"s lilt- aéduufi.’ . .. W, sun-nix. .. Swfirxjfl NE a I 0 1.6.15. . i... .l. S.‘ 3:. .. .2 3.1! .x 7 . nth?” ,Gtici J.’ 4”“. if: : ire ; n2. _ .ur Ba!!! - 21:393.. ‘ {.lfiirx‘! 2.2.3.. .0 I . .0??? t. '1 Ian“ 0) I. ¢b . .. ue:.A..v1..hr:oIl‘iIA1 L Mun-an.” ~ , . V .V ‘l ‘75. Luzuw . . . , . y) 3 r .‘Iifl‘ezs. . . ‘ ‘ n»?.fl.wl»~wfi.1§ . ‘0‘ fix... I. 1., 1. 3.. 1 .7. :3. _ . , . . . ...«.»\._ owfiot‘ ““1958 THE”? D 43>? lIBRARY Michigan State University This is to certify that the dissertation entitled A STUDY OF THE OPINIONS OF MICHIGAN RESIDENTS TOWARD GOVERNMENT BENEFITS FOR OLDER ADULTS presented by Sally Anne Rypkema has been accepted towards fulfillment of the requirements for Ph.D. mgmem Social Science LEM Major pi'ofessor ‘ Date J/--’ 7— Q! MS U is an Affirmative Action/Equal Opportunity Institution 0-12771 . ( ‘77 l 1‘0 "(f / ("o t f' (7/ PLACE IN RETURN Box to remove this checkout from your record. To AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE thy! I“) umm M1339 E’éh‘) 4. J 6/01 cJClRC/DateDuepes—p. 1 5 A STUDY OF THE OPINIONS OF MICHIGAN RESIDENTS TOWARD GOVERNMENT BENEFITS FOR OLDER ADULTS By Sally Anne Rypkema A DISSERTATION Submitted to Michigan State University In partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Social Work 2001 UMI Number: 3036742 Copyright 2001 by Rypkema, Sally Anne All rights reserved. ® UMI UMI Microform 3036742 Copyright 2002 by ProQuest Information and Learning Company. All rights reserved. This microforrn edition is protected against unauthorized copying under Title 17, United States Code. ProQuest Information and Learning Company 300 North Zeeb Road PO. Box 1346 Ann Arbor, MI 48106-1346 ABSTRACT A STUDY OF THE OPINIONS OF MICHIGAN RESIDENT S TOWARD GOVERNMENT BENEFITS FOR OLDER ADULTS By Sally Anne Rypkema There have been conflicting media claims and research results regarding tensions between age groups surrounding the issue of generational equity and the increasing allocation of society’s resources for programs for older adults. However, there are those who argue that these tensions do not exist on a societal level because of generational solidarity experienced within the family. The purpose of this study is to examine the similarities and differences between age groups regarding opinions toward programs that primarily benefit older adults. The study also investigates the role of sociodemographic factors such as sex of respondent, level of education, race, and political party affiliation. Other studies have found a significant relationship between these variables and support for such programs. Unique to this research, the study looks at the role of caregiver as a measure of generational solidarity in predicting support for programs benefiting older individuals. Data from three State of the State Surveys, designed to assess the opinions of the residents of Michigan and conducted in 1995, 1998 and 1999 by the Institute for Public Policy and Social Research at Michigan State University, are used in the study. The surveys conducted in 1995 (n = 1,133) and 1998 (n = 950) asked respondents to agree or disagree with the statement “too many tax dollars are going to support older people.” The 1999 survey (11 = 950) assessed respondents’ opinions regarding the Medicare program by asking if the age of eligibility for Medicare should be increased, if Medicare should be mllmg Io . was of M Results for older at Ital} to su were mixer poem cost taxes for dr tlr age of c msmg the a Wining 5| ”Dams be inconclmne Slimline 0; SOSSS, m3 “Ill be r Aliment“: may : - mm ”Will should be expanded to include prescription drug coverage, if respondents would be willing to pay higher taxes to finance prescription drug coverage, and if out-of-pocket costs of Medicare should be increased. Results indicate no relationship between age of respondent and support for programs for older adults in 1995; however, in 1998, younger respondents were significantly less likely to support these programs. Results regarding age and opinions regarding Medicare were mixed: there was no relationship between age and support for increasing out-of- pocket costs for Medicare, for adding prescription drug coverage, or for paying higher taxes for drug coverage. However, age was significant in predicting support for raising the age of eligibility for Medicare, with younger respondents more likely to support raising the age. Sex of respondent and political party affiliation were both significant in predicting support, as women and Democrats were consistently more likely to favor programs benefiting older adults. Results for income, race, and level of education were inconclusive. It was expected that those providing care to older persons would be more supportive of elder programs, and this proved to be the case for 8088-15 but not for 8088-3. The inconsistencies in support among age groups for programs benefiting older persons bear watching. As the population of the United States continues to age and the baby boom generation retires and becomes eligible for Medicare and Social Security, there will be more of a strain on the system and existing resources may be inadequate. As American society becomes increasingly diverse, competition for resources may become grounded in age, class, race and political ideology. There is also concern that marital disruption may erode support based on family solidarity, as changing marital patterns negatively alter the family’s role as a support network. Copyright by SALLY ANNE RYPKEMA 2001 DEDICATION To Tim Flnishi créfl- and DOI hate 3( F151 1 and [cchnlc me vhcnél filtered l 5 Imus: a ndentoocl along the W: perspeclh e ACKNOWLEDGMENTS Finishing this dissertation has been one of the most difficult tasks of my academic career, and it is clear to me why so many do not complete this process. I know I would not have achieved this goal without the help of many persons important to my life. First, I must thank my husband, Tim, whose love, support, strength, infinite patience, and technical expertise helped me accomplish the many tasks involved. He never stopped believing I could do this, even when I doubted my abilities most. He heartened me whenever I became discouraged and gently nudged me toward my goal whenever I faltered I shall be forever grateful. I must also thank my mother and my brother Jim, who, although neither really understood what I was up to all this time, have stood by me and offered encouragement along the way. Jim’s unfailing sense of humor has helped me keep everything in perspective. My mother and brother, in concert with Tim, have been my very own cheerleading squad My dissertation committee, composed of advisor Diane Levande, and faculty Ellen Strommen, Margaret Nielsen, and Barbara Ames, have stood by me and offered their constructive criticisms, insights, support, and encouragement. Diane Levande has read this document more times than I can count, and her willingness to serve and determination to get me through this process deserve special recognition and special thanks. Gary Anderson, director of the School of Social Work, and Rena Harold, coordinator of the doctoral program have not only shown their support but have also helped me manage the bureaucracy of the university. Other staff members and students have offered vi their assl: moviedg ammg lma [Million 1 mean Mann their assistance. One other doctoral student, Peg Whalen, helped immeasurably with her knowledge of research and statistics, and I am very grateful for our association, her encouragement, and her enthusiasm for my work. I must also thank other friends, neighbors, and family members, too numerous to mention individually, who also have offered their help, applauded my efforts, expressed interest in my project, and stood by me throughout this entire process. I am blessed to have so many special people in my life. vii ldclhe September were dyr'rré sars the ne World Tn suddenly Yet 0 finally ea conclusr’c A fin HOWexe] months thi5 ilme Social 5 51155:de aSseCuI PREFACE I delivered one of the last drafts of this dissertation to my advisor on the morning of September 11, 2001. While we sat in her office and chatted about next steps, thousands were dying in the worst incident of terrorism ever to occur. Later in the morning, as I saw the news reports and watched the horrific images of jet planes slamming into the World Trade Center, my dissertation, which had become the pivotal concern in my life, suddenly seemed meaningless. Yet our lives and personal concerns continue, and I have finished the dissertation and finally earned my doctorate. However, it may be instructive to reconsider the conclusions reached here in light of the events of recent weeks. A final statement in the last chapter asserts that Social Security is financially sound. However, the economic reality of the United States has changed drastically in the last months. We appear to be in a recession, the depth and severity of which is not clear at this time, although economic reports are discouraging. Furthermore, at least some of Social Security surplus funds are being used to help finance recovery efforts and to subsidize the military offensive in Afghanistan Thus, the future of Social Security is not as secure as it appeared to be a few weeks ago. The political agenda also has changed We are caught up in war and threats of terrorism. At the same time, Congress is not in session because a letter containing anthrax was sent to a senate ofiice building and numerous areas must be tested for contamination. There has been no mention in the last few weeks of Medicare and prescription drugs and other benefits that would help older persons, in particular, those older adults who are financially at risk, a topic of major concern prior to September 11. viii Our reason lndltld‘t has blOL and my: friends a warm marked i Our psychological reality has undergone a drastic shift. Many of us are, with good reason, very anxious and worried about recent events as well as their implications for our individual and collective futures. However, there also have been reports that the terror has brought us closer to one another. Parents are hugging their children tighter, husbands and wives no longer take their marriage for granted, and other family members and friends are more sensitively attuned to each other. There have been many instances of individuals going out of their way to help one another. In this sense, there has been a marked increase in solidarity among family members and friends, and even between complete strangers. ix TABLE OF CONTENTS LIST OF TABLES ............................................................................................................ xii CHAPTER 1 THE PROBLEM .......................................... r ....................................................................... 1 Need and Importance of the Study .....'A:....::, ................................................................... 3 Statement oftheProblem......................; ........................................................................ 7 Definition of Terms ............................... ’ ........................................................................ 8 General Hypotheses ofthe Study ...................................................................... 10 ,; Limitations of the Study .............................................................................................. 11 Background Theory .............................. f ...................................................................... 13 Importance of the study for Social Work .4}... ............................................................. 16 Overview ...................................................................................................................... 18 CHAPTER 2 REVIEW OF THE LITERATURE ................................................................................... 20 Images of older adults in the United States ................................................................. 20 The Generational Equity Debate ................................................................................. 25 Intergenerational Solidarity ......................................................................................... 39 Intergenerational solidarity: Intergenerational transfers ............................................. 43 Intergenerational Solidarity: Caregiving .................................................................... 49 Generational Equity, Entitlements and the Family ...................................................... 55 Summary ...................................................................................................................... 58 CHAPTER 3 METHODOLOGY ............................................................................................................ 60 Sample Selection and Characteristics .......................................................................... 64 Introduction ...................................................................................................... 64 Sampling .......................................................................................................... 65 Sample Weights ............................................................................................... 66 Sampling Error ................................................................................................. 67 Sample Characteristics: SOSS-3 ..................................................................... 68 Sample Characteristics: SOSS-15 ................................................................... 69 Sample Characteristics: SOSS-18 ................................................................... 70 Comparing the Samples ................................................................................... 71 Data Collection Procedures ......................................................................................... 74 Hypotheses ................................................................................................................... 75 Statistical Analysis ....................................................................................................... 78 Summary ...................................................................................................................... 80 CHAPTER 4 ANALYSIS OF THE DATA ............................................................................................. 81 Data Analysis ............................................................................................................... 81 SOSS-3 and 8088-15 .................................................................................................. 82 Age ................................................................................................................... 82 Sex of Respondent ........................................................................................... 84 X CHAPl SLIM Con Political Party Affiliation ................................................................................. 85 Income ............................................................................................................. 87 Caregiver Status ............................................................................................... 88 Race ................................................................................................................. 89 Education Level ............................................................................................... 91 Elders Live with Children ................................................................................ 91 Type of Elder Care ........................................................................................... 92 SOSS-18 ...................................................................................................................... 94 Age ................................................................................................................... 94 Sex of Respondent ........................................................................................... 95 Political Party Affiliation ................................................................................. 97 Income ........................................................................................................... 100 Race ............................................................................................................... 101 Education Level ............................................................................................. 102 CHAPTER 5 SUMMARY, CONCLUSIONS AND IMPLICATIONS ................................................ 108 Summary .................................................................................................................... 108 Conclusions ................................................................................................................ 109 Age ................................................................................................................. 1 10 Sex of Respondent ......................................................................................... 113 Political Party Affiliation ............................................................................... 115 Income ........................................................................................................... 116 Status as Caregiver ........................................................................................ 118 Race ............................................................................................................... 118 Level of Education ......................................................................................... 119 Older Adults Living with Adult Children ...................................................... 120 Implications for Social Work ..................................................................................... 120 Implications for Further Research ............................................................................. 123 BIBLIOGRAPHY ............................................................................................................ 126 uh u.' .,. '. LIST OF TABLES TABLE 1A Too many tax dollars are going to support older peeple by Age groups ........................... 83 TABLE 18 Too many tax dollars are going to support older people by Age groups ........................... 83 TABLE 2A Too many tax dollars are going to support older people by Sex of respondent ................ 84 TABLE ZB Too many tax dollars are going to support older people by Sex of respondent ................ 85 TABLE 3A Too many tax dollars are going to support older people by Political party ....................... 86 TABLE 38 Too many tax dollars are going to support older people by Political party ....................... 87 TABLE 4A Too many tax dollars are going to support older peeple by Do you provide care to an older relative? .................................................................................................................... 88 TABLE 48 Too many tax dollars are going to support older people by Providing elder care ............. 89 TABLE 5A Too many tax dollars are going to support older people by Race ..................................... 90 TABLE 5B Too many tax dollars are going to support older people by Race ..................................... 90 TABLE 6 Educational Level Mean Years and Standard Deviations ................................................. 91 TABLE 7 Too many tax dollars are going to support older people by Elders who cannot live alone should live with their children ........................................................................................... 92 TABLE 8 Too many tax dollars are going to support older people by Type of care you would prefer ........................................................................................................................................... 93 TABLE 9 Raise age of eligibility for Medicare by Age group .......................................................... 95 xii u. TAB‘. Incrcz TAB'L Add p1 TABU R315: 1 TABLE Add pr: TABLE Would } parry TABLE R3156 [Ill TABLE Increase TABLE I YWS Ofl TABLE 10A Increase/decrease Medicare’s out-of-pocket costs by Sex of respondent ......................... 96 TABLE 10B Add prescription drug coverage to Medicare by Sex of respondent ................................. 97 TABLE 10C Raise the age of eligibility of Medicare by Sex of respondent .......................................... 97 TABLE 11A Add prescription drug coverage to Medicare by Political party ........................................ 98 TABLE 118 Would you be willing to pay higher taxes to add Rx coverage to Medicare by Political party ................................................................................................................................... 99 TABLE 11C Raise the age of eligibility for Medicare by Political party ............................................. 100 TABLE 12 Increase/decrease Medicare’s out-of-pocket costs by Race ............................................ 101 TABLE 13 Years of education: Means and Standard Deviations ...................................................... 103 xiii In light tensions sur resources it similarities support for additional : programs 1 discussion theoretical “*6 or in most in dISs‘ZQSCS 3 Of thSc II is “Wis CHAPTER 1 THE PROBLEM In light of conflicting media claims and research results regarding intergenerational tensions surrounding generational equity and the increasing allocation of society’s resources for programs for older adults, the proposed research is designed to examine similarities and differences between young, middle-aged, and older adults regarding support for such programs at two different points in time, using secondary data An additional set of secondary data examines support between age groups for additional programs for older adults. This chapter provides background for the study, including a discussion of the theme of intergenerational relations, followed by a review of the theoretical framework. The population of the United States is aging. Life expectancy continues to increase in most industrialized countries because of improved health care, control of infectious diseases and better nutrition (Schaie & Willis, 1996). In the United States the percentage of those individuals over 65 was less than 5% in 1900, but it grew to 12.5% in 1990 and is expected to reach 22.9% by the year 2050 (Schaie & Willis, 1996). Changes in fertility rates have had a profound effect on the age distribution of the population in the United States. While cohorts between the ages of 60 and 80 are quite small, they are followed by the so-called ”baby boom" generation. The baby boomers represent a period of high fertility rates, followed by much smaller cohorts that are sometimes referred to as the "baby bust" generation (Schaie & Willis, 1996). The increase in life expectancy and the changes in fertility rates have a significant effect on funding for programs designed for older adults, such as Social Security and Medicare. The dependency ratio is a numerical measure of the economic burden imposed on those gr: retirees and little compe different pn smaller cob baby boom pan taxes 3 ofthe mpul lSclar’e 8; V The film} “(III force c Wilded for [his contract lle Workers Witter as I lift ll'orlters 3030, In 01h: 3.3 Work 913 h “Otters and r One Of the WIDE-age a 1,. “gammy, on those gainfully employed who must support individuals not in the work force, such as retirees and dependent children (Moody, 1994). While current aged cohorts experience little competition for support and services, the "baby boom generation faces a very different prospect" (Schaie & Willis, 1996, p. 8) as they will have to depend on much smaller cohorts of younger adults for support. The generation immediately following the baby boomers is unusually small (Morris, 1996). The percentage of the population that pays taxes and contributes to old-age programs will be shrinking, just as the percentage of the population receiving those benefits will be increasing in substantial numbers (Schaie & Willis, 1996). The ”implicit contract" with Social Security is that the generation currently in the work force contributes to benefits for those currently retired and can expect to be provided for in turn by the following generation of workers (Morris, 1996). However, this contract may be in danger as succeeding generations are ever smaller. As of 1950, 100 workers provided the payroll taxes to finance benefit payments for 6.1 retirees; however, as life expectancy increased the ratio of retirees rose sharply to 24.3 for every 100 workers in 1970 (Morris, 1996). The ratio becomes 42.7 retirees per 100 workers in 2030. In other words, while 16 workers supported one retiree in 1950, there will be only 2.3 workers to support each retiree in 2030 (Morris, 1996). The 2.3 to 1 ratio between workers and retirees is expected to continue until 2060. One of the major