina, Irina, John T. Jost, and Rachel E. Goldsmith. 2010. "System Justification, the Denial of Global em - Personality and Social Psychology Bulletin 36: 326 - 38. Fisher, William F. 1993. The Road to Rio: Sustainable Development and the Nongovernmental Movement in the Third World . Westport, Connecticut: Praeger. Franzen, Axel. 2003. "Environmental Attitudes in International Comparison: An Analysis of the Issp Surveys 1993 and 2000." Social Science Quarterly 84: 297 - 308. Franzen, Axel, and Reto Meyer. 2010. "Environmental Attitudes in Cross - National Perspective: A Multilevel Analys is of the Issp 1993 and 2000." European Sociological Review 26: 219 - 34. Franzen, Axel, and Dominikus Vogl. 2013. "Two Decades of Measuring Environmental Attitudes: A Comparative Analysis of 33 Countries." Global Environmental Change 23: 1001 - 08. Geliss en, John. 2007. "Explaining Popular Support of Environmental Protection: A Multilevel Analysis of 50 Nations." Environment and Behavior 39: 392 - 415. Gelman, Andrew, and Jennifer Hill. 2006. Data Analysis Using Regression and Multilevel/Hierarchal Mode ls . New York: Cambridge University Press. Germain, Tiffany, Ryan Koronowski, and Jeff Spross. 2013. "The Anti - Science Climate Denier Caucus: 113th Congress Edition." Think Progress . 331 Gertz, Matt. 2013. "Fox Responds to New Climate Action with Global W arming Denial." Media Matters for America: http://mediamatters.org/blog/2013/11/02/fox - responds - to - new - obama - climate - action - with - g/196721 . Gid dens, Anthony. 2011. The Politics of Climate Change (Second Edition) . Malden, MA: Polity Press. Gilens, Martin. 2000. Why Americans Hate Welfare: Race, Media, and the Politics of Antipoverty Policy . Chicago: University of Chicago Press. Greenberg, Ma x, Denise Robbins, and Shauna Theel. 2013. "Media Sowed Doubt in Coverage of Un Climate Report." In Secondary Media Sowed Doubt in Coverage of Un Climate Report , ed Secondary . Reprint, Reprint. Greenly, Andrew. 1993. "Religion and Attitudes towar d the Environment." Journal for the Scientific Study of Religion 32: 19 - 28. Guber, Deborah Lynn. 2013. "A Cooling Climate for Change? Party Polarization and the Politics of Global Warming." American Behavioral Scientist 57: 93 - 115. . 2003. The Gra ssroots of a Green Revolution: Polling America on the Environment . Cambridge, MA:: MIT Press. Guth, James L., John C. Green, Lyman A. Kellstedt, and Corwin E. Smidt. 1995. "Faith and the Environment: Religious Beliefs and Attitudes on Environmental Poli cy." American Journal of Political Science 39: 364 - 82. Hamilton, Lawrence C. 2008. "Who Cares About Polar Regions? Results from a Survey of Us Public Opinion." Arctic, Antarctic, and Alpine Research 40: 671 - 78. Hamilton, Lawrence C, Matthew J Cutler, a nd Andrew Schaefer. 2012. "Public Knowledge and Concern About Polar - Region Warming." Polar Geography 35: 155 - 68. Hamilton, Lawrence C. 2011. "Education, Politics, and Opinions About Climate Change: Evidence for Interaction Effects." Climatic Change 1 04: 231 - 42. Hamilton, Lawrence C., and Barry D. Keim. 2009. "Regional Variation in Perceptions About Climate Change." International Journal of Climatology 29: 2348 - 52. Hart, P. Sol, and Erik C. Nisbet. 2012. "Boomerrang Effects in Science Communicatio n: How Motivated Reasoning and Identity Cues Amplify Opinion Polarization About Climate Mitigation Policies." Communication Research 39: 701 - 23. Hedeker, Donald. 2008. "Multilevel Models for Ordinal and Nominal Variables." In Handbook of Multilevel Ana lysis , eds. Jan De Leeuw, Erik Meijer and Harvey Goldstein. New York: Springer. 237 - 74. Hedeker, Donald, and Robert D Gibbons. 1994. "A Random - Effects Ordinal Regression Model for Multilevel Analysis." Biometrics : 933 - 44. Honaker, James, Gary King, a nd Matthew Blackwell. 2011. "Amelia Ii: A Program for Missing Data." Journal of Statistical Software 45: 1 - 47. 