JUST AND EQUITABLE ACCESS: DEVELOPING AND TESTING A METHODOLOGY THAT REPRESENTS DIVERSE PERCEPTIONS O F THE IMPORTANCE OF FORESTS By Dori Marie Hopkins A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Forestry Doctor of Philosophy 2015 ABSTRACT JUST AND EQUITABLE ACCESS: DEVELOPING AND TESTING A METHODOLOGY THAT R EPRESENTS DIVERSE PERCEPTIONS OF THE IMPORTANCE OF FORESTS By Dori Marie Hopkins The United States has committed itself to s ustainable f orest m anagement (SFM), managing its forests to meet the needs of present and future generations. R equir ing an informed, aware, and engaged public , SFM t heor etically calls attention to the social aspects of forest management reflecting the diversity of society. T he literature suggests those p articipating in forestry research are mostly middle - to senior - age d, college - educated, white males , resulting in dat a that do not accurately represent the diversity of people living in the U.S. Though many natural resource professionals commonly equate this phenomenon to a lack of interest from those that do not participate, research indicates that people do not partici pate because of the lack of opportunity and access. Understanding perceptions from the diversity of people is vital because perceptions may influence opinions and acceptability of every aspect of forest management . As the racial and ethnic composition of t he U.S. continues to diversify in the coming years, addressing this lack of access and representation becomes increasingly significant. This study attempts to address the issue of the under - representation of racial minorities and women in forestry research by testing the mixing of survey sampling and data collection modes. Following a national focus group study that identified many ways th at forests and trees are important to diverse people, a survey questionnaire was developed and administered to residents of the urban m idwestern city of Lansing, Michigan . The survey was administered either via mail or in - person, and the sampling occurred in one o f three ways: random, disproportionate random, and purposive. Following the survey data collection, interested participants were contacted to participate in an evaluative focus group or interview. Most of the scales developed for this study were successful at measuring the importance of forests to people. Positive correlations and level of agreement with scales show that participants had an overall positive attitude towards trees and forests . Many differences on the perceived importance of trees a nd forests were found based on race /ethnicity , gender, and the mediating effects of education level . Findings suggest that mixing sampling and survey data collection modes increases representation of racial minorities. Follow - up interviews and focus groups with survey p articipants provided valuable information on perceptions of survey research and tips for improving the survey research process. This study is one step on the path towards creating a more inclusive and diversified research and decision/policy - making process in the field of forestry (and for the natural resources field more broadly). It presents a unique approach to collecting data on the importance of trees and forests from a diverse audience and contributes to the limited body of literature e xamining racial and ethnic minorities, women, and the environment. Results exemplify the relevancy for participation of women and nonwhite minorities highlighting the need to consider issues of justice and equity in research methodologies. Copyright by DORI MARIE HOPKINS 2015 v To all the women who have struggled through the process. vi ACKNOWLEDGEMENTS I would like to express heartfelt gratitude for the support offered by the amazing faculty serving on my doctoral committee: Maureen McDonough, Dennis Propst, and John Schweitzer (MSU), Christine Vogt who agreed to serve even as she moved from MSU to Arizo na State, and Wayde Morse from Auburn University. All of these wonderful faculty have individually and as a group granted me patience, inspiration, and valuable insight on the research and doctoral process. I especially want to acknowledge the unceasing gu idance and encouragement from my chair, Dr. Maureen McDonough. You never fail to amaze me and I could not ask for a better mentor. This research was supported by the Gender, Justice, and Environmental Change (GJEC) Dissertation Research Fellowship, the US DA Forest Service, and the MSU Department of Forestry. Completion of the dissertation would be extremely delayed without the help of my research assistants: Alec Manaia, Rebecca Blaha, and Heather Surface. Thanks to community partners Allen Neighborhood Ce nter, Northwest Initiative, South Lansing Community offered advice and opportunities for recruiting research participants. Many other individuals deserve recognition: my co - directors, Anne Ferguson and Lisa Fine at the Center for Gender in Global Context (GenCen), the supporting staff at GenCen, my graduate student peers, many MSU faculty that provided challenging and provocative learning opportunities, my wonderful and patient husband who has had to accept many evenings and weekends of stressed writing time, my parents, sister, extended family, best friend, and pets that suffered less attention and walks through this process. This dissertation is truly the result of a d octoral student with an excellent support system. I could not have done this without any of you. vii TABLE OF CONTENTS LIST OF TABLES x LIST OF FIGURES x i i CHAPTER 1: INTRODUCTION 1 Study purpose 7 Organization of the dissertation 7 CHAPTER 2: LITERATURE REVIEW 8 The importance of understanding values 8 Studying diverse populations 15 Feminist methodology 18 P roblem statement 20 Preliminary research 21 Conceptual framework 22 Research questions 24 CHAPTER 3: METHODS 25 Introduction 25 Role of researcher and rese arch assumptions 25 Description of study area 26 Sampling and participant selection 29 Sample group 1 31 Sample group 2 32 Sample group 3 32 Metrics 34 Survey data collection 36 Focus groups and personal interviews 37 Data analysis 3 9 Addressing survey errors 39 CHAPTER 4: RESULTS AND DISCUSSION 4 3 Demographic chara cteristics of respondents 43 Discussion of research question 1 results and survey errors 47 Survey errors 49 Survey data analysis 50 Assessing reliability 51 viii Discussion of scale reliability 52 Discussion of overall importance 53 Differences by race/ethnicity 57 Item - level differences by race/ethnicity 62 Discussion of differences by race/ethnicity 63 Differences by gender 65 Part 1: Gender differences within entire sample 65 Part 2: Gender differences between racial/ethnic groups 67 Analysis 1 67 Analysis 2 70 Discussion of gender differences 73 Differences by sample group 74 Discussion of sample group differences 77 Path analysis 78 Discussion of path analysis 84 Participant evaluation of survey methodology 85 Motivation to participate 86 Opinions towards surveys in general 87 Most interesting aspects 88 Least interesting aspects 88 Tips to improve survey resear ch 89 Discussion of evaluative findings 90 Summary of results and discussion 90 CHAPTER 5: CONCLUSION 92 Summary of findings 93 Mixing methods and sample representativeness 93 Metrics 94 Racial differences 95 Weighted data analysis 96 Gender differences 96 Sample group differences 97 Mediating factors 97 Evaluative fin dings 98 Suggestions for future research 99 St udy limitations 101 Implications 102 Research methods 102 Policy and management 104 Conclusion 10 5 APPENDICES 107 APPENDIX I: Survey instrument 10 8 APPENDIX II: Pre - notice letter 1 19 ix APPENDIX III: Informed consent 12 1 APPENDIX IV: Script for evaluative focus groups/interviews 12 4 APPENDIX V: Non - response survey 127 REFERENCES 129 x LIST OF TABLES Table 1 Racial and ethnic composition of the United States 4 Table 2 Research questions and corresponding variables of interest 24 Table 3 Racial composition of Lansing, MI, MI and U.S., 2013 29 Table 4 Sampling methods 33 Table 5 Attempted and obtained surveys 34 Table 6 Value categories developed from f ocus group and literature review 35 Table 7 Description of pre - test sample (n = 20) 36 Table 8 Evaluative interview questions 38 Table 9 Characteristics of focus group/interview participants 38 Table 10 Characteristics of nonresponse sample 4 1 Table 11 Nonresponse percent agreement with importance scales 42 Table 12 Demographic characteristics of survey participants 46 Table 13 Demographic differences between sample groups (Chi - square analysis) 47 Table 14 Reliability and perce nt agreement for importance scales 55 Table 15 Pearson correlation matrix for importance scales/item 56 Table 16 Differences in perceptions of the importance of trees and forests by race/ethnicity (One - way ANOVA) 59 Table 17 Descriptive statistics for importance scales by race/ethnicity 60 Table 18 Comparing weighted and non - weighted data for racial differences 62 Table 1 9 Negative feelings item - level racial/ethnic differences 63 Table 20 D escriptive statistics for importance scales by gender 66 xi Table 21 Differences in perceptions of the importance of trees and forests by gender, Independence Samples T - Test 67 Table 22 Descriptive statistics for importance scales, white males and females 69 Table 23 Differences in perceptions of trees and forests by white males and females, Independence Samples T - Test 70 Table 24 Differences in perceptions on the imp ortance of trees and forests by race/ethnicity and gender combined (One - way ANOVA) 72 Table 25 Descriptive statistics for importance scales by sample group 76 Table 26 Differences in the perceptions of the importance of trees and for ests by sample group 77 Table 27 Pearson correlations for variables used in path analytic approach 79 Table 28 Path coefficients of a model for race, educational level, and concerns for trees and forests 81 Table 29 Characteristics of evaluative sample 85 Table 30 Focus group/interview themes 86 xii LIST OF FIGURES Figure 1 Conceptual f ramework 23 Figure 2 Map of study site: Lansing, MI 28 Figure 3 Example question from survey 51 Figure 4 Proposed path model for race, education level and concerns people had for trees and forests 79 Figure 5 Full path model for race, education level and concerns people had for trees and forests 80 Figure 6 Full path model for race, education level and concerns for trees and forests, females only 81 Figure 7 Proposed path model for education level, race and the environmental and ecological function of trees and forests 82 Figure 8 Full path model path model for education level, race and the e nvironmental and ecological function of trees and forests 83 1 C HAPTER 1: INTRODUCTION s what causes a lot (Focus group participant, 2008 1 ) - 1 Quote taken from data used to inform the USDA Forest Service 2010 National Report on Sustainable Forests 2 - - - - The United States joined the Working Group on Criteria and Indicators for the Conserv ation and Sustainable Management of Temperate and Boreal Forests in 1994 (the Working Group is now referred to as the Montreal Process, as it first met in Montreal, Quebec). 3 Twelve countries participate in the Montreal Process including Argentina, Australi a, Canada, Chile, China, Japan, Republic of Korea, Mexico, New Zealand, Russian Federation, USA, and Uruguay. Montreal Process participants agreed to the following tasks: 1) Come to consensus on what constitutes SFM; 2) develop and promote internationally agreed - upon SFM criteria and indicators; and, 3) to report every five years on progress towards SFM. This ongoing and evolving process has had some success in measuring ecological and economic variable s associated with SFM. M easuring social variables has p roven to be difficult. The first national reports on SFM by member countries were released in 2003. In 2006, the Pinchot Institute for Conservation published an analysis of the 2003 USDA Forest Service National Report on SFM (Pinchot Institute for Conserv analysis described the social section as poor, citing a lack of relevant data. In 2008, Montreal Process member countries revised the indicators, adding two new social indicators and deleting others. In preparation for the 2010 USDA Forest Service National Report on Sustainable Forest Management, indicator and measure the extent and intensity of behaviors and attitudes through which individuals and communities connect with forests. The purpose of this indicator was to collect data on the range of ways that trees and forests are important, and from a range of people in order to inform an understanding of regional or demographic differe nces on the importance of trees and forests to people (USDA Forest Service, 2011). 4 Table 1: Racial and ethnic composition of the United States Race /ethnicity Percentage White/Caucasian 7 7 . 7 % Black/African American 13.2 % American Indian 1 . 2 % Asian 5.3 % Hawaiian and other Pacific Islander 0.2% Two or more races 2.4 % Total * 100% *persons of Hispanic/Latino origin: 1 7 . 1 %, White persons non - Hispanic: 6 2 . 6 % ( U.S. Census Bureau , 2013 ) - - - - The natural resources field, in general, lacks racial and ethnic minority and gender - d iversified representation and recruitment in professional, governmental and non - governmental organization positions (Kuhns, Bragg, & Blahna, 2002) . This has been reinforced over time due to institutional factors that are slow to change (Taylor, 2008; Taylo r, 2011). With racial and ethnic minorities and women at the forefront of the environmental justice movement, it is clear 5 that a lack of interest in the environment is a poor argument (Kaalund, 2004; Parker & McDonough, 1999 ; Prindevill e, 2004; Sze, 2004 ; Taylor, 2008 ). Research shows that women express slightly greater environmental concern than men and tend to prioritize environmental issues differently ( Bell & Braun, 2010 ; Dietz , Kalof, & Stern , 2002; Hunter , Hatch, & Johnson , 2004; McCright & Xiao, 2014 ). Other studies show that women are generally the initiators and leaders in environmental justice activism (Bell & Braun, 2010; Verchick, 2004). Increasingly, women of color are leading efforts to combat environmental injustices in their commun ities (Kaalund, 2004). Still, top leaders and decision makers in forestry - related policy positions tend to be white and male (Anthony , Knuth, & Lauber , 2004; Burns , Schlozman, & Verba , 2009 ; McDonough et al., 2003; Taylor, 2002 ). Despite the racial and ge nder diversity of the U.S., forestry and natural reso urce research continues to struggle with inclusivity. It is common for n atural resource professionals to equate th e underrepresentation o f diverse voices to a lack of interest from those underrepresented , but research shows another explanation: that people do not participate because of the lack of oppor tunity and access (McDonough, 2003; McDonough et al., 20 0 3 ; Taylor, 2008 ). This then becomes a justice issue. F ail ure to include people in the research pro cess that informs policy and decision - making limits their access to that same policy and decision - making. The environmental justice literature emphasizes the need to address social issues that intersect with wicked environmental problems. Disregarding these phenomena has a negative impact on the goals of sustainability (Pearsall & Pierce, 2010; Taylor, 2011). While it is true that some people may not wish to be included in research, it is also true that some are more willing to participate in research i f the methods are more personal or compatible within their community. It is not just an issue of those who wish to participate and those who do not. By not providing 6 accessible means to participate in research that informs policy and decision - making, certa in people are privileged in the policy process and the lack of representation becomes an issue of equity (McDonough et al, 2003). An inclusive research methodology addressing how trees and forests are important to people has the potential to bring diverse voices to SFM by monitoring the importance of forests to a representative cross - order to produce quality baseline data, the research sample nee ds to be representative of the racially diverse United States populace, including traditionally under - represented groups. The sample should be comparable to the racial/ethnic and gender composition of the population. Focus groups were used to develop impo rtance of forests to people data for the U.S. 2010 National Report on Sustainable Forest Management (described in more detail below). While the 7 focus group method is an excellent tool for collecting and measuring diverse perceptions, it is a labor intensiv e approach and does not allow generalization to a broad population. One solution is represent all interests in SFM over time; however, this poses a serious methodolo gical challenge. Study purpose : On e purpose of this study is to respond t o the lack of diverse representat ion of racial/ethnic minorities and women in forestry - related research by developing a more inclusive survey research methodology and to contribute to democratizing the research process and forestry decision - making. A mix of sampling and data collection modes is explored to assess effectiveness for drawing a representative sample. The second study purpose is to examine differences by race/e thnicity and gender in the ways that trees and forests are perceived to be important. It is a step in the direction to meet the goals of sustainable forest management by collecting the range of data on the importance of trees and forests from the range of people. Ch apter 2 of this dissertation synthesizes the literature on the different ways that trees and forests are represented as important to people in various disciplines and fields of study. It also presents literature on research methodologies and addressing the challe nge of studying diverse populations. Chapter 3 provides details on the research methodology used in this study. Chapter 4 presents the results and discussion . Chapter 5 summarizes the analysis and provides conclusions and recommendations. S upporting materi al such as research instruments and related items may be found in the references and appendices. 8 CHAPTER 2 : LITERATURE REVIEW My sash is lowered when night comes on; But let there never be curtain drawn Between you and - Robert Frost ( 1916 ) The 2005 Millennium Ecosystem Assessment (MEA), Ecosystems and human w ell - b eing: Synthesis , provided a starting point for examining the importance of trees and fore sts to people. The assessment i s global in scope and provides an overview of ecosystem services in relationship to human well - being (MEA, 2005). The MEA (2005) categorized ecosystem services as provisioning, regulating, supporting, and cultural, which directed a literature review on th e importance of trees and forests relating to cultural heritage, spirituality, knowledge systems, education, aesthetics, social relations, sense of place, and recreation. The Importance of Understanding Values : In addition to the large racial and ethnic diversity of the U.S. populace (Table 1), there is a great range of reasons that trees and forests are important to people. Research has demonstrated the breadth of ways people associate importance with forests. These perceptions vary with different types of forests, groves of trees, and even individual trees. Furthermore, different groups of individuals hold diverse perceptions across these types of forests. The mix of values 9 extends from consumptive to non - consumptive uses and includes items that relate to economic, ecological and social benefits derived from trees and forests. Understanding how individual and societal values are linked to ecosystem services is central to creating a metric for measuring the importance of forests to people. Rokeach (1968) noted philosophy, education, political science, economics, anthropology, and theology, as well as gue d several points about values that relate to measuring the importance of forests to people. First, values represent single, stable beliefs that people use for evaluating attitudes and behaviors, and second, values underlie value orientations (i.e., patt erns of basic beliefs), which influence attitudes and may affect behavioral intentions (also Rokeach, 1973). Value orientations relating to natural resources, they suggest, are found along a continuum from anthropocentric to biocentric. Ford , Williams, Sm ith, and Bishop (2014) developed a model for understanding the formation of public acceptability judgments which suggested that people employ their values for the environment through psychological processes based on beliefs, aesthetic experience, and trust . Ecologist E.O. Wilson (1984) has advanced a hypothesis suggesting humans have an innate tendency to focus on life Economists focus on values which affect the way people behave in their use of limited resources, and differentiate between values which are held by people for something, and the values that people assign Power, 1980). Kempton, Boster, and Hartley (1995) found that Americans in particular are motivated by a diversity of values from religion to those rela ted to human utility, and even by a strong belief that nature itself has rights. 10 There are a number of authors who have explored values related to forests. Patel , Rapport, VanDerlinden, and Eyles (1999) studied forests and societal values held by scientists as well as public perceptions of forest health. This qualitative study used focus groups to explore ecological indicators, how society defines f orests and forest health, the values placed on forests, and relating societal and scientific views of forests and forest condition. Participants in the six focus groups ( held in London and Grand Bend, Ontario) perceived forests as complex, diverse, and mul tifaceted entities. Key elements of their definitions involved themes of: holism and interconnectedness; a balance of life forms; a diversity of species; cycles of change such as death, decay, and fire; a state of being pristine, or without human intervent ion; sensual aspects, especially visual features; and metaphors such as refuge, home, origin, sacred place, hospital, and living being or entity. Healthy forests were viewed as pristine places with high diversity that were integrated with larger ecosystems . Humans and nature were regarded as distinctly incompatible, except among First Nation participants. Meinig (1976) wrote several papers on how people look at landscapes which provide some insights on perceiving change, both natural and human induced, in n ature. Meinig suggested that people see the landscape as: habitat, artifacts, systems, problems, wealth, ideology, history, place, and aesthetics. These two works suggest it is important to gauge individual perceptions or definitions of forests. Patel et al. (1999) proposed that refuge or home. Tarrant and Cordell (2002) examined the influence of four indicators of population diversity (i.e., age, ethnicity, place of res idence, and gender) on explaining amenity values of forests, environmental attitudes, and forest value. They assessed wood production (utilitarian value), clean air (life support value), scenic beauty (aesthetic value), and heritage (spiritual 11 value). Resp ondents were asked to rank these four values from most important to least important. The data showed that younger participants favored a biocentric orientation to forests and the natural environment. Saunders, Brooks, and Myers (2006) chose to divide anthr opocentrism into egoistic and humanistic categories. Their article reference d and drew Behavior (TPB). TPB argues that people will do what they say they will do. For example, if people say they will vote in a specific way, they are likely to do so. Or, if they indicate that they plan to protect or conserve fores ts, they will do so. Kals, Schumacher, and Montada (1999) explored emotional affinity toward nature as a motivation for protecting forests which depends heavily on Theory of Reasoned Action (TRA) and TPB. Drawing from a perspective on biophilia nature and that their well - being depends, to a great extent, on the relatio nships with the perspective of affinity (i.e., emotional affinity) with nature, they explore interest which motivates the seeking of knowledge and the emotional affin ity which motivates contact and sensual experience. Considering this, they explored a willingness to commit to and the manifestation of their motivation to protec t nature. Shindler and Cramer (1999) presented an interesting perspective on understanding public values and the wicked problems associated with forest management. They, too, involved the TRA and the influence of subjective societal norms and social pre ssures. While some of the relationships to forests are individually derived and defined, consciously or unconsciously, they observed a tendency among people to identify with groups and reflect the group perspective. 12 They suggested there are contrasting nat ural resource paradigms and offer that it is not sufficient to merely ask people what they want. Such studies should have people make tough decisions that weigh costs and benefits associated with forest management decisions. To accomplish this, they call f or civil discourse to explore the wicked problems. In China, Liu , Ouyang, and Miao (2010) conducted surveys considering the New Environmental Paradigm (NEP) to examine highly conflicted protected areas. They suggest that environmental education and community participation can promote proenvironmental attitudes while alleviating conflict s. Similarly, Li , Wang, Liu, & Weng (2010) examined Taiwanese forest values through survey research and found that environmental value orientations (spiritual, utilitarian, etc.) influenced how people think forests should be managed. They also found that p eople considered the forests public property and associated a range of historical and cultural values with trees and forests beyond ecological importance. Vining and Tyler (1999) conducted a qualitative study in Indiana to evaluate values, emotions, and de sired outcomes reflected in public response to the Hoosier National Forest Management Plan. They suggest that emotions are essential for interpreting and organizing information and that emotions play a role in summarizing complex information. Moreover, the y suggest that emotions motivate action and as such reveal value conflicts. Lutts (1992) presented an analysis of the effect of the Disney movie Bambi on public perceptions of forests and hunting further demonstrating the sensitive nature of values and emo tions as they relate to contentious natural resources issues. Winter and Lockwood (2003) proposed a natural area value scale (NAV) to distinguish between and gauge the relative strengths of individual intrinsic, non - use values of nature. They identified t hree principal factors in their scale: non - use; use (non - recreation); and recreation. 