concerns of population aging is that it will cause conflicts between working-age adults and older people (F oner, 2000). Thus, intergenerational relations increasingly have become a topic of interest to social science researchers because of the larger numbers of older adults as well as the growth in the scope and the cost of programs for older persons (Hareven, 1996). Since the family is the primary arena for generational ...... relation general mterger policies other be consider ‘M is also n the mere 0:” genera question ICSOUICCS The adrotate older pen retire~ Tn [7" ma} ‘5 Williams relationships, one must also focus on this microsystem. Both social scientists and the general public have expressed concern with the state of the American family and intergenerational relationships within the family (Hareven, 1996). At the same time, policies and programs at the macrosystem level that support financial independence and other benefits for older adults, such as Social Security and Medicare, must also be considered. Need and Immrtance of the Study While there is concern about intergenerational relationships within the family, there is also interest in relationships between generations on a macrosocial level. Concern with the increasingly higher concentration of older adults in the population and the problems of generational assistance in later life have led to a debate over generational equity. The question under debate is as follows: Is the United States allocating too large a share of its resources to older adults at the expense of younger generations? The generational equity discussion centers around two positions. Those who advocate generational equity criticize the share of government resources consumed by older persons and believe that an even larger share will be consumed as baby boomers retire. They argue that it will not be possible to support retired boomers at levels enjoyed by today’s retirees without putting an enormous financial burden on younger generations (Williamson & Watts-Roy, 1999). Generational equity advocates further believe that each generation should provide for itself and should not have to support another generation, especially if it means a younger generation will not have the same level of support later in life. Thus, they argue that older Americans are getting more than their fair share of federal government resources, a problem that is only going to get worse as the nation continues to age (Williamson & Watts-Roy, 1999). rejet {EC-Ii leste. Those on the other side of the debate, the generational interdependence advocates, reject the claim that each generation should provide for itself and argue that it is impossible for succeeding generations to he promised a standard of living during retirement agaiequal tol that of previous generations (Williamson & Watts-Roy, 1999). Instead, when demographic and/or economic factors make it difficult to provide for the retirement needs of older persons, both generations should share the burden Additionally, the authors point to the fact that any cuts in older adult benefits would have an adverse effect on adult children Because of the interdependence between generations at the family level, cuts in programs aimed at older adults would mean that many adult children would have to support their parents financially or at least supplement their Social Security and Medicare benefits (Williamson & Watts-Roy, 1999). Generational interdependence advocates emphasize the benefits the different generations offer each other as many transfers take place, not only within the family but also within the larger society. Older adults are not merely a drain on the nation’s resources since they provide income, childcare, psychological support, and advice to ypunger generations. In addition, increasing numbers of older persons have become suiiiigate parents to their grandchildren, and many older adults also make major artistic, intellectual, leadership, and volunteer contributions to society (Bengtson, 2001; Williamson & Watts-Roy, 1999). Furthermore, as of 1993, 16% of older adult men and 8% of older adult women were labor force participants (Hobbs & Damon, 1996). Labor force participation tends to decline with advancing age, however, among men aged 65 to 69, 28% were in the labor force as of 1990 and the Bureau of Labor Statistics projects a slight increase in labor force participation of men aged 60 to 64 through 2005 (Hobbs & Damon, 1996). _'W 1'- Wn: 111 read: limited and the that A overly : in term mlysi effecnt the pro £0}ch WEEK cannot Ind. While generational interdependence advocates have presented their views primarily in academic journals and books published by academic presses, and thus have had a limited audience, generational equity proponents have, for the most part, used the media and the popular press to make their case. Marmor, Cook, and Scher (1997) note that what Americans learn about social policy from the television and print media tends to be overly simplistic, confuses the “anecdote with social fact” (p. 195), and presents the issue in terms of crisis or impending doom. Unfortunately, this tendency toward unreasoned analysis and discussion leads to a loud but superficial public debate that fails to promote effective policy and instead divides the public as to the source as well as the solution to the problem (Marmor et al., 1997). Framing the issue as a crisis is a strategy used by generational equity advocates to force a major policy change that would severely cut programs for older persons, and they argue that the urgency of the problem means it cannot be solved in any way short of deep cuts (Williamson & Watts-Roy, 1999). Individuals on both sides of the debate have asserted that growing concerns over the rising costs of Medicare and Social Security and the increasing percentage of the federal budget that finances programs for older adults have caused conflict to develop between the generations as resentment among younger cohorts increases. Marmor et al. (1997) claim that the generational equity debate has bred conflict between some groups, but they argue that any conflict is due to a misunderstanding of Social Security along with the political efforts of those eager to cut benefits in and expenditures for the nation’s social insurance program. Minkler (1987) asserts that media coverage of the generational equity debate has fostered a negative image of older adults as a rich, greedy and powerful voting bloc that exploits younger generations to finance its retirement and that such media coverage has increased resentment toward the older persons on the part of younger cohorts. To d conflict W. State Sum (lPPSRl at relationshi; erosion of: adults (18-: aged (40-5‘. 31?: someltl these age gr ditidcd to c Mention gfnerationa} Medicare '1 we“? the The 53, the Stare Sur :anded 10 i also asked 1' f befittir {of the cohorts. To date a number of studies have found very little evidence of intergenerational conflict within American society. By analyzing more recent data from the State of the State Surveys (SOSS) conducted by the Institute for Public Policy and Social Research (IPPSR) at Michigan State University, the proposed study intends to examine the relationship between age groups at two semrate times (1995 and 1998) and to test for any erosion of support by looking for differences between the self-reports of early young adults (18-24 years of age) and later young adults (25-39 years of age), as well as middle- aged (40-59 years of age) and older adults (age 60 and older). Although the age divisions are somewhat arbitrary, they fall within the parameters that are generally used to define these age groups (Schaie & Willis, 1996). However, the young adult group has been divided to capture the opinions of the youngest adults, sometimes referred to as “Generation X” (Williams, Coupland, Folwell, & Sparks, 1997), and who, according to generational equity advocates, are least likely to benefit from Social Security and Medicare. The older adult group has been expanded to include those 60-64 in order to increase the number of subjects for more meaningful analysis. The same age ranges will be used in an analysis of a third and more recent State of the State Survey (1999) that takes the issue a step firrther and asks if Medicare should be expanded to provide prescription drug coverage. Respondents in the 1999 survey were also asked if they would be willing to pay increased taxes to pay for this additional benefit for the older persons, if the age of eligibility for Medicare should be increased and if older adults should pay a larger amount of out-of-pocket costs for Medicare. Torres-Gil (1992) argues that age consciousness and age identification are increasing and will ultimately result in age-identified groups competing for scarce resourccf relations] ms jdua gates isl more Imp oflhe Sim programs The in between Cl anum’l‘er ( bem'een tlt budget that that there 1: Medicare a. for insIanC< generations Day, I 993; l KtOdebeek programs the the opinions r“tie-time so; nmrlannes a: ”7183th gate UT resources. However, according to Bengtson (2001) multigenerational bonds, relationships across more than two generations, “are becoming increasingly important to individuals and families in American society” (p. 2). As the population of the United States is characterized by greater age-diversity, relations between age groups will be more important than ever. The proposed study will contribute to a better understanding of the similarities and differences between age groups regarding opinions toward programs that primarily benefit older adults. /S_tajement of th_e_: Problem The most significant problem with the generational equity issue is the disparity between claimsmade by proponents of generational equity and the empirical results from a num’beryof studies. Generational equity proponents and others argue that conflict exists between the generations because younger adults resent the proportion of the federal budget that is spent on programs benefiting older persons. However, most studies show that there is support among all age groups for programs such as Social Security and Medicare and that younger cohorts do not resent the benefits allocated for older adults. For instance, while Rosenbaum and Button (1993) found some tension between generations in their study of Florida residents, other researchers (Cook & Barrett, 1988; Day, 1993; Logan & Spitze, 1995; Ponza, Duncan, Corcoran, & Groskind, 1988; Rhodebeck, 1993) have not found a relationship between age and support for government programs that benefit older adults. The purpose of the proposed research is to examine the opinions of young, middle-aged, and older adult residents of the state of Michigan regarding support for programs for older adults in order to better understand the similarities and differences between age groups. Additionally, the proposed study will investigate the role of sociodemographic variables such as sex of respondent, level of education, race and political party affiliation as well as measures of solidarity, such as provision of caregiving, in shaping opinions toward programs that benefit older adults. This study is unique 123.23% will examine the responses to the same question at two different points in time. Respondents were asked in 1995 and then again in 1998 to indicate the extent to which they agreed or disagreed with the statement “Too many tax dollars are going to support older people.” This study also differs from others i@t it will assess the responses of caregivers to the same statement. Definition of Terms “Generational equity” is defined as the idea that each generation should be responsible for itself and not responsible for supporting another generation (Williamson & Watts-Roy, 1999). This term is associated with the controversy surrounding the allocation of increasing amounts of government resources to older persons at the expense of younger generations. Some proponents of generational equity argue that too many of society’s resources are being consumed by older adults at the present time and a change is necessary as soon as possible. Others argue that it is a number of years from now, when aging baby boomers retire, that the crisis in funding will occur (Williamson & Watts-Roy, 1999). “Generational interdependence” means that the generations are dependent on one another and contribute to the well-being of each other. The focus of the generational interdependence perspective is on what the generations have to offer one another rather than what they take from each other (Williamson & Watts-Roy, 1999). “Intergenerational relations” is defined as the relationships between generations. Bengtson (2001) sees intergenerational relations as those spanning more than two generations that are becoming more common in families due to increased life expectancy probl and Ls el l'llClll Ihsou cCont Ag: ‘7‘: - in American society. Bengtson, Marti & Roberts (1991) argue that there are conceptual problems in the generational equity debate because of misuse of the words “generation” and “cohort.” For Bengtson et al. (1991) and for the purposes of this study, “generation” is defined as a position within a family hierarchy while “cohort” is defined as one’s membership in a particular age group, specified by a range of years. Thus, “generation” applies to families and microsocial dynamics and “cohort” or “age group” reflects a societal or macrolevel position (Bengtson et al., 1991). Therefore, the “generational” equity debate is actually misnarned. It should be more correctly called the “cohort” or the “age-group” equity debate. “Intergenerational transfer” is defined as the private exchange of things of economic value between the generations of a family (Kronebusch & Schlesinger, 1994). Such transfers may take the form of money, goods, or services and although exchanges may take place within the family from younger to older generations, most of the money transfers across generations are from older parents to their adult children (George, 1998). “Social Security” is defined as a specific program of social insurance, initiated through the Social Security Act of 1935, whose primary purpose was to prevent economic insecurity among the aged in the United States and today has grown into Old- Age, Survivors, and Disability Insurance (Berkowitz, 1997; Williamson & Watts-Roy, 1999). Like most social insurance programs, participation in Social Security is compulsory, it is sponsored and regulated by the federal government, its resources are raised by a tax on workers and their employers, eligibility is derived from contributions, although some may qualify without contributing, and its benefits are prescribed by law and not directly related to contributions (Thompson & Upp, 1997). “Medicare” is defined as Hospital Insurance and Supplementary Medical Insurance, both initiated in 1965 by amendment to the Social Security Act. General Hymtheses of the Study Following are the research questions: 0 Does support for programs that primarily serve older adults such as Social Security and Medicare vary by age group (young, middle, and older adults)? Do sociodemographic variables other than age, specifically; race, ethnicity, political affiliation, sex, socioeconomic status, and educational level affect support for elder programs? Does occupying the role of caregiver to an older adult family member influence support for programs for older adults? Does support for adding a prescription drug benefit to Medicare vary by age group? Does support for increasing taxes to pay for a Medicare prescription drug program vary by age group? Does support for raising the age of eligibility of Medicare vary by age group? Does support for increasing the out-of-pocket costs of Medicare vary by age group? Do those who believe older adults no longer capable of living alone should live with their adult children (a measure of normative solidarity) indicate stronger support for programs for older persons? Data fi'om three State of the State Surveys, all conducted by the Institute for Public Policy and Social Research at Michigan State University during the 1990’s, will be used to test the following hypotheses: 0 H1: Age of respondent will affect support for programs that benefit older adults. 10 exelu genen l l 3431;) .- ‘dlla 1.0fid‘- H2: Adults identifying themselves as Democrats will report more support than those identifying themselves as Republican for programs benefiting older adults. H3: Adults with higher incomes will report lower levels of support for programs that benefit older adults. H4: Adults that occupy the role of caregiver to an older adult will report more support for programs that benefit older adults. H5: Women will report more support for programs that benefit older adults. H6: African Americans, Hispanics, and other minority groups will indicate more support for programs that benefit older adults than will whites. H7: Age will affect support for adding a prescription drug benefit to Medicare and increasing taxes to pay for this added benefit. H8: Adults who believe older persons should live with or be cared for by their adult children will indicate more support for programs for older adults. H12: Those with more years of education will be more supportive of programs for older adults. Limitations of the Study One of the limitations of the study is the fact that the samples used were drawn exclusively from the population of the state of Michigan. Therefore, results cannot be generalized to the population of the United States. However, while the study is limited to samples drawn only fi'om the population of Michigan, Michigan is a state with a larger than average proportion of older adults (16.2 percent). Although this is not as large as Florida’s 20 percent, it may give a more accurate picture of relations between the generations than studies conducted in other parts of the country, providing an “augury of graying America” (Rosenbaum & Button, 1993, p. 481 ). Rosenbaum and Button (1993) 11 argue adults. heme exert ' more open no in the in prone woulc hasnc E54. ~ fink argue that states like Florida and Michigan, with larger than average populations of older adults, may provide a glimpse into the future of the implications for relationships between age groups as the nation continues to age and older segments of the population exert ever-increasing influence because of their numbers. Furthermore, the third and more recent S088 (1999) takes the issue a step further and asks if Medicare should be expanded to provide prescription drug coverage and if respondents would be willing to pay increased taxes to fund prescription drug coverage for older persons. Additionally, the first two surveys provide a measure of family solidarity by asking if participants provide care to an older family member. Hypothetically, those involved in caregiving would be even more supportive of programs that benefit older adults. Previous research has not attempted to link the two issues. Another limitation of the study is that it is a telephone survey and automatically produces noncoverage error that results in the omission of part of the intended population (Weisberg, Krosnick, & Bowen, 1996). Thus, those individuals who do not possess telephones, the poor and the homeless, for instance, are omitted fi'om the sample as are those in institutions such as prisons, nursing homes, and adult care homes. However, telephone surveys produce a better response rate (50% to 70%) than do mailed questionnaires that have a response rate of between 10% and 50% (W eisberg et al., 1996). Telephone surveys have a somewhat lower completion rate since about 5% of respondents hang up before the questionnaire is finished (Weisberg et al., 1996). The completion rate for the three surveys that will be used in this study are 56.3%, 60.8%, and 53.4%. One of the problems telephone interviewers face is the increasing reluctance of individuals to answer surveys in part because some companies pretend to conduct surveys 12 as 3 rue enthusia sun‘eys l The ; Altergot adult iss theoretic for the e' mulfiger, Cohler 8 “tthin o. Lues relatiom HOI kept mama] as a ruse to sell their products (Weisberg et al., 1996). Thus, the public is becoming less enthusiastic about cooperating with telephone surveys. The refusal rates for the three surveys that will be used in this study are 37.6%, 12.7%, and 30%. Backgrgund Theory The aging of society has had a tremendous effect on family relations (Cohler & Altergott, 1995). Unfortunately, family theory building has tended to focus on young adult issues such as family formation while ignoring families in later life. New theoretical frameworks must be developed that conceptualize the aging family, account for the expanding diversity among families in later life, and encompass the relations of multigenerational families which are becoming increasingly common (Bengtson, 2001; Cohler & Altergott, 1995; Hareven, 1996) as more members of successive generations within the family are alive at any given point in time. Luescher and Pillemer (1998) argue that while interest in adult intergenerational relations within families has generated a great deal of research, theory development has not kept up with empirical results. The most prevalent organizing framework for understanding later-life family relations may be the construct of intergenerational solidarity (Luescher & Pillemer, 1998). Intergenerational solidarity is a “multifaceted, multidimensional construct” (Bengtson & Roberts, 1991, p. 856) that consists of six elements of parent-child interaction: association, affection, consensus, resource-sharing or functional solidary, the strength of familism norms or normative solidarity, and opportunities for parent-child interaction referred to as structural solidarity. Associational solidarity is defined as the fi'equency and patterns of interaction between the generations, such as phone calls, visits, family get-togethers, or shared recreational activities (Bengtson & Roberts, 1991). Affectual solidarity refers to the type and degree 13 of positive sentiments and degree of reciprocity between family members as measured by ratings of affection, warmth, trust, and ratings of perceived reciprocity (Bengtson & Roberts, 1991). Consensual solidarity is defined by the extent to which family members agree on values and beliefs; functional solidarity by the degree to which family members help each other and exchange resources whether financial, physical, or emotional; and normative solidarity refers to the strength of family role commitment and sense of filial obligations (Bengtson & Roberts, 1991). Structural solidarity is measured by the proximity of family members as well as the health and number of family members (Bengtson & Roberts, 1991). One example of intergenerational solidarity is caregiving, and families play a central role in caring for the “noninstitutionalized impaired elderly” (Brody, 1981, p. 471), providing the majority of supportive services received by the aged (George, 1986). A recent national survey found that of those households with telephones, 23% contained at least one caregiver, 76 percent of whom are currently caring for a relative or a friend over 50 years old (NAC/AARP, 1997). Caregivers are primarily female and are caring most often for an older female relative (NAC/AARP, 1997). Caregiving activities are diverse and range from those that are not particularly demanding to those that are very difficult and time-consuming, and include assistance with many instrumental tasks as well as personal care (NAC/AARP, 1997). Caregiving is an example of functional solidarity, measured by the degree to which family members help each other and exchange resources. Caregiving involves an exchange of resources that may be financial, physical or psychological; caregiving often involves emotional as well as physical support. Caregiving is also an example of normative solidarity, measured by the strength of family role commitment and the degree 14 to which the sense of ill la cared for the aged parent Another of economic intergeneratr them in value may help the reclPtents of . Not all ag lIBtead Preset rtmarch em pl elder abme ar Importum fact Itcipmcw thl OfPTOfit and l( to which the generations feel responsible for each other. The adult child caregiver feels a sense of filial obligation to help meet the needs of an aging parent. Just as the parent cared for the adult child when she was young, the child feels duty-bound to care for the aged parent when he or she needs help later in life. Another example of support between generations is that of intergenerational transfers of economic help, goods, and services. George (1998) notes that “family-based intergenerational transfers certainly rival public transfers to the elderly and may surpass them in value” (p. 28). While adult children who can afford to make monetary transfers may help their needy parents financially, it is more likely by far that adult children are the recipients of money transfers from parents (George, 1998). Not all agree that families are characterized