332 Hooper, Troy. 2012. "Santorum and Gingrich Dismiss Climate Change, Vow to Dismantle the Epa." The Colorado Independent , February 6, 2012. I nglehart, Ronald. 1990. Cultural Shift in Advanced Industrial Society . Princeton, NJ: Princeton University Press. . 1997. Modernization and Postmodernization: Cultural, Economic, and Political Change in 43 Societies . Princeton, NJ: Princeton Unive rsity Press. . 1995. "Public Support for Environmental Protection: Objective Problems and Subjective Values in 43 Societies." PS: Political Science & Politics 28: 57 - 72. . 1977. The Silent Revolution: Changing Values and Political Styles among W estern Publics . Princeton, NJ: Princeton University Press. Inglehart, Ronald, and Wayne E. Baker. 2000. "Modernization, Cultural Change, and the Persistence of Traditional Values." American Sociological Review 65: 19 - 51. Inglehart, Ronald, and Christia n Welzel. 2005. Modernization, Cultural Change, and Democracy: The Human Development Sequence . New York: Cambridge University Press. Jacobs, Alan M., and J. Scott Matthews. 2012. "Why Do Citizens Discount the Future? Public Opinion and the Timing of P olicy Consequences." British Journal of Political Science 42: 903 - 35. Jacoby, William G. 2000. "Issue Framing and Public Opinion on Government Spending." American Journal of Political Science : 750 - 67. Jacoby, William G. 2008. "The Dimensionality of P ublic Attitudes toward Government Spending." Political Research Quarterly 61: 158 - 61. Jacques, Peter J., Riley E. Dunlap, and Mark Freeman. 2008. "The Organisation of Denial: Conservative Think Tanks and Environmental Scepticism." Environmental Politic s 17: 349 - 85. Jones, Robert Emmet, and Riley E Dunlap. 1992. "The Social Bases of Environmental Concern: Have They Changed over Time? 1." Rural sociology 57: 28 - 47. Jost, John T., Sally Blount, Jeffrey Pfeffer, and György Hunyady. 2003a. "Fair Market Ideology: Its Cognitive - Motivational Underpinnings." Research in Organizational Behavior 25: 53 - 91. Jost, John T., Jack Glaser, Arie W. Kruglanski, and Frank J. Sulloway. 2003b. "Political Conservatism as Motivated Social Cognition." Psychological Bull etin 129: 339 - 75. Jost, John T., Brian A. Nosek, and Samuel D. Gosling. 2008. "Ideology: Its Resurgence in Social, Personality, and Political Psychology." Perspectives on Psychological Science 3: 126 - 36. Kahan, Dan M., Ellen Peters, Maggie Wittlin, Paul Slovic, Lisa Larrimore Ouellette, Donald Braman, and Gregory Mandel. 2012. "The Polarizing Impact of Science Literacy and Numeracy on Perceived Climate Change Risks." Nature Clim. Change 2: 732 - 35. 333 Kanagy, Conrad L., and Fern K. Willits. 1993. "A "Gr eening" of Religion? Some Evidence from a Pennsylvania Sample ". Social Science Quarterly 74: 675 - 83. Kellstedt, Paul M., Sammy Zahran, and Arnold Vedlitz. 2008. "Personal Efficacy, the Information Environment, and Attitudes toward Global Warming and Cl imate Change in the United States." Risk Analysis 28: 113 - 26. Kemmelmeier, Markus, Grzegorz Król, and Young Hun Kim. 2002. "Values, Economics, and Proenvironmental Attitudes in 22 Societies." Cross - Cultural Research 36: 256 - 85. Key, V.O. 1961. Public Opinion and American Democracy . New York: Knopf. Kidd, Quentin, and Aie - Rie Lee. 1997. "Postmaterialist Values and the Environment: A Critique and Reappraisal." Social Science Quarterly (University of Texas Press) 78: 1 - 15. King, Gary, James Honaker, Anne Joseph, and Kenneth Scheve. 2001. "Analyzing Incomplete Political Science Data: An Alternative Algorithm for Multiple Imputation." American Political Science Review 95: 49 - 69. Klein, Naomi. 2011. "Capitalism Vs. The Climate." The Nation . Knight, Kyle W., and Benjamin L. Messer. 2012. "Environmental Concern in Cross - National Perspective: The Effects of Affluence, Environmental Degradation, and World Society*." Social Science Quarterly 93: 521 - 37. Krosnick, Jon A., and Bo MacInnis. 