13 Several authors reported on scales considering the importance of nature or specific forests to people. These generally had two primary constructs: anthropocentric and bio centric (M anning, Valliere, & Minteer , 1999; Vaske , Donnelly, Williams, & Jonker , 2001). Mann ing et al. (1999) also looked at national forests (i.e., Green Mountains National Forest) relative to a range of forest values. The items they explored were one - di mensional and included anthropocentric and biocen tric items. Tarrant and Cordell (2002) suggested a scale to measure public values of forests (PVF). They note d importance assigned to economi c forest values among the general public and a concomitant found respondents valued the role forests could have in protecting ecosystems. Following their hypothesi s that environmental values develop from a sense of connectivity with nature, Dutcher , Finley, Luloff, and Johnson (2007) surveyed Pennsylvania landowners to examine the extent to which connectivity to nature may predict and illuminate environmental concer n and behavior. Through multiple regression models, they found that those reporting a high level of connectivity with nature did maintain a significant, positive relationship to environmental concern and environmental behavior. Beyond those mentioned abov e, a diverse collection of studies from many disciplines describes ways in which trees and forests affect and are affected by people. Cultural aspects relating to trees and forests, and ranging from traditional knowledge systems, to those associated with e thnicity, to different uses within groups have been documented (Allison, 1992; Chavez, 1993; Emery & Pierce, 2005; Flood, 2007; Gordon, Barton , & Adams, 2013 ; Hamilton, 2012; Johnson, Bowker , Green, & Cordell, 2007; Leatherberry, 2000 ; MEA, 2005; Mitchell & Hobby, 2010; Murphy, Chretien, & Brown & Njambi, 2012; Parker & McDonough, 14 1999; Roberts & Chitewere, 2011; Taylor, 2002; UNEP, 2007;). Social relations and forests have been studied as well in relation to maple syrup production, occupatio nal communities, knowledge systems, and informal social institutions (Carroll , Lee, & McLain , 2005; Richard & Burns, 1999; Whitney & Upmeyer, 2004). The sense of place literature provides examples of research demonstrating how people attach meaning to part icular trees and/or forested areas (Chiesura, 2004; Gustafson, 2001; Jorgensen & Stedman, 2001; Kil , Holland, & Stein , 2013; Lewicka, 2011; Perkins, 2010 ; Spartz & Shaw, 2011; Tuan, 1979; Williams & Vaske, 2003 ). Though the dominant, modern western view of nature has been to see forests as secular landscapes, there are notable studies that discuss the spiritual values of forests that can be organized in four categories: forests as intrinsically sacred places (Castleden , Garvin, & Nation , 2009 ; Glowacka , Was hburn, & Richland , 2009; LaVelle, 2001; Pemberton, 1985), forests as the loci of significant spiritual history or culture (Votrin , 2005), forests as a reflection of a higher power, and forests as a place to commune with a higher power and/or experience tra nscendence (Kamitsis & Francis, 2013; Leatherberry, 2000 ; Swearer, 1998; Trigger & Mulcock, 2005). Phenomena dealing with development, cognition, and contact with nature and forests are explored in a large set of diverse studies relating this context to q uality of life issues (Bratman, 2012; Hammitt, 2000; Heerwagen & Orians, 1993; Kaplan & Kaplan, 1989; Kellert, 198 4 ; Kellert, 1996; Kellert, 2012; Nakau et al . , 2013; Orians, 1980; Orians, 1986; Quantz, 1897; Shin , Shin, & Yeoun , 2012; Tyrväinen et al . , 20 12 ; Ulrich, 1993; Ulrich et al., 1991a; Ulrich, Dimbeg, & Driver, 1991b; Van Wieran & Kellert, 2013). The importance of the aesthetic value of trees and forests is documented in studies linking preferences to a biological underpinning or a cognitive desire for diversity, patterns, and complexity (Carroll, 2007; Dandy & Van Der Wal, 2011 ; Hauru , Koskinen, Kotze, & Lehvävirta , 2014; Kaplan 1987 ; Kovacs , LeRoy, Fis cher, 15 Lubarsky, & Burke , 2006; Rolston , 1998). A multitude of studies on the importance of recreational forest use and the contribution to personal benefits exist as well (Bradley, 2011 ; Brown, 1984; Driver, 1996; Driver , Nash, & Haas, 1987; Driver , Brown, & Peterson , 1991; Duerdan , Taniguchi, & Widmer , 2012; Kaplan & Talbot, 1983; Manfredo & Driver, 1996; Norton, 2010; Norton et al., 2014; Shellman, 2014; Stein & Anderson, 2002; Van Wieren & Kellert, 2013; White & Hendee, 2000; Young & Crandall, 1984). This review outlines diverse literature s panning the natural and social sciences to provide a foundation for understanding the breadth of ways that trees and forests are important to people. There are multiple ways that people, both individually and as groups, utilize and value the many benefits derived from forests. The literature documents the benefits more frequently discussed among forestry professional s including environmental and biological values, economic values, recreational values, and the important products provided by trees and forests . It also illustrates the ways in which people interact with trees and forests that provide a difficult to define special meaning. An understanding of the various ways that trees and forests are valued by people helps to inform and create policy that bette Studying diverse populations: The diverse literature reviewed above shows that there are a variety of ways in which people use and attach meaning to trees and forests. Few studies address the various ways that trees and forests are important representing racially diverse voices. Most of the l iterature represents voices of white, educated males, and is lacking in data collected from racial and ethnic minorities and women. Trees provide a broad range of benef its to communities and may be utilized or considered important (or unimportant) for different reasons (Donovan & Mills, 2014; 16 Flocks, Escobedo, Wade, Varela, & Wald, 2011; Jennings, Johnson Gaither, & Schulterbrandt Gragg, 2012; Poe, McLain, Emery, & Hurle y, 2013). Research conducted with diverse populations is necessary to broaden an understanding of how trees and forests are perceived to be important in order to provide a useful contribution to policy development. Environmental sustainability should serv e the community as a whole, with a conscious effort to address the social inequalities that often divide our communities (Pearsall & Pierce, 2010; Warner, 2002). Environmental justice scholars emphasize the importance of including multiple perspectives in research about natura l resources with special attention given to the structural inequalities that impact perspectives (Berland, Schwarz, Herrmann, & Hopton, 2015; Donovan & Mills, 2014). There is a lack of relevant literature within the field of natural re source management on successful ways to study racially and ethnically diverse populations . O ther academic discipline s have pursued various techniques to study traditionally underrepresented audiences. Racial and ethnic minorities respond differently to res earch recruitment attempts (McLea n & Campbell , 2003). Recruitment of minority participants (racial/ethnic and women) has been documented in health and medical studies. Yancey , Ortega, and Kumanyika (2006) examined 95 studies published between 1999 and 2005 that suggested effective recruitment strategies for underserved populations. These included: strong sampling approach/identification of targeted participants, community involvement/nature and timi ng of contact, incentives and logistical issues, and cultural adaptations. Feldman , Radermacher, Browning, Bird, and Thomas (2008) identified three recruitment methods to increase participation in research studies with diverse and underrepresented groups: identifying and working with key informants, using bilingual interviewers, and supporting research partners in recruitment activities. Flexibility is 17 necessary as these populations are traditionally more difficult to recruit and employing different strateg ies may be necessary. Effective methods for teaching computer literacy to underserved audiences included using a mobile computer lab brought into communities of interest. Participants were more open ronstad , Teegerstrom, & Osgood , 2004). Reed , Foley, Hatch, and Mutran (2003) used a community church - based strategy to develop trust and recruit older African Americans to participate in survey research. Garber and Arnold (2006) suggest ed that in the medic al field, the researcher would be more effective at recruiting racial minority participants if s / he focused the study around what the minority community wants from the research. Parker and McDonough (1999) tailored a sampling strategy to elicit survey resp onses from African Americans who traditionally have low survey response rates, resulting in a 52% response rate. Satterfield (2001) used a variety of non - traditional methods to solicit environmental values. She then suggested a few alternatives in methodol ogy, belief statements. The literature describing social science research methods for use with racial and ethnic minority populations is sparse. It does reit erate the importance of using alternative methods to increase participant diversity. Studies examining forest values among racial and ethnic minority populations do exist (Allison, 1992; Chavez, 1993; Flood, 2007; Johnson et al ., 2007; Leatherberry, 2000; Parker & McDonough, 1999; Swearer, 1998 ; Taylor, Kuo, & Sullivan, 2002 ). R elatively speaking, there are very few studies on trees and forests that have included raci ally and ethnically diverse perspectiv es and none of them have accomplished this at the nat ional level or over time. In order to successfully include diverse citizens in forest policy 18 research, it is necessary to approach the study of how trees and forests are important to people with a research methodology that considers issue s of justice and f airness. One way to achieve this is through feminist - inspired mixed methods research. Feminist Methodology : Feminist scholars focus on the problems of representation and interpretation of people and groups that are traditionally isolated from cultural, political, and economic power (Kirsch, 1999). In response to the positivist paradigm, feminist principles of research include a focus on reflexivity (Fonow & Cook, 1991; Kirsch, 1999; Reinharz, 1992), a commitment to balancing the inequalities in research, conducting ethical research that is also emancipatory, an understanding that science is not value - neutral, and a multiplicity of research methods ( Fonow & Cook, 1991; Harding, 1991; Hardin g, 2001; Jayarantne & Stewart,1991; Kirsch, 1999; Reinharz, 1992). There is no one single feminist research method protocol; rather, most feminist researchers focus on the methods most suitable to address the research question (Fonow & Cook, 1991; Jayarantne & Stewart, 1991; Reinharz, 1992). Utilizing mixed methods is pow erful for thoroughness in research and addressing inequalities as different people respond to different research recruitment strategies (Hodgkin, 2008; Nightingale, 2003 ; Reinharz, 1992 ). Feminist epistemology and the alternative research paradigm known as pragmatism have inspir ed the methodology used in the research reported here . These scientific worldviews combine the methods from qualitative and quantitative worlds to produce mixed methods research (Creswell, 2009; Fonow & Cook, 1991; Jayarantne & Stewa rt, 1991; Johnson et al., 2007 ; Reinstein, 1992 ; Tashakkori & Teddlie, 1998; Tedd lie & Yu, 2007 ). Mixed methods research is focused on developing the methods that best answer important research questions. 19 Pragmatism is especially helpful for the production 2010, p.6 - 7), in this case, measuring the importance of forests to people. The strength lies in the combination of methods. Pragmatists and feminist researchers acknowledge that power and privilege determine which (or whose) reality will be prioritized in a research context (Hodgkin, 2008; Kirsch, 1999 ; Mertens, 2007). In mixed methods research, it is acceptable to strategically use different kinds of sampling (Hodgkin, 2008 ; Jayarantne & Stewart, 1991; Tashakkori & Teddlie, 1998). Often, sample, or no sampling procedure, that is suitable for all research ( Tashakkori & Teddlie, 1998, p.74). Pragmatis ts and feminist researchers posit that sampling should be based more on the nature of the question. The traditional survey methodology is challenged by the continued decline in response rates in affluent countries like the U.S. (Groves & Peytcheva, 2008) . Also, urbanicity shows an effect on response as response rates tend to be lower in urban areas. This is especially relevant for the research reported here because there is also a relationship between where racial and ethnic minorities live and urbanicity . Racial and ethnic minorities tend to have lower response rates but this can be changed if the material is focused on minority subcultures (Groves & Peytcheva, 2008). Researchers can utilize a mixed - modal survey for collecting data when it is difficult t o achieve the desired responses through one mode alone. For the purposes of this research, the focus is collecting responses from traditionally under - represented groups, including racial and ethnic minorities and women. T raditional mail surveys continue to encounter challenges as people face endless piles of junk mail and survey solicitations, busy lives, and phon e solicitations 20 (Dillman, Smyth, & Christian, 2009 a; Dillman et al. , 2009b). Switching to another su rvey data collection mode is an effective way to increase response rate and achieve a representative sample (Dillman et al., 2009b; de Leeuw , Hox, & Dillman , 2008). Problem statement: An abundance of literature is available on the different ways that people use and value trees and forests. T he lite rature suggests those p articipating in forestry research are mostly middle - to senior - age d, college - educated, white males , resulting in data that do not accurately represent the diversity of people living in the U.S. Though many natural resource profession als generally interpret this phenomenon to a lack of interest from those that do not participate, research indicates that people do not participate because of the lack of opportunity and access. Low participation may instead be the result of utilizing rese arch methodologies that are not compatible with diverse participants . As the U.S. moves toward a more inclusive decision - making process to fulfill the goals of sustainable forest management, action is needed to address the lack of diverse representation i n forestry research. Studies suggest that the U.S. population will continue to diversify in the coming decades, making it imperative that racial minorities and women are included in forestry research that has the potential to effect policy and decision - mak ing. Understanding perceptions from the diversity of people is vital because perceptions may influence opinions and acceptability of every aspect of forest management . This study attempts to address the issue of the under - representation of racial minoritie s and women in forestry research by using a feminist - inspired mixed methods research methodology. The mixing of survey sampling and data 21 collection modes are tested to assess for sample representativeness, and data are analyzed for differences by race and gender. Preliminary research : The first step in examining the importance of trees and forests to people in the U .S. was exploratory and involved conducting 26 focus groups with different demographic populations (USDA Forest Service, 2011). The focus group s uncovered a wide range of values grounding the importance of forests as well as descriptions of changes in interactions with trees and forests over time and negative f eelings and concerns people had about tre es and forests. The method targete d and successfully engaged diverse populations, including those who are less likely to respond to traditional mail surveys such as African Americans, Asian Americans, Hispanic Americans/Latinos, and Native Americans (Feldman et al., 2008; McLean & Campbell 2003; Tronstad et al., 2004; Yancey et al., 2006). However, focus groups are time consuming, and do not provide the survey - type data favored, for better or worse, by policy - and decision - makers. F ocus group data illustrated the majority of the ways that trees and forests are important similar to those identified in the literature review. T he focus group methodology greatly increased the diversity of research participants and identified important differences and nuances. Focu s group participants in multiple regions of the U . S . , representing many different ethnic/cultural backgrounds, discussed the importance of trees and forests in terms of environmental services, cultural values, recreation, forest products (timber and non - ti mber), education, economic values, and in ways related to sense of place, spirituality, and health and well - being. The focus groups provided realistic data and captured distinctions based on race, location (urban vs. rura l), age, and more . Some differences uncovered by the focus groups 22 include : feelings of exclusion and fear associated with forests among African Americans ; rural respondents exhibiting greater concern with forest degradation, policy and management issu es while urban respondents were more lik ely to show concern for damage to their home and the quality of local parks and street trees ; and, younger respondents regularly interacted with forests w hile older respondents expressed greater appreciation for aesthetics , such as viewing trees from their windows. In addition, responses suggested that these values are not static and through the focus group discussions it was possible to explore how perceptions of trees and forests have changed over time for these diverse groups of people (USDA Forest Serv ice, 2011) . Conceptual Framework: The conceptual framework (Figure 1) for this study is focused on the sampling, methods, and data collection approach to examine the importance of trees and forests to people. Inclusivity in sampling is critical. Beginnin g on the left, the first column assumes that people from all demographic categories in the study area are included in the sampling frame. Moving to the second column there is a choice for sampling and data collection methods including traditional survey sa mpling, purposive sampling and face - to - face methods, and mixed or multi - methods of sampling and data collection. However, the choice of sampling and data collection method affects the level of inclusivity, which affects an overall contribution to the goals of SFM. The assumption is that mixing sampling and data collection modes will result in a more inclusive and representative sample which leads to a more diverse and illustrative understanding of the importance of forests to people. 23 Figure 1 : Conceptual framework ( The dotted arrow line signifies that this method has other caveats, such as financial and time constraints, which make it less efficient for measuring the importance of trees and forests at a national level. ) Citizen demographics: Race Income Age Gender Sampling/Data Collection Method: Mixed/multi - methods sampling/data collection Purposive sampling/ face - to - face methods Traditional mail survey sampling Inclusivity, Power and Voices: High | | | | | | | | | | | | | Low Contribution to objectives of SFM: Strong | | | | | | | | | | | | | Weak 24 Research questions : (Table 2 2 ): 1. Does the pairing of different survey sampling (including probability and purposive) and data collection strategies result in a representative sample ? a. What are the demographic characteristics of the respondents for probability sampling paired with a mail survey data collection? b. What are the demograph ic characteristics of the respondents for stratified or disproportionate sampling paired with a mail survey data collection? c. What are the demographic characteristic s of the respondents for purposive sampling paired with an in - person survey data collection? 2. How are trees and forests perceived to be important by different racial and ethnic groups? a. Are there differences based on racial and ethnic group? b. Are there diffe rences based on gender both within and between races? c. Are there differences based on the different sampling and data collection methods? 3. What study design factors influenced the research participants to participate? Why? Table 2: Research questions and corresponding variables of interest Research Question Variable 1 - a Demographic characteristics ( probability sampling, mail survey ) 1 - b Demographic characteristics ( stratified or disproportionate sampling, mail survey ) 1 - c Demographic characteristics ( purposive sampling, in - person survey ) 2 - a Differences in perceptions (race) 2 - b Differences in perceptions (gender) 2 - c Differences (based on sampling/data collection) 3 Participation influence 2 Measured variables for each research question 25 C HAPTER 3: METHODS Introduction: Following a philosophy inspired by feminist epistemology and pragmatism, this study used a flexible methodology in an attempt to best address the research questions and problem. Sustainable forest management involves understanding how trees and forests are important to people and calls for meaningful dialogue with the diversity of people at all levels of society. In reality, that dialogue currently occurs with a rather homogenous group of people, namely, those that are educated, older, male, and white (McDo nough et al., 2003; Smith & McDonough, 2001). A survey instrument developed from exploratory data on the importance of forests was tested in a m idwestern city, administered in three different sampling and data collection combinations. Descriptive statistics, Chi - square, Independent Samples t - t est, Analysis of Variance (ANOVA), and a path analytic approach utilizing Multiple Regression models were used to a nalyze differences in perceptions on the importance of trees and forests. A f ocus group and semi - structured interviews were used to determine study design factors that influence d survey participation. Details of each step of this research can be found in s ubsequent sections of this chapter. Feminist epistemology serves as the foundation for the research presented here . One tenet of feminist epistemology is the belief that researchers all carry some or many forms of bias, and that there is truly no such thing as value - free, or objective research. Science has a strong historic 26 white, androcentric bias, and has undervalue d the voices, perspectives, and experiences of women, people of color, and other traditionally marginalized societal groups. Science should be emancipatory and more focused on developing an inclusive scientific process, and valuing the many ways, qualitati ve and quantitative, that data can be collected. e presented research was conducted by a white, educated, privileged woman with strong personal connections and emotional attachment to tree s, forests, and other natural resources. The researcher also serves as must be considered, along with ethics, when interacting with research participant s, in the handling of the data they produce, and in the manner in which the research is represented. One way that feminist researchers attempt to monitor their biases is through critical reflection throughout each step of the research process, which is att empted here. This research assumes that a traditional survey research methodology will not achieve a representative sample. A representative sample may be attained by using a mix of sampling and data collection methods. Description of study area: Due to its diverse population and proximity to the research institution, the city of Lansing, Michigan served as the study area (Figure 2 ). With a population of approximately 115,000 people, the city of Lansing hosts a diversity of families and individuals from different racial/ethnic backgrounds (Table 3), income levels, and education levels, and is also 51.6% female (U.S. Census, 2013). Lansing provided a research site with opportunities to include racial and ethnic minorities as it comparatively has a la rger proportion of non - whites than Ingham County (where it is located), Michigan, and the United States. The city focus also allowed for 27 sampling down to the census block level rather than the tract, thus allowing for more concentrated sampling frames. The population of Lansing does not represent the U.S. in its totality, but it provided a starting point to test the effectiveness of sampling techniques and a survey measuring the importance of trees and forests to people across diverse populations. Lansing a lso has a variety of municipal parks and tree lined streets providing natural spaces for its residents. 28 Figure 2 : Map of study site: Lansing, MI 29 Table 3: Racial composition of Lansing, MI, MI and U.S , 2013 Race Lansing, MI Percentage* Michigan Percentage** United States percentage*** White/Caucasian 61.2% 78.9% 77.7% Black/African American 23.7% 14.2% 13.2% American Indian 0.8% 0.6% 1.2% Asian 3.7% 2.3% 5.3% Hawaiian and other Pacific Islander Z Z 0.2% Other 4.4% 1.9% Z Two or more races 6.2% 2.1% 2.4% Total 100% 100% ~100% *Hispanic/Latino origin: 12.5%, White persons non - Hispanic: 55.5% (2013 U.S. Census Bureau) ** Hispanic/Latino origin: 4.4%, White persons non - Hispanic: 76.6% (2013 U.S. Census Bureau) *** Hispanic/Latino origin: 17.1%, White persons non - Hispanic: 62.6% (2013 U.S. Census Bureau) Z= greater than zero but less than unit of measurement Sampling and Participant Selection: The two goal s of this research w ere to collect data on and analyze differences in perceptions of the importance of trees and forests to people, and to provide data that are representative of the study population. This required careful sampling techniques and a consideration of the appropriateness of sampli ng approaches for the various demographic groups. One size generally does not fit all. A representative sample is defined as one that accurately reflects the population that is being sampled, or closely approximates the same collective characteristics of t he sampled population (Babbie, 2005). The greater the representativeness of a sample, the more likely that the study has external population validity (Tashakkori & Teddlie, 1998). Many social scientists will argue that a population sample can only be consi dered representative if probability sampling 30 is used, giving each person in the population an equal chance of being selected. Probability sampling is used to avoid bias, and can enhance representativeness. But samples rarely, if ever, perfectly represent t he populations from which they are selected (Babbie, 2005). Survey samples often are especially lacking in their representation of non - white participants. Representativeness can also be influenced by the sample size. Sampling error can be reduced by incr easing the sample size. Sample size does not guarantee representativeness, especially when considering inclusiv ity . Error may be reduced by decreasing the homogeneity of population elements, or increasing diversity. Subpopulations may also be sampled dispr oportionately to ensure sufficient representation for analysis. This involves giving a subpopulation a disproportionately better chance of selection than those located in categories and then allows for samples to be analyzed separately or comparatively (Ba bbie, 2005). Considering the above factors, probability sampling was combined with disproportionate sampling to solicit participation from racial/ethnic minorities and women. Additionally, purposive sampling was used to increase the sample size, assuming lower response from subpopulations. The pairing of different sampling and data collection strategies was used strategically with the goal to develop a representative sample from which to study the importance of trees and forests to people. A sample that co mbines elements from both purposive and probability sampling is likely to produce population external validity, or transferability (Tashakkori & Teddlie, 1998; Teddlie & Yu, 2007). If researchers can collect quantitative data from audiences who generally d o not respond to surveys, the research has the ability to make visible the traditionally invisible on a large scale (DeVault, 1999; Jayarantne & Stewart, 1991). Census blocks served as the initial sampling frame. There are approximately 1400 census blocks in Lansing. Specific criteria were then used to determine which census blocks were to be 31 included in the actual sampling frame . F or example, blocks had to contain a certain number of addresses for inclusion. T o address research questions 1a and 1b (Tabl e 2), using the census blocks, address - based sampling occurred in one of two ways utilizing a Delivery Sequence File (DSF). The DSF is an electronic file which contains all deliverable addresses serviced by the USPS, though the addresses are not always ass ociated with names (Dillman et al., 2009a). Recent studies have found address - based sampling using DSF to produce higher response rates, though an overrepresentation of non - Hispanic whites is still documented (Dillman et al., 2009a; Link, Battaglia, Franke l, Osborn, & Mokdad, 2008). A DSF that contained names and phone numbers i f available was purchased. A bout 44% of the addresses also had phone numbers, and approximately 90% had names attached. The file was purchased from Marketing Systems Group (MSG). Se veral studies found MSG to have reliable address lists (DiSogra, Dennis, & Fahimi, 2010; Iannacchione, 2011). The address lists provided by MSG are updated monthly for accuracy. Sample Group 1: To address research question 1a 3 (Table 2), 100 census blocks were randomly chosen from a sample stratified by population density. Using Statistical Package for the Social Sciences (SPSS), the 1400 census blocks were first ranked by population size. Those containing less than 15 persons were eliminated from the sample to ensure that there were at least five houses in the census blocks. Then, using Microsoft Excel, the remaining census blocks were ranked based on cumulative population in order to ensure that census blocks with varying popu lation sizes had the 3 What are the demographic characteristics of the respondents for probability sampling paired with a mail survey data collection? 