by intergenerational solidarity and instead present a more negative view of family relations in later life; another line of research emphasizes the stress and isolation caregivers experience, family conflict, and elder abuse and neglect (Luescher & Pillemer, 1998). George (1986) argues that an important factor underlying the stress of caregiving is the conflict between the norm of reciprocity, the belief that members of a relationship should experience equal exchanges of profit and loss, and the norm of solidarity, the belief that family members should be given as much help as they need without consideration of return on one’s investment in the helping relationship. Caregiving can lead to a “long-term imbalance in adherence to these norms and creates personal discomfort and the conclusion that one has behaved badly—regardless of which norm is adhered to the most strongly” (George, 1986, p. 68). Luescher and Pillemer (1998) propose ambivalence between generations as an alternative to the previous two views. Although they believe that ambivalence is theoretically and empirically useful when studying intergenerational relations, it is 15 currently 2 (lumber perspectit model or t encompas ambivalen aliened h' should be feelings at Warmth an tesentrnen The co this SILICA microsygre for progm family and necessani) Mitten to l o 35“ dchat 1%“ 0f 2 'l "Arno, er ‘ currently a general orientation rather than a comprehensive theoretical formulation (Luescher & Pillemer, 1998). The authors argue that ambivalence offers a more complex perspective on parent-child relations in later life than do either the generational solidarity model or the family conflict model. They propose that the notion of ambivalence encompasses both sociological and psychological ambivalences. Sociological ambivalence is found in ambiguous role expectations; major societal changes have affected human relationships and, as a result, there is confusion over how social relations should be conducted Psychological ambivalence refers to the reality of having opposing feelings at the same time. For instance, regarding the caring relationship, one might feel warmth and tenderness toward the parent needing care while feeling frustration and resentment over the added burden and stress in one’s life (Luescher & Pillemer, 1998). The construct of intergenerational solidarity provides the theoretical framework for this study. As Foner (2000) argues, intergenerational bonds are forged within the family microsystem and are reflected on a societal level by support among younger generations for programs benefiting older persons and vice versa Benefits are exchanged within the family and within the society. However, Foner (2000) also notes that past support is not necessarily a guarantee of future support and the issues that an aging nation must confront may ultimately “become serious enough to drive a wedge throughout the society between young and old” (p. 275). Irnmce 9f the st_u9y for Social Work The generational equity debate is important to social work because it is part of the larger debate in American politics over social policy. “It can best be understood as an instance of a broader dialogue about the proper role and scope of the federal government” (Marmor et al., 1997, p.195). In other words, it is a debate about who is responsible for 16 extra Stud} Wom usur Interc- grand maintaining older adults; should it be the federal government or private sector, including individuals being responsible for their own retirement income. Ultimately, the decision affects social policy considerations. Those who believe persons should be responsible for themselves tend to see programs such as Social Security and Medicare as “entitlement” programs, welfare programs to support the aged. Currently it is politically fashionable to be anti-welfare and welfare recipients; when programs for older adults are framed as welfare, it becomes more acceptable to consider cutting them. Social Work is also concerned with microlevel issues concerning the family. Close relationships between the generations contribute to the well-being of both children and adults, are a source of social support, and enrich the lives of all. For instance, in a study of 73 men and women between the ages of 50 and 60, Bergquist, Greenberg and Klaum (1993) found that those with grandchildren reported grandparenting to be an extraordinary experience, one they enjoyed immensely. For many of the men in the study, grandparenting was more satisfying than parenting (Bergquist et al., 1993). The women in the study reported taking “particular delight in watching their own offspring assume the new and significant role of parenting” (Bergquist et al., 1993, p. 102). Interestingly, Silverstein and Parrott (1997) found that childhood contact with grandparents lessened opposition among young adults to contributory policies favoring older adults. This would suggest that microlevel relationships affect macrolevel policies. There is also evidence that the reverse is true; macrolevel policies affect intergenerational relationships on a microlevel. In a study of five different nations with social security systems for older adults, Kunemund and Rein (1999) found that the most generous systems which gave the greatest resources to older adults helped to increase family solidarity. Kunemund and Rein (1999) conclude that cuts in spending for older 17 mlest‘ -- adults “or As the all? group: and poliC} macro are mtergencr The p: intergener The quest ashare of PTOblCm l. WPOnem aimed an age gTOUp Proposed generallol bm'een a ‘3de adults would ultimately harm families. As the nation ages and as more families become multi-generational, relations between age groups as well as within families will continue to be salient issues for social workers and policy makers. Social policy affects families as well as age groups; the micro and macro are closely related. It also may be that solidarity within the family translates into intergenerational solidarity throughout the larger society. 04.613123! The preceding chapter introduced the generational equity debate and the intergenerational dynamics, both on microsocial and macrosocial levels, that surround it. The question under debate also was introduced: Isthe United States allocating too large a share of its resources to older adults at the expense of younger generations? The problem under consideration is the result of claims by the media and generational equity proponents who suggest that conflict exists between age groups as a result of programs aimed at older persons. These claims conflict with empirical evidence that suggests all age groups support programs benefiting older adults. The theoretical framework of the proposed study is based on the premise that families are marked by solidarity between generations, andjt is this solidarity on the microsocial level that prevents conflict between age groups on a macrosocial level. The general hypotheses and procedures of the study also were outlined in the Preceding chapter. Using secondary data from three State of the State Surveys conducted by the Institute for Public Policy and Social Research at Michigan State University, the purpose of the study is to examine the opinions of young, middle-aged, and older adult residents regarding support for programs for older adults. Additionally, the study will investigate the role of sociodemographic variables such as sex of respondent, level of 18 educ plOV education, race and political party affiliation as well as measures of solidarity, such as provision of caregiving, in Manions toward ”25895.93? benefit older‘persons. The “WWI will review the literature related to the generational equity debate and will discuss in more depth the issue of family solidarity both as it relates to individual families and the larger society. Thegthird chapter will provide an overview of the research design and statistical procedures that will be used, a description of the $088 process, and a description of the characteristics of the samples involved as well as sampling procedures. Chapter} will present the findings and Chapter 5 will discuss the implications of those findings. 19 Firs {Olll CHAPTER 2 REVIEW OF THE LITERATURE The following is a review of the literature related to the generational equity debate. First, the changing image of older adults in the United States will be discussed This is followed by a discussion of the generational equity debate as well as studies assessing the strength of support for programs for older adults. The issue of intergenerational solidarity also will be examined, as will the examples of solidarity of family caregiving and intergenerational transfers within the family. Images of older adults in the United States, Minkler (1987) argues that there is a "growing mass media phenomenon" (p. 48) that is reshaping our society's views of older persons from those who are needy and deserving to those who are greedy and powerful individuals for whom government programs flourish at the expense of younger generations. The charge of inequitable distribution of scarce resources is leading to conflict between the generations (Minkler, 1987). Minkler describes the mission of Americans for Generational Equity (AGE) as promoting the interests of younger generations by lobbying for reducing the federal budget deficit and reforming Social Security and Medicare. AGE generated a great deal of media attention because of its ability to capitalize on American society’s growing concern over increasing poverty among children, a huge federal deficit, and declining income among baby boomers, all of which coincided with an increasingly large older population AGE presented older Americans as part of the problem and argued that the economic difficulties of younger generations are directly linked to a system that disproportionately benefits older adults regardless of their financial status (Minkler, 1987). While AGE, founded in 1985, is no longer in existence, another group, Third Millennium 20 [MW] Mt reason natal only h. popul; Third it is a! Palm ret‘ers “"3114 they 5 P331 “We Mill-e l 0 k; (www.thirdmil.org), was launched in 1993 and has essentially the same mission. Minkler (1987) claims that the logic behind AGE’s message is flawed for a number of reasons and promotes, rather than averts, intergenerational conflict. Older Americans are not a homogeneous group. There is tremendous income variation among older persons and severe poverty does exist, especially among older women and minorities, who not only have high rates of poverty but are the fastest growing segments of the older population (Minkler, 1987). AGE presents an image of older adults as financially secure. Third Millennium acknowledges poverty among some groups of older adults and claims it is attempting to avoid conflict between the generations by advocating benefits based on need rather than age. Referring to Social Security as a “generational scam” which is “fiscally unsound and generationally inequitable,” Third Millennium (1993) argues that deep cuts in programs are necessary to prevent the revolt of younger generations weary of paying for programs from which they are not likely to benefit. Third Millennium (1993) refers to Social Security and Medicare as middle class entitlement programs that benefit well-off retirees regardless of financial need, financed by workers less affluent than those they support They recommend raising the retirement age to at least 70, an end to “paying the greens fees for well-heeled retirees” and instead basing benefits on need, and eXploring alternative retirement systems to ultimately replace Social Security (Third Millennium, 1993). Wisensale (1988) claims that an organized backlash against older adults has begun, and the general public is learning about complex policy from the media who “reduce it to a kids versus canes issue” (p. 774). Wisensale (1988) also mentions the formation of AGE as evidence of an orchestrated effort against programs benefiting the aged He notes, however, that critics of AGE point out that 75% of its budget comes from 21 cor corporate sponsors that are eager to have the political debate moved away from defense spending and corporate tax breaks and onto older persons and the percentage of the federal budget spent unfairly on them. Binstock (1991) claims that until the late 1970’s older adults were stereotypically characterized as poor and frail and in need of positive image-building; impotent as a political force and in need of advocacy, and, because of a youth-oriented society that has I“ forced them into retirement, disadvantaged and in need of public assistance. However, in f?‘ o- the last two decades, the stereotype of the aged has been reversed and the media, public policy papers, politicians and scholars have presented a new image of older persons. This view presents older Americans as financially well-off; a potent political force that always votes in its own self-interest; and because of demographic changes, becoming more numerous, politically powerful, while demanding more benefits and an ever-increasing share of the federal budget (Binstock, 1991). Thus, American popular culture depicts older adults as the new elite, more healthy, richer, and more powerful than previous generations (Binstock, 1991). Furthermore, these accounts claim that their greed is ruining the nation. The aged have become the scapegoat “for an impressive list of American problems”, including the “plight of youngsters who have inadequate nutrition, health care, education, and supportive family environments” (Binstock, 1991, p. 328). However, the media picture of older adults is distorted since many do live in poverty and many can be classified as near-poor, in particular older adult minorities and older women (T orres-Gil, 1992). Torres-Gil (1992) agrees with Binstock (1991) that our views of older Americans have shifted in the last few decades. At one time older adults were portrayed as needy, and this perception served the purpose of advocates for the aged However, the new 22 image of older persons as well-off, even greedy, “serves the purpose of those who want to reduce or dismantle hard-won entitlement programs” (Torres-Gil, 1992, p. 69). The problem with both perceptions of older adults is that they fail to take into account the tremendous diversity of the aged and thereby ignore individual differences. Torres-Gil (1992) sees the challenge as acknowledging this diversity and incorporating it into social policy. MacManus (1996) argues that politicians, the media, marketers, and manufacturers have become acutely aware of the nation’s changing age profile and have begun to target their messages to distinct age groups. “Ours is now an age-conscious nation in which perceptions of a widening generation gap and even outright generational conflict prevail” (MacManus, 1996, p. 3). MacManus (1996) claims that age will continue to take on even greater importance in the near future and that age consciousness and age identification will increase in the coming decades. However, MacManus (1996) also points out that the media tends to categorize both the youngest and the oldest generations as self-centered, with the yOtmgest referred to as “spoiled and selfish Generation X” and the oldest called “greedy geezers.” Price (1997) sees older adults as having gone, within a single generation, from being the poorest and most vulnerable age group in American society to being the richest and most politically powerful. Older adults are wealthy because of government entitlements and powerful because of their ever-increasing numbers. However, Price (1997) disagrees that conflict exists between the generations. Instead, he observes widespread support for programs for the aged and believes this support is due to self-interest. In other words, since everyone is aging, it is in everyone’s future self-interest to maintain programs for the aged Thus, talk of a war between the generations is sensationalism on the part of the 23 fror met you ch11 Saf: old. gm late in ado] in a mic; media. Older adults are a universal interest group, one to which we shall all someday belong, and programs such as Social Security and Medicare ultimately benefit everyone (Price, 1997) However, groups such as AGE and Third Millennium argue that only the aged benefit from Social Security and other income transfers to older adults. Minkler (1987) notes that while the direct benefits of Social Security paid to those under 65 (such as survivor’s benefits) are substantial, the “indirect crossgenerational benefits of the program are significant” (p. 50). For instance, adult children are often freed from financial support of aged parents because of Social Security benefits. Younger working people continue to prefer to have their parents indirectly supported than to shoulder this burden; in spite of media claims to the contrary, taxpayers support Social Security and Medicare, even at high cost to themselves (Minkler, 1987). Social Security is a program that has intergenerational benefits. Yet Haveman (1993) claims that the newly poor in the United States are primarily young adults with little education and few marketable skills who, along with their children, feel the government has abandoned them by no longer offering an economic safety net during hard times. At the same time, Haveman (1993) argues, the aged and older workers have risen to the top of the economic heap primarily because the government has instituted policies that have favored entitlements to the aged As a result, intergenerational animosity has increased and there is growing resentment among young families against older persons because of the publicly supported prosperity of older adults. Thus, government policy is the primary source of the intergenerational conflicts in which the country is embroiled (Haveman, 1993). One of Haveman’s recommendations is a drastic revision in policy in which Social Security would become a 24 nerd-b: uoddl Suc equity] immdc Bengts gtncrat the not: nosed suffer f approat From I} stun“ Ger adults 1: mflrec Persons mtmd‘ no“ of 19”) Be Built: need-based benefit program for adults over 65, thereby freeing funds for programs that would benefit younger adults and children. The Generational Equity Debra}; Such differences in opinion have spawned the “generational equity debate” in which equity proponents argue that the old demand too much from the young and the interdependence proponents point out the positive exchanges between generations. Bengtson, Marti, and Roberts (1991), however, identify four approaches to the generational equity debate. The first, and the one often presented in media accounts, is the notion of ‘greedy elders’ versus ‘powerless children.’ In this scenario, older adults are viewed as using federal funds to further enhance affluent lifestyles while younger cohorts suffer from impoverishment (Bengtson et al., 1991). The assumption of this first approach is that older persons and the young compete for society’s scarce resources. From this perspective, the aged are ultimately the scapegoats and are blamed for many of society’s problems because of their greed. Generational equity advocates argue that because of their increasing numbers, older adults have successfully lobbied for additional benefits for themselves and, as a result, their economic status and access to health care have improved markedly. Thus, older persons are better off than younger cohorts. Furthermore, as benefits have increased to the old, the flow of resources to children has decreased proportionally. The inequity of flow of resources will ultimately result in intergenerational conflict (Bengtson, et al., 1991). Bengtson et al., (1991) challenge the four assumptions underlying this argument First, it is not clear that increasing benefits for the young must come at the expense of programs for older adults. Second, not all older adults are well-off financially and 25 goo lssr in: deb thzr therefore not deserving of public assistance. Third, older adults are not merely a drain on society but also provide many benefits to younger generations in the form of transfers of goods and services. Fourth, Bengtson et al., (1991) challenge the notion that equity issues must be flamed as cross-sectional rather than longitudinal as “shared needs occur in everyone’s life cycle” (p. 260). Bengtson et al., (1991) identify the second approach to the generational equity debate as one of distributive justice in which the issue is presented in terms of class rather than age. From this perspective, inequality is the result of the growing disparity between rich and poor in the United States. Thus, inequities are less a function of generational interests than they are the result of self-serving interest groups that are age heterogeneous (Bengtson et al., 1991). From this perspective, inequality must be resolved through programs based on need rather than on the age of the recipient. The third approach to the generational equity debate looks at values concerning aging, dependency/independency, productivity and obligation to other generations (Bengtson et al., 1991). Those taking this particular stance argue that because of increasing longevity it is impossible for society to fulfill its obligations to the aged Since societal resources have been severely taxed, we must revise our goals. Perhaps the most controversial part of this perspective is that concerning the reevaluation of the goals of medicine and the acceptance of death. Because medical technology has extended life expectancy to such a degree, society’s resources have been pushed to the limit. Instead, proponents suggest a return to a belief in a natural life span, the rationing of health care to the aged, and a radical revision of the culture’s moral standards regarding life and death (Bengtson et al., 1991). The fourth, and last, approach identified by Bengtson et al. (1991) is the life-course 26 perspective. Proponents of this point of view argue that there is no causal relationship between improved well-being on the part of the nation’s older adults and the increasing poverty among the country’s children Instead, the improved status of older adults in the United States is the result of increased government action in favor of older persons, while adverse labor market conditions have affected the earning power of younger adults. Children are thereby more vulnerable to poverty since the economic condition of children is dependent on the earning of their parent or parents. The life-course approach also incorporates the shifts in income throughout individual’s work history with those at younger and older ends of the income cycle reporting less disposable income than those in the middle (Bengtson et al., 1991). Those who advocate this perspective also argue that political leaders should recognize the interdependence and obligation between the generations, as society’s members are dependent on one another to various degrees throughout the life cycle (Bengtson et al., 1991). Yet polls conducted over the last 50 years consistently show strong public support by all generations for Social Security (Reno & Friedland, 1997). During the 1970’s, researchers began to collect data regarding the public’s confidence in the future of Social Security after long-range projections indicated impending deficits in the system. While confidence in the future of Social Security dropped “precipitously... .Americans wanted to maintain the level of benefits provid ” with 54 to 68 percent opposed to cutting benefits or raising the retirement age (Reno & Friedland, 1997, p. 181). Confidence in Social Security dropped again in the early 1980’s with predictions of insolvency, yet proposed cuts in benefits for early retirees and those with disabilities were met with alarm and were thereby abandoned. More recently, critics of the system “describe the future retirement of the baby boomers as an unaffordable demographic 27 bulge that threatens the very sustainability of Social Security” (Reno & Friedland, 1997, p. 183), further eroding public confidence in the system. Yet the majority of Americans continue to express strong support for the Social Security program, believing that society as a whole benefits from the program and that Social Security contributes to the financial independence of older adults (Reno & Friedland, 1997). Most Americans are connected in some way to an individual or individuals who receive Social Security benefits; such benefits may relieve adult children of the burden of contributing to the financial support of their parents or having dependent parents moving in with them. Thus, recent polls (1991 and 1994) show that the majority of Americans (73-80 percent) support Social Security, do not mind paying into the program, and are opposed to cutting benefits, while less than half believe they will receive benefits when they retire (Reno & Friedland, 1997). There is clearly a relationship between age of respondent and his or her confidence in Social Security. In a 1994 poll, individuals were asked to rank their confidence levels in employer-sponsored pensions, personal savings plans and Social Security. All age groups under age 55 rated Social Security the lowest and personal savings (actually the most risky given the very low United States savings rate) the highest; individuals 55 and over ranked Social Security the highest and pension and savings plans the lowest (Reno & Friedland, 1997). While younger adults have most confidence in employer-sponsored pensions and personal savings plans in providing income for their retirement years, the reality is that Social Security is the principal source of income for most retirees today and is projected to continue to be so (Reno & Friedland, 1997). Ironically, even those who express low confidence in the future viability of Social Security expect to receive it, even if only as a minor source of income. Thus, the fear is not that Social Security will be 28 discontinr was never because 0 dmwmc Reno the news because t between 1 mfonnat: disagreer 192), Th be one of make the OfAmen‘ Adm the arena flambo} a discontinued but that it will be inadequate as a source of income. While Social Security was never intended to be the sole source of income, for older adults with low incomes because of lack of pensions and income from savings, it has become the primary source of income (Moon, 1997). Reno and Friedland (1997) emphasize the fact that many survey respondents report the news media as their source of information about Social Security. This is problematic because the media tend to focus on problems in the system that need reform or on conflict between policy makers. “Counting on the news media as the primary source of information about Social Security means that it is likely to be cast in terms of crisis, disagreement, and debate or be overlooked altogether” (Reno and Friedland, 1997, p. 192). The media’s tendency to describe the situation with Social Security as a crisis may be one of the factors in the eroding public confidence in Social Security. The authors make the important observation that Social Security will survive as long as the majority of Americans support it. Advocates of the generational equity perspective have focused on the mass media as the arena in which to express their point of view. “That arena is supportive of flamboyance, simplification, polarization, and the related styles that emphasize the crisis nature of social problems and issues” (Williamson & Watts-Roy, 1999, p. 26). As a result, the public is never exposed to the complex nature of the issues involved, nor are they made aware of the fact that the aged are not a monolithic but a very diverse group socially, politically, and economically. Increasingly, older adults are presented as a group entirely dependent on the support of younger generations to the detriment of the future of those generations. However, Ponza et al. (1988) analyzed data from a nationwide sample drawn for the 29 1973 an: respond: older adt mfers children. 