2013. "Does the American Public Support Legislation to Reduce Greenhouse Gas Emissions?". Daedalus, the Journal of the American Academy of Arts & Sciences 141: 26 - 39. Kunda, Ziva. 1990. "The Case for Motivated Reasoning." Psychological Bulletin 108: 480 - 98. . 1987. "Motivated Inference: Self - Serving Generation and Evaluation of Causal Theories." Journal of Personality and Social Psychology 53: 636 - 47. A Cros s - National Study of 47 Countries." Journal of Peace Research 49: 11 - 22. Lawrence, Jill. 2011. "Jon Huntsman Goes 'Crazy'." The Daily Beast , (August 19, 2011). Layman, Geoffrey C., and Thomas M. Carsey. 2002. "Party Polarization and "Conflict Extensio n" in the American Electorate." American Journal of Political Science 46: 786 - 802. Layman, Geoffrey C., Thomas M. Carsey, and Juliana Horowitz. 2006. "Party Polarization in American Politics." Annual Review of Political Science 9: 83 - 110. Levendusky, M atthew. 2009. The Partisan Sort: How Liberals Became Democrats and Conservatives Became Republicans . Chicago: University of Chicago Press. Lodge, Milton, and Ruth C. Hamill. 1985. "A Partisan Schema for Political Information Processing." American Pol itical Science Review 80: 505 - 19. 334 Lodge, Milton, and Charles S. Taber. 2013. The Rationalizing Voter . New York: Cambridge University Press. Lupia, Arthur, and Mathew D. McCubbins. 1998. The Democratic Dilemma: Can Citizens Learn What They Need to Know . New York: Cambridge University Press. Malka, Ariel, Jon A. Krosnick, and Gary Langer. 2009. "The Association of Knowledge with Concern About Global Warming: Trusted Information Sources Shape Public Thinking." Risk Analysis 29: 633 - 47. Marquart - Pyatt, Sandra T. 2008. "Are There Similar Sources of Environmental Concern? Comparing Industrialized Countries." Social Science Quarterly 89: 1312 - 35. . 2012. "Contextual Influences on Environmental Concerns Cross - Nationally: A Multilevel Investigation." Social Science Research 41: 1085 - 99. Marquart - Pyatt, Sandra T., Rachael L. Shwom, Thomas Dietz, Riley E. Dunlap, Stan A. Kaplowitz, Aaron M. McCright, and Sammy Zahran. 2011. "Understanding Public Opinion on Climate Change: A Call for Research." Envir onment: Science and Policy for Sustainable Development 53: 48 - 52. Maslow, Abraham K. 1954. Motivation and Personality . New York: Harper and Row. McCright, Aaron M., and Riley E. Dunlap. 2003. "Defeating Kyoto: The Conservative Movement's Impact on U.S . Climate Change Policy." Social Problems 50: 348 - 73. . 2011. "The Politicization of Climate Change and Polarization in the American Public's View of Global Warming, 2001 - 2010." The Sociological Quarterly 52: 155 - 94. McCright, AaronM. 2011. "Polit Climate Change." Climatic Change 104: 243 - 53. Mooney, Chris. 2005. "Some Like It Hot." Mother Jones . Mulaik, Stanley A. 2010. Foundations of Factor Analysis . Second ed. Boca Raton, FL: Chapman & Hall/CRC. Nawrotzki, Raphael J. 2012. "The Politics of Environmental Concern: A Cross - National Analysis." Organization & Environment 25: 286 - 307. Nordhaus, William. 2013. The Climate Casino: Risk, Uncertainty, and Economics for a Warming Wo rld . New Haven: Yale University Press. Oreskes, Naomi, and Erik M. Conway. 2010. Merchants of Doubt: How a Handful of Scientists Obscured the Truth on Issues from Tobacco Smoke to Global Warming . New York: Bloomsbury Press. Painter, James, and Teresa A she. 2012. "Cross - National Comparison of the Presence of Climate Scepticism in the Print Media in Six Countries, 2007 - 2010." Environmental Research Letters 7: 1 - 8. 335 Pampel, Fred C. 2014. "The Varied Influence of Ses on Environmental Concern." Social S cience Quarterly . Pierce, John C, and Nicholas P Lovrich Jr. 1980. "Belief Systems Concerning the Environment: The General Public, Attentive Publics, and State Legislators." Political Behavior 2: 259 - 86. Popkin, Samuel L. 1991. The Reasoning Voter: Co mmunication and Persuasion in Presidential Elections . Chicago: University of Chicago Press. Rahn, Wendy M. 1993. "The Role of Partisan Stereotypes in Information Processing About Political Candidates." American Journal of Political Science 472 - 496: 2. Robbins, Denise. 