32 same chance at being chosen for the sample. Using www.randomizer.org , 100 blocks were then drawn (Figure 2) . From this sample, five addresses were randomly chosen from each of the 100 census blocks (Table 4). Sample Group 2: To incre ase response rate, some researchers have oversampled difficult to reach populations (i.e., racial minorities, when conducting survey research) (Morrissey & Manning, 2000; Parker & McDonough, 1999). The second sampling method utilized disproportionate sampl ing following the same format as the first group mentioned above. To address research question 1b 4 - included in the sampling frame. Those blocks containing less than 15 persons w ere eliminated from the sample to ensure that there were at least five houses in the census blocks. Then, using Microsoft Excel, the remaining census blocks were ranked based on cumulative population in order to ensure that census blocks with varying popul ations had the same chance at being chosen for the sample. Using www.randomizer.org , 100 additional blocks were drawn (Figure 2) . From this sample, five addresses were randomly chosen from each of the 100 census blocks (Table 4). Sample Group 3: Continui ng to use the same survey sampling methods because of comfort and familiarity can be dangerous as it likely poses a limited view of the studied phenomenon (Knap & Propst, 2001). It has been documented across many disciplines that racial and ethnic minoriti es and women respond better with more personal data collection methods, such as face - to - face 4 What are the demographic characteristics of the respondents for stratified or disproportionate s ampling paired with a mail survey data collection? 33 interviews or focus groups. Despite this, researchers continue to try and collect information from underrepresented audiences using telephone and mail surveys desp ite evidence that response rates remain low. Even in instances where response rates are higher, there is sure to be a response bias, whether based on misrepresentation of the population, or the difference between those who choose to respond and those who choose to not respo nd (Stoop, 2004). Knowing that underrepresented groups feel more comfortable with face - to - face data collection methods, the third approach to sampling (addressing research question 1c 5 ) was more purposive in nature, and one that decreased the distance bet ween the researcher and the researched. It was hypothesized that the address - based sampling would yield low response rates for underrepresented groups and a non - representative sample. In areas near the 200 chosen census blocks, the researcher made contact with different interest groups (i.e., community groups, clubs, churches, neighborhood organizations) and brought the survey instrument directly to the people (approximately 50 surveys per interest group) (Table 4). Table 4: Sampling methods Sample Number of Census Blocks Attempted Number of Surveys Random, Address - based Disproportionate, Address - based Purposive, In - person 100 100 n/a 500 500 n/a Total 200 1000+ 5 What are the demographic characteristics of the respondents for purposive sampling paired with an in - person survey data collection? 34 It may seem fitting to have explored a fourth sampling method using the face - to - face data collection with probability sampling in the spirit of comparing the different methods. D ue to the salience of the research, the time and labor commitment, and the ten dency for probability sampling in a public place to not be representative, it was unrealistic to expect any significant contribution from this additional method, and thus it was not explored. Even with a high quality address list, several addresses were un usable (either returned as addresses in the random sample, and 449 usable addresses in the disproportionate sample. Of the attempted and delivered surveys, a total of 287 were completed: 124 from the random sample, 103 from the disproportionate sample, and 60 delivered and collected in person in the purposive sample (Table 5). Table 5: Attempted and obtained surveys Test m ethod Attempted number of s urveys Us able a ddresses Obtained number of s urveys Address - based random sampling/ Mail s urvey 500 458 124 (27% response) Address - based disproportionate sampling/ Mail s urvey 500 448 103 (23% response) Purposive sampling/ In - person survey n/a n/a 60 Total 1000+ 932 n=287 Metrics: A comprehensive list of values associated with the importance of forests was developed from the focus group data (Table 6 ) and the reviewed literature . Grouped by relevance to the value categories, o ver 150 s tatements were extracted from the focus group data and used to form 35 survey questions . Value statements were reviewed carefully to reduce the quantity resulting in 70 s tatements that were simple and easy to understand , with three or more statements clustered together to m easure one factor . This resulted in the design of a survey instrument (Appendix I) to test if the statements measure the forest values that are important to people (validity test). The survey instrument also asked questions about participant demographic ch aracteristics in order to better understand the sample. Table 6 : Value categories developed from f ocus group and literature review Value category Environmental & biological values Social & cultural values Forest products Recreation benefits Sense of place Health and well - being Aesthetic values Spiritual & religious values Economic values Educational values Concer n s for trees & forests Interactions with trees & forests Negative feelings associated with trees & forests Likert scales assessed the degree of agreement or disagreement with the value statements in reference to how trees and forests are important (Babbie, 2005). Likert scales provide the ability to ask about several diverse items (Alreck & Settle, 2004; Babbie, 2005) which aligns well with this study because of the range of ways that trees and forests are important to people. Experts recommend the use of both positive and negative statements in Likert scales which was exercised in this s t udy (Dillman et al., 2009a). 36 Once the research instrument was developed, the value statements were pre - tested across different groups before data collection. Pre - test groups consisted of college students and Michigan citizens una ffili at ed with th e research . Twenty people of varying ages, racial/ethnic identity, and genders participated in the pre - test (Table 7) . Survey items were revised based on feedback from pre - test participants. Table 7 : Description of pre - test sample (n = 20) Gen der Race/Ethnicity Age Undergraduate students n = 4 Female = 2 Male = 2 White = 3 Other = 1 18 - 21 Graduate students n = 3 Female = 2 Male = 1 White = 1 African American = 2 25 - 38 Michigan residents n = 13 Female = 7 Male = 6 White = 6 African American = 4 Other = 3 27 - 71 Survey Data Collection: The survey data collection further addressed the first 6 and second 7 research questions. Data were collected using a mixed - modal se lf - administered questionnaire. F ollowing the - notice letter was first sent to all addresses in the sample (Appendix II). Multiple motivational features were utilized in an attempt to increase participation from traditionally under - represented groups. The survey and informational materials targeted the specific receiving community. Inner city residents received 6 Does the pairing of different survey sampling (including probability and purposive) and data collection strategies result in a representative sample? 7 How are trees and forests perceived to be important by different racial and ethnic groups? a) Are there differences based on racial and ethnic group? b) Are there difference s based on gender both within and between races? c) Are there differences based on the different sampling and data collection methods? 37 appropriate motivational/advertising features (for example, focused on parks, street trees, and clean air and water). Dillman and others (2009a ) outlined ways to establish trust, increase benefits of participation and decrease costs of participation. Th ey recommend making the task appear important and the survey interesting, providing social validation, supporting group values, making it convenie nt to respond, using neutral language (avoiding subordinate language), minimizing solicitation of personal information, and making the questionnaire short and simple to complete. Best practices for questioning and administering the survey which consider ed both clarity and readability were used. Some best practices for question development include: making sure the question applies to the respondent, asking one question at a time, keeping the language simple and familiar while using specific and concrete term s, using as few words as possible, and ensuring the question specifies the response task (Dillman et al., 2009a). This study followed Dillman and others advice to the maximum extent possible given the available resources. All participants were entered into a lottery to win one of four $25 VISA gift cards. Focus group and personal interviews: The third research question 8 is evaluative and required feedback from willing research participants. As part of the survey instrument, the researcher asked participants if they would be willing to partake in a focus group or short telephone or in - person interview regarding the data collection mode . Due to low response from participants, one focus group, personal interviews, and phone interviews were arranged. After consenting to participate (Appendix III), participants were asked why they decided to take part in the research and about their percept ions of the 8 What study design factors influenced the research participants to participate? Why? 38 particular sampling and data collection mode through which they responded (Table 8 , Appendix IV). A total of 11 people participated in the evaluative interview portion of this study. Three participated in the focus group, three were interviewed in person, and five were interviewed by telephone. Four of the participants were male, seven were female . Two of the participants were B lack/African American and the re st were W hite/Caucasian (Table 9 ). All participants were entered into an additional lot tery to win one $25 VISA gift card. Table 8 : Evaluative interview questions Did you receive the survey on the importance of trees and forests in the mail or in - person? What motivated you to complete the survey on the importance of trees and forests ? Were there any aspects of the survey that were especially interesting or relevant to you? Were there any aspects that were especially disinteresting or irrelevant? Have you completed other surveys in the past? (In the mail or in - person?) How do you decid e whether or not you will complete a survey? How can survey research on topics such as the imp ortance of trees and forests be improved? Are there particular methods that you think are more user - friendly than others? uture research on the importance of trees and forests? Table 9 : Characteristics of focus group/i nterview participants Black/African American White/Caucasian Female 2 5 Male 0 4 39 Data Analysis: One intention of this study was to collect data from a diverse, racially - and gender - representative sample. The second intention was to examine differences in perceptions of the importance of trees and forests. Data analysis sought to determine value differences and similarities defining the importance of fore sts to people across the sample. The analysis explored which sampling and data collection strategies were effective across sociodemographic variables by both examining the data, and then conducting follow - up focus groups and interviews with willing partici pants. Analysis further included comparing the different sampling modes: the address - based sampling with oversampling, and the face - to - face sampling, to see if one wa s more effective at yielding a representative sample than the others, or if they work ed i n tandem. Differences in the substantive data on the importance of trees and forests based on the different sampling and data collection methods, and if the data differ from that found in the preceding focus group research were also examined. The evaluati ve focus group and interviews were analyzed qualitatively by the researcher . Notes and transcripts were read through carefully first to explore emerging ideas. Then the ideas were documented and separated into categorical themes. The notes and transcripts were read through a second time and assigned themes as appropriate. Addressing survey e rror s: The Tailored Design Method focuses on developing survey procedures that work in tandem to motivate different types of people to respond to the survey. This method attempts to minimize multiple sources of survey error including coverage, sampling, measurement, and nonresponse, by customizing survey procedures for particular circumstances such as the topic, sponsor, or type of respondent (Dillman et al., 2009a; de Le euw et al., 2008). Accordingly, some 40 people received the survey by USPS mail while othe rs received it in person with an introduction from the researcher. As d iscuss ed in the subsequent chapters, each sampling group contributed differently to survey errors. L ow sample representativeness found in sample group 1 increased sampling error and coverage error . Sampling error is understood as the degree to how a sample is limited to describe a population because only some (not all) elements i n the population are sampled while c overage error is the discrepancy between the target population and the group of individuals in cluded in the sampling frame (Vaske, 2008; Dillman et al., 2009). S ample group 2 was more representa tive of Lansing residents , decreasing coverage and sampling errors. However, sample group 2 had less demographic and geographic exposure which tends to increas e sampling error and coverage error . A purposive sample paired with an in - person survey, as utilized for s ample group 3 , increased the likelihoo d for more sampling and coverage errors as the element of ra ndomness is missing . I n this context, the presence of the researcher may have influenced participant response resulting in increasing the likelihood of a m easurement error. Measurement error is ability to be compared to answers provided by other participants (Vaske, 2008 ; Dillman et al., 2009 ) . C ollectively the combined sampling methods and survey modes provide d a more diverse sample and improved coverage and sampling error with a focus on inclusivity . The goal was to reduce coverage errors by m ixing modes (Dillman et al., 2009). In an attempt to address non - response bias/error, or for differences between survey participants and those in the sample that did not respond (Vaske, 2208 ; Dillman et al., 2009 ) , two samples of 40 potential participants (no n - responders) were chosen from s ample group 1 and s ample g roup 2 to be contacted and administered a shortened version of the survey instrument 41 (Appendix V). Of the 80 potential participants, contact was established with 28 people. Eleven participants agreed to complete the non - response survey. Non - response data w ere collected by telephone. Of the 11 participants, all were White/Caucasian with the exception of one person who identified as Black/African Amer ican , and 7 participants were female (Table 10 ). Table 10 : Characteristics of nonresponse sample Black/African American White/Caucasian Female 1 6 Male 0 4 The data from the nonresponse survey, though collected from a much smaller sample, did not differ from results collected in the original survey data collection (see Chapter 4) , suggesting low nonresponse error (Vaske, 2008 ; Dillman et al., 2009 ) . Partici pants expressed a large items) with all scales except that measuring spiritual importance in which participants expressed 46 % agreement (Table 1 1 ). 42 Table 1 1 : N onresponse percent agreement with importance scales Importance scale Percent Agreement Environment/Ecological function 93.9% Economic 72.7% Culture 81.8% Education 90.9% Heath & well - being 90.9% Aesthetics 90.9% Products 90.9% Sense of place 72.7% Spiritual 45.5% Recreation 100.0% Item - level 43 CHAPTER 4: RESULTS AND DISCUSSION C hapter 4 summarizes and discusses results from the survey questionnaire, focus groups, and interviews. Sections are organized by results for each research question (RQ). The first section describes the demographic characteri stics of survey respondents, presents st atistically significant differences by sample group , and an explanation of errors . The second section describes the substantive survey data and differences based on race/ethnicity, gender, and sample group. The third section encapsulates the qualitative da ta collected from focus groups and interviews evaluating the survey methodology. A brief summary of the results concludes the chapter. All statistical analyses were conducted using the software package IBM ® SPSS ® Statistics, Version 21. Demographic charact eristics of respondents : A total of 285 Lansing residents completed the survey questionnaire. The sample was 6 1 % female and 39% male. More women than men participated in each sample group. For simplification in the statistical data analysis, race was cate gorized as White/Caucasian, Black/ African American, or Other. The sample consisted of 70% participants iden tifying as White/Caucasian, 18 % identifying as Black/African American, and 12% identifying as a race other than white or black. Sample groups varied in their racial representativeness with groups 2 and 3 having more non - white participants 9 . Participants were from varying educational 9 For review, sample group 1 was ch osen randomly; sample group 2 was chosen disproportionately random (census blocks with greater than 50% non - white racial minorities); and sample group 3 was purposively chosen (in - person). 44 backgrounds, though most (8 2 degree or higher. Participa nts reported a variety of income levels with 38 % earning up to $25,000, 30 % earning $25,000 - 50,000, and 32% earning more than $50,000. All adult age levels were re presented, though 52 % of participants were 55 years and older. The largest percentage of participants (45 %) were employed full - time, 10% were employed part - time, 10% were unemployed, and 35 % were retired (Table 12 ) RQ 1 - A: What are the demographic characteristics of the respondents for probability sampling paired with a mail survey data colle ction (sample group 1)? Sample group 1 (randomly sampled) was 5 9 .5% female and 41.5% male for a total of 123 participants . The majority of the group (72 %) was White, 13% was Black, and 15% identified as another race. All education l evels were represented, with 21 % having high school or less, 40 % having some college, 22% being a college graduate, and 1 8 % holding a graduate or professional degree. The income le vels for sample group 1 were 28 % making up to $25,000, 3 3 % making $25,000 to $50,000, and 40 % makin g over $50,000. All age groups were represented, with 3 % being 18 - 24 years, 12 % being 25 - 34 years, 17% being 35 - 44 years, 16% being 45 - 54 years, 29 % being 55 - 64 years, and 2 4 % being 65 and older. The largest percentage of participants (4 9 %) were employed f ull time, 8% employed part time, 7% were unemployed, and 36% were retired (Table 12 ). RQ 1 - B: What are the demographic characteristics of the respondents for stratified or disproportionate sampling paired with a mail survey data collection (sample group 2) ? Sample group 2 was 6 8% female and 32% male for a total of 102 participant s. The majority of the group (64 %) was White, 21 % was Black, and 1 7% identified as another race. All 45 education le vels were represented, with 12 % having high school or less, 4 5% havi ng some college, 2 5% b eing a college graduate, and 18 % holding a graduate or professional degree. The income levels for sample group 2 consisted of 3 4 % making up to $25,000, 3 5 % making $25,000 to $50,000, and 3 2 % making over $50,000. All age groups were re presented, with 1% being 18 - 24 years, 12 % being 25 - 34 years, 17% being 35 - 44 years, 1 5 % being 45 - 54 years, 22 % being 55 - 64 years, and 32 % being 65 and older. The largest group of participants (4 6%) were employed full time, 14 % employed part time, 4% were unemployed, and 36% were retired (Table 12 ). RQ 1 - C: What are the demographic characteristics of the respondents for purposive sampling paired with an in - person survey data collection (sample group 3)? Sample group 3 was 60 % female and 36 % male for a total of 60 participants . The majority of the group (63 %) was White, 20% was Black, and 17% identified as another race. All education le vels were represented, with 22 % having high school or les s, 35% having some college, 26 % being a college graduate, and 1 6 % ho lding a graduate or professional degree. The income levels for sample group 3 were 6 7% making up to $25,000, 1 5 % mak ing $25,000 to $50,000, and 19 % making over $50,000. All age groups were represented, with 9% being 18 - 24 years, 25 % being 25 - 34 years, 9 % b eing 35 - 44 years, 14% being 45 - 54 years, 28% being 55 - 64 years, and 16 % being 65 and older. The largest group of participants (3 5%) were employed full time, 9 % employed part time, 2 6 % were unemployed, and 3 1 % were retired (Table 12 ). 46 Table 12 : Demograph ic characteristics of survey participants Combined samples Sample g roup 1 Sample g roup 2 Sample g roup 3 Gender Female 60.9% 58.5% 67.8% 60.4% Male 39.1% 41.5% 32.2% 36.2% n = 285 n = 123 n = 102 n = 60 Race White 70.1% 71.5% 63.7% 63.3% Black 17.5% 13.0% 21.3% 20.0% Other 12.4% 15.4% 16.7% 16.7% Education High school or less 18.3% 20.7% 12.2% 22.4% Some college 40.6% 39.7% 44.9% 35.0% College graduate 17.3% 22.3% 24.5% 25.9% Grad/professional degree 17.3% 17.6% 18.4% 15.5% Income Up to $25,000 38.30% 28.1% 33.7% 66.7% $25,000 - $50,000 29.50% 32.5% 34.8% 14.8% Over $50,000 32.20% 39.5% 31.5% 18.5% Age 18 - 24 years 3.2% 2.5% 1.0% 8.8% 25 - 34 years 14.3% 11.5% 12.1% 24.6% 35 - 44 years 15.4% 17.2% 17.2% 8.8% 45 - 54 years 15.4% 16.4% 15.5% 14.0% 55 - 64 years 26.3% 28.7% 22.2% 28.1% 65 and older 25.4% 23.8% 32.3% 15.8% Employment Part time 10.4% 8.4% 13.8% 9.1% Full time 44.6% 48.7% 45.7% 34.5% Unemployed 9.7% 6.7% 4.3% 25.5% Retired 35.2% 36.1% 36.2% 30.9% Because individual sample groups did have varying demographic representation, Chi - square analyses from cross - tabulations were computed for each demographic category. Statistically significant differences between sample groups were found for the demographic variables e mplo yment status and income (Table 13 ). Sample group 3 had more participants with lower incomes and more participants that were underemployed or unemployed. 47 Table 13 : Demographic differences between sample groups (Chi - square analysis) Demographic variable df Value p Gender 2 0.630 0.730 Race/ethnicity 4 2.619 0.623 Education level 6 4.046 0.670 Employment status 6 21.698 0.001 * Income 4 25.196 0.000 * * p < .0 1 Discussion of RQ 1 results and survey errors: One objective of this study was to explore if the pairing of different survey sampling and data collection strategies would result in a racially - and gender - representative sample. A comparison of the demographic characteristics of each sample group provide d valuable details about the usefulness of mixing sampling and data collection methods. Sample group 1, which received the survey via traditional methods (probability sampling paired with a mailed survey) provided the least racial representativeness of the three sample groups, as hypothesized. Both sample groups 2 (disproportionately random) and 3 (purposive) provided greater racial and gender representativeness than sample 1. None of the sample groups provided exact representati on of the population (this w ould be very difficult to achieve), but sample groups 2 and 3 increased African American responsiveness by 7 8 .1 percentage points (21% and 20% h). For context, Lans ing is 24 O ther . of mixing survey sampling and data collection methods to elicit a racially representative sample ( Table 12 ) as recommended by Di inist and pragmatist mixed methods researchers (de Leeuw et al., 2008; Hodgkin, 2008; Jayarantne & Steward, 1991; Nightingale, 2003; Reinharz, 1992; Tashakkori & Teddlie, 1998). 48 Women were overrepresented in each sample group with the most representation in sample group 2 ( Table 12 ) . While the overrepresentation of women may be common in other types of survey research (Curtin, Presser, & Singer, 2000; Moore & Tarnai, 2002; Singer, van Hoewyk, & Maher, 2000), this is not always the case with forestry and natural resources - related research where participants tend to be educated, older, white males (McDonough, Russell, Burban, & Nancarrow, L., 2003; Smith & McDonough, 2001) . Since women are genera lly under - represented in forestry research and decision - making processes, it is important to note that this methodology was successful at recruiting participation by women . The methodology was Other demographic characteristics of survey participants show representation of a variety of educational backgrounds, income levels, age (though most participants were 55 and older), and employment levels. There were significant differences between sample groups based on income level and employment status. This is likely due to the fact that many participants (but not all) in sample group 3 were recruited at community events 10 geared towards lower income persons. Differences in age representation by sample groups, though not statistically s ignificant, are worth noting. While sample groups 1 and 2 had the highest percentage of persons aged 55 and older, sample group 3 had the highest percentage of persons aged 18 - 34 years. This may be due to the community events geared towards low income pe rsons, or, it may signify that younger persons are more likely to complete a survey given in - person instead of received via postal mail (Table 12 ). 10 food giveaways, and neighborhood center events. 