1986 the educatio program 1973 and 1986 General Social Surveys to test the hypothesis that the age of the respondent would determine his or her support for various governmental policies i.e., older adults would be more supportive of Social Security and Medicare and income transfers to the aged, and younger adults would be more supportive of programs for children While older adults were more supportive of increased educational spending in 1986 than they were in1973, their support for education was still less than support for education among other age groups; older adults also were less supportive of welfare programs. However, age did not affect support for Social Security spending. Vignettes included in the 1986 General Social Survey allowed more specific analysis of how age and income levels affected support for means-tested transfers of income (Ponza et al, 1988). Analysis of the vignettes did not support the notion that older adults in general are less supportive of transfers to families with children and more supportive of transfers to the aged. Instead, income level complicated the picture; those older adults with high incomes, for whom the tax cost of such transfers is highest and who benefit the least from them, are the least supportive of them. Those older adults with low incomes are more supportive of transfers to low-income families with children than are individuals in almost any other age group except the very young (Ponza et al., 1988). Interestingly, the older adults in this study were less supportive than younger cohorts of transfers to low-income older persons. The authors also note that while there is increasing activism on the part of older adult interest groups, the success of these groups, which results in funding for programs benefiting older adults, is due to the "willingness of individuals in all age groups to support the elderly" (Ponza et al, 1988, p. 463). Using data collected over a number of years in National Election Study surveys, Rhodebeck (1993) examined the ways age affects political attitudes and behaviors. The 30 ht Ci ft results show that all older adults do not necessarily "express policy preferences consistent with our expectations of their age group interests" (Rhodebeck, 1993, p. 361). Instead, it is those older adults who have low incomes, or those who perceive their financial situation deteriorating, and/or those with Democratic or liberal political views who tend to express policy preferences that benefit the aged While older adults tend to be more supportive of government health insurance options than younger age groups, since 1972 older adults have become more supportive of private options while becoming increasingly more likely "to complain about low levels of spending on Social Security or health care" (Rhodebeck, 1993, p. 361). However, Rhodebeck concluded that while presently there is little evidence of intergenerational conflict over issues such as health care and Social Security, "the aging of the baby boom cohort may put enough pressure on entitlement programs that politicians will respond in ways that set the stage for confrontations between an increasingly large older adult population that expects to receive public assistance and a nonelderly population that resents the strain these expectations place on government budgets" (1993, p. 362). In a national study of 1,209 Americans aged 18 and older, Cook and Barrett (1988) found that 81% of the respondents said they were satisfied that a tax on the incomes of working people ftmds Social Security benefits; furthermore, 90% of those satisfied respondents "would oppose spending cuts (in Social Security)...and 71% would be willing to pay higher taxes to save the Social Security system from any cuts in benefits” (p. 353). They also found that age had little to do with support for Social Security, as there were few differences in support across age groups. In fact, demographic variables in general did little to explain supportiveness; more important were perceptions of the program itself. 31 Two factors outweigh all others in perceptions about Social Security: "...the belief that society benefits" because of the Social Security program and the "perception that Social Security recipients have no alternative sources of income other than Social Security" (Cook & Barrett, 1988, p. 354). By far the majority (89%) of respondents believed that Social Security recipients needed that income, and 63% believed that recipients had no other sources of income. The authors concluded with the observation that support for Social Security might change if the knowledge that there are those who receive those benefits who do not really need them and/or there are many older adults with alternative sources of income became widespread. "...if media portrayals [of the elderly]...lean toward a stereotype of economic well-being, then support could decline (Cook & Barrett, 1988, p. 354). In fact, using data from a Florida public opinion poll concerning community images and activities of aging, Rosenbaum and Button (1993) found some cause for concern Noting that Florida is the state with the highest proportion of retirees among its residents (1 in 5), the authors interviewed 535 Floridians in 1990 and 556 Floridians in 1991. The researchers were particularly interested in probing generational differences and similarities between age categories (Rosenbaum & Button, 1993). Although most residents felt that older adults were an economic benefit to the community, 42% of younger adults (those under age 55) and 25% of older adults believe that older persons are more of an economic burden than a benefit Furthermore, 35% of the younger respondents and 10% of older ones did not agree that older persons improved the quality of life of the community. Most surprising to the authors was the finding that a majority or near majority of both younger (48%) and older respondents (46%) agreed that older adults do not want to pay for local services benefiting groups other than themselves, 32 yet interestingly, the younger adults (51%) were somewhat more opposed to increasing school taxes than were their elders (43%) (Rosenbaum & Button, 1993). While the authors are careful to note that their study measured intergenerational tensions at a community level and therefore was not indicative of generational conflict on a larger scale, they conclude by asserting that they “think there is good reason for continuing concern about generational tensions in the US.” (Rosenbaum & Button, 1993, p. 489). However, the authors also admit that while there may be cause for concern, the generations interviewed in their study shared complimentary and beneficial images of older adults, as 57% of younger adults and 70% of older adults believed that older persons improved the quality of life of the community. However, Rosenbaum and Button (1993) found that age was significant with a number of attitude items. A substantial proportion (one third to one half) of younger (those 55 and younger) respondents agreed that older adults were an economic burden, tended to oppose paying for programs that did not benefit them, and fostered political opposition from younger residents because of their selfishness. However, the one independent variable most significantly related to critical attitudes toward older persons was the proportion of older adults in the respondent’s county. In other words, the higher the percentage of older adults where the respondent lived, the more likely the respondent felt resentment toward them. Race was not a significant variable in attitudes toward older adults, and the significance of educational level and socioeconomic status and hostility toward the aged was inconclusive. However, women were more likely than men to judge older persons and their civic activities favorably. It may be that women of all ages felt a natural empathy toward an older population that is predominantly female (Rosenbaum & Button, 1993). 33 Rhodebeck (1993) found that age was not “particularly effective in structuring preferences on age-related issues” (p. 361 ), and in fact, variation in opinion among older persons often equaled or exceeded variation in opinion across generations. Older adults with lower incomes or more pessimistic appraisals of current financial well-being as well as older adults who identified themselves as Democrats or liberals were more supportive of governmental health insurance and of increases in federal spending on programs that benefit the aged Rhodebeck (1993) did not examine the effects of race, gender, and educational level on attitudes toward programs for older adults. In a recent study that examined adults’ images of aging, the AARP (1995) found little evidence of intergenerational conflict. Of the national sample of 1,200 adults aged 18 and over, the majority of those aged 18 to 64 (87%, n = 983) believed that older persons get too little respect from younger people while only 60% of the older adults in the study (those aged 65 and over, 11 = 202) believed they received too little respect (AARP, 1995). Fifty-five percent of adults aged 18-64 believed older adults have too little influence, and 37% said older adults have the right amount of influence. Only 43% of older adults believed they had too little influence, and 45% said they had the right amount. Eighty-six percent of those aged 18-64 and 86% of those 65 and over agreed that older residents generally improved the quality of life in their communities (AARP, 1995). A generational conflict index was constructed from a number of survey items and included the statement “older people are getting more than their fair share of government benefits” as well as the other items mentioned above. Possible conflict scores could range from 1 to 14, yet the distributions in the sample produced a mean score of 2.7. Thus, very little evidence of generational conflict existed in the AARP samme, with the majority (52%) of the respondents scoring 2 or less on the conflict index and the 34 response Researcl com‘ltct anxiety : anxiety 1 I995). Ustn 1200 re: metropo attitudes Their 53] inch ees c Children mmmh financial Sfindi n c‘ to afford responses of the rest of the samme being dispersed along the rest of the scale. Researchers found no evidence of a relationship between demographic characteristics and conflict scores. Instead, those who indicated higher conflict scores also indicated higher anxiety about aging, with 55% of those scoring high on conflict also being in the highest anxiety group as compared with 38% of those scoring low on the conflict index (AARP, 1995 ). Using random sampling procedures, Logan and Spitze (1995) conducted a survey of 1,200 residents age 40 and over in the Albany-Schenectady-Troy, New York metropolitan area The researchers were interested in the effects of age on respondent attitudes toward parent/child relations and toward government programs for older adults. Their sample was primarily White (93.7%) and female (63.8%). The authors constructed indices of family obligation (should aged parents move in with adult children, do adult children provide as much care to parents as in past generations, should adult children stay in touch with parents on a regular basis, should adult children depend on their parents for financial help, should aged parents provide babysitting help) and public policy (is the US spending the correct amount on programs for the aged, anyone over 65 should be entitled to affordable health care, programs should be developed which keep older adults in their homes, public transportation should be provided for older adults). Logan and Spitze (1991) found no evidence of generational self-interest. In fact, relations across age groups appeared to have an altruistic character, with older people giving greater weight to the needs of younger generations, and vice versa (Logan & Spitze (1995). Furthermore, there was a great deal of consensus on most of the items by all age groups, with almost 100% of respondents indicating they agreed or strongly agreed that anyone over 65 should be entitled to affordable health care and a strong 35 majority in every category on all other items (Logan & Spitze, 1995). One of the greatest limitations of this study is that it did not include the attitudes of those in their twenties and thirties. However, Third Millennium (2001) conducted a nationwide telephone survey of 500 adults aged 18 to 34 in order to assess attitudes regarding Medicare and expanding Medicare to include prescription drug coverage. Thirty percent of respondents were single without children and 42% were married with children; 6% were single with children and 16% were married without children. By far the majority (90%) had health insurance, and the sample was fairly evenly divided by political affiliation, with 29% labeling themselves as Democrats, 31% were Republicans, and 34% were Independents. Forty-eight percent of the sample was male and 52% was female. While the majority agreed that prescription drugs should be a benefit under Medicare, only 38% felt the benefit should be given to all older adults regardless of income. Seventy-eight percent were opposed to prescription drug coverage for those older adults with incomes of $100,000 or more while 87% were in favor of a prescription drug benefit for those with incomes of $16,000 or less. Political party affiliation had little effect on attitudes toward prescription drugs, but female respondents were twice as likely (45% to 23%) to support a prescription drug benefit to all older adults, regardless of income, whose out-of-pocket drug bills are more than $6,000 a year (Third Millennium, 2001). While 78% of Third Millennium (2001) respondents thought that the amount of money older adults pay for Medicare should be tied to their household incomes, a majority (54%) supported increasing the payroll tax rate on workers in order to fund Medicare in the future. Yet 55% opposed raising the Medicare eligibility age to 67, and 75% opposed raising the Medicare eligibility age to 70 over the next 20 years. 36 Interestingly, only 18% of respondents thought that the amount of the federal budget allocated to older adults is excessive, while 52% thought it was just right and 21% believed it was not high enough (Third Millennium, 2001). However, 38% of respondents felt the federal government should spend the same amount on children as on senior citizens, 40% believed more should be spent on children than on older adults, and 12% indicated they thought programs for children should receive more funds than those for elders. Silverstein and Parrott (1997) were interested in whether younger adults were less supportive of benefits for older persons and whether past contact with grandparents moderated attitudes toward programs for older adults. Using data from the 1990 Study of Intergenerational Linkages, Silverstein and Parrott examined a sample of 1,431 respondents, 61% of whom were female, 84% White, 8% Black, and 5% Hispanic. Age of respondent was collapsed into five categories: 14% between the ages of 18 and 24, 28% between 25 and 34, 23% between 35 and 44, 21% between 45 and 64, and 13% at least 65 years of age. Respondents also were asked to agree/disagree with statements regarding government spending for programs for older adults. Frequency of past contact with grandparents was assessed using a scale of 1 to 12, with 0 indicating no childhood contact, 11 indicating daily contact, and 12 indicating having lived with a grandparent during childhood. Silverstein and Parrott (1997) generated predicted values from estimated regression coefficients for the two age groups of interest (those between the ages of 18 to 24 and those between the ages of 25 and 34) and found that among respondents with no childhood contact with grandparents, young adults between 18 and 25 felt more strongly than those 25 and older that benefits to the aged are too costly (3.34 vs. 2.77). Young 37 adults with no contact also indicated that Social Security benefits should be taxed (3.35 vs. 2.03) and that older persons should pay more for health care (2.53 vs. 1.630). However, the differences in opinion were sharply reduced among those young adults with the most frequent childhood contact with grandparents: cost of benefits (2.73 vs. 2.74), taxing Social Security (2.54 vs. 2.25), and older adults paying more of their medical costs (1.80 vs. 1.69). Thus, the authors conclude that childhood exposure to grandparents moderates support for programs for older adults in that “greater past contact causes young adults to hold increasingly similar opinions to those of older age groups” (Silverstein & Parrott, 1997, p. 125). Cook and Barrett (1988) measured public support for Social Security and found that age differences in support for programs that benefit older adults were not statistically significant However, Black respondents were somewhat more supportive than Whites, while low-income Whites were more supportive than upper-income Whites. Income had no effect on support among Black respondents. Women were more supportive of Social Security than were men, but there were no differences in support among women with different education levels. Those respondents labeling themselves Democrats were more supportive than Republicans, liberals more satisfied with current programs than those who saw themselves as politically conservative, and liberal Democrats were most supportive. However, the authors conclude that “although some differences exist between sociodemographic groups, these differences account for very little of the variance in support scores” (Cook & Barrett, 1988, p. 349). However, F oner (2000) argues that it is class rather than age that is at the core of disagreements over programs for older persons and believes that the focus on age conflicts masks class differences and interests regarding Social Security and Medicare. 38 She notes that middle and working class individuals indicate strong support for Social Security and Medicare, while “high-income classes see potential tax... benefits in curtailing Social Security and other social welfare programs for the old” (Foner, 2000, p. 275). Integggnergional Solidagty Foner (2000) also asserts that microlevel connections within the family form the intergenerational bonds that ultimately explain the lack of intergenerational conflict on a societal level. Thus, it is family solidarity as evidenced by intrafamily exchanges that prevents social conflict between age groups over public policies that favor older adults. During the summer of 1990 the AARP, the Andrus Gerontology Center, and Harvard University conducted the Intergenerational Linkages Survey, a national telephone survey of 1,500 adults between the ages of 18 and 90. The purpose of the study was to assess intergenerational relationships and the attitudes of the generations toward one another, the strength of solidarity between generations, and the nature of the exchanges between them (Harootyan & Bengtson, 1994). The weighted sample was 48% male and 52% female; 65% of respondents were between the ages of 18 and 49 while 19% were 50 to 64, and 16% were 65 and over; 9% of the sample was Black, 6% were Hispanic, and 85% were White. Using data from the 1990 AARP Intergenerational Linkages Survey, Lawton, Silverstein and Bengtson (1994) examined the six dimensions of family solidarity (structural, associational, affectual, consensual, functional and normative) across intergenerational pairs of family members. Structural solidarity refers to living with or nearby other family members, associational solidarity is measured by the amount of contact between family members, affectual solidarity refers to feeling close to family 39 (9 Hit lif- members emotionally, consensual solidarity is measured by the extent to which family members share the same opinions, functional solidarity refers to family members helping each other, and normative solidarity involves family members feeling responsible for other generations within the family (Bengtson & Roberts, 1991; Lawton et al., 1994). Family differences on the solidarity dimensions were examined by the age, gender and race of the respondent as well as position held within the family structure as indicated by marital and childbearing statuses (Lawton et al., 1994). While it is not unusual for unmarried young adults to live with their parents, it is uncommon in the United States for adults from different generations to be living together (Lawton et al., 1994). Instead, once adult children marry, they form households away from the parental home. However, when young adults leave home they tend to move fairly close to their parents; more than half of all survey respondents reported living within an hour’s drive of their mothers and fathers (Lawton et al., 1994). Such structural solidarity or proximity increases the opportunities for contact between the generations. The AARP Intergenerational Linkages Survey found there was a great deal of contact between adult children and their parents; 69 percent reported contact with their mothers at least one time per week and 20 percent reported daily contact. Fifty-six percent reported weekly contact with their fathers and 12 percent reported daily contact (Lawton et al., 1994). Not only did the survey find a high degree of associational solidarity, but also affectual solidarity, although respondents reported feeling closer to their mothers than their fathers. While 72% of all respondents reported feeling very close to their mothers, only 55% reported feeling very close to their fathers (Lawton et al., 1994). Eighty-eight percent of African Americans reported feeling very close to their mothers, while 57% felt close to their fathers. Parents also reported feeling close to their children, 40 with 81% stating they felt very close. In terms of consensual solidarity, similarity in opinions and perspectives, the majority (50% to 70%) of adult children reported their opinions were similar or very similar to their parents. However, respondents also indicated less similarity with fathers’ opinions if the father was unmarried (48%) or remarried (50%) than when the father was married (65%). While mothers’ marital status did not affect perceived similarity of opinions, parents (both mothers and fathers) reported more similarity of opinion with married children than with unmarried children (Lawton et al., 1994). Race, age, and gender of respondent did not account for any of the variance in consensual solidarity. Functional solidarity refers to the exchange of help between parents and children, and such help is motivated by sentiment as well as a sense of obligation between generations (Lawton et al., 1994). “Help” in the AARP survey was limited to hands-on assistance such as running errands, babysitting, and helping with household repairs. The survey found modest levels of assistance between parents and adult children, with no more than 35% of any category of child or parent giving or receiving help. The survey did not provide an opportunity for detailed analysis of levels of personal care given by either adult children or parents. Normative solidarity is the expectation that one generation should be responsible for the other and represents an ideal of what should be rather than what actually occurs in families (Lawton et al., 1994). Normative solidarity was measured on the survey by a scale that assessed the extent of respondents’ agreement that parents should help adult children financially if needed, that parents should leave an inheritance for their children, that parents should assist adult children in paying for health care if financial help is needed, that parents should assist adult children with childcare, and that grown children 41 should not be expected to support their parents (Lawton et al., 1994). Results indicated that those with adult children agreed more with norms of parental responsibility (30%) than did childless respondents (19%), and older respondents (65-74) agreed more (39%) than did somewhat younger (SS-64) respondents, only a quarter of whom strongly agreed with norms of parental responsibility (Lawton et al., 1994). Those respondents with less than a high school education also agreed more with normative solidarity (40%) than did those with more education (17% of college graduates), and the authors hypothesize that this might be due to the least educated having fewer economic resources and relying more on parents for assistance than the better educated (Lawton et al., 1994). Interestingly, the reverse was true for responses to the measure of adult children’s obligations to parents. Less than one third of those with high school educations or less endorsed the norm of obligation to parents, while 50% of the college graduates agreed that adult children should help older parents (Lawton et al., 1994). However, only 30% of parents of adult children endorsed norms of children’s obligation to parents while 52% of the childless did. This, according to the authors, may reflect the parents’ desire not to become a burden to their children (Lawton et al., 1994). The AARP Intergenerational Linkages Survey also examined relationships between grandparents and grandchildren. In terms of structural solidarity, 54% of grandchildren reported living within an hour’s distance of at least one grandparent, and 26% reported weekly contact (associational solidarity) with a grandparent via a visit or telephone call. In terms of measures of affectual solidarity, grandparents were asked about their relationship with their eldest adult grandchild Seventy-three percent of grandparents reported feeling “very close” to that grandchild and only 6% reported feeling “not close 42 at all.” However, when adult grandchildren were asked how close they felt to their grandparent, only 44% reported that relationship as “very close,” 41% said it was “somewhat close” and only 15% said it was “not close at all.” Results also differed by race, with 64% of Afiican Americans reporting feeling very close to their grandparents and 41% of all others reporting the same (Lawton et al., 1994). Ten percent of grandchildren reported that their opinions were ‘Very similar” to their grandparents, and 36% said their opinions were “somewhat similar.” Again, race had an effect, as 56% of Afiican Americans reported having very or somewhat similar opinions to their grandparents, while 45% of all others reported the same. Regression analysis also showed that influence of the grandparents in childhood affected perceptions of similarity (Lawton et al., 1994). The authors conclude that there are strong generational ties in many families and strong connections within families. There also are many examples of generational solidarity between parents and adult children and between grandparents and grandchildren Many of their respondents lived close to one another and saw each other often The generations felt emotionally close to one another and felt they shared Opinions. The generations also felt a sense of obligation to one another and often helped each other. Interggnergtiongl sol id_§_ri_ty: Intergenerational transfers One element of intergenerational solidarity is functional solidarity, the degree to which family members help each other and exchange resources. Economic resources that are exchanged are called intergenerational transfers. George (1998) claims that intergenerational transfers within the family may even exceed in value public transfers to older adults. The dependency ratio notwithstanding, the most common flow of money 43 within the family is from older parents to their adult children, either in equal disbursements to each adult child regardless of need or in response to the financial needs of a particular adult child (George, 1998). While some adult children (estimated to be between 3% and 10% of middle-aged children) do give money to their parents either sporadically or on a regular basis, the vast majority of intergenerational transfers within the family are from parents to their middle-aged children (George, 1998). Using data from the 1987-1988 National Survey of Families and Households (N SFH), Lee and Aytac (1998) examined the financial assistance that adult children give to and receive from their parents and also looked at racial and ethnic differences in financial assistance. The NSFH sample included 13,017 cases, 9,643 main respondents plus double samples of African Americans and Latinos (Lee & Aytac, 1998). The authors limited their sample to Whites (81%), Latinos (8%), and Afiican Americans (11%) between the ages of 25—65, with at least one surviving parent not living with the respondent Respondents were asked if they had received financial assistance of more than $200 from their parents during the past five years. Respondents also were asked if they had provided financial assistance of more than $200 to their parents during the past five years. The results showed that financial transfers between the two generations were quite low, except for exchanges between White parents and their children. While 20% of the White respondents reported receiving more than $200 from their parents, only 6% of African American and Latino respondents received such assistance (Lee & Aytac, 1998). Financial transfers from child to parents occurred in 4%-5% of each group (Lee & Aytac, 1998). While children gave financial support to their parents based on need, parents in all groups gave financial assistance to their children based on the parents’ “investment 44 motives” (Lee & Aytac, 1998, p. 439). In other words, the authors found that parents gave money to those children who were more highly educated and who therefore appeared to offer a better return on their investment. This finding was especially true for Afiican American and Latino parents who were more likely to give financial help to adult children of higher socioeconomic status than to those children not performing as well (Lee & Aytac, 1998). Thus, in minority families the financial need of the child did not determine parents’ helping behaviors; instead, availability of resources as well as the child’s potential appeared to determine the intergenerational transfer of money. Lee, Netzer, and Coward (1994) hypothesized that aged parents’ norms of filial responsibility (normative solidarity) would be positively related to intergenerational exchange of both aid given by their children and aid given by the parents to their children Parents who endorsed norms of filial responsibility also would give more aid to their children and receive more aid from them (Lee et al., 1994). Their sample was drawn from eight counties in Florida, and respondents without children were excluded from this analysis. The final sample consisted of 387 older adults, 138 of whom were males and 249 were females. “Aid” was measured as advice, financial assistance, gifts other than money, help with household tasks, help with transportation, and, for aid given to children, help with childcare. Filial responsibility expectations were measured by a scale with response options of “agree” to “strongly disagree” on the following statements: grown children and their parents should share as many activities as possible; if grown children live nearby, they should visit their parents once a week; grown married children should live close to their parents in order to help each other; a family should be willing to sacrifice some things intended for their children in order to help support aging parents; older people should be able to depend upon their grown children to help them; parents are 45 entitled to some return for sacrifices made for their children Parents’ filial responsibility expectations were unrelated to either aid given to or received from children However, there was a moderately strong correlation (r = .310) between aid given to children and aid received from them, suggesting reciprocity (Lee et al., 1994). Parents with more resources (being married, better educated, having a higher income, being in good health) had lower expectations than others, which suggested that those who need less help expect less help (Lee et al., 1994). Those aged parents with more resources gave more to their adult children but received less from them. Lee et al. (1994) also found that those parents who believe children should provide more assistance gave more to their children The authors concluded that “parents’ expectations for assistance may not always be realized, but their helping behaviors are clearly reciprocated” (Lee et al., 1994, p. 565). Using information from the AARP Intergenerational Linkages Survey, Kronebusch and Schlesinger (1994) attempted to ascertain the magnitude and pattern of private transfers as well as how the pattern of private transfers compared to the pattern of government benefits. The authors claimed that private transfers are very large, more than double government transfers, yet are not regularly measured or reported in government statistics. Kronebusch and Schlesinger (1994) argued that it is important to consider the implications of private transfers as they may be an indication of the strength of the bond between generations, serving as a measure of caring. Intergenerational transfers may also foster a sense of obligation between generations as receiving help might make one more likely to give it in the future. Kronebusch and Schlesinger (1994) claimed that the generational equity debate always focuses on government-based transfers that are assessed out of context because the private transfers between age groups that public 46 transfers engender are largely ignored because they are less visible. In order to present two different pictures of private intergenerational transfers, Kronebusch and Schlesinger (1994) used two definitions of age groups: “generation” refers to family position or family relationship while “policy age group” is based on chronological age (under 20, 20-59, and 60+). The authors divided transfers into three categories: time (devoted to personal assistance or formal volunteering), financial (gifts, charitable contributions, inheritances) and housing (sharing living arrangements). Each type of transfer was expressed in monetary terms: the survey provided the monetary value of inheritances and gifts and the monetary value of the other transfers were estimated using the prices of goods and services that would have to be purchased if the transfers were not available. The magnitude of all private transfers was high; the average level of giving was the equivalent of $9,100 per adult per year. However, the level of giving changed with age, reaching a peak at midlife and declining with age. Child-rearing costs predominated for those 25-44; the peak for gifts and financial assistance was found among those 55-64 (Kronebusch & Schlesinger, 1994). Adults reported receiving an average of $2,700 per year with young adults (18-24) receiving $6,200 per year, midlife adults received the least (less than $2,000 per year), while receiving increased somewhat for older adults because of increases in personal assistance. Private transfers within the family tended to flow downward from parents to children; nonfanrily transfers (help to neighbors and friends, volunteering, charitable contributions) were valued at $2,000 per year. Within- family transfers were valued at $7,100 per year (Kronebusch & Schlesinger, 1994). Based on their findings, Kronebusch and Schlesinger (1994) put the aggregate of private transfers at about $1.6 trillion and noted that the giving was more frequently 47 between rather than within age groups. Of this total 70% or $1.1 trillion were transfers between policy age groups. The authors then compared these totals to government transfer payments (Social Security, unemployment, veterans’ and welfare benefits) in 1990 which amounted to just under $700 billion, less than half of the total of private transfers (Kronebusch & Schlesinger, 1994). Downward flows of resources were the most common in intergenerational private transfers from one policy age group to another: over $1 trillion flowed from adults and elders to youth and $129 billion from elders to adults. “Overall, about 56% of private giving goes to youth, 35% to adults, and 9% to elders” (Kronebusch & Schlesinger, 1994, p. 139). The same downward flow predominated in families, with most transfers to adjacent generations. Elders, the authors note, were not a dependent group: as a policy age group, older adults gave to other generations 50% more than they received, and as a family age group, grandparents gave 75% more than they received. Pezzin and Schone (1999) hypothesized that parents’ marital disruption negatively affected the process of intergenerational transfers within the family. Drawing from the first wave of the Assets and Health Dynamics of the Elderly (AHEAD) survey, an Ongoing stratified panel survey of a nationally representative sample of those 70 and older, the authors excluded from their sample those elders who were childless and/or never-married Their final sample consisted of 2,840 individuals with an average age of 78.8 years, 19% of whom were male, 16% Black, and 6% Hispanic. Pezzin and Schone (1999) found that intergenerational transfers varied by family structure in that a larger Percentage (24%) of widowed parents made more transfers than divorced parents (17.1%) and a150 made transfers that averaged twice as large ($6,764 as compared to $3,379). Furthermore, remarried parents made significantly smaller transfers to their children than 48 parents who never remarried ($4,061 as compared to $7,061 ). These results were consistent with the hypothesis that marital disruption affected the intergenerational transfer of financial resources. Since intergenerational transfers are a measure of functional solidarity, one could also conclude that marital disruption has a negative effect on family solidarity. Pezzin and Schone (1999) discussed their findings in light of the high divorce rate in the United States and suggested that changing marital patterns may alter the family’s role as a support network. Their concern was with older adults of the future who have experienced significantly higher rates of divorce and remarriage than the subjects in their study and with the implications for their well-being. Additionally there may exist the potential for excessive demands upon public programs to fill in the gaps left by fewer intergenerational economic transfers between family members. Intergenerational Solidarigz: Caiegflrg Caregiving is an example of both functional solidarity and normative solidarity and is measured by the degree to which family members help each other and exchange resources. Informal care refers to the care provided by family members rather than professionals to others in the family and involves an exchange of resources that may be financial, physical or psychological. Caregiving often involves emotional as well as physical support. Caregiving is also an example of normative solidarity and is measured by the strength of family role commitment and the degree to which the generations feel responsible for each other. Brody (1981) argued that it is a myth that families do not take care of the aged. Instead, Brody (1981) noted, intergenerational ties are strong; filial responsibility remains intact; families, rather than professionals, most often provide health and social services to their older members; and families also go to great lengths to 49 avoid placing their aged relatives in institutions. The National Alliance for Caregiving (NAC) and The American Association of Retired Persons (AARP) conducted a telephone survey of 1,509 family caregivers, 623 of whom were Whites, 306 were Blacks, 307 were Hispanics, and 264 were Asians. Caregivers included in this study were at least 18 years of age and were either currently providing informal care to a relative or fiiend aged 50 or older or had provided such care during the previous year. The major purpose of the study was to assess the effects of caregiving on the caregiver’s life and attitudes (NAC/AARP, 1997). The study found that just over 23% of all American households contain at least one caregiver (NAC/AARP, 1997). The prevalence of informal caregiving was higher among Asian (31.7%) and Black (29.4%) households than it was among Hispanic (26.8%) or White (24%) households (NAC/AARP, 1997). The typical caregiver was a married woman in her mid-forties who had a full-time job outside the home. Forty-one percent of all caregivers in the study reported having children under age 18 in the household, but 51% of Blacks, 58.3% of Hispanics, and 51.1% of Asians reported having children in the home. Eighty-five percent of the caregivers reported taking care of a relative; 31% were taking care of their own mothers, 9% were caring for their mothers-in-law, and 12% were caring for a grandmother (NAC/AARP, 1997). In terms of the kinds of help given to care recipients and nature of the resource exchanges, 98% of the caregivers reported helping with Instrumental Activities of Daily Living (IADLs) which included transportation, grocery shopping, household chores, meal Preparation, financial management, and administering injections and medications (NAG/AARP, 1997). Fifty-one percent of caregivers reported helping with at least one Activity of Daily Living (ADLs) and 29% helped with at least three ADLs, including 50 helping the care recipient in and out of chairs, dressing, bathing, toileting, feeding, and continence or diapers. Although the demands of providing such help could be taxing, the majority (85%) of the caregivers reported that they had not suffered any physical or mental health problems as a result of caregiving, and 76% said that providing care was not a financial hardship for them. While 55% of all caregivers reported finding caregiving not very stressful, those providing the most intense levels of care were more likely (53%) to frnd caregiving emotionally stressful. However, caregivers also mentioned the rewards of giving care: knowing that the care recipient was cared for well (16%), personal satisfaction from a good deed (16%), and experiencing the appreciation of the care recipient (16%). Eleven percent of caregivers also reported satisfaction fi'om watching the recipient’s health improve, and another 11% found rewards in experiencing family loyalty and fulfilling family obligations, measures of normative solidarity. A final 10% of caregivers reported satisfaction from spending time with the care recipient, a measure of associational solidarity (NAC/AARP, 1997). Using time transfers from adult children to aging parents as the measure of caregiving, Wong, Kitayama, and Soldo, (1999) hypothesized that they would find differences among racial and ethnic groups on patterns of time transfers. Their sample consisted of 2,496 White families with 8,643 adult children, 554 Black families with 2,234 adult children, and 316 Hispanic families with 1,376 adult children. The sample was from the 1992 Health and Retirement Study and included only unmarried parents since married older adults tend to prefer spouses as caregivers when possible (Wong et al., 1999). The adult child respondents (and their siblings) were the unit of analysis for each family, and they were asked if they or their siblings had spent 100 or more hours in the past Year helping parents or stepparents with basic personal needs (dressing, eating, 51 bathing) (Wong et al., 1999). Wong et al. (1999) found that overall, 7% of adult children provided time assistance to their unmarried parents; 6.6% of Whites, 8.2% of Blacks, and 9.2% of Hispanics provided assistance. Thus, consistent with previous research, Blacks and Hispanics were more likely than Whites to provide time assistance to parents. Furthermore, consistent with other research, Wong et al. (1999) found that adult daughters, married or single, were far more likely than adult sons or daughters-in-law to provide assistance. However, being a married son is “protective” of White children, while among Blacks and Hispanics, being a married son increases the likelihood that the son or his spouse will provide time assistance. Wong et al. (1999) found no evidence that wealth of the parent influenced time transfers. Peek, Coward, and Peek (2000) compared and contrasted the receipt of care from different family members among older (65 years and older) African American and White adults. Their study explored the effects of family and household structure on differences by race in informal long-term care by examining the extent to which older African American and White elders received informal help with ADLs and IADLS over a 30- month period Respondents were 1,200 community-dwelling older adults from four counties in northern Florida that contained relatively large proportions of older African Americans. Respondents were interviewed every 6 months over the 30-month period and the final sample (n = 742) included respondents who reported difficulties performing ADL or IADL tasks at any of the six data collection points (Peek et al., 2000). Of the final sample, 402 were Afiican Americans with a mean age of 75, and 340 were Whites with a mean age of 75.9. Not surprisingly, the sample was primarily female, as 75.4% of the African Americans and 77.9% of the Whites were female. 52 Peek et al. (2000) found some statistically significant differences between the races in providers of care. Older African Americans were more likely to receive care from children (32.03% compared with 23.26% for Whites), grandchildren (7.5% compared with 4.19% for Whites), and other family members (7.99% compared with 4.4% for Whites). Whites were somewhat more likely (31.78%) than African Americans (26.98) to receive care from spouses, but the difference here was not statistically significant. While there were no significant differences in receipt of care from friends, there were statistically significant differences by race in the receipt of formal care, with Whites more likely to receive formal care than African Americans (Peek et al., 2000). While there were differences, the authors noted that adult children are a major source of care for both older Whites and Afiican Americans, but Afiican American elders were more likely to have more children and more likely to be living with an adult child. Furthermore, Afiican American women were the least likely to be married and therefore did not have a spouse available for caregiving. African Americans also were more likely to be living with grandchildren, and coresidence with a grandchild accounted for a large portion of the differences by race in receipt of care from a grandchild (analyses not shown in article) (Peek et al., 2000). While Peek et al., (2000) pointed out the advantage of older Afiican Americans compared to older Whites in the receipt of informal care from family, they noted that it is primarily due to the greater probability of receiving care from children and, to a lesser extent, from grandchildren. However, the authors did remark that their findings should not be interpreted as evidence of greater filial responsibility on the part of Afiican Americans, but rather due to older Afiican Americans having larger numbers of children and having a greater likelihood of living with children. 