2014. "13 Dumbest Things the Media Said About Climate Change in 2013." Media Matters For America: http://mediamatters.org/bl og/2014/01/01/13 - dumbest - things - media - said - about - climate - chan/197331 . Rosa, Eugene A, and Riley E Dunlap. 1994. "Poll Trends: Nuclear Power: Three Decades of Public Opinion." Public Opinion Quarterly : 295 - 324. Ross, Michael L. 2013. The Oil Curse: Ho w Petroleum Wealth Shapes the Development of Nations . Princeton, NJ: Princeton University Press. Sabbagh, Clara. 2005. "Environmentalism, Right Wing Extremism, and Social Justice Beliefs among East German Adolescents." International Journal of Psycholo gy 40: 118 - 31. Sandvik, Hanno. 2008. "Public Concern over Global Warming Correlates Negatively with National Wealth." Climatic Change 90: 333 - 41. Schuldt, Jonathon P., Sara H. Konrath, and Norbert Scharz. 2011. ""Global Warming" or "Climate Change"? W hether the Planet Is Warming Depends on Question Wording." Public Opinon Quarterly 75: 115 - 24. Schultz, P. Wesley, Lynnette Zelezny, and Nancy J. Dalrymple. 2000. "A Multinational Perspective on the Relation between Judeo - Christian Religious Beliefs and Attitudes of Environmental Concern." Environment and Behavior 32: 576 - 91. Scruggs, Lyle, and Salil Benegal. 2012. "Declining Public Support About Climate Change: Can We Blame the Great Recession?". Global Environmental Change 22: 505 - 15. Siddiqui, Sab rina. 2014. "Scott Brown, Cory Gardner Shift Stance on Climate Change in First Senate Debates." Huffington Post , October 6, 2014. Siegrist, Michael, and Vivianne H. M. Visschers. 2013. "Acceptance of Nuclear Power: The Fukushima Effect." Energy Polic y 59: 112 - 19. Sniderman, Paul M, Richard A Brody, and James H Kuklinski. 1984. "Policy Reasoning and Political Values: The Problem of Racial Equality." American Journal of Political Science : 75 - 94. Sniderman, Paul M., Richard A. Brody, and Philip E. Tet lock. 1993. Reasoning and Choice: Explorations in Political Psychology . New York: Cambridge University Press. 336 Snjiders, Tom A. B., and Roel J. Bosker. 2012. Multilevel Analysis: An Introduction to Basic and Advanced Multilevel Modeling Edited by T. T housand Oaks, CA: Sage. Steenbergen, Marco S., and Bradford S. Jones. 2002. "Modeling Multilevel Data Structures." American Journal of Political Science 46: 218 - 37. Stegmueller, Daniel. 2013. "How Many Countries for Multilevel Modeling? A Comparison o f Frequentist and Bayesian Approaches." American Journal of Political Science 57: 748 - 61. Taber, Charles S., and Milton Lodge. 2006. "Motivated Skepticism in the Evaluation of Political Beliefs." American Journal of Political Science 50: 755 - 69. Van Hi el, Alain, and Malgorzata Kossowska. 2007. "Contemporary Attitudes and Their Ideological Representation in Flanders (Belgium), Poland, and the Ukraine." International Journal of Psychology 42: 16 - 26. Van Liere, Kent D, and Riley E Dunlap. 1980. "The S ocial Bases of Environmental Concern: A Review of Hypotheses, Explanations and Empirical Evidence." Public Opinion Quarterly 44: 181 - 97. White, Lynn Jr. 1967. "The Historical Roots of Our Ecological Crisis." Science 155: 552 - 68. Will, George F. 2010. "Global Warming Advocates Ignore the Boulders." Washington Post , February 21, 2010. Wood, B. Dan, and Arnold Vedlitz. 2007. "Issue Definition, Information Processing, and the Politics of Global Warming." American Journal of Political Science 51: 552 - 6 8. Zajko, Mike. 2011. "The Shifting Politics of Climate Science." Society 48: 457 - 61. Zaller, John R. . 1992. The Nature and Origins of Mass Opinion . New York: Cambridge University Press.