49 For the purpose of RQ 1, to investigate the potential of achieving a representative sample by mixing survey sampling and data collection modes, this explorative study provided useful results. Borrowing from feminist and pragmatist mixed methods (Hodgkin, 2008; Jayarantne & Stewart, 1991; Nightingale, 2003; Reinharz, 1992; Tashakkori & Teddle, 1998), c ombining a sample with random, disproportionate, and purposive elements resulted in a more diverse sample of participants. N one of the utilized sampling methods were extremely useful for achieving representativeness on their own which reflects earlier studies calling for the use of mixed methods to appropriately answer research questions (Tashakkori & Teddle, 1998) . Survey e rrors Sample group 1 was the least representative of Blacks/African Americans resulting in increased sampling error and coverage error (Vaske, 2008 ; Dillman et al., 2009 ) . Sample group 2 was more representative of Blacks/African Americans, but with less demographic coverage due to the disproportionate sampling in census blocks with 50% or greater nonwhite inhabitants , a lso increasing sampling err or and coverage error (Vaske, 2008 ; Dillman et al., 2009 ) . Sample group 3 was more representative of Blacks/African Americans, residents from lower income brackets, and unemployed residents but due to the purposive sampling and in - person data collection th e element of randomness is missing . For sample group 3, the likelihood of s ampling, c overage, and measurement error increase d (Vaske, 2008 ; Dillman et al., 2009 ) , and it is quite time consuming to obtain a large sample for statistical analysis. C ollectively the combined sampling methods and survey modes provided a more diverse sample for examining the importance of trees and forests and improved coverage error with a focus on inclusivity . Mixing modes was useful for reducing coverage error while s ingle modes did not sufficiently cover the studied population (Dillman et al. , 2009). 50 Survey data analysis : Data organization: The Likert - type items in the survey questionnaire measured responses on a 5 - point scale 3 ). T he decision had to be made regarding how to analyze the data categories N and DK combined. A ll analyses were repeated with the N category left as is, and the DK category coded as missing data. The two groups of analyses were then compared. There were statistically significant differences between analyses conducted as N and DK combined, and DK coded as missing data. For consistency, all analyses discussed in this dissertation are based on the data with the DK category coded as missing. Because there were differences in the analyses, it is not accurate to say that N and DK are similar enough to consider as the same response. sed the importance of trees and forests for cleaning and filtering water (19% DK response), the importance of trees and forests for providing fuel (12 % DK response), and a question addressing the mana gement of trees and forests (19 % DK response). These thr ee questions had a significant phrasing was confusing for people, or, that participants genuinely did not know the answer. If used in future surveys, t hese quest ions need to be explored further for clarity and revised or reworded . 51 Figure 3 : Example question from survey (for full survey see Appendix I) The fifth question is asking about the ways in which trees and forests influence human health and a sense of we ll - being. Please describe the extent to which you agree or disagree with the following statements. Question 5 1 2 3 4 5 6 Strongly Disagree Neutral Strongly Agree Know 1. Trees and forests are relaxing 2. Trees and forests provide a place for privacy/solitude 3. Trees and forests provide a place for 3. Trees and forests provide a connection to a Higher Power Assessing reliability: The survey questionnaire presented Likert - type items organized by category of importance including: environmental/ecological function, economic, culture, education, health and well - being, aesthetics, products, sense of place, and spiritual importance. One item asked about the recreational importance of trees and forests. Additionally, categories related to interactions people had with trees and forests, and concerns and negative feelings people had for trees and forests were included in the survey. Scales were created by computing the mean responses for each category of questions. coefficient was computed for each scale of importance. The closer the alpha coefficient is to 1.0 th e greater the internal consistency of the items in the scale. It is commonly accepted in the social sciences that an alpha coefficient greater than 0.70 is acceptable, greater than 0.80 is good, 52 and greater than 0.90 is excellent (George and Mallery, 2003) . For this study, all but two of the scales had good or excellent alpha coefficients for reliability. Two scales had questionable alpha coefficients: those measuring economic importance ( = 0.69) and negative feelings ( = 0.64) (Table 14 ). Discussion of scale reliability: Overall, the Likert - type items developed into scales were successful at measuring the importance of trees and forests. Two scales from the survey showed questionable reliability via those measuring economic importance , and those measuring negative feelings people had towards lpha can give an estimate of internal consistency or the inter - relatedness of items in a scale. Questionable reliability can signify that the questions comprising the scale may not have been measuring that which they were intended to measure (increasing the likelihood of measurement error) . Many participants seemed to be confused about the question measuring negative feelings towards trees and forests. In several inst ances, participants wrote in the comments section of the mail survey that s/he agreed that there are several negative phenomena that accompany trees and forests but that the positive aspects significantly outweighed the negatives. The evaluative focus grou p/interviews (reported below) also found similar results in regards to the negative feelings scale. Additionally, this particular scale was the only scale that was mostly negatively correlated with all other scales. If participants agreed with importance s cales, they tended to disagree with the negative feelings scale. This is quite interesting because most questions comprising the negative feelings scale were based on realistic scenarios, not questions based on opinion. Questions addressing economic import ance and negative feelings towards trees and forests need to be reconsidered and revised in any future studies with this instrument. 53 RQ 2: How are trees and forests perceived to be important by different racial/ethnic groups? Percent agreement with importance scales was reported as a measurement combining the . most scales of importance (Table 14) , and, all scales but one were positively correlated with on e another (Table 15 ). Positive correlations were statistically significant for all scales except that measuring the importance of forest products which was only statistically significantly correlated with environmental/ecological function, economics, and edu cation. The scale measuring negative feelings towards trees and forests was, in most cases, weakly and negatively correlated with all other scales. The scales with the lowe st percent agreement were those meas uring spiritual importance (49 % agreement) and the negative feelings people may have ha d towards trees and forests (36 % agreement) (Table 14 ) . Discussion of overall importance: Percent agreement with importance scales along with the finding that scales were positively correlated with one another indica tes an overall positive attitude towards trees and forests. The scale measuring spiritual importance showed lower overall a greement than other scales (49 %), with responses ranging from strongly disagree to strongly agree. This s uggests that the spiritual i mportance of trees and forests is quite salient for some, and insignificant for others. In the focus group discussions, one participant explained that what another considered a spiritual experience when near trees or forests, he considered not related to s pirituality, but rather, that which enhances his sense of health and well - being. Both participants seemed to be discussing a 54 similar feeling, but with different labels. This is reminiscent of literature documenting the human need to affiliate with nature f or both spiritual and health reasons related to quality of life (Kamitsis & Francis, 2013; Kaplan, 1995; Kellert, 1985; Kellert, 1997; Kellert, 2012; Nakau et al., 2013; Quantz, 1897; Tyrv ä inen et al., 2014; and others) . Such overall agreement with importance scales underscores that trees and forests are valued for many reasons, as discussed in the literature. The data addressed benefits more frequently discussed among forestry professionals including environmental and ecological function, economic i mportance, recreational use, and the important products provided by trees and forests. The data also show that people interact with trees and forests in ways that provide a special meaning that is not easily defined. Trees and forests help contribute to on health and well - being including spiritual and religious feelings. Trees and forests are valued for their beauty. They also seem to play an important role in human cognitive development and education. A deeper understanding of the many ways tha t trees and forests are important to preferences. 55 Table 14 : Reliability and percent agreement for importance s cales Importance factor Size Min Max SD Mean Cronbach's Alpha Items per scale Percent Agreement* Environment/Ecological Function 285 2.7 5 0.4 4.7 0.91 12 89.0% Economic 283 2 5 0.6 4.5 0.69 3 85.1% Culture 283 2. 7 5 0.6 4.4 0.81 3 83.3% Education 285 2. 3 5 0.5 4.7 0.85 3 91.7% Health & well - being 285 2. 7 5 0.5 4.7 0.86 3 93.3% Aesthetics 285 2. 3 5 0.4 4.9 0.89 3 97.3% Products 282 1 5 0.8 4.4 0.83 3 80.9% Sense of place 281 1 5 0.8 4.2 0.87 5 72.3% Spiritual 276 1 5 1.1 3.6 0.84 3 48.5% Recreation 285 2 5 0.6 4.7 n/a 1 93.9% Interactions 284 1.8 5 0.7 4.1 0.84 10 73.3% Concerns 281 1 5 0.8 4.3 0.92 6 77.8% Negative feelings 281 1. 4 5 0.6 2.9 0.64 9 36.1% * Item - level 56 Table 15 : Pearson correlation matrix for importance scales/item EF EC CU ED HW AE PR SOP SP RE IN CO Environment/ Ecological Function (EF) 1.0 Economics (EC) .43** 1.0 Culture (CU) .51 ** .50 ** 1.0 Education (ED) .55 ** .38 ** .68 ** 1.0 Health & well - being (HW) .50 ** .39 ** .50 ** .49 ** 1.0 Aesthetics (AE) .42** .41 ** .46 ** .44 ** .63 ** 1.0 Products (PR) .12* .19 ** 0.08 .12 * 0.0 3 0.11 1.0 Sense of Place (SOP) .4 5** .42 ** .57 ** .54 ** .39 ** .45 ** 0.0 8 1.0 Spiritual (SP) .31** .2 5 ** .46 ** .41 ** .32 ** .27 ** .18 ** .60 ** 1.0 Recreation (RE) .47 ** .39 ** .53 ** .58 ** .65 ** .63 ** .1 1 .38 ** .33 ** 1.0 Interactions (IN) .49 ** .36 ** .56 ** .56 ** .53 ** .49 ** .13 * .68** .56 ** .51** 1.0 Concerns (CO) .54 ** .36 ** .46 ** .42 ** .51 ** .45 ** 0.06 .56 ** .46 ** .512** .58 ** 1.0 Negative feelings (NF) - .17 ** .01 - .11 - .0 5 - .09 - .01 .1 2 - .0 7 - .02 - .0 7 - .1 5 * - .10 * p < .05, ** p < .01 , Item - level 57 Differences by race/ethnicity : RQ 2 - A Are there differences based on racial/ethnic group? One - way Analysis of Variance (ANOVA) was conducted to test for differences in perceptions on the importance of trees and forests by race/ethnicity. All ANOVA results were confirmed with the non - parametric equivalent test, Kruskal Wallis, because of the use of Likert - type and non - normally distributed data. There was a statistically significant difference in perceptions on the environmental/ ecological importance of trees and forests for racial/ethnic groups as determined by the one - way ANOVA (F (2, 271) = 7 .608, p = .001) (Table 16 ). Post - hoc analyses (Tukey - Kramer test) showed statistically significant differences between Black/African Americans and both White/Caucasians and Others. African Americans had slightly lower mean scores (4.53) for the environment al/ecological function scale than both Whites (4.78) and Others (4.76) (Table 17 ) . A statistically significant difference between racial/ethnic groups was found for perceptions on the economic importance of trees as determined by the one - way ANOVA (F (2, 268) = 4.311, p = .014) (Table 16 ). Post - hoc analyses (Tukey - Kramer test) showed statistically significant differences between Black/African Americans and both White/Caucasians and Others. African Americans had slightly lower mean scores (4.25) for the ec onomic importance scale than both Whites (4.53) and Others (4.53) (Table 17 ) . Data from the health and well - being importance scale demonstrated a statistically significant difference between racial/ethnic groups determined by the one - way ANOVA (F (2, 269 ) = 5.285, p = .006) (Table 16 ). Post - hoc analyses (Tukey - Kramer test) showed statistically significant differences between Black/African Americans and both White/Caucasians and 58 Others. African Americans had slightly lower mean scores (4.52) for the health and well - being importance scale than both Whites (4.78) and Others (4.80) (Table 17 ) . A statistically significant difference between racial/ethnic groups was found for the sense of place importance scale determined by the one - way ANOVA (F (2, 266) = 8.2 82, p = .000) (Table 16 ). Post - hoc analyses (Tukey - Kramer test) showed statistically significant differences between Black/African Americans and both White/Caucasians and Others. African Americans had slightly lower mean scores (3.72) for the sense of plac e importance scale than both Whites (4.21) and Others (4.39) (Table 17 ) . Another statistically significant difference for racial/ethnic groups was found for perceptions on concerns for trees and forests as determined by the one - way ANOVA (F (2, 266) = 5.57 9, p = .004) (Table 16 ). Post - hoc analyses (Tukey - Kramer test) showed statistically significant differences between Black/African Americans and both White/Caucasians and Others. African Americans had slightly lower mean scores (3.94) for the concerns scale than both Whites (4.36) and Others (4.44) (Table 17 ) . It is noteworthy that African Americans did not produce lower mean scores for all importance scales (Table 17 ) . The differences in means scores were only statistically significant for the importance sc ales mentioned above. Even considering statistical significance, it is important to see that in most cases African American mean scores were still quite favorable, or (5)). 59 Table 16 : Differences in perceptions o f the importance of trees and forests by race/ethnicity (One - way ANOVA) Importance scale/item df SS MS F p Environment/Ecological Function 2 2.412 1.206 7.608 0.001 Economic 2 2.971 1.485 4.311 0.014 Culture 2 0.747 0.374 0.925 0.398 Education 2 0.495 0.247 0.87 0.42 Health & well - being 2 2.622 1.311 5.284 0.006 Aesthetics 2 0.504 0.252 1.655 0.193 Products 2 1.817 0.909 1.605 0.203 Sense of place 2 10.992 5.496 8.282 0.000 Spiritual 2 6.138 3.069 2.645 0.073 Recreation 2 0.644 0.322 0.880 0.416 Interactions 2 2.363 1.182 2.777 0.064 Concerns 2 6.949 3.475 5.579 0.004 Negative feelings 2 0.943 0.472 1.236 0.292 Item level question 60 Table 17 : Descriptive statistics for importance scales by race/ethnicity Importance scale/item Race/Ethnicity N Mean Std. Dev. Environment/Ecological Function Black/African American 48 4.53 0.505 White/Caucasian 192 4.78 0.368 All others 34 4.76 0.395 Economic Black/African American 46 4.25 0.696 White/Caucasian 191 4.53 0.562 All others 34 4.53 0.563 Culture Black/African American 46 4.40 0.694 White/Caucasian 190 4.41 0.625 All others 34 4.56 0.612 Education Black/African American 47 4.56 0.567 White/Caucasian 191 4.67 0.509 All others 34 4.66 0.617 Health & well - being Black/African American 47 4.53 0.601 White/Caucasian 191 4.78 0.488 All others 34 4.80 0.386 Aesthetics Black/African American 47 4.77 0.417 White/Caucasian 191 4.87 0.399 All others 34 4.91 0.288 Products Black/African American 47 4.59 0.56 White/Caucasian 189 4.38 0.782 All others 34 4.46 0.812 Sense of place Black/African American 46 3.72 0.934 White/Caucasian 191 4.21 0.782 All others 32 4.39 0.824 Spiritual Black/African American 46 3.68 1.26 White/Caucasian 186 3.48 1.05 All others 32 3.93 0.957 Recreation Black/African American 44 4.61 0.722 White/Caucasian 190 4.75 0.572 All others 34 4.74 0.618 Interactions Black/African American 46 3.92 0.81 White/Caucasian 192 4.15 0.613 All others 34 4.22 0.657 Concerns Black/African American 44 3.94 1.07 White/Caucasian 191 4.36 0.713 All others 34 4.44 0.771 Negative feelings Black/African American 45 3.04 0.771 White/Caucasian 190 2.93 0.571 All others 34 2.82 0.644 Item - level 61 Weighting is a statistical tool used to account for oversampling and undersampling bias in survey research (Vaske , 2008). Often researchers will use weighting as a substitution for sample representativeness, giving more weight to variables undersampled, and less weight to variables oversampled. Because sample group 1 (SG1) was random with low racial representation an d sample group 2 (SG2) was disproportionately random with better racial rep resentation, the data for SG1 were weighted 11 and the ANOVA test was repeated. Results for weighted SG1 were compared to results for unweighted SG2. This process explored if weightin g data really does result in achieving a higher degree of statistical representativeness. In other words, the process examined if weighting data compensates for the lack of racial representativeness in SG1 . A comparison of ANOVA results for two factors showed differences between the weight ed and unweighted data for sample groups 1 and 2 suggesting that weighting data was not successful in compensating for the lack of racial representativeness (Table 18). A comparison of ANOVA results for a n oth e r factor s howed that the statistically significant differences were found for both sample groups 1 and 2 indicating that weighting data may have compensated for a lack of racial representativeness (Table 18 ). These inconsistent r esults demonstr ate that, in this stud y, weighting data was only marginally successful in compensating for the lack of racial representativeness. 11 Formula for weighting: Weighted factor = percent in population/percent in sample group 62 Table 18: Comparing weighted and non - weighted data for racial differences Factor: Importance scale/item Weighted sample group 1 F Sample group 2 F Full s ample F Environment/Ecological Function 0.233 6.525** 7.608** Economic 3.659* 0.002 4.311* Culture 0.868 0.808 0.925 Education 0.103 0.331 0.87 Health & well - being 0.187 2.016 5.284** Aesthetics 1.438 0.603 1.655 Products 2.507 0.716 1.605 Sense of place 4.821** 3.652* 8.282** Spiritual 2.37 1.563 2.645 Recreation 0.513 0.397 0.880 Interactions 0.515 1.938 2.777 Concerns 0.241 2.759 5.579** Negative feelings 0.28 2.579 1.236 * p < .05, ** p < .01, = item level Item - level differences by race/ethnicity: Previous findings from the focus group study that informed this research prompted item - level analysis of the negative feelings scale. African Americans in the focus group study discussed feeling unwelcome in forested areas or parks, and expressed a safety concern when near trees or forests. Similar results were found for this study. There was a statistically significant difference in negative feelings for trees and forests for racial/ethnic groups as determined by the one - p = .002) (Ta ble 1 9 ). Post - hoc analyses (Tukey - Kramer test) showed statistically significant differences between Black/African Americans and both White/Caucasians and Others. African Americans had slightly higher mean 63 scores (2.00) than both Whites (1.43) and Others (1 .42). Further statistical significance was p = .002) (Table 19 ). Post - hoc analyses (Tukey - Kramer test) showed statistically significant differences between Black/African Americans and both White/Caucasians and Others. African Americans had slightly higher mean scores (2.34) than both Whites (1.66) and Others (1.71). Table 1 9 : Negative feelings item - level racial/ethnic differences , One - way ANOVA Item df SS MS F p I do not feel welcome in forests or parks 2 12.037 6.019 6.459 0.002 I have a safety concern when I am near trees or forests 2 16.52 8.26 6.566 0.002 Discussion of differences by race/ethnicity: This study documented significant differences based on race and gender in how people slightly higher mean responses than African Americans signifying a higher level of agreement with the scale questions. These differences are of particular significance for several scales. The exceptions to this trend are that African Americans had higher mean scores for the products scale (timber and non - timber forest products) and the negative feelings scale. Another exception is that whites had the lowest mean scores for the spiritual importance scale with African Americans and all other non - white racial minorities holding the highest mean scores, and this was statistically s ignificant at p < .10 (this difference was not highlighted in Table 16 ). Leatherberry (2000) provide d a n overview of African American spiritual and historical ties to the forest. Many other studies show the spiritual significance of trees and forests to nonwhite racial groups, mostly Native Americans in particular (Castleden et al., 2009; 64 Glowacka et al., 20 09; LaVelle, 2001; Pemberton, 1985; Swearer , 1998). Differences found here provide an example underscoring the need to sample diverse audiences as the small details sometimes can carry a great weight. Since most forest policy is representing views of older , more affluent, white males ( Anthony et al., 2004; Burns et al., 2009; McDonough et al., 2003; Taylor, 2002) , the spiritual importance of trees and forests may not be considered a priority for management objectives. Similarly, special attention is merited for the item - level racial differences found for the negative feelings scale . With higher mean scores than W hites and O thers, African Americans expressed feeling unwelcome in forests or parks and having a safety concern when near trees or forests. T e for these questions, there was much more variance in responses than for Whites and Others (Table 17) . These results concur with previous focus group findings on the imp ortance of trees and forests, along with literature on the lingering historical effects of racism, slavery, and civil rights (Leatherberry, 2000; Meraji, 2015; Taylor, 2002). D ata show that one size does not fit all, giving even more strength to the argu ment that diversity in sampling is critical for measuring the importance of trees and forests and understanding the nuances among people groups that may affect forest use. This is especially imperative if data will be used to develop policy affecting all p eople. Analyses also showed that, for this study in particular, weighting data does not necessarily compensate for lack of racial representativeness. 65 Differences by gender : RQ 2 - B: Are there differences based on gender both within and between racial/ethnic groups? Part 1: Gender differences within entire sample: An Independen t Sample s t - t est was conducted to assess differences in perceptions on the importance of trees and forests by gender. All results were compared with those from the non - para metric equivalent, Mann - Whitney U test, because of the use of Likert - type and non - normally from men, indicating a higher level of agreement (Table 20 ) . The two exceptions to this include the products scale, and the negative feelings scale. Differences in scores were statistically significant for the following (Table s 20 and 2 1 ) : Environmental/ecological function scale: The difference in scores for females (4.79 ± 0 .355) and males (4.67 ± 0.449) was significant with t(269) = 2.350, at p = .019. Cultural scale: The difference in scores for females (4.52 ± .569 ) and males (4.33 ± .685) was significant with t(266) = 2.50, at p = .013. Education scale: The differenc e in scores for females (4.77 ± .414 ) and males (4.48 ± .622) was significant with t(267) = 4.679, at p = .000. Health and well - being scale: The difference in scores for females (4.82 ± .411 ) and males (4.67 ± .551) was significant with t(267) = 2.625, at p = .009. Aesthetics scale: The difference in scores for females (4.91 ± .328 ) and males (4.79 ± .440) was significant with t(267) = 2.463, at p = .014. 66 Sense of place scale: The difference in scores for females (4.24 ± .784 ) and males (4.02 ± .846) was significant with t(266) = 2.148, at p = .033. Spiritual scale: The difference in scores for females (3.72 ± 1.07 ) and males (3.32 ± 1.09) was significant with t(261) = 2.943, at p = .004 . Table 20 : Descriptive statistics for importance scales by gender Importance scale Gender N Mean Std. Dev. Environment/Ecological function Female 165 4.79 0.355 Male 106 4.67 0.449 Economic Female 161 4.50 0.570 Male 106 4.48 0.616 Culture Female 163 4.52 0.569 Male 105 4.33 0.685 Education Female 164 4.77 0.414 Male 105 4.48 0.622 Heath & well - being Female 163 4.82 0.411 Male 106 4.67 0.551 Aesthetics Female 163 4.91 0.328 Male 106 4.79 0.440 Products Female 160 4.42 0.783 Male 106 4.42 0.736 Sense of place Female 162 4.24 0.784 Male 106 4.02 0.846 Spiritual Female 159 3.72 1.074 Male 104 3.32 1.089 Recreation Female 161 4.80 0.459 Male 104 4.64 0.709 Interactions Female 163 4.20 0.616 Male 106 4.05 0.672 Concerns Female 163 4.34 0.797 Male 104 4.26 0.804 Negative feelings Female 162 2.91 0.644 Male 104 2.97 0.601 Item - level 67 Table 2 1 : Differences in perceptions of the importance of trees and forests by gender, Independent Sample s t - t est Importance scale t df p Environment/Ecological function 2.350 269 0.019 Economic 0.200 265 0.842 Culture 2.500 266 0.013 Education 4.679 267 0.000 Heath & well - being 2.625 267 0.009 Aesthetics 2.460 267 0.014 Products 0.060 264 0.952 Sense of place 2.148 266 0.033 Spiritual 2.943 261 0.004 Recreation 2.189 263 0.029 Interactions 1.927 267 0.055 Concerns 0.800 265 0.425 Negative feelings - 0.761 264 0.447 Item - level Part 2: Gender differences between racial/ethnic groups: Analysis 1: In order to address the second part of research question 2 - B 12 , two analyses were conducted. First, the data file was split by racial/ethnic group, and then Independence Samples t - t ests were conducted for each group to analyze gender differences between race/ethnic groups. All results were compared with those from the non - parametric equivalent, Mann - Whitney U test, because of the use of Likert - type and non - normally distributed data. For this analysis, there were 12 RQ 2 - B: Are there differences based on gender both within and between racial/ethnic groups? 