53 Using a sample of older (aged 65 and over) disabled parents living in a community setting, Peek, Coward, Peek, and Lee (1998) examined the effects of the social norm of filial responsibility and the specific expectations of older parents about the receipt of care from adult children on the receipt of actual care. The sample was drawn from four Florida counties and consisted of 334 elders who reported ADL and IADL impairments. The majority of the sample (80.2%) was female, 69.5% not married (mostly widowed), with 51.8% Afiican Americans. The social norm of filial responsibility (normative solidarity) was assessed by agreement on a scale with the following global expectation items: grown children and their parents should share as many activities as possible, adult children should visit their parents at least once a week if they live nearby, married adult children should live close to their parents so that they can help each other, families should be willing to sacrifice in order to help aging parents, older people should be able to depend on adult children to help them do the things they need to do, parents are entitled to some return for the sacrifice they have made for their children (Peek et al., 1998). In order to assess specific elder expectations, respondents were asked to whom they would turn for help if they felt lonely and needed to talk, if they needed help getting to the doctor, if they realized they did not have enough money to pay a big medical bill, and if they found they could no longer live independently. Forty-five percent of the sample indicated they would turn to an adult child if they needed to talk, 49.4% would ask an adult child for help getting to the doctor, 31.7% would ask a child for money, and 28.4% indicated they would live with an adult child if they could no long live on their own (Peek et al., 1998). Peek et al., (1998) found that specific expectations were significantly correlated with amounts of help received and global expectations were positively correlated with amount 54 of care received, but it was not a statistically significant correlation However, gender, marital status, education, number of children, and physical health limitations all were significantly correlated with the amount of care received with unmarried women who were less educated, had more children, and experienced more ADL and IADL limitations receiving more care fi'om adult children. Furthermore, specific expectations of help had a direct positive effect on the amount of care received, while global expectations had a significant indirect effect on the amount of care received. The findings suggested that parents who intended to seek help from their adult children actually received more help than other parents. Thus, expectations of assistance were antecedents of helping behavior, and this was consistent with the norms of intergenerational assistance that were learned within the family context and were shared by the generations (Peek et al., 1998). Generational Eguity, Entitlements and the Family Walker (1996) argues that the generational equity debate is really not about justice between age cohorts but instead is about political ideology where the goal is to restructure the welfare state that is seen as a burden to the private sector. There always has been ambivalence on the part of society toward welfare with those receiving entitlements viewed negatively. As Western populations age and increasing numbers of older adults receive entitlements, the aged also are viewed as a burden and “population aging is regarded as a threat to accumulation” (Walker, 1996, p. 24). Walker (1996) also claims that the generational equity debate cannot be separated from the caring relationship in the family, as is generally the case in the literature. However, since the family is viewed as a private domain, most governments have failed to develop policies that would share care between the family and the state. This lack of policy also reflects the state’s goal of avoiding funding for social care and sustaining the 55 primacy of the family as the source of caregiving. Yet Walker (1996) notes that the lack of social policy regarding care creates unnecessary burden on the family and on the caring relationship within the family as families respond to the changes in intergenerational patterns brought about by p0pulation aging. While the state may not intervene directly in the provision of family care, Walker (1996) claims that it intervenes ideologically by reinforcing the expectation that care will be provided by a family member, primarily female, to an aging relative. In other words, the norm of reciprocity, the notion that one generation cares for another, is as much a cultural value as it is a family value. In fact, Walker (1996) argues, not providing care is regarded as deviant behavior both by families and society. Thus, the state avoids making policies that would help caregivers, financially or otherwise, in order not to corrupt the norm of reciprocity, a measure of family solidarity. In fact, one argument used against all entitlements for older adults is that they interfere with intergenerational transfers within the family and thereby damage family solidarity. Kunemund and Rein (1999) challenged the assumptions of the “crowding out” hypothesis which states that a generous welfare state that provides for the needs of older adults crowds out family obligations to help and thus has a negative effect on family solidarity. Instead the authors argued for the “crowding in” hypothesis and maintained that the more generous the welfare system, the stronger the exchanges between the generations within the family (Kunemund & Rein, 1999). In other words, a growing welfare state would not cause a decline in family solidarity but rather would enhance family solidarity. When older adults have adequate resources of their own, they are not forced to rely on their adult children for economic help and therefore are able to focus on developing family relationships based on intimacy and emotional closeness rather than 56 need (Kunemund & Rein, 1999). An extensive welfare state also may generate intergenerational transfers in unintended ways. “Adult children transfer their present earnings to elderly people in the form of a public pension, and elderly people give back their savings as a monetary gift or a bequest to their children” (Kunemund & Rein, 1999, p. 102). Using data from five developed nations, Kunemund and Rein (1999) compared countries with different levels of welfare spending in order to assess levels of giving and receiving between generations within the family and to measure evidence of a reciprocity effect. Their sample consisted of 4,628 individuals 65 and older, representing the United States (it = 910), Canada (it = 930), the United Kingdom (n = 940), West Germany (n = 948), and Japan (11 = 900). Using both telephone and face-to-face interviews, researchers assessed the type (help around the house, help with transportation, help with money, help when sick), level and intensity (often, occasionally, never) of family assistance given to respondents. Their results indicated that older adults in Germany received the most types of help from children, while elders in the United Kingdom received help most often Elders in the United States received help least often, and the United States has the highest percentage of elders who received no help whatsoever from their children (33.3% reported never receiving help when ill, 66.6% reported no help taking care of their home, 63.6% reported no help with transportation, and 82.2% reported no financial help from adult children). Older adults in Canada and Japan fell in the middle of these extremes, except for money transfers in Japan: 40% of older adults in Japan received financial help from their children. Elders in Germany, with the highest welfare state spending for older adults and the lowest poverty rates among older persons as compared with the other four countries, 57 reported the most help from their adult children (Kunemund & Rein, 1999). The most striking contrast was between Germany and the United States, the country with the weakest welfare system: older Americans also reported the least help fi'om their children. Therefore, the authors concluded that the crowding out hypothesis cannot be supported. A generous welfare state did not undermine family solidarity but instead reinforced family solidarity by promoting a pattern of crowding in (Kunemund & Rein, 1999). Unfortunately, the study failed to assess in detail what kinds of help elders gave to their children with the exception of babysitting help, a variable the authors used as a proxy for family solidarity (Kunemund & Rein, 1999). In all countries, those elders who provided help (specifically with babysitting) were more likely to receive help from their adult children Walker (1996) argues that those countries looking at changes in entitlements to older adults such as raising the age of eligibility or basing Social Security benefits on need are on a “collision course” (p. 33) with the generational changes occurring at the family level. In fact, governments that in some way restrict entitlements to elders and refuse to enact policies that will help families share the burden of care with the state can expect increased conflict between the generations, at least on the level of the family. Without some help from the government, it will be impossible for families alone to care for the increasing numbers of older adults (Walker, 1996). In other words, governments that attempt to limit benefits to older adult family members will ultimately harm family solidarity. Summgy The preceding review of the literature described the change in the image of older adults in the United States from one that presents them as needy and deserving to one 58 where older persons are well-off, enjoying life as never before, and perhaps even greedy. However, both stereotypes fail to take into account the tremendous diversity that exists among older adults in the United States today. The changing image of older persons has helped fuel the generational equity debate as has the claim by some that the newly poor in this country are primarily younger persons who feel the government has abandoned them while continuing to support prosperity for older persons. This controversy was examined in more depth and a review of the various studies assessing the strength of support for programs for older adults was included in the chapter. Most of these studies indicate support among all age groups for programs benefiting older persons. In fact, Social Security is a very popular government program, supported even by those who believe they will never benefit from it. The concept of intergenerational solidarity was developed further and research that assessed the strength of the relationship between generations was presented. Additionally, studies regarding family caregiving and intergenerational transfers also were reviewed in order to better understand the concept of intergenerational solidarity. Finally, the connection between the generational equity debate, a macrolevel phenomenon, and family solidarity, a microlevel occurrence, was discussed. It is argued that the reason conflict between age groups does not exist on a macrolevel is because of intergenerational solidarity which exists on a microlevel, within the family. The next chapter will outline the study and will provide a detailed discussion of the research methodology, including sampling, data collection, and statistical analysis. 59 CHAPTER 3 METHODOLOGY In the fall of 1994 the Institute for Public Policy and Social Research (IPPSR) at Michigan State University launched the State of the State Survey (SOSS) in order to provide systematic monitoring of the opinions of the residents of the state of Michigan on a variety of issues (Hembroff & Silver, 1995, 1998, 1999). SOSS is conducted quarterly and uses Computer Assisted Telephone Interviewing (CAT 1) to interview a random sample of the residents of Michigan. SOSS has a number of objectives. The first is to capture and disseminate via the mass media and special studies the opinions of the residents of Michigan. Another objective is to provide data for scientific and policy research by the faculty at Michigan State University and also information for use by university programs and offices involved in extension and Outreach (Hembroff & Silver, 1995, 1998, 1999). Another objective is to develop survey methods by experimenting in wording, formatting, and ordering of questions and the/telly contribute to the scientific literature on survey methodology (Hembroff & Silver, 1995, 1998, 1999). A final objective of SOSS is to provide opportunities for Michigan State University faculty and students to use recent survey data in their research classes, theLebqucreasing the relevance of educational projects (Hembroff & Silver, 1995, 1998, 1999). SOSS uses a specific calendar in order to capture continuities and fluctuations in opinions. In the fall of even-numbered years, SOSS concentrates on elections and political opinions, and in the fall of odd-numbered years the topic is health and the environment (I-Iembroff & Silver, 1995, 1998, 1999). Each winter the SOSS focus is on the state of the state of Michigan, including governmental performance, perceived quality 60 of life of Michigan residents (satisfaction with education, work, crime protection, environmental protection), and attitudes toward the political economy (Hembroff & Silver, 1995, 1998, 1999). In the spring the emphasis changes to the state of Michigan’s families and the status of women and children, and the summer round looks at the state of ethnic lVIichigan by assessing the strength of ethnic ties and identities, tolerance for diversity and experience of discrimination (Hembroff & Silver, 1995, 1998, 1999). The questionnaire for each round of SOSS is designed by a different group of principle investigators made up of faculty and students at Michigan State University. Each survey has three main parts: the demographic core, a nondemographic core, and the main substantive theme or themes (Hembroff & Silver, 1995, 1998, 1999). The demographic core assesses the background of respondents by capturing age, sex, education level, occupation and employment status, marital status, household size and number of children, income, and ethnic identity and is included in each survey round (Hembroff & Silver, 1995, 1998, 1999). The nondemographic core questions, also repeated in each round, assess consumer confidence, self-identification on a liberal- conservative scale, political party identification, and assessments of presidential and gubernatorial performance (Hembroff & Silver, 1995, 1998, 1999). The third part of the survey varies by theme according to the calendar, consists of questions composed by the principle investigators, and reflects the specific research interests of those investigators. If a pressing issue or current event makes it necessary, SOSS may depart from its thematic structure in order to address the pmticular issue (Hembroff & Silver, 1998, 1999). Three SOSS data sets will be used for the study. The first data set is from a telephone survey of 1,133 adult (age 18 and older) residents of the state of Michigan conducted 61 during the late spring and summer of 1995. The focus of the spring, 1995 survey (SOSS- 3) was the state of Michigan families. Although assessing the generational equity issue was not the purpose of this survey, one question specifically targeted the issue of generatiorml conflict: respondents were asked if they agreed or disagreed with the statement “Too many tax dollars are going to support older people” (Hembroff and Silver, 1995). This question is at the heart of the generational equity debate and directly assesses Michigan residents’ opinion toward government spending on programs benefiting older adults. A measure of intergenerational solidarity also was included in the survey. Respondents were asked if they provided assistance to an older family member and also were asked to indicate frequency of assistance. Another survey question assessed normative solidarity by asking respondents if adult children should allow aging parents to live with them when these parents were no longer capable of caring for themselves. The second data set is also a State of the State Survey (SOSS-15), conducted during the summer of 1998 (n = 950). In the same manner as SOSS-3, SOSS-15 respondents were asked if they thought too many tax dollars were going to support older persons. Additionally, respondents were asked if they provided care to an older adult family member. The major difference in the third set of data (SOSS-1 8, conducted during the summer of 1999, n = 950) is that the survey questionnaire included items that asked respondents if they supported limiting somewhat or expanding benefits to older adults. The first item asked if the age for Medicare eligibility should be raised, the next item asked respondents if Medicare recipients should pay more or fewer out—of-pocket costs, and the last two items asked if Medicare should be expanded to include prescription drug coverage and if 62 respondents would be willing to pay increased taxes to fund such coverage. Concerns surrounding funding and eligibility for Medicare are also central to the generational equity debate as Medicare is a program primarily benefiting older adults. Specifically, the following survey items, by survey number and year, have been chosen for analysis: DATA SET SOSS-3 1995 SOSS-15 1998 SOSS-18 1999 YEAR CASES SURVEY QUESTIONS n = 1,133 Are you currently providing regular assistance to an older n=950 n=950 family member such as providing physical care, transportation, home upkeep or car repairs? (Yes, No, Do not know, No answer) Children ought to let aging parents live with them when the parents can no longer live by themselves. (Would you say you strongly agree, somewhat agree, somewhat disagree, or strongly disagree?) Too many tax dollars are going to support older people. (Would you say you strongly agree, somewhat agree, somewhat disagree, or strongly disagree?) Are you currently providing regular assistance to an older family member such as providing physical care, transportation, home upkeep, car repairs, or financial support? (Yes, No, Do not know, refused) Too many tax dollars are going to support older people. (Would you say you strongly agree, somewhat agree, somewhat disagree, or strongly disagree?) To help save the Medicare program from bankruptcy, the federal government should raise the age of eligibility for Medicare above 65 years. (Would you say you strongly agree, somewhat agree, somewhat disagree, or strongly disagree?) Do you think that Medicare coverage should be expanded to include prescription drugs? (Yes, No, Do not know, Refused) 63 Would you be willing to pay higher taxes so that all elderly people would have prescription drug coverage under Medicare? (Yes, No, Do not know, Refused) In your opinion, should the out-of-pocket costs of Medicare (that is costs that people are responsible for themselves) for those who are covered be increased, decreased, or stay about the same? The following independent variables from all three of the surveys will be included in the analysis: 0 0 Income Do you think of yourself as a Republican, a Democrat, or Independent? Are you currently providing assistance to an older adult family member? Age Sex of respondent Years of education Race §_ample Selection grid Characteristics Introduction The sampling design for all of the SOSS data sets was a stratified random sample based on state regions (East Central, Southwest, combined Upper Peninsula and Northern Lower Peninsula, West Central, Southeast, and the City of Detroit); regions were sampled somewhat disproportionately to assure a sufficient minimum number of respondents. Respondents were contacted by telephone using random-di git dial sampling procedures. Cases were weighted according to race, sex, age and by region (according to 1990 census data) so that the sample is proportionate to the general population (Hembroff 64 & Silver, 1995). Sampling From the list of active phone numbers in the state of Michigan, those numbers known to be unused and those known to be used only by institutions were eliminated (Hembroff & Silver, 1995, 1998, 1999). From the remaining numbers, telephone numbers are selected at random in proportion to the number of households in each county until the total number of numbers needed within a certain regional grouping is obtained. These numbers are then screened against directory-based information so that, on average, about 60-75% of phone numbers generated at random will be working household numbers (Hembroff& Silver, 1995, 1998, 1999). The SOSS-3 (spring of 1995) sampling design specified 150 interviews from the East Central Region, the Southwest Region, and the combined Upper Peninsula and Northern Lower Peninsula Regions. Two hundred interviews were to be conducted from both the West Central Region and the Southeast Region. Another 350 interviews were to be completed fi'om the City of Detroit. The total sample was 1,200 (Hembroff & Silver, 1995). The SOSS-15 (summer of 1998) sampling design specified approximately 150 interviews from the East Central Region, the Southwest Region, and the combined Upper Peninsula and Northern Lower Peninsula Regions. Approximately two hundred interviews were to be conducted in both the West Central Region and the Southeast Region. The final 150 interviews were to be completed from the City of Detroit. The total sample was expected to be approximately 1,000 (Hembroff & Silver, 1998). The SOSS-18 (summer of 1999) sampling design also called for approximately 150 interviews from the East Central Region, the Southwest Region, and the combined Upper 65 Peninsula and Northern Lower Peninsula Regions. Approximately two hundred interviews were to be conducted in both the West Central Region and the Southeast Region The final 150 interviews were to be completed from the City of Detroit. The total sample was expected to be approximately 1,000 (Hembroff & Silver, 1999). Sample Weights Stratification and unequal sampling rates across the strata make it necessary to use sample weights to bring the characteristics of the sample into line with those of the region or with those of the state as a whole. As a result, the data files contain weights for the separate regions as well as the state as a whole. The weights have been constructed from characteristics of the p0pulations of the regions based on 1990 census data (Hembroff & Silver, 1995, 1998, 1999). Since households with more than one phone line have a higher probability of being sampled, respondents were asked how many separate phone numbers exist in their households. Each case is weighted by the reciprocal of the number of phone numbers and then adjusted so the total number of cases matches the actual number of completed interviews. This weight is labeled “PHWT” in the data sets (Hembroff & Silver, 1995, 1998, 1999). Additionally, an adult in a two-adult household would have half the chance of being selected for interviewing as would an only adult in a single-adult household. Thus, each case is weighted by the inverse of its probability of selection within the household, or by the number of adults in the household and then adjusted so the total number of weighted cases matches the actual number of completed interviews. This weight is labeled “ADLTWT” in the data sets (Hembroff & Silver, 1995, 1998, 1999). Since some groups of individuals are more difficult to reach or more likely to refuse 66 in random-digit dialing surveys, accuracy of results can be distorted by non-response patterns. As a result, it is common to weight cases in the sample to adjust for non- response by weighting each case so that cases of each type appear in the sample proportionately to their representation in the general population. Thus, cases were weighted so the proportions of White males and females, Afiican American males and females and other racial group males and females matched the proportions that each represent in the adult population of each region based on 1990 census data This weighting factor is labeled “RACGENCT’” in the data sets. Again, based on 1990 census data, cases are weighed by age (AGEWT) which has been adjusted somewhat by region (ADJWT) (Hembroff & Silver, 1995, 1998, 1999). Since the sample was drawn disproportionately across regions of the state, estimates of the opinions of the population of the state require post-stratification weights to adjust for the over-sampling of some regions and the under-sampling of other regions. Again, each case is weighted proportionately according to 1990 census data The weighting factor for this post-stratification weighting in the data set is called “STATEWT” and is for developing statewide results (Hembroff & Silver, 1995, 1998, 1999). Hembroff & Silver (1995, 1998, 1999) also note that weight factors are constructed sequentially so that each builds on