68 no statistically significant differences found for African Americans and all other non - white racial groups. This may be due to the low sample size for each of those racial categories (n = 41 and n = 31 respectively). However, statistically significant differences between genders were found for Whites. White women had higher mean scores than white men for all i mportance scales except negative feelings (Table 2 2 ) . Differences were statistically significant as follows (Table 2 3 ) : Environmental/ecological function scale: The difference in scores for white females (4.83 ± 0 .302) and white males (4.70 ± 0.446) was s ignificant with t(190) = 2.363, at p = .019. Cultural scale: The difference in scores for white females (4.50 ± .554 ) and white males (4.46 ± .635) was significant with t(188) = 2.418 , at p = .017. Education scale: The difference in scores for white females (4.78 ± .408 ) and white males (4.48 ± .598) was significant with t(189) = 4.121, at p = .000. Health and well - being scale: The difference in scores for white females (4.85 ± .401 ) and white males (4.67 ± .589) wa s significant with t(189) = 2.458, at p = .015. Aesthetics scale: The difference in scores for white females (4.92 ± .338 ) and white males (4.79 ± .473) was significant with t(189) = 2.209, at p = .028. Sense of place scale: The difference in scores for white females (4.30 ± .737 ) and white males (4.05 ± .834) was significant with t(189) = 2.120, at p = .035. Spiritual scale: The difference in scores for white females (3.66 ± 1.012 ) and white males (3.18 ± 1.042) was signific ant with t(184) = 3.126, at p = .002. Interactions scale: The difference in scores for white females (4.23 ± .950 ) and white males (4.04 ± .635) was significant with t(190) = 2.072, at p = .04 69 Table 2 2 : Descriptive statistics for importance scales, white males and females Importance Scale/item N Mean Std. Dev. Environment/Ecological Function White females 119 4.83 0.302 White males 73 4.70 0.446 Economic White females 118 4.57 0.510 White males 73 4.46 0.635 Culture White females 118 4.50 0.554 White males 72 4.28 0.709 Education White females 119 4.78 0.408 White males 72 4.48 0.598 Health and well - being White females 118 4.85 0.401 White males 73 4.67 0.589 Aesthetics White females 118 4.92 0.338 White males 73 4.79 0.473 Products White females 116 4.39 0.785 White males 73 4.35 0.782 Sense of Place White females 118 4.30 0.737 White males 73 4.05 0.834 Spiritual White females 115 3.66 1.012 White males 71 3.18 1.042 Recreation White females 118 4.84 0.413 White males 72 4.60 0.744 Interactions White females 119 4.23 0.590 White males 73 4.04 0.635 Concerns White females 119 4.42 0.646 White males 72 4.26 0.806 Negative feelings White females 118 2.88 0.576 White males 72 3.02 0.557 Item - level 70 Table 2 3 : Differences in perceptions of trees and forests by white males and females, Independence Samples t - t est Importance scale t df p Environment/Ecological function 2.363 190 0.019 Economic 1.399 189 0.163 Culture 2.148 188 0.017 Education 4.121 189 0.000 Heath & well - being 2.458 189 0.015 Aesthetics 2.209 189 0.028 Products 0.303 187 0.762 Sense of place 2.120 189 0.035 Spiritual 3.126 184 0.002 Recreation 2.879 188 0.004 Interactions 2.072 190 0.040 Concerns 1.516 189 0.131 Negative feelings - 1.570 188 0.118 Item - level Analysis 2: A combined gender and race/ethnicity variable was created for the second analysis to address the second part of research question 2 - B 13 . This variable categorized participants as a) black females, b) black males, c) white females, d) white males, e) all other females, and f) all other males. One - way Analysis of Variance (ANOVA) was conducted to assess differences by the combined gender an d race/ethnicity category. All ANOVA results were confirmed with the non - parametric equivalent test, Kruskal Wallis, because of the use of Likert - type and non - normally distributed data. 13 RQ 2 - B: Are there differences based on gender both within and between racial/ethnic groups? 71 A statistically significant difference between combined gender and rac ial/ethnic groups was found for the environmental/ecological function importance scale determined by the one - way ANOVA (F (5, 257) = 2.906, p = .014) (Table 2 4 ). Post - hoc analyses (Tukey - Kramer test) showed weakly ( p < .10) statistically significant differences between black females and white females, and between black males and white females. Black females (4.60) and black males (4.54) had slightly lower mean scores for the environmental/ecological function importance scale than white females (4.79). Another statistically significant difference between combined gender and racial/ethnic groups was found for the economic importance scale determined by the one - way ANOVA (F (5, 254) = 2.547, p = .029) (Table 2 4 ). Post - hoc analyses (Tukey - Kramer test) show ed statistically significant differences ( p < .05) between black females and white females. This difference was weakly ( p < .10) confirmed with the non - parametric equivalent Kruskal Wallis test. Black females had slightly lower mean scores (4.16) for the e conomic importance scale than white females (4.57). Differences between combined gender and racial/ethnic groups were found statistically significant for the educational importance scale determined by the one - way ANOVA (F (5, 255) = 4.749, p = .000) (Table 2 4 ). Post - hoc analyses (Tukey - Kramer test) showed statistically significant differences ( p < .01) between white females and white males. White females had slightly higher mean scores (4.78) for the educational importance scale than white males (4.48). A s tatistically significant difference between combined gender and racial/ethnic groups was found for the sense of place importance scale determined by the one - way ANOVA (F (2, 254) = 4.287, p = .001) (Table 2 4 ) . Post - hoc analyses (Tukey - Kramer test) showed s tatistically significant ( p < .05) differences between black males and both white females and all other 72 females. Black males had slightly lower mean scores (3.51) for the sense of place importance scale than both white females (4.30) and all other females (4.45). Differences between combined gender and racial/ethnic groups were found statistically significant for the spiritual importance scale determined by the One - way ANOVA (F (5, 255) = 2.748, p = .019) (Table 24 ). Post - hoc analyses (Tukey - Kramer test) we re unclear. Similarly, differences were found for the concerns scale determined by the One - way ANOVA (F (5, 259) = 3.121, p = .009 ( Table 24 ) . P ost - hoc analyses were unclear. Table 24 : Differences in perceptions on the importance of trees and forests by ra ce/ethnicity and gender combined (One - way ANOVA) Importance scale df SS MS F p Environment/Ecological function 5 2.26 0.452 2.906 0.014 Economic 5 4.303 0.861 2.547 0.029 Culture 5 2.935 0.587 1.506 0.188 Education 5 6.208 1.242 4.749 0.000 Heath & well - being 5 2.19 0.438 1.937 0.089 Aesthetics 5 1.104 0.221 1.539 0.178 Products 5 2.508 0.502 0.872 0.501 Sense of place 5 13.431 2.686 4.287 0.001 Spiritual 5 15.589 3.118 2.748 0.019 Recreation 5 2.832 0.566 1.715 0.132 Interactions 5 3.379 0.676 1.641 0.149 Concerns 5 9.584 1.917 3.121 0.009 Negative feelings 5 1.401 0.28 0.754 0.584 Item - level Discussion of gender differences: In concordance with literature that women tend to exhibit greater environmental concern (Bell & Braun, 2010; Dietz et al . , 2002; Hunter et al., 2004; McCright & Xiao, 2014), women 73 had higher mean scores (indicating a higher level of agreement) for all but two importance scales. This was significant for environmental/ecological function, cultural importance, educa tional importance, health and well - being, aesthetics, sense of place, and spirituality. Women and men agreed equally on the importance of trees and forests for providing timber and non - timber products. Men showed slightly higher agreement with the negative feelings scale, and forests. In this study, women had higher levels of agreement with many facets of trees and forests, and were in disagreement with the inhere nt negative aspects of trees and forests. Gender differences between racial/ethnic group were difficult to establish with certainty. This may be due to the low representation of non - white participants. The first analysis found significant differences bet ween white females and white males; however, no significant gender differences were found for African Americans and other non - white minorities. The second analysis examined differences between participants based on their gender and race combined into one v ariable. A significant finding from this analysis is that white females, overall, expressed higher levels of agreement with most importance scales than white males, black females and males, and all other females and males. These results demonstrate the imp act of having a diverse sample in order to capture the range of perceptions on the importance of trees and forests, and the necessity to analyze racial and gender differences if interested in fully understanding these perceptions. With women of color leadi ng environmental justice activism in their communities (Kaalund, 2004; Verchick, 2004), it is necessary to investigate the gendered racial differences further for clarity and understanding. 74 Differences by Sample Group : RQ 2 - C Are there differences based on the different sampling and data collection methods? For all scales except that measuring negative feelings, sample group 3 (purposive/in - person) mean scores were higher than those from sample groups 1 (random) and 2 (disproportionate), indicating a high er level of agreement (Table 2 5 ) . One - way Analysis of Variance (ANOVA) was conducted to test for differences in perceptions on the importance of trees and forests by sample group. All ANOVA results were confirmed with the non - parametric equivalent test, Kruskal Wallis, because of the use of Likert - type and non - normally distributed data. There was a statistically significant difference in perceptions on the cultural importance of trees and forests for sample groups as determined by the one - way ANOVA (F (2 , 278) = 3.973, p = .02) (Table 26 ) . Post - hoc analyses (Tukey - Kramer test) showed statistically significant differences between sample group 1 and sample group 3. Sample group 1 had slightly lower mean scores (4.35) for the cultural importance scale than b oth sample group 2 (4.44) and sample group 3 (4.63) ( Table 2 5 ) . A statistically significant difference in perceptions on the educational importance of trees and forests for sample groups was also found as determined by the one - way ANOVA (F (2, 280) = 3.0 94, p = .047) (Table 26 ) . Post - hoc analyses (Tukey - Kramer test) showed weakly statistically significant differences ( p < .10) between sample group 1 and sample group 3. Sample group 1 had slightly lower mean scores (4.62) for the educational importance sca le than both sample group 2 (4.63) and Sample Group 3 (4.81) (Table 2 5 ) . However, this result was weakly confirmed ( p < .10) with the non - parametric equivalent, Kruskal Wallis test. 75 Difference s in perceptions on the sense of place importance of trees and f orests for sample groups was also found statistically significant as determined by the one - way ANOVA (F (2, 276) = 3.61, p = .028) ( Table 26 ). Post - hoc analyses (Tukey - Kramer test) showed statistically significant differences ( p < .05) between sample group 1 and sample group 3. Sample group 1 had slightly lower mean scores (4.07) for the sense of place importance scale than both sample group 2 (4.11) and sample group 3 (4.41) ( Table 2 5 ) . 76 Table 2 5 : Descriptive statistics for importance scales by sample group Importance Scale/Item N Mean Std. Dev. Environment/Ecological Function Group 1 123 4.72 0.417 Group 2 102 4.69 0.435 Group 3 60 4.84 0.286 Economic Group 1 121 4.47 0.595 Group 2 101 4.49 0.606 Group 3 59 4.52 0.566 Culture Group 1 123 4.35 0.658 Group 2 98 4.44 0.636 Group 3 60 4.63 0.538 Education Group 1 123 4.62 0.563 Group 2 100 4.63 0.550 Group 3 60 4.81 0.364 Health and well - being Group 1 122 4.73 0.508 Group 2 101 4.70 0.529 Group 3 60 4.86 0.418 Aesthetics Group 1 122 4.83 0.438 Group 2 101 4.84 0.393 Group 3 60 4.93 0.236 Products Group 1 119 4.43 0.707 Group 2 101 4.37 0.834 Group 3 60 4.50 0.722 Sense of Place Group 1 121 4.07 0.850 Group 2 99 4.11 0.874 Group 3 59 4.41 0.682 Spiritual Group 1 118 3.50 1.045 Group 2 98 3.56 1.178 Group 3 58 3.76 1.082 Recreation Group 1 121 4.68 0.622 Group 2 98 4.72 0.622 Group 3 58 4.86 0.476 Interactions Group 1 122 4.13 0.621 Group 2 101 4.05 0.731 Group 3 59 4.28 0.557 Concerns Group 1 122 4.20 0.794 Group 2 100 4.35 0.812 Group 3 57 4.45 0.790 Negative feelings Group 1 121 2.98 0.681 Group 2 101 2.91 0.636 Group 3 57 2.89 0.550 Item - level 77 Table 2 6 : Differences in the perceptions of the importance of trees and forests by sample group Independent variable df SS MS F p Environment/Ecological Function 2 0.844 0.422 2.631 0.074 Economic 2 0.084 0.042 0.119 0.887 Culture 2 3.122 1.561 3.973 0.02 Education 2 1.69 0.845 3.094 0.047 * Health & well - being 2 0.992 0.496 1.996 0.138 Aesthetics 2 0.469 0.235 1.566 0.211 Products 2 0.603 0.302 0.524 0.593 Sense of place 2 4.931 2.465 3.61 0.028 Spiritual 2 2.643 1.322 1.089 0.338 Recreation 2 1.344 0.672 1.9 0.152 Interactions 2 1.994 0.997 2.355 0.097 Concerns 2 2.575 1.288 2.014 0.135 Negative feelings 2 0.415 0.208 0.508 0.603 Item - level *This finding was confirmed weakly ( p < .10) with the Kruskal Wallis test Discussion of sample group differences: A purpose of this study was to obtain a representative sample through mixing sampling and data collection methods. O ne aspect of this methodological exploration was to examine for differences between the sample groups. Sample group 3 had a higher level of a greement with all scales except that measuring negative feelings (statistically significant for three scales). This does lead to question ing if the in - person survey mode combined with purposive sampling had some effect on participant response, which is oft en discussed in survey research literature (Dillman et al., 2009a). For example, participants in sample group 3 had a personal interaction with the researcher before and after completing the survey. This may have affected response. Also, the presence of a researcher may have influenced a participant to respond more favorably 78 than if they had received the survey anonymously in the mail. There is also the issue of bias. Perhaps several participants with a more favorable attitude towards trees and forests chos e to participate because of the salience of the topic. A deeper explorati on of the sample group differences may be explored in the future to provide a better understanding of these potential nuances . Path Analysis : A path analytic approach was used to understand more complex relationships between influence, or direct and indirect effects when there are several predicting independent variables. It is a method t hat is powerful for assessing the strengths of complex relationships in hypothesized models (Agresti & Finlay, 1997; Propst & Koesler, 1998; Streiner, 2005), such as if variable X influences variable Y, which in turn influences variable Z. Path analysis ca nnot prove causality, but it can help make more sense of complex relationships (Streiner, 2005). Endogenous and exogenous variables were chosen for this analysis based on correlated relationships ( Table 2 7 ) and t o examine if education level (ex ogenous) was influencing, or mediating relationships between variables. 79 Table 2 7 : Pearson correlations for variables use d in path analytic approach RA EL GE EF CO RA Race 1.0 EL Education level 0.108 1.0 GE Gender 0.039 - 0.043 1.0 EF Eco. Function 0.208** 0.198** - 0.142* 1.0 CO Concerns .195** .153* - 0.049 0.536** 1.0 * p < .05, ** p < .01 Earlier ANOVA analyses showed that there was a statistically significant relationship between race/ethnicity and concerns for trees and forests. Participant education level and race/ethnicity were both correlated with concerns, though education level and r ace were not correlated with each other (Table 2 7 ) . Attempting to examine strengths of the relationships and potential spurious correlations, the first path model explored if participant education level (EL) mediated the relationship between race/ethnicity (RA), and concerns (CO) people had for trees and forests (Figure 4 ) . Gender was excluded from the initial proposed model because gender was not significantly correlated with any of the other endogenous or exogenous variables. Figure 4 : Proposed path model for race, education level and concerns people had for trees and forests CO = x 1 RA + x 2 EL + e 1 80 Path coefficients showed that race/ethnicity had both a direct and an indirect effect on concerns. The indirect effect was mediated by the influence of ed ucation level in a positive direction, indicating that as education increases, concerns also increase. Education level m ediat ed the effects of race for measuring concerns people had for trees and forests (Figure 5 ). S tatistically significant at p values le ss than .05 and .01 , t his effect accounts for 5.3% (R 2 = .053) of the variance in the model. The correlation between race and concerns may be somewhat spurious as education level appears to mediate the effects of race on concerns. Figure 5 : Full path model for race, education level and concerns people had for trees and forests * p < .01, ** p < .01 To further investigate if there were any possible gender differences despite the lack of significant correlat ion with other variables, the data file was split into two groups (males and females) and the model (Figure 4) was tested a second time. Splitting the data into male and female groups resulted in statistically significant findings related to gender difference s. Path coefficients showed that for women, but not men, race/ethnicity had both a direct and an indirect effect on concerns (Table 28 ) . The indirect effect was mediated by the influence of education level in a positive direction, ind icating that, for wome n , as education increases, concerns also increase. Education level mediat ed the effects of race on concerns people had for trees and 81 forests (Figure 6 ). This was statistically significant at p values less than .01 and the effect a ccounts for 10 . 8 % (R 2 = . 1 08 ) of the variance in the model. Table 2 8 : Path coefficients of a model for race, educational level, and concerns for trees and forests Endogenous (Dependent) Variable RA EL R 2 F Significance Concerns (CO) Full sample 0.174 .137* 0.053 7.343 ** Males 0.065 0.032 0.005 0.262 ns Females 0.26 0.197 0.108 9.339 ** * p < .05 ** p < .01 Figure 6: Full path model for race, education level and concerns for trees and forests, females only Earlier ANOVA tests showed that there was a statistically significant relationship between race/ethnicity and the perceived importance of environmental and ecological function of trees and f orests. An Independent Samples t - t est also showed statistically significant gender differences in the perceived importance of environmental and ecological function of trees and 82 forests. Education level, race/ethnicity, and gender were correlated with environment/ecological function, though education level, race, and gender were not significantly correl ated with each other (Table 2 7 ) . Attempting to examine strengths of the relationships and potential spurious correlations, the second path model assessed if education level (EL) mediated the relationship between race/ethnicity (RA) and environment/ecologic al function (EF) of trees and forests (Figure 7 ) . The proposed model includes the relationship between gender (GE) and perceptions of the importance of environmental and ecologica l function, and shows the weak but positively correlated relationship between gender and race/ethnicity. Figure 7 : Proposed path model for education level, race and the environmental and ecological function of trees and forests EF = x 1 RA + x 2 EL + e 1 EF = x 1 GE + e 2 EL = x 1 RA + e 3 EF = x 1 RA + e 4 83 Path coefficients showed that race/ethnicity had both a direct and an indirect effect on environment/ecological function. The indirect effect was mediated by the influence of education level in a positive direction, indicating that as education increases, perceived importance of the en vironmental and ecological function of trees and forests also increase. The finding was statistically significant at p values less than .05 and .01 (Figure 8 ) . This effect accounts for 7.6% (R 2 = .076) of the variance in the model. The correlation between race and environment/ecological function may be somewhat spurious as education level appears to mediate the effects of race on concerns. The path coefficients also showed that gender had a direct effect on perceived importance of the environmental and eco logical function of trees and forests. This effect appears to be unmediated by other variables. There were no additional statistically significant findings based on gender when analyzing the date file split into groups of males and females. Figure 8 : Full path model path model for education level, race and the environmental and ecological function of trees and forests * p < .01, ** p < .01 84 Discussion of path analysis: A path analytic approach was implemented to further understand the complex relationships between variables and if there was any effect on perceptions. Specific variables were chosen based on relationships with race and gender, which are the focus of this s tudy. E arlier analyses in this chapter may lead one to wonder why the income level or employment status variables were not included in the path models. Income level and employment status variables were not correlated with environmental and ecological funct ion or concerns (the end ogenous variables in the path models). However, income level and employment were correlated with education level, and educational level may have an effect on income level and employment status. All these things considered , education al level was chosen as a potential mediating variable for the path model. In these analyses, participant education level mediated the relationship between race and concerns for trees and forests, and also race and perceptions of the environmental/ecological importance of trees and forests. When analyzed further, the finding that education level mediated the relationship between race and concerns for trees and forests was found to be stronger for women as the model explained 10.8% of t he variance, compared to 5.3% for the full sample. These findings provide an additional and interesting nuance to understanding the importance of environment al perceptions , and for women, it seems to play a stronger role for environmental concern . For men, another variable may mediate this relationship. These findings resonate with o ther research studies that suggest education level is associated with higher l evels of environmentalism (Arcury & Christianson, 1990; Arcury, Johnson, & Scollay, 1986; McMillan, Hoban, Clifford, & Brant, 1997; Milbrath, 1984; Van Liere & Dunlap, 1980) and studies 85 showing that women exhibit greater environmental concern ( Bell & Braun , 2010 ; Dietz , Kalof, & Stern , 2002; Hunter , Hatch, & Johnson , 2004; McCright & Xiao, 2014 ) . There are many possible ways to conduct a path analysis with these data. A simple exploration was conducted to see if there were some connections related to the li terature. Other relationships between variables need to be further explored. Research suggests that environmental value orientations (spiritual, utilitarian, etc.) do influence perceptions related to forest management (Dutcher et al., 2007; Li et al., 2010 ). Future path models may investigate if race, gender, age, etc., on perceived importance of trees and forests. Participant evaluation of survey methodology : RQ 3: What study design factors influenced the research participants to participate? A total of eleven people participated in the evaluative portion of this study, via personal interviews and focus group. Six participants received the survey via mail and five received the survey in person. Of the 11 participants, seven were female, four we re male, two were Black/African American, and nine were White/Caucasian (Table 2 9 ). Table 2 9 : Characteristics of evaluative sample n = 11 Black/African American White/Caucasian Female 2 5 Male 0 4 86 Data were analyzed qualitatively. First, notes and transcripts were read through carefully to explore emerging ideas. The ideas were then documented and separated into categorical themes. Then, the notes and transcripts were read through a second time and assigned themes as appropriate. Themes included par ticipant motivation, opinions towards surveys in general, the most and least interesting asp e cts of the survey, and tips offered by the participants to improve survey research (Table 30 ). Table 30 : Focus group/interview themes Received survey by mail 6 participants (n = 11) Motivation to participate Interest in topic Opportunity to express opinion Influence decision - making Opinions towards surveys in general Mostly positive Most interesting aspect Spiritual value questions Health & well - being questions Environment/ecological function questions Least interesting aspect Questions about negative feelings associated with trees and forests Tips Shorten survey Survey in - person at baseball stadiums, shopping malls Survey in - person with option to take home with postage paid envelope Motivation to participate: Participants were asked about their motivation to participate in the survey on the importance of trees and forests. Consistently, participants answered that they had an interest in the topi c, or that they wanted to take the opportunity to express their opinion. One participant 87 even offered that they were interested in the opportunity to influence decision - making. The following quotes are directly from the interviews and focus group: I thought it was important to have an opinion about it and to express that opinion. If I Be hopefully in a way Opinions towards surveys in general: When asked about their opinions towards surveys in general, participants expre ssed an overall positive view. Similar to their motivation to participate, they explained that they would participate in a survey if they felt the topic was important, relevant to them, or contributing to something positive. The exception to this was any s urvey that was also asking for a financial contribution or excessive personal information. The quotes that follow provide direct examples: the mail, because you have nonsense surveys I just throw away more for business or for their capitalization. They 88 Most interesti ng aspects: Participants were asked if there were any aspects of the survey that were especially interesting. Their favorite aspects of the survey included those addressing the spiritual value of trees and forests, questions related to health and well - being, and questions about the environmental and ecological function of trees and forests. In their view, the spiritual value and the contribution to a sense of health and well - being are often undervalued aspects of trees and forests. Additionally, th e environmental and ecological function of trees and forests was viewed as a priority. The following quotes are examples directly from the participants: s piritual. I re alance of life and relaxing than I was to the spiritual aspect of it, making me closer to the balance aspect better than the spiritual aspect, coming from a spiritual - wise, being able to relax and have that down because that is where I feel God, where I sense the presence of a creat or, not in a church. I only not being amazed how they can possibly think it came from nowhere just kind of magically appeared like that, it is so complex down to the tiniest smallest thing, the exquisite colors of an insect, the color of a butterfly, the many, many different colors of a Least i nteresting aspects: The participants were also asked about their least favorite aspects of the survey. The main response from this question addressed the discomfort they felt about the questions on negative 89 feelings associated with trees and forests. Part icipants felt this question was odd because it seemed out of place, and in contrast to the rest of the survey. They also felt the question was confusing because the value of trees and forests cannot be overshadowed by the perceived negative associations. T he following quotes provide examples: ty things that come along with trees, they are worth so much more than those pesty things! I Tips to improve survey research: Lastly, participants were asked if they had any advice, tips, or comments about how to improve survey research in general, how to increase representation, and specifically for studying the importance of trees and forests. A few participants suggested shortening the survey instrument, especially if surveying in person. For surveys in general , participants recommended surveying at baseball stadiums (Comerica Park in Detroit) that attract diverse people from all across the state of Michigan, or even shopping malls. They also encouraged surveying in person, at shopping malls or farmers markets, with the option to take the survey home with a postage paid envelope for return. The following quotations serve as direct examples from the data: Rapids and Flint and all ar get anybody interested. Could you just have them take the survey home, not necessarily 90 can tell you. If you had it all set up with a return address stamped and all that and they were interested enough to ask you or you promoted it that way. Discussion of evaluative findings: The evaluative data collection revealed that participants chose to complete the survey because they had an interest in the topic, or were taking the opportunity to express their opinion. None of the focus group participants discussed the VISA gift card lottery incentive as a reason for completing the survey. Focus group and interview participants had mostly positive feelings towards surveys in general, and provided practical tips to improve the survey research process. Some of the tips at first do not seem useful or feasible for the goals of this study or future studies on the importance of forests . I t may be worthwhile to explore the practicality of conducting surveys in a shopping mall or baseball stadium to contribute to the development of a representative sample. Focus group and interview participants also elaborated on the survey questions that we re most interesting and those that were least interesting, which is helpful for further survey development. From the results of the focus group and interviews, it seems clear that only those interested in the topic of trees and forests continued their part icipation through the evaluative feedback, presenting an additional bias. Lastly, the sample for this phase of the research was not racially representative. Rather it was over - representative of white Lansing residents and females. Summary of results and d iscussion: This study had two purposes. It examined the usefulness of mixing survey sampling and data collection mod e s to achieve a racially/ethnically - and gender - representative sample for 91 understanding the importance of trees and forests. The results i ndicate that mixing sampling methods and data collection modes does increase the racial representativeness of survey participants, thus providing a contribution to the goals of sustainable forests management (collecting data on the diversity of ways the tr ees and forests are important from the diversity of people). Findings show that there are important differences based on race/ethnicity and gender in how people perceive trees and forests to be important , addressing the second purpose of this study . Data also show that there may be mediating factors accompanying race (and perhaps gender) influencing perceptions on the way trees and forests are valued. The findings presented provide strong support to collect data from diverse audiences if committed to susta inable forests management. Survey findings concurred with qualitative data from an earlier study informing this research showing that the developed metrics were acceptable for measuring the importance of trees and forests. Some aspects of the survey need r evisions for participant clarity. Responses from evaluative focus groups and interviews provided valuable feedback on this particular study and the survey research process in general. 92 CHAPTER 5: CONCLUSION The United States has committed itself to su stainable forests management (SFM) which considers the needs of present and future generations across the ecological, economic, and social dimensions of forest use. SFM calls attention to the social aspects of forest management by requiring the participati on of an informed, aware, and engaged public. Th is presents a need for research on the social dimensions of forestry, including understanding perceptions of the importance of trees and forests, to support the goals of SFM. However, the literature suggests those participating in forestry research are mostly middle - to senior - aged, educated, white males inaccurately representing the diversity of the United States populace. Many natural resource pr ofessionals attribute the absence of diverse participants in forestry research to a lack of interest from those that do not participate. However, data suggest that people do not participate because of the lack of opportunity and access, resulting in justic e and equity issues. Understanding perceptions from the diversity of people is vital because perceptions may influence support (or lack of) for forest management. The significance of this issue only grows as the U.S. continues to diversify. Responding to the justice and equity issue of low representation of racial/ethnic minoriti es and women in SFM research, this study explored the mixing of survey sampling and data collection techniques with the intention to achieve a representative sample. The exploratio n resulted in a racially representative sample and collected diverse data on the importance of trees and forests uncovering nuanced differences based on race, gender, and sample group. This 93 chapter provides a brief summary of the findings, directions for f uture research, study limitations, implications , and a conclusion . Summary of the findings: Mixing methods and sample representativeness One study objective was to examine the usefulness of mixing survey sampling and data collection methods to achieve a ra cially/ethnically - and gender - representative sample for understanding the importance of trees and forests. More specifically, three different pairings of sampling and data collection methods were tested: Sample group 1 was randomly chosen and received a s urvey questionnaire in the postal mail. Sample group 2 was chosen disproportionately random and also received a survey questionnaire in the postal mail. Lastly, sample group 3 consisted of participants chosen purposively and received a survey questionnaire in person with an introduction to the study including relevant background information. The survey data collection was followed by interviews and a focus group with willing survey participants to discuss the survey research process. Combining methods provided a useful methodology for achieving a racially representative and gender - diverse survey sample. A total of 285 Lansing residents participated in the survey data collection. Women are often underrepresented in forestry and natural resource - related research, yet for this study w omen were over represented in each sample group: Sample group 1 consisted of 59% females, sample group 2 had 68% females, and sample group 3 achieved 60% female representation . I ndividual sample groups varied i n their racial representativeness with sample group 1 containing the least diversity (72% White, 13% Black, 94 15% Other) and the most diversity found in sample groups 2 (64% White, 21% Black, 17% Other) and 3 (60% White, 20% Black, 17% Other). The complete sample was 61% female, 39% male, 70% White, 18% Black, and 12% Other, with a variety of ages, income, education, and employment levels. Combining a sample with random, disproportionate, and purposive elements resulted in a more diverse sample of participan ts. None of the utilized sampling methods were entirely useful for achieving representativeness on their own: Sample group 1 was randomly chosen, but lacked diversity/representativeness. Sample group 2 was disproportionately chosen with more diverse racial representation, but was restricted to specific geographic areas, limiting coverage. Sample group 3 was purposively chosen and more diverse, but lacked the element of randomness. This idea of representation is reflected in literature calling for the use of mixed methods to appropriately answer research questions (Tashakkori & Teddlie, 1998) and the mixing of survey modes may help to reduce survey errors (Dillman et al., 2009). Metrics Previous qualitative research on the importance of trees and forests pro vided rich data used to create the survey metrics for this study. Twelve scale factors and one item level question addressed many ways in which trees and forests may be important to people in a Likert style format. M etrics developed to understand the perce ived importance of trees and forests were successful overall . Ten of the twelve scale factors demonstrated good or excellent reliability cultural, educational, aes thetic, and spiritual importance, health and well - being, forest products, sense of place, interactions, and concerns related to trees and forests. Two scale factors, measuring economic importance and negative feelings associated with trees and forests, had 95 questionable alpha coefficients and require revisions for future use. One item from the economic importance scale was inconsistent with other items. Participants seemed confused about the question on negative feelings associated with trees and forests. M ore details on this surfaced in the evaluative focus group and interviews. was 70% or higher for most scales, and all scales except one were positively correlated wi th one another. These findings indicate an overall positive attitude towards trees and forests. Findings on the perceived importance of trees and forests reflect those found in previous studies informing this research. The developed survey instrument provi des a foundation for future data collections on the importance of trees and forests. Racial differences The second study objective was to analyze data on the importance of trees and forests . S urvey data showed important differences based on race/ethnicity and gender in how people perceive the importance of trees and forests. express higher levels of agreement with scale questions compared to African Americans. These findings were statistically significant for scales measuring environmental/ecological function, economic importance, sense of place, health and well - being, and concerns for trees and forests. African Americans survey scores were still quite favorable overall, but with more variation in response. Whites had the lowest mean scores for the spiritual importance scale compar ed to An examination of item level questions from the negative feelings scale showed that African Americans had higher levels of agreement with statements measuring feeling unwelcome in forests or parks, and having a s afety concern when near trees or forests. These 96 findings concur with previous qualitative data on the importance of trees and forests and the literature on the lingering historical effects of racism, civil rights, and slavery (Leatherberry, 2000; Meraji, 2 015; Taylor, 2002). Findings give more strength to the argument that diversity in sampling is critical for truly understanding the importance of trees and forests to people. Otherwise, nuances such as these differences related to feelings of safety and unw elcome may go unnoticed and unaddressed. Weighted data analysis Weighting as a statistical tool was tested for its effectiveness to account for data that lacked racial representativeness, such as that found in sample group 1 . Data from sample group 1 were weighted to reflect the racial characteristics of the study population and the ANOVA test assessing differences based on race was repeated. The weight ed sample group 1 ANOVA data were compared to unweighted data from sample group 2. This analysis showed in consistent results when comparing the weighted and unweighted data indicating that for this study, weighting data was only marginally successful in compensating for the lack of racial representativeness. Gender differences Women expressed higher levels of agreement with all importance scales which is supported by literature showing that women tend to exhibit greater environmental concern (Bell & Braun, 2010; Dietz et al., 2002; Hunter et al., 2004; McCright & Xiao, 2014). These findings were statistically s ignificant for the scales measuring environmental/ecological function, cultural, educational, aesthetic, and spiritual importance, health and well - being, and sense of place. An ANOVA test examining gender and race showed that white women had the highest le vels of 97 agreement with importance scales, overall, compared to white men, and nonwhite men and women, though these findings should be interpreted cautiously due to differing sizes of the analyzed groups (white men, white women, black men, black women, othe r men, other women). Women (and people) of color are at the forefront of the environmental justice movement engaging in community activism (Kaalund, 2004; Verchick, 2004) and as such, it is necessary to investigate the gendered racial differences further. Sample group differences Sample group differences on the perceived importance of trees and forests were found through an ANOVA test. Sample group 3, which consisted of participants taking the survey in - person, demonstrated higher levels of agreement with a ll scales except that measuring negative feelings associated with trees and forests. These differences were statistically significant for factors measuring cultural and educational importance and sense of place. Participants in sample group 3 had a persona l interaction with the researcher along with an introduction to the research. It is unclear if this interaction affected response. Mediating factors A path analysis approach suggest ed that there may be mediating factors accompanying race that influence pe rceptions on the way trees and forests are valued , and that the mediating factors may differ for men and women . Specifically , one path model showed that participant education level mediated the relationship between race and perceived importance of the envi ronmental/ecological function of trees and forests. Previous studies suggest that education level is associated with higher levels of environmentalism (Arcury & Christianson, 1990; Arcury et al., 1986; McMillian et al., 1997; Milbrath, 1984; Van Liere & Du nlap, 1980). 98 Another path model showed that for women, but not men, education level mediated the relationship between race and concerns for trees and forests. This relationship was significantly weaker for men. Again, these results agree with literature s howing that women exhibit greater environmental concern (Bell & Braun, 2010; Dietz et al., 2002; Hunter et al., 2004; McCright & Xiao, 2014). The path analyses offer additional support for the importance of collecting data from diverse audiences instead of meeting the status quo, or collecting data from the same kinds of people over time. Again, without a diverse sample, certain details may go unnoticed and unaddressed. Evaluative Findings Responses from the evaluative focus group and interviews provided practical feedback on this study and survey research more generally. Participants reported an interest in the topic as the motivating factor for completing the survey and focus group/intervi ew process. They answered questions about their opinions towards survey research in general, and questions about the most and least interesting aspects of the survey questionnaire. The most interesting aspects were described as questions addressing the spi ritual importance of trees and forests, how trees and forests contribute to a sense of health and well - being, and questions about the environmental and ecological function of trees and forests. The least favorite aspect was the question addressing negative feelings towards trees and forests. Participants concluded the evaluative portion of this study by offering suggestions for recruiting a diverse sample of people. The evaluative portion of this s tudy helped to better understand motivations to participate , question comprehension, development, and revision, and approaches to sampling and data collection. Collectively, the findings from this study exemplify the importance of being flexible with research methodologies (if the goal is to have a diverse, repres entative sample), the 99 significant results a diverse sample can uncover (in this case, racial and gender differences), and the usefulness of engaging research participants in the evaluation process. Suggestions for further research: This study only begins t o answer some questions and simultane ously raises new questions. The research provided an exploration of mixing sampling and data collection methods and forest s. Results suggest that mixing methods contributes to recruiting a representative sample. The mixing of methods in this manner needs to be further developed. Since this study was exploratory in nature, it may be beneficial to repeat the stud y with a larger sample size and at a larger scale (regional, state or national level) to see if similar trends in sample representativeness emerge. Recommendations for further rese arch with in - person data collection s (sample group 3) include working with more key inform ants to access a broader range of community events and organizations for recruitment. Small - scale studies could examine if there are particular seasons of the year that are more successful for this type of research. The in - person data collection for this s tudy occurred in t he summer months which had benefits (Michigan residents tend to spend more time outdoors at community events in the summer) and costs (Michigan residents travel in the summer months, so many people may not be present for research activiti es). The primary concern for in - person data collections should be focused on making connections and building meaningful relationships with a variety of community organization s. This necessitates taking time to learn how the researcher can give back to the community group, or uncovering ways that the research will or can be directly beneficial to participants, which is a 100 goal of emancipatory research methodologies such as those used in many feminist, pragmatist, and environmental justice studies. Future re search can explore if different survey modes (in - person, mail, online) are more appealing to different age groups. In this study, younger persons were more likely to respond to in - person data collection. Is this simply due to chance, or is there a connecti on between data collection mode and age/generational preferences? O nline surveys were not explored . W ith the increase of internet and web media in the live s of the younger generation, the s e mode s require additional attention. The use of online surveys need s to be mixed with other modes because this method does not generally result in a racially, gendered, aged, or socioeconomically representative sample (Dillman et al., 2009a). Those responding to online surveys are mostly younger, white, and male ( Kwak & Radler, 2002; Saxon, Garratt, Gilroy, & Cairns, 2003 ) . Research shows many people still pref er mail surveys ( Dillman et al., 2009a). Some focus group and interview participants expressed similar feelings about preferences for mail surveys. Mail survey rese arch is here to stay at least for the immediate futur e, but it may be most representative if mixed modal, or by providing multiple options for survey participation (mail, in - person and /or online, for example). Much can be learned from assessing relationshi ps between socioeconomic variables to see if there are direct and indirect effects on perceptions of the importance of trees and forests. This s tudy touched upon mediating variables through the interpretation of path coefficients. A deeper analysis is meri place and perceptions on the importance of trees and forests? W omen had higher levels of agreement with importance scales. Are there indirect effects of other variables influencing this level of agreement? Are there additional differences based on gender, if path models are 101 analyzed separately for males and females? D oes race play a role in gendered differences on the perceived importance of trees and forests? A feminist intersection al research approach can help uncover the importance of relationships between variables. Study Limitations: L ike all research methods, survey research ha s limitations. Limitations for this study include: The study used multiple sampling and data collect ion strategies . Combining sampling and data collection modes is often perceived as incompatible with inferential statistical analysis. This limitation is also assumed to be the strength for creating a more representative sample. The researcher chose to us e a Delivery Sequence File (DSF) because the literature identifies this file as being reliable . The purchased list, though likely superior to other alternatives, still resulted in survey returns due to an unoccupied home or incorrect address . Lansing residents that only use a post office box did not receive a mailed survey. If the household receiving the survey did not speak English, they likely could not participate . The survey was given face - to - face in Spanish to two participants. T he sampl ing focused on race/ethnicity and gender, though those are only two indicators of representativeness. Education, income, socioeconomic status (SES) and other indicators of representativeness were not critically assessed for sampling purposes in the interes t of simplicity. 102 T he in - person data collection wa s biased towards low - income persons because of the lack of response from clubs/organizations with more income diversity. S ample group 3 (in - person survey) was much smaller than sample groups 1 and 2. Hund reds of people received this survey in the mail and chose not to complete and return. The data reported here only capture perceptions of a small proportion of the popul ation. Examining the issue of nonresponse and sampling error needs to be critically cons idered when making any inferences from these data. Implications: Research methods The most significant implication for this study is that mixing sampling and data collection methods did provide a more racially representative sample of survey participants. Traditional survey (probability/random) sampling provided the least demographic re presentation while disproportionate and purposive sampling provided a more realistic demographic representation. Weighting data to substitute for underrepresentation of some racial groups and overrepresentation of others proved problematic for this study a s weighted results were inconsistent with results from the racially representative sample achieved with disproportionate sampling. This suggests that weighting data may be only marginally successful at accounting for a lack of racial representativeness. Us ing different sampling and data collection methods presents a tradeoff. Traditional science requires strict adherence to rigid guidelines for research methods and any diversion from those guidelines is perceived to introduce bias. However , traditional surv ey sampling methods 103 generally do not produce representative samples, which is another type of bias. Traditional survey methodologies, especially used in forestry research , are successful for collecting data from mostly white, educated, older males. These s amples tend to not be representative yet still the data are used as if they were. In many forestry and natural resource publications, the absence of nonwhites from research respondents is not even acknowledged. C onsidering the rapidly changing U.S. racial and ethnic demographics (Collins, Hall, & Neuhaus, 1999; Ortman & Guarneri, 2009; Toosi, 2002; Waddington & Welkoff, 2010) , inclusivity is imperative when working toward ability of future generations to meet their own needs . suggests it is a violation of sustainability to exclude diverse participants in the research process as they will comprise a larger proportion of future generations. Research method s should be chosen based on the usefulness in relation to the research questions. T o collect data from traditionally underr epresented groups (racial/ethnic minorities and women), it is necessary to use mixed methods for succe ssful sampling . It is comparable to an employer asking an employee to complete a task and the employee is unable to complete the task because s/he did not know where to begin or the steps to take to accomplish the task. It seems obvious that if one method does not accom plish the task to switch to another method and keep trying. If the employee is not showing progress, eventually s/he will lose the job. It is not good enough to ignore or s imply accept low representation. I t is important to investigate why people do not pa rticipate, or if there are other more culturally - friendly methods that encourage people to participate if made available. In other words, it is important to understand the motivations and barriers to participation for different types of people. 104 The U.S. p ublishes a national report on the progress of SFM every five years. As the U.S. continues working towards SFM, the methodology developed for this study may be used to s uccessful at capturing perception s of the perceived importance of trees and forests, and can be revised and improved over time. Mixing methods provided an example of how to achieve a representative sample. The methodology and survey instrument combined off er a useful tool to address the call for diversity described by Indicator 6.44 and to inform the five - year national reports. Policy and management The current study identified differences based on race/ethnicity and gender . These findings require addition al exploration and attention by both researchers and professionals . management objectives at different scales (neighborhoods to state and federal levels), region s, and for different types of forests. Racial and gender differences found in this study further emphasize the ne cessity to seek diversity in sampling to tease out differences and improve the ource managers can better meet the needs of society. Decision - makers and/or resource managers will need to find ways to address differences in the ways that people perceive the importance of tr ees and forests. For example, in this study nonwhite racial mi norities expressed stronger agreement with the spiritual importance of trees and forests. How can managers and natural resource professionals utilize this information for prioritizing management objectives? Similarly, how can decision - makers and natural resource professionals deal with gendered differences? What is the meaning behind these dif ferences ? 105 The differences identified here present new challenges that will require natural resource professionals to engage more with diverse groups of people. This requires a shift in the current culture of the natural resources field in general, as there is a lack of racial and gender diversity across all institutional levels, types of organizations, public participation activities, and research. While ongoing diversity training with a focus on inclusiveness may assist some professionals to address this challenge, more effective approaches need to be developed. Establishing and maintaining relationships with community organizations may help address the challenge as various organizations (neighborhood groups, non - profits, faith - based, etc.) are often the gatekeepers for diverse communities. Diverse communities are less receptive to communicating with a stranger conducting research or soliciting participa tion in natural resource - related activities. Establishing relationships often requires meeting various community leaders face to face. Maintaining relationships entails continuing the communication through a variety of ways. Bringing interact ive and educat ional activities and holding specific events relevant for a particular community may help foster relationships. These activities may help to build trust and understanding between resource professionals and the community of interest. The key is to focus on building a mutually beneficial relationship. Conclusion: It is critical for the quality of trees, forests, and other envir onmental resources that diverse voices are sought, heard, and integrated into the decision - making process. Understanding eptions has the additional implication of understand ing their level of acceptance and support for management decisions (Ford et al., 2014). Furthermore, participating in forestry - 106 related research and activities can prom ote proenvironmental attitudes (Li u e t al., 2010) which is foundational for the success of sustainable forest management. Over the last century f orest management in the U.S. has begun to transition to more inclusiv ity. Despite some progress, there is still im portant work to be done for SFM to truly be considered inclusive The research presented here is one step on the path towards creating a more inclusive and diversified research and decision/policy - making process in the field of forestry (and for the natural resources field mo re broadly ). It presents a new approach to collecting data on the importance of trees and forests from a diverse audience and contributes to the scarce and growing literature examining racial and ethnic minorities, women, and the environment. Results exemplify the r elevancy for participation of women and nonwhite minorities highlighting the need to consider issues of justice and equity in research methodologies. 