the preceding factor. Thus, STATEWT weights cases by all of those adjustments implied by AGEWT (number of phone lines, number of adults in the household, gender x race category proportions within each region, and age category proportions within regions). STATEWT also adjusts the proportions of cases across regions of the state. Thus, STATEWT is the variable that must be consistently used when computing the statistical analysis of the various data sets. Sampling Error 67 The margin of error for the entire statewide sample for SOSS-3 is i 2.8%, with a sample size of n = 1 ,202. For SOSS-15, the sample size is n = 950 and the margin of error is :1: 3.2%. For SOSS-18, the sample size is also n = 950 with a margin of error of i 3.2°/o (Hembroff & Silver, 1995, 1998, 1999). Sample Characteristics: SOSS-3 The final sample for the first data set, SOSS-3, consisted of 1,202 individuals fi'om the state of Michigan, 570 (47.4%) of whom were male and 632 (52.6%) female. The mean age of the sample was 43.7 years with 173 (14.6%) respondents between the ages of 18 and 24, 405 (34%) between the ages of 25 and 39, 348 (29.3%) between the ages of 40 and 59, and 263 (22.1%) 60 and over. Whites (85.3%) made up the majority of the sample, 12.2% were African American, .3% were Asian or Pacific Islanders, 1.4% were Native Americans and .7% classified themselves as “other.” Hispanics made up 2.4% of the sample. The majority (54.3%) reported being better off financially than the year before, and an even greater majority (65.9%) felt they would be better off financially the following year. Almost a third (29.7%) reported their financial status was the same as a year ago but a lower percentage (17.7%) believed their financial status would be the same the next year. Sixteen percent reported being worse off than a year ago, and 12.1% believed they would be worse off the following year. Asked to rate their current overall financial status, 7.5% reported it to be excellent, 54.9% said it was good, 30.8% rated it just fair, 39% claimed it was not so good, and 2.8% rated their financial status as poor. The largest percentage (29.5%) of respondents reported their incomes between $35,000 and $59,999, while 14.1% reported incomes of $19,999 or under, 18.9% had incomes of $20,000 to $34,999, 19.2% reported incomes of $60,000 to $99,999 and 6.1% had 68 incomes of $100,000 or more. While 34.1% of respondents were high school graduates, 28.4% had some college, 17.8% were college graduates, 9.4% had pursued post-graduate education, and 3.1% had attended either technical school or junior college. When asked their political party affiliation, 30.1% identified themselves as Republicans, 36.3% saw themselves as Independents, and 26.4% were Democrats. Of those, 14.4% saw themselves as strong Republicans and 13.3% identified themselves as strong Democrats. Respondents also were asked if they provided care to an older adult family member. Such care could include physical care, transportation, home upkeep or car repairs. Almost a fourth of the sample (24.3%) reported providing such care to an older relative. Sample Characteristics: SOSS-15 The final sample for the second data set, SOSS-15, consisted of 950 individuals from the state of Michigan, 450 (47.4%) of whom were male and 499 (52.6%) female. The mean age of the sample was 44 years, and 158 ( 17%) respondents were between the ages of 18 and 24, 296 (31.8%) were between the ages of 25 and 39, 273 (29.3%) were between the ages of 40 and 59, and 205 (22%) were 60 and over. Whites (86.2%) made up the majority of the sample, 12.1% were African American, .7% were Asian, .9% were Native Americans or Alaskan natives. Hispanics made up 3.4% of the sample. The majority (53.1%) reported being better off financially than the year before, and an even greater majority (70%) felt they would be better off financially the following year. Almost a third (31.8%) reported their financial status was the same as a year ago but a lower percentage (23%) believed their financial status would be the same the next year. Fifteen percent reported being worse off than a year ago, and 7.1% believed they would be worse off the next year. Asked to rate their current overall financial status, 69 11.9% reported it to be excellent, 49.1% said it was good, 31.9% rated it just fair, 5.2% claimed it was not so good, and 1.9% rated their financial status as poor. About one fourth (25.8%) of respondents reported their incomes between $35,000 and $59,999, while 11.4% reported incomes of $19,999 or under, 19% had incomes of $20,000 to $34,999, 26.3% reported incomes of $60,000 to $99,999 and 4.1% had incomes of $ 100,000 or more. While 29.7% of respondents were high school graduates, 29.3% had some college, 19.1% were college graduates, 11.8% had pursued post-graduate education, and 3.2% had attended either technical school or junior college. When asked their political party affiliation, 24.6% identified themselves as Republicans, 38.9% saw themselves as Independents, and 28.8% were Democrats. Of those, 10.8% saw themselves as strong Republicans and 12.7% identified themselves as strong Democrats. Respondents also were asked if they provided care to an older adult family member. Such care could include physical care, transportation, home upkeep or car repairs. A fourth of the sample (25.6%) reported providing care to an older relative. Sample Characteristics: SOSS-I 8 The final sample for the last data set, SOSS-18, was made up of 950 individuals from the state of Michigan, 450 (47.3%) of whom were male and 501 (52.7%) female. The mean age of the sample was 44.1 years with 137 ( 14.6%) respondents between the ages of 18 and 24, 319 (34%) between the ages of 25 and 39, 278 (29.5%) between the ages of 40 and 59, and 206 (21.9%) 60 and over. Whites (86.4%) made up the majority of the sample, 11.9% were African American, .3% were Hawaiians or Pacific Islanders, .1% were Asians, .7% were Native Americans or Alaskan natives. Hispanics made up 3.3% of the sample. 70 The majority (58.5%) reported being better off financially than the year before, and an even greater majority (67.8%) felt they would be better off financially the following year. Less than a fourth (22.6%) reported their financial status was the same as a year ago but a lower percentage (14.8%) believed their financial status would be the same the next year. Almost 18% (17.9%) reported being worse off than a year ago, and 13.8% believed they would be worse off the following year. Asked to rate their current overall financial status, 9.8% reported it to be excellent, 52.9% said it was good, 30.9% rated it just fair, 3.8% claimed it was not so good, and 2.5% rated their financial status as poor. The largest percentage (31.2%) of respondents reported their incomes between $35,000 and $59,999, while 10.9% reported incomes of $19,999 or under, 20.3% had incomes of $20,000 to $34,999, 30.9% reported incomes of $60,000 to $99,999 and 6.7% had incomes of $100,000 or more. While 26% of respondents were high school graduates, 33.6% had some college, 20.7% were college graduates, 12.2% had pursued post-graduate education, and 2% had attended either technical school or junior college. When asked their political party affiliation, 27.8% identified themselves as Republicans, 35.2% saw themselves as Independents, and 31.9% were Democrats. Of those, 12.3% saw themselves as strong Republicans and 14.2% identified themselves as strong Democrats. SOSS-18 respondents were not asked if they cared for an older relative. Comparing the Samples The three samples, SOSS-3, SOSS-15, and SOSS-18, were similar along a number of dimensions. Almost exactly the same percentage of males (47.4%, 47.4%, 47.3%) and females (52.6%, 52.6%, 52.7%) existed in each. While the mean age (43.7 years, 44 years, 44.1 years) was almost the same for each sample, the percentage of individuals 71 within the specified age ranges was somewhat different. Only 14.6% of SOSS-3 respondents fell with the age 18 to 24 range, while 17% of SOSS-15 respondents were within that range. For SOSS-18, the percentage was again 14.6%. The larger percentage of SOSS-15 respondents in the 18 to 24 age range was reversed in the 25 to 39 years range as a smaller percentage of SOSS-15 respondents fell in that range: 34% of SOSS-3 respondents were between the ages of 25 to 39, 31.8% of SOSS-15 respondents fell within that range, and 34% of SOSS-18 respondents were between the ages of 25 and 39. However, the samples were almost identical for the 40 to 59 years range: 29.3%, 29.3%, and 29.5%. Again, for the oldest age range, 60 and over, the samples were very much alike: 22.1%, 22%, and 21.9%. Whites made up the majority of all three samples in almost the same percentages (85.3%, 86.2%, 86.4%) and African Americans were the next largest group, again with almost the same percentages: 12.2%, 12.1%, and 11.9%. The SOSS-15 and SOSS-18 surveys classified the other racial categories somewhat differently than SOSS-3 in that “Hawaiian/Pacific Islander” became a category separate from “Asian,” and the Native American category mentioned Alaskan natives. In SOSS-3, .3% of the sample were Asians or Pacific Islanders, 1.4% were Native Americans, and .7% of respondents classified themselves as “other.” For SOSS-15 and SOSS-18, .7% and .l% were Asians, .9% and .7% were Native Americans/Alaskan Natives, and 0% and .3% were Hawaiians/Pacific Islanders. In terms of ethnicity, the percentage of Hispanics participating in each of the surveys increased somewhat for the last two surveys with 2.4% of SOSS-3, 3.4% of SOSS-15, and 3.3% of SOSS-18. Self-reports of financial status were mixed across the three samples, and no clear pattern emerged. More than half (54.3%, 53.1%, 58.5%) of respondents in all three 72 samples reported their financial status was better than the year before and by far the majority (65.9%, 70%, 67.8%) in all three samples believed their financial status would be even better the next year. Less than a third (29.7%, 31.8%, 22.6%) of each of the samples reported their financial status the same as last year and even fewer (17.7%, 23%, 14.8%) believed it would be the same next year. About the same percentage (16%, 15%, 17.9%) in each sample thought they were worse off financially than a year ago. The largest percentage differences were in self-repOrts of future financial status: 12.1% of SOSS-3 respondents thought they would be worse off financially next year, only 7.1% of SOSS-15 respondents thought they would be worse off, but that percentage almost doubles with 13.8% of SOSS-18 respondents believing they would be worse off financially the following year. Again, no clear pattern emerged in self-reports of current financial status with only a few percentage points difference between the samples. The minority of each sample (7 .5%, 11.9%, 9.8%) rated their financial status as “excellent,” most (54.9%, 49.1%, 52.9%) rated it “good,” about a third (30.8%, 31.9%, 30.9%) saw it as “just fair,” and much smaller percentages rated their financial status “not so good” (3.9%, 5.2%, 3.8%) or “poor” (2.8%, 1.9%, 2.5%). Income levels increase somewhat over the time period (1995 to 1999) in which the three surveys were conducted. Each year a smaller percentage (14.1%, 11.4%, 10.9%) reported incomes of $19,999 or less and a larger percentage reported incomes of $20,000 to 34,999 (18.9%, 19%, 20.3%). Each year a substantially larger percentage (19.2%, 26.3%, 30.9%) reported incomes of $60,000 to $99,999. However, the reports of incomes between $35,000 to $59,999 (29.5%, 25.8%, 31.2%) and $100,000 or more (6.1%, 4.1%, 6.7%) did not follow the same pattern The mean years of education was computed for each SOSS sample. For SOSS-3, the 73 mean was 14.31 years. For SOSS-15, the mean was 14.07 years. Finally, for SOSS-18, the mean years of education was 13.91 years. A larger percentage (26.4%, 28.8%, 3 1 .9%) of respondents identified themselves as Democrats with each round, about a third of respondents in each survey (36.3%, 38.9%, 35.2%) saw themselves as Independents, and less than a third (30.1%, 24.6%, 27.8%) identified themselves as Republicans. There was no pattern in the percentages (14.4%, 10.8%, 12.3%) of those calling themselves “stro ” Republicans or in the percentages (13.3%, 12.7%, 14.2%) of those identifying themselves as “strong” Democrats. SOSS-3 and SOSS-15 respondents were asked if they provided care to an older relative. About a quarter (24.3% and 25.6%) of each sample reported providing care. SOSS-18 respondents were not asked the caregiving question. Data Collection Ptocedyres Interviews for all data sets were conducted using the Computer Assisted Telephone Interviewing system (CATI) of IPPSR’s Survey Research Division (SRD). The completed interview is scripted and programmed so that each question or instruction is presented on the computer in the proper order to the interviewer. The program indicates numeric codes or text the interviewer is allowed to enter as responses to the questions, and, as they are entered, the responses are stored directly into the data set (Hembroff & Silver, 1995, 1998, 1999). New interviewers received about 20 hours of training as well as practice interviewing. The trainees received training manuals, c0pies of interview forms, and descriptions of operations. Experienced interviewers received two hours of training to acquaint them with the study protocols, each new survey instrument, and question objectives. Forty-six interviewers were involved with data collection for SOSS-3, 74 interviewers were 74 involved with 8088-15, and 79 interviewers were involved with SOSS-18. Most interviewers had had prior experience (Hembroff & Silver, 1995, 1998, 1999). SOSS-3 interviewing began on May 10, 1995 and continued through June 20, 1995. Interviewing began for SOSS-15 on July 14, 1998 and continued through September 4, 1998. 8088-1 8 interviewing commenced on July 8, 1999 and continued through August 31, 1999. Telephone numbers were called across times of day as well as days of the week If a number was tried at least 6 times across a variety of time periods and no response received, the number was retired as nonworking. If contact was established, the number was tried until the interview was completed, refused, or the case was judged to be ineligible. The median for SOSS-3 interviews was 22 minutes, for SOSS-15 interviews the median was 20 minutes, and for SOSS-18 interviews the median was 21 minutes (Hembroff& Silver, 1995, 1998, 1999). A total of 1,202 interviews were completed for SOSS-3, with a completion rate of 56.3% and a refusal rate of 37.6%. For SOSS-15, 950 interviews were completed, with a completion rate of 60.8% and a refusal rate of 12.7%. A total of 950 interviews were also completed for SOSS-18, with a completion rate of 53.4% and a refusal rate of 30% (Hembrofi‘& Silver, 1995, 1998, 1999). Hypgtheses Nonparametric, rank-order based statistical analysis of the three SOSS data sets will be used to test the following null hypotheses: Null Hypothesis SOSS-3 H01. There is no relationship between age of respondent and support for programs for older adults. 75 H02. There is no relationship between sex of respondent and support for programs for older adults. H03. There is no relationship between political party afiiliation and support for programs for older adults. H04. There is no relationship between income and support for programs for older adults. H05. There is no relationship between status as a caregiver and support for programs for older adults H06. There is no relationship between race and support for programs for older adults. H07. There is no relationship between educational level and support for programs for older adults. H08. There is no relationship between those who believe older adults should live with adult children and support for programs for older adults. SOSS-15: H09. There is no relationship between age of respondent and support for programs for older adults. H010. There is no relationship between sex of respondent and support for programs for older adults. H011. There is no relationship between political party affiliation and support for programs for older adults. H012. There is no relationship between income and support for programs for older adults. H013. There is no relationship between status as a caregiver and support for programs for older adults. H014. There is no relationship between race and support for programs for older adults. H015. There is no relationship between educational level and support for programs for older adults. H016. There is no relationship between age and support for increasing out-of-pocket costs for Medicare. H017. There is no relationship between sex of respondent and support for increasing out- of-pocket costs for Medicare. 76 H018. There is no relationship between political party affiliation and support for increasing out-of-pocket costs for Medicare. H019. There is no relationship between income and support for increasing out-of-pocket costs for Medicare. H020. There is no relationship between race and support for increasing out—of-pocket costs for Medicare. H021. There is no relationship between years of education and support for increasing out- of-pocket costs for Medicare. H022. There is no relationship between age and support for adding prescription drug coverage to Medicare. H023. There is no relationship between sex of respondent and support for adding prescription drug coverage to Medicare. H024. There is no relationship between political party affiliation and support for adding prescription drug coverage to Medicare. H025. There is no relationship between income and support for adding prescription drug coverage to Medicare. H026. There is no a relationship between race and support for adding prescription drug coverage to Medicare. H027. There is no relationship between years of education and support for adding prescription drug coverage to Medicare. H028. There is no relationship between age and support for paying higher taxes for prescription drug coverage for Medicare. H029. There is no relationship between sex of respondent and support for paying higher taxes for prescription drug coverage for Medicare. H030. There is no relationship between political party affiliation and support for paying higher taxes for prescription drug coverage for Medicare. H031. There is no relationship between income and support for paying higher taxes for prescription drug coverage for Medicare. H032. There is no relationship between race and support for paying higher taxes for prescription drug coverage for Medicare. 77 H033. There is no relationship between years of education and support for paying higher taxes for prescription drug coverage for Medicare. H034. There is no relationship between age and support for raising the age of eligibility for Medicare. H035. There is no relationship between sex of respondent and support for raising the age of eligibility for Medicare. H036. There is no relationship between political party affiliation and support for raising the age of eligibility for Medicare. H037. There is no relationship between income and support for raising the age of eligibility for Medicare. H038. There is no relationship between race and support for raising the age of eligibility for Medicare. H039. There is no relationship between years of education and support for raising the age of eligibility for Medicare. Note: The research questions included both race and ethnic group; however, the sample numbers related to ethnicity were too small for meaningful analysis. Statistical Analysis A number of nonparametric statistical tests, used with rank-ordered data, were employed to analyze SOSS-3 and SOSS-15 survey data. Kendall’s tau-b (r-b), one of the most frequently used gem measures of association, tests the extent to which a change in one variable is associated with a change in another, or how the two variables vary together (Weisberg et al., 1996). A tau value of :t1 .0 would indicate a perfect relationship, while a value of 0.0 would indicate no relationship. A tau value above .7 is considered high and indicates a strong relationship between the variables. A tau value between .3 and .7 is considered moderate, and a tau between 0 and .3 is considered small. However, “values higher than .3 are rare in survey research, so correlations of even .1 are reported as important” (Weisberg et al., 1996, p. 266). Therefore, tau values of .1 and 78 above will be reported as significant, while values of less than .1 will not be considered significant in the analysis that follows. The Mann-Whitney U, another nonparametric test, is used to compare two independent samples (male/female, African American/White) on Edgaltlevel data. This statistic assesses the equality of megians rather thanqpeans and is an altematiye to the independent-samples t test (P ink, 1995). The Kruskal-Wallis (H) test is analogous to one-way analysis of variance for independent samples and is used to compare three or more groups (Republican/Democrat/Independent) with rank-ordered data (Kerlinger, 1986). The Kruskal-Wallis H is approximately distributed as x2jChi-square) (Kerlinger, 1986) A substantial amount of the SOSS-18 data is ngminal, and the Chi-square (x2) statistic was used in its analysis. The Chi-square statistic tests whether there is a relationship between variables by comparing the obtained count with the expected count. The expected count is computed as if there were no relationship. The Chi-square statistic does not measure the strength of the relationship between variables; however, inspection of percentages provides a relatively clear description of the nature of the variables’ relationship. Kendall’s tau-b, Mann-Whitney U, and Kruskal-Wallis also were used where appropriate with the SOSS-18 data. Additionally, the independent samples t-test was used to compare the means of two groups for interval-level data. The most elementary assumption of regression analysis is that the variables are measured at an interval or ratio level (Weisberg et al., 1996). Regression analysis of ordinal data has become more common, but it is controversial (Weisberg et al., 1996). While there has been some discussion that there is too much emphasis on the distinction 79 between numeric and ordinal measurement, Weisberg et al. (1996) recommend against regression analysis of ordinal data. Summary The preceding sections outlined the nature, background, and objectives of the State of the State Surveys (SOSS), the purpose of which is to assess the Opinions of the residents of the state of Michigan four times a year. SOSS also provides data for scientific and policy research by the faculty and students at Michigan State University. Each survey asks a core set of questions as well as questions composed by principle investigators that reflect their individual research interests. Three SOSS data sets were used with this study, one from the spring and summer of 1995 (SOSS-3), one from the summer of 1998 l (SOSS-15), and one from the summer of 1999 (SOSS-18). All three surveys assessed Michigan residents’ opinions regarding programs benefiting older persons. SOSS uses a stratified random sampling design based on state regions and contacts residents by telephone, employing random-digit dial sampling procedures. The sample is weighted according to race, sex, age, and region according to census data in order to assure the sample is proportionate to the general population of the state. The preceding chapter also described the sample characteristics from the three SOSS data sets used for this particular study and compared and contrasted the samples. Data collection procedures were outlined and the study hypotheses were presented in null form. Finally, a discussion of the various nonparametric statistical tests, used with rank- ordered data and employed to analyze the SOSS data sets, was introduced. Chapter 4 will present the analysis of the data from SOSS-3, SOSS-15, and SOSS-18 and the final chapter, Chapter 5, will include a discussion of the various findings. 