107 APPENDICES 108 APPENDIX I: Survey Instrument 109 Greetings from Michigan State University! The United States has committed itself to managing forests for long term benefits for everyone, and one important aspect of that is listening to residents about the ways in which they value trees and forests. We are very interested in hearin g from everyone, even if you do not think that your opinions about trees and forests are important. We can learn best about these issues by asking a variety of people to share their thoughts and opinions. You have been selected as part of a small sample so it is important to hear back from nearly everyone. By taking a few minutes to share your thoughts and opinions, you will be helping us out a great deal. The information you share with us will be used to better understand the different ways that trees an d forests are important to people and can be used to enhance forest management in areas near you and throughout the country. The return of this survey is your consent to participate in the research. Please return this survey using the postage - paid envelope in the next 10 days. The questions should take about 15 - 20 minutes to complete. Your responses are voluntary and will be kept confidential. Your answers will never be associated with your mailing address or your name. Please make sure that an adult (age 18 or older) in your household fills out the questionnaire. If you have any questions about this study, such as scientific issues, how to do any part of it, or to report an issue please contact the Principal Investigator Maureen McDonough at 126 Natural Resources Building, Michigan State University, East Lansing, MI 48824 , phone (517) 432 - 2293, email mcdono10@msu.edu. If you have questions about your role and rights as a research participant, or would like to register a complaint Human Resear ch Protection Programs, Kristen Burt, by phone: (517) 355 - 2180, fax: (517) 432 - 4503, email: irb@msu.edu, or regular mail: 202 Olds Hall, East Lansing, MI 48824. Upon receipt of your completed survey, you will be entered into a drawing to win one of four $25 Visa Gift Cards to thank you for your time. Sincerely, Dr. Maureen McDonough Dori Pynnonen Hopkins, M.S. Professor and Researcher Research and Doctoral Candidate Department of Forestry Department of Forestry 110 Importance of Trees and Forests i. How long have you lived in your current residence? ____________ MONTHS or ____________ YEARS ii. What do you like most about your neighborhood? iii. Did you grow up (ages 1 - 16) primarily in a(n): (check more than one, if applicable) Urban setting Suburban setting Rural setting iv. What do you like most about being outside? Nothing. I do not like being outside. v. Please check: I own my home. Other:_________________ I rent my home. Many people find trees and forests to be important for a variety of reasons. The following several sections of this survey are asking you about the different ways trees and forests are important or not important to you. 111 The first question is asking how tr ees and forests are important to you in relation to the environment or ecological function. Please describe the extent to which you agree or disagree with the following statements. Question 1 1 2 3 4 5 6 Strongly Disagree Neutral Strongly Agree Know 1. Trees provide clean air 2. Trees clean and filter water 3. Trees provide shelter for wildlife 4. Trees provide oxygen 5. Trees improve air quality 6. Trees influence temperature 7. Trees provide protection from weather 8. Trees are important for wildlife 9. Trees provide shade 10. Trees help prevent erosion 11. Trees are important for insects 12. Trees are important for the water cycle 112 The second question is asking about the ways in which trees and forests provide financial benefits to you or your community (or other communities). Please describe the extent to which you agree or disagree with the following statements. Question 2 1 2 3 4 5 6 Strongly Disagree Neutral Strongly Agree Know 1. Trees provide a monetary value 2. Trees and forests provide jobs 3. Trees and forests are important for the tourism industry Some people think that trees and forests are important culturally. For the third question, please describe the extent to which you agree or disagree with the following statements. Question 3 1 2 3 4 5 6 Strongly Disagree Neutral Strongly Agree Know 1. Trees and forests can be landmarks of a community 2. Trees and forests can tie a community together 3. Trees are important for community building Question 4 is asking about the ways in which trees and forests are important for learning and education. Please describe the extent to which you agree or disagree with the following statements. Question 4 1 2 3 4 5 6 Strongly Disagree Neutral Strongly Agree Know 1. Trees help us to learn about nature 2. Trees provide an educational place 3. Trees and forests are important for environmental education 113 The fifth question is asking about the ways in which trees and forests influence human health and a sense of well - being. Please describe the extent to which you agree or disagree with the following statements. Question 5 1 2 3 4 5 6 Strongly Disagree Neutral Strongly Agree Know 1. Trees and forests are relaxing 2. Trees and forests provide a place for privacy/solitude 3. Trees and forests provide a place for Some people find trees and forests to be important for their beauty. For Question 6, please describe the extent to which you agree or disagree with the following statements regarding the aesthetic importance of trees and forests. Question 6 1 2 3 4 5 6 Strongly Disagree Neutral Strongly Agree Know 1. Trees are pleasing to view 2. Trees add beauty to parks 3. Trees add beauty to neighborhoods Trees provide a variety of products. In response to Question 7, please describe the extent to which you agree or disagree with the following statements, regarding the importance of the products that come from trees and forests. Question 7 1 2 3 4 5 6 Strongly Disagree Neutral Strongly Agree Know 1. Trees are important for providing paper products 2. Trees are important for providing building materials 3. Trees are important for providing fuel 114 The eighth question is asking about your attachment to trees and forests. Please describe the extent to which you agree or disagree with the following statements. Question 8 1 2 3 4 5 6 Strongly Disagree Neutral Strongly Agree Know 1. I feel a connection to a specific tree or trees 2. Trees provide a feeling of connection or rootedness 3. Trees and forests are part of cultural identity 4. Trees serve as markers of history 5. I have a special memory of an individual tree or trees For some people, trees and forests have a spiritual and/or religious significance. For Question 9, please describe the extent to which you agree or disagree with the following statements regarding the spiritual importance of trees and forests. Question 9 1 2 3 4 5 6 Strongly Disagree Neutral Strongly Agree Know 1. Trees represent the balance of life 2. Trees are connected to religious or spiritual feelings for me 3. Trees and forests provide a connection to a Higher Power 115 The following question addresses some of the ways people have stated that they enjoy or interact with trees and forests. For Question 10, please describe the extent to which you agree or disagree with the following statements. Question 10 1 2 3 4 5 6 Strongly Disagree Neutral Strongly Agree Know 1. I enjoy the smells of trees 2. I enjoy the sounds of trees 3. I enjoy spring blooms 4. I enjoy touching trees 5. I enjoy sitting under trees 6. I enjoy autumn color change 7. I have fond memories related to trees and forests 8. I have planted trees 9. I have participated in a community project related to trees 10. I like to pray or meditate near trees For Question 11, please describe the extent to which you agree or disagree with the following statement regarding the recreational importance of trees and forests. Question 11 1 2 3 4 5 6 Strongly Disagree Neutral Strongly Agree Know 1. Trees and forests are important for outdoor recreation Please list the outdoor recreation activities that you currently do or have ever done around trees and forests: 116 We are also interested in how your interactions or experiences with trees and forests have changed over the course of your lifetime. For Question 12, please indicate if you agree or disagree with the following statements about the changes that have occurred since you were younger. Question 12 Disagree Agree No Difference Know 1. There are fewer jobs associated with trees and forests 2. There is an increase in pollution around trees and forests 3. There are more trees and forests 4. There are positive changes in forest policies or laws 5. I interact more with trees and forests 6. I interact with trees and forests differently The following question addresses concerns that some people have expressed about trees and forests. For Question 13, please describe the extent to which you agree or disagree with the following statements regarding possible concerns you may have about trees and forests. Question 13 1 2 3 4 5 6 Strongly Disagree Neutral Strongly Agree Know 1. I am concerned about forest degradation, or the conditions of trees and forests 2. I am concerned about environmental sustainability 3. I feel concerned about the ways trees and forests are managed 4. I am concerned about a lack of trees in urban areas 5. I am concerned about new developments affecting trees and forests 6. I am concerned with lost connections between people and forests 117 People sometimes have negative feelings about trees and forests. Question 14 asks about ways in which you might feel negatively about trees and forests. Please describe the extent to which you agree or disagree with the following statements. Question 14 1 2 3 4 5 6 Strongly Disagree Neutral Strongly Agree Know 1. Trees may damage my house 2. Trees may damage the sidewalk 3. There are always insects and spiders around trees 4. Trees may bring animal pests 5. I am allergic to trees or other plants near trees managed 7. I do not feel welcome in forests or areas where there are trees, such as parks 8. I have a safety concern when I am near trees or forests 9. I feel the positive benefits of trees are greater than the negative aspects Please add any additional thoughts or comments here: 118 Demographic Information For statistical purposes, we need to ask you a few demographic questions. Please remember that the information you provide is confidential! 15. What year were you born? 16. What is your gender? _____________________________ 17. Are you of Hispanic, Latino or Spanish descent? Yes -- If yes, from which country -- No 18. What is your ethnicity? American Indian Asian Black/African American White/Caucasian Other 19. What is your marital status? Single Divorced Other _______________________ Married Widowed 20. What is your highest degree or level of school completed? Did not complete high school Associate degree High School Diploma or GED Bachelor degree Some college, but no degree Graduate or professional degree Other 21. Please check the box that corresponds to your income for 2013. Less than $14,999 $25,000 - $34,999 $75,000 - $99,999 $15,000 to $19,999 $35,000 - $49,999 $100,000 - $149,999 $20,000 to $24,999 $50,000 - $74,999 $150,000 or more 22. Please check the box that best describes your current employment status: I work part time I am unemployed I work full time I am retired 23. We would like to follow up with the participants in this study. Would you be willing to participate in an interview or focus group session about this research? ( Your continued participation will enter you in an additional drawing for a $25 Visa Gift Card) If yes, please write your name and the best way to contact you: 1 9 119 A PPENDIX II: Pre - notice letter 120 Dear Lansing Resident, We are writing to ask your help with an important study being conducted by Michigan State University to understand the many ways in which trees and forests are important to people. The best way we have of learning about this topic is by asking all differen t kinds of people to share their thoughts and opinions. In the next few days you will receive a request to participate in this project by answering questions about your perceptions of how and why trees are (or are not) important to you personally, and to y our community. We would like to do everything we can to make it easy and enjoyable for you to participate in the study. We are writing in advance because many people like to know ahead of time that they will be asked to fill out a questionnaire. This rese arch can only be successful with the generous help of people like you. To say thanks, in exchange for your completed survey, you will be entered into a drawing to win a $25 VISA gift card. Several study participants will be randomly chosen to receive this gift card token of appreciation. I hope you will take 15 - 20 minutes of your valuable time to help us. Most of all, I hope that you enjoy the questionnaire and the opportunity to voice your thoughts and opinions about how trees and forests are important. Respectfully, Dr. Maureen McDonough Dori Pynnonen Hopkins, M.S. Professor & Research Researcher & Doctoral Candidate Department of Forestry Department of Forestry 121 APPENDIX III: I nformed Consent 122 Consent to Participate In An Interview Title: Just and equitable access: Developing and testing a methodology that represents diverse perceptions on the importance of forests. Description The objectives of this research study are to identify the ways in which trees and forests are personally and socially important to people, and what motivates citizens to participate in research about trees and forests. We are contacting specific groups of people that are often left unheard in the natural resource decision - making process to participate in an eval uative interview on this topic. Your input is crucial to the development of social indicators in sustainable forest management Risks and Benefits: There are no serious risks to you from participating in this interview (see confidentiality statement). One benefit is that your ideas will contribute to a better understanding of the importance of forests to all types of people as well as contribute to the use of more inclusive research methods. Your suggestions are necessary in order to improve sustainable n atural resource management. Time Commitment, Cost and Payments: The interview will take about 30 minutes to complete. There are no other costs to you for helping us with this study. If you choose, your name may be entered a second time into the drawing fo r one of $25 Visa Gift Cards. Confidentiality: Although we will record our discussion, we will not put your name on the audio file or transcript. The only information that will be on the file will be a code number, which will be stored in a separate locat ion from the interview material. Therefore, we do not believe that you can be identified. Your privacy will be protected to the maximum extent allowable by law. 123 Right to Withdraw: Participation in this study is voluntary. You may choose not to partici pate at all. Furthermore, you may refuse to answer certain questions. If you begin, you may discontinue your participation at any time. Contact Information: If you have any questions about this study, please contact the Principal Investigator Maureen McDo nough at 126 Natural Resources Building, Michigan State University, East Lansing, MI 48824, phone (517) 432 - 2293, e - mail mcdono10@msu.edu. If you have questions about your role and rights as a research participant, or would like to register a complaint abo ut this study, Programs, Kristen Burt, by phone: (517) 355 - 2180, fax: (517) 432 - 4503, email: , or regular mail: 202 Olds Hall, East Lansing, MI 48824. S tatement of Consent: I voluntarily agree to participate in the study. _________________________________________________ Signature........................................................................................ Date I also consent to be recorded for this study: _________________________________________________ Signature........................................................................................ Date _______________________________________________ Signature of Investigator:......... ................................................... Date 124 APPENDIX IV: Script for evaluative focus group/interviews 125 Script for Evaluative Focus Groups/Interviews Greet participant(s) . Thank you for agreeing to participate in an interview to evaluate the research with which you participated on the importance of trees and forests to people. Oftentimes, the vast majority of people that participate in natural resources - related research are white, middle - aged, middle - to upper - class males. However, the United States is an extremely diverse country and the research concerning natural resources management needs to reflect this diversity. Through our discussion today, I hope to learn more about what prompted you to participate in the survey research, what future research on this topic. The information we receive from you will be very useful to natura l resource policy - and decision - makers as well as those interested in more inclusive research methods. Before we begin, I need to have you read and sign a form giving your consent to participate and be recorded. The consent form describes the purpose of the interview, risks and benefits to you, time commitment, how we maintain confidentiality and your right to withdraw at any time. Your signed consent is required by law and enforced by Michigan State University. Collect signed consent form. Turn on the d igital recorders. What do you like about living in Lansing? Did you receive the survey on the importance of trees and forests in the mail? Or did you receive the surve y in - person? What motivated you to complete the survey on the importance of trees and forests to people? Have you completed other surveys in the past? Did you receive the surveys in the mail or in - person? How do you decide whether or not you will complet e a survey? Were there any aspects of the survey on the importance of trees and forests that were especially interesting to you? Were there any aspects that were especially disinteresting? Were there aspects that were especially relevant or irrelevant? H ow can survey research on topics such as the importance of trees and forests be improved? Are there particular methods that you think are more user - friendly than others? 126 p eople? Do you have any other questions or comments about this topic? Thank them for their time. 127 APPENDIX V: Non - response survey 128 If you agree to consent to participate in this study, I will begin the questions. I am going to ask you about a variety of ways that trees and forests may or may not be important to you. Please describe the extent to which you agree or disagree with the following questions. Question 1 2 3 4 5 6 Strongly Disagree Neutral Strongly Agree Know 1. Trees provide clean air 2. Trees clean and filter water 3. Trees provide shelter for wildlife 4. Trees and forests are important for their economic value, such as providing jobs or financial benefits. 5. Trees and forests have cultural importance, such as ties to a community. 6. Trees provide educational value, such as helping us to learn about nature. 7. Trees and forests provide a place for relaxing). 8. Trees add beauty to parks and neighborhoods. 9. Trees are important for providing paper products, building materials, and other resources. 10. I feel a connection to a specific tree, trees, or a particular forest. 11. Trees are connected to religious or spiritual feelings for me. 12. Trees are important for outdoor recreations. 129 REFERENCES 130 REFERENCES Agresti, A. and Finlay, B. 1997. Statistical methods for the Social Sciences, 3 rd Ed. Upper Saddle River, NJ: Prentice Hall. Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50, 179 - 211. Allison, T. Maria, (1992). Fostering cultural diversity: Problems of access and ethnic boundary maintenance. Proceedings of the Symposium on Social Aspects and Recreation Research, February 19 - 22, Ontario, California. USDA Forest Service Gen. Tec. Rep. PSW - 132. pp 22 - 24. Alreck, P.L., & Settle, R.B. (2004). The survey research handbook, 3 rd Ed. New York, NY: McGraw - Hill/Irwin Anthony, M., Knuth, B., & Lauber, T.B. (2004). Gender and citizen participation in wildlife management d ecision making. Society and Natural Resources , 17(5), 395 - 411. doi: 10.1080/08941920490430179 Arcury, T., & Christianson, E. (1990). Environmental worldview in response to environmental problems. Environment and Behavior , 22(3), 387 - 407. doi:10.1177/00139 16590223004 Arcury, T., Johnson, T., & Scollay, S. (1986). Ecologi cal worldview and environmental knowledge: The "New Environmental Paradigm." Journal of Environmental Education , 17(4), 35 - 40. doi:10.1080/00958964.1986.9941424 Babbie, E. (2005). The basi cs of social research , 3 rd Ed. Belmont, CA: Wadsworth. Bell, S.E., & Braun, Y.A. (2010). Coal, identity, and the gen dering of environmental justice activism in central Appalachia. Gender & Society , 24(6), 794 - 813. doi:10.1177/0891243210387277 Berland , A., Schwarz, K., Herrmann, D.L., & Hopton, M.E. (2015). How environmental justice patterns are shaped by place: Terrain and tree canopy in Cincinnati, Ohio, USA. Cities and the Environment, 8(1), article 1. Available at: http://digitalcommons.lmu.edu/ca te/vol8/iss1/1/ Bradley, S. H. (2011). Outdoor girls: Incorporating an ongoing connection with the outdoors into the prevention of negative body image in girls Woman's University, Denton, TX. 131 Brown, P. J. (1984). Benefits of outdoor recreation and some ideas for valuing recreation opportunities. In G.L. Peterson and A. Randall (Eds.), Valuation of wildland resource benefits (pp. 209 - 220). Boulder, CO: Westview Press. Burns, N., Schlozman, K.L. , & Verba, S. (2009). The private roots of public action: Gender, equality, and political participation. Cambridge, MA: Harvard University Press. Carroll, M.S., Lee, R.G., & McLain, R.J. (2005). Occupational community and forest work: Three cases from t he Pacific Northwest. In R.G. Lee and D.R. Field (Eds), Communities and forests: Where people meet the land, (pp 159 - 175). Corvallis, OR: Oregon State University Press. one/ connected): A Huu - ay - aht worldview for seeing forestry in British Columbia, Canada. Society & Natural Resources , 22(9), 789 804. doi: 10.1080/08941920802098198 Chavez, D.J. (1993). The wildland/urban interface: Hispani cs in the national forests. In. P.H.Gobster (Eds.), Managing urban and high - use recreation settings. (Gen. Tech. Rep. NC - 163, 107 - 110). St. Paul, MN: USDA Forest Service, North Central Forest Experiment Station. Chiesura, A. (2004). The role of urban parks for the sustainable city. Land scape and Urban Planning, 68(1) , 129 - 138. doi:10.1016/j.landurbplan.2003.08.003 Clark, T.W. (2002). The policy process: A practical guide for natural resource professionals. New Haven, CT: Yale University Press. Collins, K.S., Hall, A.H., & Neuhaus, C. ( 1999). U.S. minority health: A chartbook. The Commonwealth Fund. Retrieved July 28, 2012 from: http://www.commonwealthfund.org/usr_doc/collins_usminority.pdf Creswell, J.W. (2009). Research design: Qualitative, quantitative, and mixed methods approaches , 3 rd Edition. Thousand Oaks, CA: Sage. Curtin, R., Presser, S., & Singer, E. (2000). The effects of response rate changes on the index of consumer sentiment. Public Opinion Quarterly, 64(4), 413 428. doi: 10.1086/318638 Dandy, N. & Van Der Wal , R. (2011). Shared appreciation of woodland landscapes by land management professionals and lay people: An exploration through field - based interactive photo - elicitation. Landscape and Urban Planning , 102(1), 43 - 53. doi:10.1016/j.landurbplan.2011.03 .008 132 Derr, V. (2001). Growing up in the Hispano homeland: The interplay of nature, family, culture, (Doctoral dissertation ). School of Forestry and Environmental Studies, Yale University: New Haven. de Leeuw, W.D., Hox, J.J., & Dillman, D.A. (2008). International Handbook of Survey Methodology. New York, NY: Lawrence Erlbaum Associates. DeVault, M.L. (1999). Liberating Me thod: Feminism and Social Research. Philadelphia, PA: Temple University Press. Dietz, T., Kalof, L., & Stern, P.C. (2002). Gender, values, and environmentalism. Social Science Quarterly, 83, 353 - 364. doi: 10.1111/1540 - 6237.00088 Dillman , D.A., Smyth, J.D., & Christian, L.M. (2009a). Internet , mail, and mixed - mode surveys: The tailored design method, 3 rd Ed. Hoboken, NJ: John Wiley and Sons. Dillman, D.A., Phelps, G., Tortora, R., Swift, K., Kohrell, J., Berck, J., & Messer, B.L. (2009b) . Response rate and measurement differences in mixed - mode surveys using mail, telephone, interactive voice response (IVR) and the internet. Social Science Research, 39, 1 - 18. doi:10.1016/j.ssresearch.2008.03.007 DiSogra, C., Dennis, M., & Fahimi, M. (2010 ). On the quality of ancillary data availab le for address - based sampling. Proceedings of the American Statistical Association, Section on Survey Research Methods, 4174 - 83. Donovan, G.H. & Mills, J. (2014). Environmental justice and facto rs that influence participation in tree planting programs in Portland, Oregon, U.S. Arboriculture & Urban Forestry , 30(2), 70 - 77. http://actrees.org/files/Events/donovan_and_millsenvironmentaljustice.pdf Driver, B. L., Brown, J. P., & Peterson, G.L. (1991). Benefits of Lei sure . State College, PA: Venture Publishing. Driver, B. L., Nash, R., & Haas, G. (1987). Wilderness benefits: A state - of - knowledge review. Proceedings from the National Wilderness Research Conference: Issues, state - of - en, UT. Duerden, M. D., Taniguchi, S., & Widmer, M. (2012). Anteceden ts of identity development in a structured recreation setting: A qualitative inquiry. Journal of Adolescent Research, 27(2), 183 - 202. doi:10.1177/0743558411417869 Dutcher, D.D., Finley, J.C., Luloff, A.E., & Johnson, J.B. (2007). Conn ectivity with nature as a measure of environmental values. Environment and Behavior, 39(4): 474 - 493. doi:10.1177/0013916505283421 133 Feilzer, M.Y. (2010). Doing mixed methods research pr ag matically: Implications for the rediscovery of pragmatism as a research paradigm. Journal of Mixed Methods Research, 4(1), 6 - 16. doi:10.1177/1558689809349691 Feldman, S., Radermacher, H., Browning, C., Bird, S., & Thomas, S. (2008). Challenges of Recru itment and Retention of Older People from Culturally Diverse Communities in Research. Ageing and Society , 28(4), 473 - 493. doi:10.1017/S0144686X07006976 Flocks, J., Escobedo, F., Wade, J., Varela, S., & Wald, C. (2011). Environmental justice implications o f urban tree cover in Miami - Dade County, Florida. Environmental Justice, 4(2), 125 - 134. doi:10.1089/env.2010.0018 Flood, P. J. & McAvoy, L.H. (2007). Voices of my ancestors, their bones t alk to me: How to b alance US Forest S ervice r ules and r egulations with t raditional v alues and v ulture of American Indians. Human Ecology Review, Society of human Ecology Vol. 14, No. 1, pp 76 - 88. http://www.humanecologyreview.org/pastissues/her141/floodandmcavoy.pdf Fonow, M.M. & Cook, J.A. (1991). A look at the second wave of feminist e pistemology and methodology. In M.M. Fonow & J.A. Cook (Eds.), Beyond methodology: Feminist scholarship as lived research, Bloomington & Indianapolis, IN: Indiana University Press. Ford, R. M., Williams, K. J. H., Smith, E. L. , & Bishop, I. D. (2014). Beauty, belief, and trust: Toward a model of psychological processes in public acceptance of forest management. Environment and Behavior , 46(4), 476 506. doi:10.1177/0013916512456023 Frost, R. (1916). Tree at my window. In R. Fro st * L. Untermeyer (Eds.), Garber, M., & Arnold, R.M. (2006). Promoting the participation of minorities in research. The American Journal of Bioethics , 6(3): w14 - w20. doi:10.1080/15265160600686331 George, D. & Mallery, P. (2003). SPSS for Windows step by step: A simple guide and reference. 11.0 update (4 th ed.). Boston, MA: Allyn & Bacon. Balancing western economies with Native American spiritualities. Current Anthropology , 50 (4), 547 561. doi:10.1086/599069 Gordon, J., Barton, A., & Adams, K. (2013). An exploration of African American forest landowners in Mississippi. Rural Sociology, 7 8 (4), 473 - 497. doi: 10.1111/ruso.12014 Groves, R.M., & Peytcheva, E. (2008). The impact of nonresponse rates on nonresponse bias: A meta - analysis. Public Opinion Quarterly . 72(2): 167 - 189. doi:10.1093/poq/nfn011 134 Grumbine, R.E. (1994). What is ecosystem management? Conservation Biology. 8(1): 27 - 38 doi:10.1046/j.1523 - 1739.1994.08010027.x Gustafson, P. (2001). Meanings of place: Everyday experience and theoretical conceptualizations. Journal of Environmental Psychology , 21(1), 5 - 16. doi:10.1006/jevp.2 000.0185 Hamilton, E. (2012). Non - timber forest products in British Columbia: Policies, practices, opportunities, and recommendations. Journal of Ecosystems and Management, 13 (2), 1 24. http://jem.forrex.org/index.php/jem/article/viewFile/165/111 Hammi tt, W.E. (2000). The relation between being away and privacy in urban forest recreation environments. Environment and Behavior, 32 (4), 521 - 540. doi:10.1177/00139160021972649 Harding, S. (1991). lives. Ithaca, NY: Cornell University Press. Harding, S. (2001). Feminist standpoint epistemology. In M. Lederman and I. Bartsch (Eds.), (pp 145 - 168). The gender and science reader . London and New York: Routledge. Harshaw, H.W., & Tindall, D.B. (2005) . Social structure, id entities, and values: a network approach to understanding people's relationships to forests. Journal of Leisure Research, 37 (4), 426 - 449. http://socialnetworks.soci.ubc.ca/SocNets/DavidTindall/Files/Tindall%20 - %20Social%20Structure.pdf Hauru, K., Koskinen, S., Kotze, D.J., & Lehvävirta, S. (2014). T he effects of decaying logs on the aesthetic experience and acceptability of urban forests Implications for forest management. Landscape and Urban Planning 123, 114 - 123. doi:10.1016/j.landurbplan.2013.12.014 Heerwagen, J. H. & Orians, G.H. (1993). Humans, habitats, and aesthetics. In Ke llert, S.R. & Wilson, E.O. (Eds.). (pp 138 - 172). The biophilia hypothesis . Washington, D.C.: Island Press/Shearwater Books. approach. Journal of Mixed Methods Res earch, 2(4), 296 - 316. doi: 10.1177/1558689808321641 Hunter, L.M., Hatch, A., & Johnson, A. (2004). Cross - national gender variation in environmental behaviors. Social Science Quarterly , 85, 677 - 694. doi:10.1111/j.0038 - 4941.2004.00239.x 135 Iannacchione, V.G . (2011). The changing role of address - based sampling in survey research. Public Opinion Quarterly , 75(3), 556 - 575. doi:10.1093/poq/nfr017 Jayarante, T.E., and Stewart, A.J. (1991). Quantitative and qualitative methods in the social sciences: Current femi nist issues and practical strategies. In M.M. Fonow & J.A. Cook (Eds.), Beyond methodology: Feminist scholarship as lived research. Bloomington and Indianapolis, IN: Indiana University Press. Jennings, V., Johnson Gaither, C., Schulterbrandt Gragg, R. (2 012). Promoting environmental justice through urban green space access: A synopsis. Environmental Justice, 5(1), 1 - 7. doi:10.1089/env.2011.0007 come Journal of Forestry , 105(5), 257 - 265. http://www.srs.fs.usda.gov/pubs/28797 Jorgensen, B.S. & Stedman, R.C. (2001). Sense of place as an attitude: Lakeshore owners attitudes toward their properties. Journal of Environmental Psychology. 21(3), 233 - 248. doi:10.1006/jevp.2001.0226 Kaalund, V.A. (2004). Witness to truth: Black women heeding the call for environmental justice. In. R. Stein (Ed.), New Perspectives on Environmenta l Justice. (pp 78 - 92). New Brunswick, NJ: Rutgers University Press. Kals, E., Schumacher, D., & Montada, L. ( 1999). Emotional affinity toward nature as a motivational basis to protect nature. Environment and Behavior, 31(2), 178 - 2002. doi: 10.1177/00139 169921972056 Kamitsis, I., & Francis, A. J. P. (2013). Spirituality mediates the relationship between engagement with nature and psychological wellbeing. Journal of Environmental Psychology , 36 , 136 143. doi:10.1016/j.jenvp.2013.07.013 Kaplan, S. & Talbot, J. (1983). Psychological benefits of a wilderness experience. In I. Altman & J. Wohlwill (Eds.). Behavior and the natural environment. New York, NY: Plenum Press. Kaplan, S. (1987). Aesthetics, affect, and cognition: Environmental preference f rom an evolutionary perspective. Environment and Behavior, 19(1), 3 - 32. doi:10.1177/0013916587191001 Kaplan, S. (1995). The restorative benefits of nature: Toward an integrative framework. Journal of Environmental Psychology, 15 , 169 - 182. http://is.muni.cz/el/1423/jaro2014/HEN597/um/47510652/Kaplan__S._1995.pdf 136 Kaplan, R., & Kaplan, S. (1989). The experience of nature: A psychological perspective . New York, NY: Cambridge University Press. Kellert, S.R. (1984). Attitudes toward animals: A ge - related development among children. In M.W. Fox & L.D. Mickley (Eds.), Advances in animal welfare science 1984/85 (pp 43 - 60). Washington, DC: The Humane Society of the United States. Kellert, S.R. (1996). The value of life: Biological diversity and hum an society. Washington, D.C.: Island Press. Kellert, S.R. (1997). Kinship to mastery: Biophilia in human evolution and development. Washington, D.C.: Island Press/Shearwater Books. Kellert, S.R. (2012). Birthright: People and nature in the modern world. New Haven and London: Yale University Press. Kempton, W., Boster, J.S., & Hartley, J.A. (1995). Environmental Values in American Culture. Cambridge, MA: The MIT Press. Kil, N., Holland, S. M., & Stein, T. V. (2014). Place Meanings and Participatory Pl anning Intentions, Society & Natural Resources: An International Journal, 27 (5), 475 - 491. doi:10.1080/08941920.2013.842274 Kirsch, G.A. (1999). Ethical Dilemmas in Feminist Research . Albany, NY: State University of New York Press. Knap, N.E., & Propst , D.B. (2001). Focus group interviews as an alternative to traditional survey methods for recreation needs assessments. Journal of Park and Recreation Administration , 19(2), 62 - 84. http://js.sagamorepub.com/jpra/article/view/1577 Kovacs, Z.I., LeRoy, C.J., Fischer, D.G., Lubarsky, S., & Burke, W. (2006). How do aesthetics affect our ecology? Journal of Ecological Anthropology. 10, 61 - 65. http://scholarcommons.usf.edu/cgi/viewcontent.cgi?article=1049&context=jea Krueger, R.A. (1994). Focus Group Int erviews: A Practical Guide for Applied Research, 2 nd Edition. Thousand Oaks, CA: Sage. Krueger, R.A. & Casey, M.A. (2000). Focus Group Interviews : A Practical Guide for Applied Research, 3rd Edition. Thousand Oaks, CA: Sage. Kruger, L.E. (2003). Understa nding community - forest relations. (General Technical Report PNW - GTR 566). Portland, OR: U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station. 137 Kuhns, M.R., Bragg, H.A., & Blahna , D.J. (2002). Involvement of women and minorities in the urban forestry profession. Journal of Arboriculture, 28(1), 27 - 34. https://forestry.usu.edu/files/uploads/Research%20Pubs/JofA28(1)W&M1.pdf Kwak, N. & Radler, B. (2002). A comparison between mail and web surveys: response pattern, respondent profile, and data quality. Journal of Official Statistics, 18(2), 257 - 273. http://www.barold.com/www/JOS%20article.pdf LaVelle, J.P. (2001). Rescuing Paha Sapa: Achieving environmental justice by restoring the great grasslands and returning the sacred Black Hills to the Great Sioux Nation. Great Plains Natural Resources Journal, 40, 41 - 100. spiritual ties to forests. Proce eding of the American Society of Foresters 1999 National Convention (Portland, OR 11 - 15 Sep 1999), 452 - 457. Li, C., Wang, C., Liu, S., & Weng, L. (2010). Forest value orientations and importance of forest recreation services. Journal of Environmental Man agement, 91 (11), 2342 - 2348. doi: 10.1016/j.jenvman.2010.06.020 Link, M.W., Battaglia, M.P., Frankel, M.R., Osborn, L., & Mokdad, A.H. (2008). A comparison of address - based sampling (ABS) versus random digit dialing (RDD) for general population surveys. Public Opinion Quarterly, 72(1), 6 - 27. doi: 10.1093/poq/nfn003 Liu, J., Ouyang, Z., & Miao, H. (2010). Environmental attitudes of stakeholders and their perceptions regarding protected area - Journal of Environme ntal Management 91 (11), 2254 - 2262. doi:10.1016/j.jenvman.2010.06.007 Lohr, S. (1999). Sampling: Data and analysis. Pacific Grove, CA: Brooks/Cole Publishing Co. Lutts nature. Forest & Conservation History, 36 (4), 160 - 171. doi: 10.2307/3983677 Manfredo, M.J. & Driver, B.L. (1996). Measuring leisure motivation: a meta - analysis of the recreation preference scales. Journal of Leisure Research, 28 (3), 188 - 213. http://www.researchgate.net/publication/250614022_Measuring_leisure_motivation_a_ eta_analysis_of_the_Recreation_Experience_Preference_Scales Manning, R., Valliere, W., & Minteer, B. (1999). Values, ethics, and attitudes toward national forest management: An empirical study. Society and Natural Resources. 12:421 - 436. doi:10.1080/089419299279515 138 McCright, A.M. & Xiao, C. (2014). Gender and Environmental Concern: Insights from Rec ent Work and for Future Research. Society and Natural Resources , 27(10), 1109 - 1113. doi:10.1080/08941920.2014.918235 McDonough, M.H. (2003).Understanding the meaning and value of trees. In Kruger, L.E. (tech. ed.) Understanding community - forest relations . Gen. Tech. Rep. PNW - GTR - 566. Portland, OR: U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station. McDonough, M.H., Russell, K., Burban, L., & Nancarrow, L., (2003). Dialogue on diversity: Broadening the voices in urban and community forestry. NA - IN - 04 - 03, USDA Forest Service, Urban and Community Forestry Program, Northeastern Area. http://www.nrs.fs.fed.us/pubs/8567 McLean, C.A., & Campbell, C.M. (2003). Locating research informants in a multi - ethnic community: Ethnic iden tities, social networks and recruitment methods. Ethnicity and Health , 8(1), 41 - 61. doi:10.1080/13557850303558 McMillan, M.B., Hoban, T.J., Clifford, W.B., & Brant, M.R. (1997). Social and demographic influences on environmental attitudes. Southern Rural Sociology, 13(1), 89 - 107. http://www.ag.auburn.edu/auxiliary/srsa/pages/Articles/SRS%201997%2013/SRS%2019 97%2013%2089 - 107.pdf Meinig, D.W. (1976). The beholding eye: Ten versions on the same scene. Landscape Architecture, 66, 47 - 54. http://student.agsci. colostate.edu/rdfiscus/tenviews.pdf Meraji, S.M. (Narrator). (2015, July 12). Outdoor afro: Busting stereotypes that black people Weekend Edition Sunday. Washington DC: National Public Radio. Mertens , D.M. (2007). Transformative paradigm: Mixed methods and social justice. Journal of Mixed Methods Research, 1(3), 212 - 225. doi:10.1177/1558689807302811 Milbrath, L. (1984). Environmentalists: Vanguard for a New Society. Albany, NY: State University of Ne w York Press. Millennium Ecosystem Assessment (MEA). (2005). Ecosystems and human well - being: Synthesis. Washington, D.C.: Island Press. http://www.millenniumassessment.org/documents/document.356.aspx.pdf Mitchell, D., & Hobby, T. (2010). From rotati ons to revolutions: Non - timber forest products and the new world of forest management. BC Journal of Ecosystems and Management, 11 (1&2), 27 38. http://jem.forrex.org/index.php/jem/article/view/58/22 139 Moore, D.L. & Tarnai, J. (2002). Evaluating nonresponse error in mail surveys. In: Groves, R. M., Dillman, D. A., Eltinge, J. L., and Little, R. J. A. (eds.), Survey Nonresponse (pp. 197 211), New York, NY: John Wiley & Sons. Morrissey, J., & Manning, R. (2000). Race, residence and environmental co ncern: New Englanders and the White Mountain National Forest. Human Ecology Review, 7(1), 12 - 24. https://www.uvm.edu/parkstudieslaboratory/publications/Race_Residence_and.pdf Murphy, B.L., Chretien A.R., & Brown L.J. (2012). Non - timber forest products, maple syrup and climate change. Journal of Rural and Community Development, 7 (3), 42 64. Nakau, M., Imanishi, J., Imanishi, J., Watanabe, S., Imanishi, A., Baba, T., & Morimoto, Y. (2013). Spiritual care of cancer patients by integrated medicine in urb an green space: A pilot study. Explore: The Journal of Science and Healing, 9 (2), 87 - 90. doi:10.1016/j.explore.2012.12.002 Nightingale, A. (2003). A feminist in the forest: Situated knowledges and mixing methods in natural resource management. ACME: An International E - Journal for Critical Geographies, 2(1), 77 - 90. http://www.geos.ed.ac.uk/homes/rgroves/nightingalepub3.pdf Norton, C. L. (2010). Into the wilderness A case study: The psychodynamics of adolescent depression and the need for a holistic intervention. Clinical Social Work Journal, 38 (2), 226 - 235. doi:10.1007/s10615 - 009 - 0205 - 5 Norton, C. L., Tucker, A., Russell, K. C., Bettmann, J. E., Gass, M. A., Gillis, H. L., & Behrens, E. (2014). Adventure therapy with youth. Journal of Experiential Education, 37 (1), 46 59. doi:10.1177/1053825913518895 The Journal of American Culture, 35.1, 15 - 25. doi:10.1111/j.1542 - 734X.2011.00794.x Orians, G. H. (1980). Habitat selection: General theory and applications to human behavior. In J.S. Lockard (ed.). The evolution of human social behavior. New York, NY: Elsevier. Orians, G.H. (1986). An ecological and evolutionary approach to landscape aesthetics. In E. C. Penning - Rowsell & D. Lowenthal (eds). London, UK: Allen and Unwin. Ortman, J.M. & Guarneri, C.E. (2009). United States Population Projections: 2000 t o 2050 Online July 28, 2012. Retrieved from: http://www.census.gov/population/www/projections/analytical - document09.pdf 140 Parker, J.D. & McDonough, M.H. (1999). Environmentalism of African Americans: An analysis of the subculture and barriers theories. Environmental and Behavior, 31(2), pp. 155 - 177. doi:10.1177/00139169921972047 Patel, A., Rapport, D.J., VanDerlinden, L., & Eyles, J. (1999). Forests and societal values: Comparing scientific and public perception of forest health. Environmentalists. 19(3), 239 - 249. doi:10.1023/A:1026402812084 Patton, M.Q. (2002). Qualitative Research and Evaluative Methods, 3 rd Edition. Thousand Oaks, CA: Sage. Pearsall, H., and Pierce, J. (2010). Urban sustainability and environmental justice: evaluating the linkages in public planning/policy discourse. Local Environment, 14 (6), 569 - 580. doi:10.1080/13549839.2010.487528 Pemberton, R. (1985). I saw that it wa s holy: the Black Hills and the concept of sacred land. Law and Inequality: A Journal of Theory and Practice 3, 287 - 342. Perkins, H. (2010). Measuring love and care for nature. Journal of Environmental Psychology, 30 , 455 - 463. doi:10.1016/j.jenvp.2010.0 5.004 Pinchot Institute for Conservation. (2006). Advancing sustainable forest management in the United States. V.A. Sample, S.l. Kavanaugh, & M.M. Snieckus (Eds). Washington, D.C. http://www.pinchot.org/pubs/10 Poe, M.R., McLain, R.J., Emery, M. & Hurl ey, P.T. (2013). Urban forest justice and the rights to wild foods, medicines, and materials in the city. Human Ecology, 41, 409 - 422. doi:10.1007/s10745 - 013 - 9572 - 1 Power, T.M. (1980). The Economic Value of the Quality of Life. Boulder, CO: Westview Press, Inc. environmental justice. In. R. Stein (Ed.), New Perspectives on Environmental Justice (pp. 93 - 108). New Brunswick, NJ: Rutgers University Press. Propst, D.B., & Koesler, R.A. (1998). Bandura goes outdoors: Role of self - efficacy in the outdoor leadership development process. Leisure Sciences, 20(4), 319 - 344. doi:10.1080/01490409809512289 Pynnonen Hopkins, D., McDonough, M.H., and Blaha, R. (2014). A literature review on the importance of trees and forests. Report prepared for USDA Forests Service. Quantz, J.O. ( 1898). Dendro - psychoses. The American Journal of Psychology. 9(4), 449 - 506. doi:10.2307/1412187 141 Reed, P.S., Foley, K.L., Hatch, J., & Mutran, E.J. (2003). Recruitment of Older African Americans for Survey Research: A Process Evaluation of the Community and Church Based Strategy in the Durham Elders Project. The Gerontologist, 43(1), 52 - 61. doi:10.1093/geront/43.1.52 Reinharz, S. (1992) . Feminist Methods in Social Research . London & New York: Oxford University Press. Richard, T. & Burns, S. (1999). The Ponderosa Pine Forest Partnership: Forging new relationships to restore a forest. Office of Community Services, Fort Lewis College, Dur ango, Colorado. Roberts, N. S. & Chitewere, T. (2011). Speaking of justice: Exploring ethnic minority perspectives of the Golden Gate Recreation Area. Environmental Practice, 13 (4), 354 - 369. doi: http://dx.doi.org/10.1017/S1466046611000378 Rokeach, M. (1968). Beliefs, attitudes and values: A theory of organization and change. San Francisco, CA: Jossey - Bass. Rokeach, M. (1973). The Nature of Human Values . New York, NY: The Free Press. Rolston, H. III. (1998). Human values and natural systems. Society and Natural Resources, 1(1), 269 - 283. doi:10.1080/08941928809380658 Satterfield, T. (2001). In search of value literacy: Suggestions for the elicitation of environmental values. Environmental Values, 10(3), 331 - 359. http://www.jstor.org/stable/30301824 Saunders, C.D., A.T. Brooks, O.E. Myers, Jr. (2006). Using psychology to save biodiversity and human well - being. Conservation Biology, 20(3), 702 - 705. doi:10.1111/j.1523 - 1739.2006.00435.x Saxon, D., Garratt, D., Gilroy P., & Cairns C. (2003). Collecting data in the Information Age. Research in Education, 51. Shellman, A. (2014). Empowerment and experiential education: A state of knowledge paper. Journal of Experiential Education, 37 (1), 18 - 30. doi:10.1177/1053825913518896 Shin, W.S., Shin, C.S., & Y eoun, P.S. (2012). The influence of forest therapy camp on depression in alcoholics. Environmental Health and Preventive Medicine, 17 (1), 73 76. doi: 10.1007/s12199 - 011 - 0215 - 0 Shindler, B., L.A. Cramer. (1999). Shifting public values for forest management: Making sense of wicked problems. Western Journal of Applied Forestry, 14(1), 28 - 34. http://andrewsforest.oregonstate.edu/pubs/pdf/pub2465.pdf 142 Singer, E., van Hoewyk, J., & Maher, M.P. (2000). Experiments with incenti ves in telephone surveys. Public Opinion Quarterly, 64(2), 171 188. doi: 10.1086/317761 Smith, P.D., and McDonough, M.H. (2001). Beyond public participation: Fairness in natural resource decision making. Society and Natural Resources . 14(3), 239 - 249. doi:10.1080/08941920120140 Spartz, J. T. & Shaw, B. R. (2011). Place meanings surrounding an urban natural area: A qualitative inquiry. Journal of Environmental Psychology, 31 (4) 344 - 352. doi:10.1016/j.jenvp.2011.04.002 Stein, T.V. & Anderson, D.H. (2002). Combining benefits based management with ecosystem management for landscape planning: Leech Lake watershed Minnesota. Landscape and Urban Planning, 60(3), 151 - 161. doi:10.1016/S0169 - 2046(02)00053 - 1 Stein, T.V., Denny, C.B., & Pennisi, L.A. (2003). Using visitors' motivations to provide learning opportunities at water - based recreation areas. Journal of Sustainable Tourism, 11 (5), 404 - 425. doi:10.1080/09669580308667213 Stoop, I.A. (2004). Surveying Nonrespondents. Field Methods, 16(1), 23 - 54. doi:1 0.1177/1525822X03259479 Streiner, D.L. (2005). Finding our way: An introduction to path analysis. Canadian Journal of Psychiatry , 50(2), 115 - 122. Swearer, D.K. (1998). Buddhism and ecology: Challenge and promise . Earth Ethics 10(1) Center for Respect of Life and Environment. http://fore.yale.edu/religion/buddhism/ Sze, J. (2004). Gender, asthma politics, and urban environmental justice ac tivism. In. R. Stein (Ed.), New Perspectives on Environmental Justice (pp. 177 - 206). New Brunswick, NJ: Rutgers Uni versity Press. Tarrant, M.A., & Cordell, H.K. (2002). Amenity values of public and private forests:Examining the value attitude relationship. Environmental Management , 30(5), 692 703. doi:10.1007/s00267 - 002 - 2722 - 7 Tashakkori, A., & Teddlie, C. (1998). Mixed Methodology: Combining Qualitative and Quantitative Approaches. Thousand Oaks, CA: Sage. Taylor, A. F., Kuo, F.E. & Sullivan, W.C. (2002). View of nature and self - discipline: Evidence from inner city children. Journal of Environmental Psychology , 22 , 49 - 63. doi:10.1006/jevp.2001.0241 143 Taylor, D.E. (2002). Race, class, gender, and American environmentalism. Gen. Tech. Rep. PNW - GTR - 534. Portland, OR: U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station. http://www.fs.fed.us/pnw/pubs/gtr534.pdf Taylor , D.E. (2008). Diversity and the environment: Myth - making and the status of minorities in the field. Research in Social Problems and Public Policy , 15, 89 - 148. doi:10.1016/S0196 - 1152(07)15003 - 1 Taylor, D.E . (2011). The evolution of environmental justice activism, research, and scholarship. Environmental Practice, 13(4), 280 - 301. doi:10.1017/S1466046611000329 Teddlie, C., & Yu, F. (2007). Mixed methods sampling: A typology with examples. Journal of Mixed Methods Research, 1(1), 77 - 100. doi:10.1177/2345678906292430 Toossi, M. (2002). A century of change: The U.S. labor force, 1950 - 2050. Monthly Labor Review, 125(5), 15 - 28. doi:10.2307/41845363 Trigger, D., Mulcock, J. (2005). Forests as spiritually signif icant places: Nature, culture and 'belonging' in Australia. The Australian Journal of Anthropology , 16(3), 306 - 320. doi: 10.1111/j.1835 - 9310.2005.tb00313.x Tronstad, R., Teegerstrom, T., & Osgood, D.E. (2004). Role of technology for reaching underserved audiences. American Journal of Agricultural Economics , 86(3), 767 - 771. http://www.jstor.org/stable/3697820 Tuan, Y.F. (1977). Space and Place: The Perspective of Experience. Minneapolis, MN: University of Minnesota Press. Tyrväinen, L., Ojala, A., Korpela, K., Lanki, T., Tsunetsugu, Y., & Kagawa, T. (2014). The Journal of Environmental Psychology, 38 , 1 - 9. doi:10.1016/j.jenvp.2013.12.005 Ulrich, R. S., Simons, R. , Losito, B.D., Fiorito, E., Miles, M.A., & Zelson, M. (1991a). Stress recovery during exposure to natural and urban environments. Journal of Environmental Psychology, 11(3), 201 - 230. doi:10.1016/S0272 - 4944(05)80184 - 7 Ulrich, R.S., Dimbeg, U. & Driver, B.L. (1991b). Psychophysiological indicators of leisure benefits. In B. L. Driver, P. J. Brown, and G. L. Peterson (eds.). Benefits of Leisure (pp. 73 - 89). State College, PA: Venture Publishing. Ulrich, R. S. (1993). Biophilia, biophobia, and n atural landscapes. pp. 73 - 137 In S.R. Kellert & E.O. Wilson. The biophilia hypothesis . Washington, D.C.: Island Press/Sherwater Books. 144 United Nation Environmental Program (UNEP). (2007). Global Environment Outlook: Environment for Development, Biodiversity (Chapter 5) United Nation Environment Program, Nairobi, Kenya, 158 - 192. U.S. Census Bureau. (2013). USA Quickfacts . (Data file). Retrieved August 1, 2015, from: http://quickfacts.census.gov. U.S. Department of Agriculture Forest Service. (2 011). National Report on Sustainable Forests 2010. FS - 979. http://www.fs.fed.us/research/sustain/national - report.php U.S. Department of Agriculture Forest Service. (2005). FIA fact sheet series: Forest inventory and analysis sampling and plot design. Acc essed online, June 24, 2011: http://www.fia.fs.fed.us/library/factsheets/datacollections/Sampling%20and%20Plot%2 design.pdf Van Liere, K., & Dunlap, R. (1980). The social bases of env ironmental concern: A review of hypotheses, explanations, and empirical evidence. Public Opinion Quarterly , 44(18), 1 - 197. http://www.jstor.org/stable/2748427 Van Weiren, G. & Kellert, S. (2013). The origins of aesthetic and spiritual values in children's experience of nature. Journal for the Study of Religion, Nature & Culture, 7 (3), 243 - 264. doi:10.1558/jsrnc.v7i3.243 Vaske, J.J., Donnelly, M.P., Williams, D.R., & Jonker, S. (20 01). Demographic influences on environmental value orientations and normative beliefs about national forest management. Society and Natural Resources, 14(9), 761 - 776. doi:10.1080/089419201753210585 Vaske, J.J. & Donnelly, M.P. (1999). A value - attitude - beh avior model predicting wildland preservation voting intentions. Society and Natural Resources. 12(6), 523 - 537. doi:10.1080/089419299279425 Verchick, R. (2004). Feminist theory and environmental justice. In. R. Stein (Ed.), New Perspectives on Environmental Justice (pp. 63 - 77). New Brunswick, NJ: Rutgers University Press. Vining, J. & Tyler, E. (1999) Values, emotions, and desired outcomes r eflected in public responses to forest management plans. Human Ecology Review, 6(1), 21 - 34. http://www.humanecologyreview.org/pastissues/her61/61viningtyler.pdf Votrin, V. (2005). The orthodoxy and sustainable development: A potential for broader involv ement of the orthodox churches in Ethiopia and Russia. Environment, Development and Sustainability , 7(1), 9 21. doi:10.1007/s10668 - 003 - 5053 - 9 145 Waddington, D.G., & Velkoff, V.A. (2010). Projecting race and Hispanic origin for the U.S. Populationand an examination of the impact of net international migration. Working paper for United Nations Statistical Commission & Economic Commission for Europe Conference of European Statisticians. Retrieved online July 28, 2012: http://www.unece.org/fileadmin/DA M/stats/documents/ece/ces/ge.11/2010/wp.34.e.pdf Warner, K. (2002). Linking local sustainability initiatives with environmental justice. Local Environment , 7 (1), 35 - 47. doi:10.1080/13549830220115402 Wellman, J.D. & Propst, D.B. (2004). Wildland Recreati on Policy, 2 nd Edition. Malabar, FL: Krieger Publishing Co. White, D.D. & Hendee, J.C. (2000). Primal hypothesis: the re lationship between naturalness, solitude, and the wilderness experience benefits of development of self, development of community, and s piritual development (Proceedings No. RMRS - P - 15 - VOL - 3): USDA Forest Service. http://www.fs.fed.us/rm/pubs/rmrs_p015_3/rmrs_p015_3_223_230.pdf Whitney, G.G. & Upmeyerm M.M. (2004). Sweet trees, sour cir cumstances: The long search for sustainability in the North American maple products industry. Forest Ecology and Management , 200(1 - 3), 131 - 333. doi:10.1016/j.foreco.2004.07.006 Williams, D.R. & Vaske, J.J. (2003). The measurement of place attachment: Validity and generalizability of a psychometric approach. Forest Science, 49(6), 830 - 840. http://www.fs.fed.us/rm/pubs_other/rmrs_2003_williams_d001.pdf Wilson, E.O. (1984). Biophilia. Cambridge, MA: Harvard University Press. Winter, C. & Lockwood, M. (2004). The natural area value scale: A new instrument for measuring natural area values. Australian Journal of Environmental Management, 11(1), 11 - 20. doi:10.1080/14486563.2004.10648594 World Commission on Environment and Development (WCED). (1987). Our Common Future. Oxford: Oxford University Press. Yancey, A.K., Ortega, A.N., & Kumanyika, S.K. (2006). Effective recruitment and retention of minority research participants. Annual Review of Public Health , 27, 1 - 28. http://toniyancey.com/publications_files/aypub2006d.pdf Young, R.A. & Crandall, R. (1984). Wilderness and self - actualization. Journal of Leisure Research, 16 (2), 149 - 160.