80 CHAPTER 4 ANALYSIS OF THE DATA Data will first be presented by survey (SOSS-3, SOSS-15, SOSS-18) and by the rejection status of the null hypotheses per survey. Next, each of the variables will be discussed by survey, and data from SOSS-3 and SOSS-15 will be compared and contrasted, again, by variable. Both SOSS-3 and SOSS-15 data sets contain questions in common. Data from SOSS-18 will then be discussed by variable since questions from this survey, while not the same as those used in SOSS-3 and SOSS-15, may provide additional insight into the related issues of the addition of prescription drug coverage to Medicare, raising taxes to pay for this benefit, increasing out-of-pocket costs of Medicare, and raising the age of eligibility for Medicare. Finally, findings will be summarized in a chart listing each of the null hypotheses and its status. Data Analysis A number of nonparametric statistical tests, used with rank-ordered data, were employed to analyze SOSS-3 and SOSS-15 survey data. They include Kendall’s tau-b (t-b ), which tests the extent to which a change in one variable is associated with a change in another; the Mann-Whitney U, used to compare two independent samples; and the I(l‘l-lSkal-Wallis (H) test, which is analogous to one-way analysis of variance for independent samples and is used to compare three or more groups. The Chi-square (x2) Stati Stic was used in the analysis of nominal data, and the independent samples t-test was used to compare the means of two groups for interval-level data. The Kendall’s tau-b assigns a value to the extent to which two variables vary togetimer. A tau value above .7 is considered high, a tau value between .3 and .7 is considered moderate, and a tau between 0 and .3 is considered small (Weisberg et al., 81 1996). Since it is rare to find values higher than .3 in survey research, correlations of .1 are declared important (Weisberg et al., 1996). Therefore, while tau values of .1 or less may be statistically significant, in this analysis only values of .1 or above will be reported as consequential. W Age Using SOSS-3 data, the relationship between the variables “age” and “too many tax dollars are going to support older people” was tested using Kendall’s tau-b. Results (r-b = .058, p = .009) indicate a significant relationship between the variables; however, it is very weak (less than .1). Thus, the null hypothesis “There is no relationship between age of respondent and support for programs for older adults” was not rejected. Table 1A presents frequencies of responses to the statement “too many tax dollars are going to support older people” by age group from SOSS-3 survey results. Eighty-one percent (n = 140) of those age 18 to 24 somewhat or strongly disagreed with the statement “too many tax dollars are going to support older people” while 81.3% (n = 317) of those age 25 to 39 somewhat or strongly disagreed, 80.2% (n = 273) of those 40 to 59 somewhat or strongly disagreed, and of those 60 and older, 73.3% (n == 184) somewhat or strongly disagreed In fact, the youngest group tended to be more supportive of programs for older adults than older adults themselves, 81% versus 73.3%. Using SOSS-15 data, the relationship between the variables “age” and “too many tax dollars are going to support older people” also was tested using Kendall’s tau-b. Results (1‘6 = .106, p = .000) indicate a significant and stronger relationship between variables than was found in 3033-3. A .106 tau value indicates only a weak correlation between the Variables, but the null hypothesis “There is no relationship between age of respondent 82 and support for programs for older adults” was rejected. 3.? Eli-1’ A Too many tax dollars are going to support older people by Age group Neither Szongly Somewhat Agree nor Somewhat Strongly Total gree Agree Di ee Dlsagree Dlsagree 18-24 2.3% 15.6% 1.2% 48.6% 32.4% 100% (n = 4) (n = 27) (n = 2) (n = 48) (n = 56) (n = 173) 25_39 4.4% 12.3% 2.1% 34.6% 46.7% 100% (n = 17) (n = 48) (n = 8) (n = 135) (n = 182) (n = 390) 40_59 4.4% 13.8% 1.5% 22.6% 57.6% 100% (n = 15) (n = 47) (n = 5) (n = 77) (n = 196) (n = 340) 86131 7.2% 14.3% 5.2% 20.7% 52.6% 100% (n=18) (n=36) (n= 13) (n=52) (n= 132) (n=251) over 22:12-11; Too many tax dollars are going to support older people by Age group Neither Strongly Somewhat Agree nor Somewhat Strongly Total Agree Agree Disagree Dlsagree Dlsagree 18-24 3.8% 25.0% 0.6% 50.6% 19.9% 100% (n=6) (n=39) (n=-1) (n=79) (n=31) (n= 156) 25_39 2.4% 13.2% 3.8% 37.6% 42.9% 100% (n = 7) (n = 38) (n = 11) (n = 108) (n = 123) (n = 287) 40_59 4.1% 10.1% 3.0% 28.4% 54.5% 100% (n = 11) (n = 27) (n = 8) (n = 76) (n = 146) (n = 268) £31 6.7% 11.9% 1.5% 37.6% 42.3% 100% over (11 = 13) (n = 23) (n = 3) (n = 73) (n = 82) (n = 194) As indicated in Table 13, support among the youngest respondents (age 18 to 24) has declined from 31% (n = 104) in SOSS-3 to 70.5% (n = 110) who somewhat or strongly disagreed with the statement. Of those age 25 to 39, 80.5% (n = 231) somewhat or $131’<>ng]y disagreed, 82.8% (n = 222) of those age 40 to 59 somewhat or strongly 83 disagreed, and of those age 60 and older, 80% (n = 155) somewhat or strongly disagreed. Sex of Respondent The Mann-Whitney U statistic was used to test for differences between responses of males and females to the statement “too many tax dollars are going to support older people.” Results indicate a significant difference between the two groups (U = 153758.00, p = .000). The mean rank for males was 555.09, and the mean rank for females was 635.08. Therefore, the null hypothesis of no difference between males and females on support for programs for older adults was rejected. As Table 2A shows, twice as many men (6.9%) as women (2.8%) strongly agreed with the statement “too many tax dollars are going to support older people.” Women (54.5%) were more likely to strongly disagree with the same statement than were men (43.3%). SOSS-3 Too many tax dollars are going to support older people by Table 2A Sex of respondent Neither are an“ A... sage; 3:2"er n... Disagree Male 6.9% 14.9% 2.4% 32.5% 43.3% 100% (n = 38) (n = 82) (n = 13) (n = 179) (n = 238) (n = 550) 2.8% 12.3% 2.4% 27.9% 54.5% 100% Female (n=17) (n=76) (n=15) (n=172) (n=336) (n=616) The Mann-Whitney statistic also was used with SOSS-15 data to test for differences between males and females on the statement “too many tax dollars are going to support 01 der- people.” Again, results indicate a significant difference between the two (U = 789 1 8.500, p = .0000), resulting in a mean rank for males of 400. 15, and a mean rank for fenf‘lales of 518.24. The null hypothesis was thus rejected. 84 SOSS-15 Too many tax dollars are going to support older people by Table 2B Sex of respondent Neither 3:3 5°33“ .... 533:3 3333 Disagree Male 6.2% 21.0% 3.0% 37.6% 32.3% 100% (n = 27) (n = 92) (n = 13) (n = 165) (n = 142) (n = 439) 1.9% 8.5% 9.0% 36.5% 51.2% 100% Female (n=9) (n=4l) (n=9) (n=176) (n=247) (n=482) As indicated by Table 28, SOSS-15, conducted three years later in 1998, found even stronger support among women for programs for older adults. Less than 2% of women as compared to 6.2% of men strongly agreed with the statement “too many tax dollars are going to support older people.” Significantly more women (51.2%) than men (32.3%) strongly disagreed with the statement. Political Party A fliliation The Kruskal-Wallis statistic was used to test the difference between political party affiliation and responses to the statement “too many tax dollars are going to support older people” using Republican, Democrat, and Independent as the three groups. Results (12 = 68.04, df= 2, p = .000) indicated a statistically significant difference between respondents’ political affiliations and support for programs benefiting older persons. The Mann-Whitney statistic also was used to test for differences between Republicans and Democrats. Results indicated a statistically significant difference (U = 36952.50, p = - 000) such that the mean rank for Republicans equaled 280.88 and the mean rank for Democrats was 388.82. Therefore, the null hypothesis “There is no relationship between P01 itical party affiliation and support for programs for older adults” was rejected 85 SOSS-3 Too many tax dollars are going to support older people by Political party Table 3A Strongly Somewhat Neither Somewhat Strongly Agree Agree Agree nor Disagree Disagree Total Disagree Other 3.0% 10.4% 4.5% 31.3% 50.7% 100% (n=2) (n=7) (n=3) (n=21) (n=34) (n=67) Re “bum 8.5% 19.4% 2.4% 38.8% 30.9% 100% P (n = 29) (n = 66) (n = a) (n = 132) (n = 105) (n = 340) In dc “dam 2.3% 12.2% 3.3% 29.1% 53.1% 100% p" (n = 10) (n = 52) (n = 14) (n = 124) (n = 226) (n = 426) De 3.9% 9.8% 0.3% 23.3% 62.6% 100% mocrat (n=12) (n=30) (n=l) (n=7l) (n=191) (n=305) In SOSS-3, as shown in Table 3A, respondents identifying themselves as Republicans (8.5%) were three times as likely as Independents (2.3%) to strongly agree with the statement “too many tax dollars are going to support older people” and almost twice as likely as Democrats (3.9%) to strongly agree. However, 62.6% of Democrats strongly disagreed with the same statement as compared with 53.1% of Independents and 30.9% of Republicans. With SOSS-15 data, the Kruskal-Wallis statistic was again used to test the difference between Republicans, Democrats, and Independents as to their responses to the same statement. Results indicated a statistically significant difference (x2 = 30.68, df= 2, p = - 000) between respondents’ political affiliations and support for programs benefiting ol der- persons. The Mann-Whitney U test was also used to test this difference, identifying I{epilblicans and Democrats as the two groups. Again, results indicate a statistically Significant difference (U = 18450.00, p = .000) such that the mean rank for Republicans Was 1 93.20 and the mean rank for Democrats equaled 244.13. Therefore, the null 86 hypothesis of no difference between political party affiliations and support for programs for older adults was rejected Table 3B shows that for SOSS-15, the percentages of those strongly agreeing with the statement were similar (3.3% of Republicans, 3.8% of Independents, and 2.9% of Democrats). Again, Democrats (57.2%) were more likely to strongly disagree with the statement than were Independents (40.9%) or Republicans (31%). Republicans (49%) were more likely to somewhat disagree with the statement “too many tax dollars are going to support older people.” $2133; Too many tax dollars are going to support older people by Political party Strongly Somewhat Neither Somewhat Strongly Agree Agree Agree nor Disagree Disagree Total Disagree Size” 11.8% 41.2% 23.5% 11.8% 11.8% 100% Party (n=2) (n=7) (n=4) (n=2) (n=2) (n=17) Re ublican 3.3% 11.0% 5.7% 49.0% 31.0% 100% p (n=7) (n=23) (n= 12) (n=103) (n=65) (n=210) In de 11 dent 3.8% 19.1% 1.2% 35.0% 40.9% 100% pe (n = 13) (n = 65) (n = 4) (n = 119) (n = 139) (n = 340) D ocrat 2.9% 12.3% 0.4% 27.2% 57.2% 100% (n=7) (n=30) (n=1) (n=66) (n=-139) (n=243) Income For SOSS-3 data Kendall’s tau-b was used to measure the association between l°Il<>orrle of respondents and the variable “too many tax dollars are going to support older People.” Results indicated there was not a significant relationship between the two (r-b = ‘- 03 2, p = .193), and the null hypothesis “There is no relationship between income and Support for programs for older adults” was not rejected. However, for SOSS-15 data, the 87 results were significant (r-b = -.059, p = .024), but the relationship, a negative one, was very weak, less than .1. Therefore, the null hypothesis was again not rejected. Caregiver Status The Mann-Whitney U statistic was used to test the null hypothesis that there is no difference between those who provide care to an older relative and those who do not and the variable “too many tax dollars are going to support older people.” Results were not significant (U = 123036.50, p = .815) for SOSS-3, and the null hypothesis was not rejected. However, results for SOSS-15 were significant (U = 70107.50, p = .001) and the null hypothesis was rejected. Caregivers had a mean rank of 503.93, while those not providing care had a mean rank of 443.25. As is evident in Table 4A, the responses of caregivers and those not providing care are fairly similar. However, as Table 48 indicates, among SOSS-15 respondents, those providing elder care were more likely to strongly disagree (52.3%) with the statement “too many tax dollars are going to support older people” than those not providing care (38.8%). Only 2.9% of caregivers strongly agreed with the statement, while 4.3% of those not providing care strongly agreed. SOSS-3 Too many tax dollars are going to support older people by Table 4A Do you provide care to an older relative? Neither Strongly Somewhat Somewhat Strongly Agree nor . _ Total Agree Agree Disagree Dlsagree Dlsagree Yes 6.6% 16.4% 1.7% 24.8% 50.3% 100% (1‘ =19) (n = 47) (n= 5) (n = 71) (n=144) (n = 286) No 4.0% 12.5% 2.6% 31.9% 48.9% 100% (n = 35) (n = 110) (n = 23) (n = 280) (n = 429) (n = 877) 88 SOSS-15 Too many tax dollars are going to support older people by Table 43 Do you provide care to an older relative? Neither 3:3 33:3 .... 3333 3333 w Disagree Yes 2.9% 10.9% 2.1% 31.8% 52.3% 100% (n = 7) (n = 26) (n = 5) (n = 76) (n = 125) (n = 239) No 4.3% 15.7% 2.5% 38.8% 38.8% 100% (n=29) (n=107) (n=17) (n=264) (n=264) (n=681) Race There were so few Native American, Asian, Pacific Islander, and Other respondents (11 = 29) that they were not included in the statistical analysis of race. Therefore, the analysis of differences between the races was conducted using only African American (11 = 147) and White (11 = 1023) respondents. SOSS-3 data were analyzed using the Mann- Whitney U statistic to test for differences between races regarding the statement “too many tax dollars are going to support older people.” Results indicated a significant difference (U = 58500.50, p = .000), and the null hypothesis “There is no relationship between race and support for programs for older adults” was rejected Afiican American respondents had a mean rank of 481.00, and White respondents had a mean rank of 43 7.27. As seen in Table 5A, 64.1% of Afiican Americans as opposed to 47.4% of Whites strongly disagreed with the statement that “too many tax dollars are going to support older people.” However, when the categories “somewhat disagree” and “strongly disagree” were collapsed, 83.4% (n = 121) of African Americans somewhat or strongly disagreed, while 79% (n = 782) of Whites somewhat or strongly disagreed. 89 SOSS-3 . Table 5 A Too many tax dollars are gorng to support older people by Race Strongly Somewhat Agar Somewhat Strongly Agree Agree Disagree Dlsagree Dlsagree 4.0% 14.3% 2.6% 31.6% 47.4% 100% (n=40) (n=141) (n=26) (n=313) (n=469) (n=989) African 7.6% 7.6% 1.4% 19.3% 64.1% 100% Americans (n=11) (n= 11) (n=2) (n=28) (n=93) (n=145) Total Whites Again, there were so few Native American, Asian, Pacific Islander, and Other respondents (11 = 31) in the SOSS-15 sample that they were not included in the statistical analysis. Therefore, the analysis of differences between the races was made using only the African American (11 = 112) and White (11 = 798) respondents. The Mann-Whitney U test also was used with SOSS-15 data, and here results were not significant (U = 40193.50, p = .065) and the null hypothesis was not rejected As seen in Table SB, Afiican Americans (86.5%, n = 96) were more likely than Whites (78.4%, n = 593) to somewhat or strongly disagree with the statement “too many tax dollars are going to support older people.” However, African Americans also were more likely (6.3%) than Whites (3.4%) to strongly agree with the statement. SOSS-15 . Table SB Too many tax dollars are gorng to support older people by Race Strongly Somewhat Neither Somewhat Strongly A A Agree nor Di ee Di Total gree gree Disagree 538’ 3381’“ . 3.4% 15.3% 2.8% 38.5% 39.9% 100% L L hm” (n=26) (n=116) (n=21) (n=291) (n=302) (n=756) Af‘rican 6.3% 7.2% 0.0% 29.7% 56.8% 100% Americans (n=7) (n=8) (n=0) (n=33) (n=63) (n=111) 9O Education Level Kendall’s tau-b was used to test the association between education level (as measured by years of education) and the variable “too many tax dollars are going to support older adults.” SOSS-3 results indicated a significant correlation (r-b = -.069, p = .004), however, it was less than .1 and therefore very weak. Thus, the null hypothesis “There is no relationship between educational level and support for programs for older adults” was not rejected SOSS-15 results were not Significant (r-b = -.008, p = .769) and the null hypothesis was again not rejected. Table 6 presents the mean years of education and standard deviations for the two samples. Table 6 Educational Level Mean Years and Standard Deviations Data Set N M SQ SOSS - 3 1201 14.31 6.71 SOSS—15 943 14.07 2.48 Elders Live with Children Respondents to SOSS-3 were asked if older adult family members, unable to live independently, should live with their adult children. This question was viewed as a measure of family solidarity (normative solidarity), and its relationship to support for programs for older adults was assessed. Kendall’s tau-b indicated the association between variables was not significant (r-b = -.042, p = .101). Therefore, the null hypothesis “There is no relationship between those who believe older adults should live With adult children and support for programs for older adults” was not rejected. Table 7 Sh<3VVS that those strongly disagreeing that elders should live with their adult children 91 were just as likely (57.1%, n == 32) to strongly disagree with the statement “too many tax dollars are going to support older people” as those strongly agreeing that elders should live with their adult children (58.1%, n = 248). SOSS-3 Too many tax dollars are going to support older people by Table 7 Elders who cannot live alone should live with their children Should live with their children Strongly Somewhat Ami; Somewhat Strongly Tax Dollars Agree Agree Disagree Dlsagree Disagree Strongly 6.8% 3.5% 2.9% 3.4% 3.6% Agree (n=29) (n=17) (n=1) (n=5) (n=2) Somewhat 8.2% 18.4% 8.6% 1 1.5% 10.7% Agree (11 = 35) (n = 88) (n = 3) (n = 17) (n = 6) 1232:“ 3.5% 1.3% 11.4% 0.0% 5.4% Disagree (n = 15) (n = 6) (n = 4) (n == 0) (n = 3) Somewhat 23.4% 34.9% 34.3% 35.1% 23.2% Disagree (n = 100) (n = 167) (n = 12) (n = 52) (n = 13) Strongly 58.1% 42.0% 42.9% 50.0% 57.1% Disagree (n = 248) (n = 201) (n = 15) (n = 74) (n = 32) Total 100% 100% 100% 100% 100% (n = 427) (n = 479) (n = 35) (n = 148) (n = 56) Type of Elder Care Respondents to SOSS-15 were asked what type of care they would prefer for an older family member who required long-term special care: home care by a family member, nursing home care, home health aides, adult day care, or assisted living. The number of cases was frequently too small for meaningful statistical analysis. However, frequencies have been included below in Table 8. While 15.4% (n = 10) of those who would choose a nursing home for their relative indicated they strongly agreed with the statement “too many tax dollars are going to support older people, only 2.1% (n = 7) of those who would provide home care strongly agreed. However, 76.8% (n = 256) of those preferring home care as well as 76.9% (n = 50) of those preferring nursing home care indicated that they somewhat or strongly disagreed with the statement “too many tax dollars are going to support older people.” SOSS-15 Too many tax dollars are going to support older people by Table 8 Type of care you would prefer Strongly Somewhat ANcézhfigr Somewhat Strongly Total Agree Agree gr Disagree Disagree Dlsagree $3253” 2.1% 18.0% 3.0% 39.0% 37.8% 181°? Member (n = 7) (n = 60) (n = 10) (n = 130) (n = 126) 333) Nursing 15.4% 6.2% 1.5% 27.7% 49.2% 100% Home Care (n = 10) (n = 4) (n = 1) (n = 18) (n = 32) (n = 65) :23; 1.0% 19.4% 0.0% 31.6% 48.0% 100% Aides (n=1) (n=19) (n=0) (n=31) (n=47) (n=98) Da Car 10.0% 40.0% 0.0% 30.0% 20.0% 100% y e (n=1) (n=4) (n=0) (n=3) (n=2) (n= 10) ’ 0 ffji‘fitged 2.6% 13.2% 2.6% 50.0% 31.6% 1810 5’ Residential (11 = 3) (n = 15) (n = 3) (n = 57) (n = 36) 114) Do Not 11.9% 5.1% 6.8% 42.4% 33.9% 100% Know (11 = 7) (n = 3) (n = 4) (n = 25) (n = 20) (n = 59) It should be noted that of those responding to the question “if you had an elderly family member who required long term special care, which one of the following would you choose,” 48.5% (n = 342) indicated they would prefer home care by a family member. Almost 10% (n = 67) indicated they would choose a nursing home, 14.7% (n = 1 04) indicated they would prefer home health aides, 1.5% (h = 10) said they would choose some sort of elder day care, and 16.1% (n = 114) said they would prefer assisted 93 living residential care. _S_0_S_S;1_8. SOSS-18 was conducted in the summer of 1999 and posed a different set of questions to interviewees. Respondents were asked if out-of-pocket costs for which Medicare recipients are responsible should be increased, decreased or stay the same, if prescription drug coverage should be added to Medicare, if they would be willing to pay higher taxes for a prescription drug benefit, and if the age of eligibility for Medicare should be raised above 65 years to prevent its bankruptcy. Each of these dependent variables was analyzed in relation to demographic characteristics of the sample. While these questions are different from the SOSS-3 and SOSS-15 question that assessed whether or not respondents agreed or disagreed that too many tax dollars were going to support older persons, the SOSS-18 Medicare questions are issues related to generational equity and may offer additional insights. Questions regarding prescription drug coverage for Medicare are particularly salient at this time, as social policy makers are considering adding that benefit to the federal health insurance program for older adults. Age Kendall’s tau-b indicated a significant relationship (r-b == -.O86, p = .00) between age of respondent and support for increasing out-of-pocket costs for Medicare. However, the relationship, less than -.1, was very weak. Therefore, the null hypothesis “There is no relationship between age and support for increasing out-of-pocket costs for Medicare” was not rejected The Mann-Whitney U test indicated that age groups did not differ significantly (U = 42 62 l - 00, p = .726) in support for adding a prescription drug benefit to Medicare. The 94 Mann-Whitney U test also indicated that age groups did not differ significantly (U = $3503.00, p = .647) in support for paying higher taxes to finance adding the prescription drug benefit to Medicare. Thus, both null hypotheses were not rejected. However, Kendall’s tau-b indicated that age was significant (t-b = .155, p = .000) in predicting support for raising the age of eligibility for Medicare. Since the relationship was more than .1, the null hypothesis “There is no relationship between age and support for raising the age of eligibility for Medicare” was rejected. As Table 9 indicates, those 18 to 24 were more likely to strongly agree (11.2%) or somewhat agree (41.8%) with raising the age of eligibility for Medicare. Only 14.2% of those 18 to 24 strongly disagree with raising the age of eligibility, while 41.3% of those 25 to 39 strongly disagree, 60.4% of those 40 to 59 strongly disagree, and 52.2% of those 60 and older strongly disagree. SOSS-l8 . . . . . . Table 9 Rarse age of ellglbrllty for Medicare by Age group Strongly Somewhat Somewhat Strongly T . . otal Agree Agree Dlsagree Dlsagree 18-24 11.2% 41.8% 32.8% 14.2% 100% (n=15) (n= 56) (n=44) (n=19) (n=134) 2 5_39 11.1% 19.4% 28.3% 41.3% 100% (n=35) (n=-61) (n=98) (n=130) (n=315) 40_59 7.6% 9.8% 22.2% 60.4% 100% (n=21) (n=27) (n=61) (n= 166) (n=275) 60 and 10% 16.9% 20.9% 52.2% 100% over (11 = 20) (n = 34) (n = 42) (n = 105) (n = 201) Sex of Respondent There were significant differences (U = 86065.00, p = .000) between males and 95 ' m7, - u ‘ females in support for increased out—of-pocket costs for Medicare. The mean rank for males was 509.35, while the mean rank for females was 421.72. Therefore, the null hypothesis was rejected. As Table 10A indicates, men were almost twice as likely (19.3%) as women (10.3%) to support increased costs. Women were more likely to support decreased costs (35.4%) than were men (20.5%). SOSS-l8 Increase/decrease Medicare’s out-of-pocket costs by Table 10A Sex of respondent Decrease Stay the same Increase Total Male 20.5% 60.1% 19.3% 100% (n = 87) (n = 255) (n = 82) (n = 424) Female 35.4% 54.3% 10.3% 100% (n = 169) (n = 259) (n = 49) (n = 477) Sex of respondent also was significant (12 = 7.69, df= 1, p = .006) in predicting support for adding the prescription drug benefit to Medicare. The null hypothesis “There is no relationship between sex of respondent and support for adding prescription drug coverage to Medicare” was rejected. As seen in Table 108, women (91.4%) were more likely than men (85.6%) to support adding the benefit. However, sex of respondent was not significantly (x2 = .427, df= l, p = .513) related to willingness to pay higher taxes for this benefit, as the majority of both men and women were willing to pay higher taxes. Thus, the null hypothesis was not rejected. 96 SOSS-l8 Add prescription drug coverage to Medicare by Sex of respondent Table 10B Yes No Total Male 85.6% 14.4% 100% (n = 368) (n = 62) (n = 430) Female 91.4% 8.6% 100% (n = 446) (n = 42) (n = 488) 12 = 7.69 df= 1 p = .006 However, there was a significant difference (U = 101 196.00, p = .044) at the p <.05 level between males and females and support for raising the age of eligibility for Medicare. Males had a mean rank of 450.43 and females had a mean rank of 483.82. Thus, the null hypothesis was rejected Table 10C shows that more women (75.9%, n = 377) than men (65.7%, n = 288) either strongly or somewhat disagreed with raising the age of eligibility. 3.? 512.11 3C Raise the age of eligibility of Medicare by Sex of respondent Strongly Somewhat Somewhat Strongly Total Agree Agree Disagree Disagree Mal 9.1% 25.1% 23.5% 42.2% 100% e (n = 40) (n = 110) (n = 103) (n = 185) (n = 438) Female 10.5% 13.7% 27.6% 48.3% 100% (n = 52) (n = 68) (n = 137) (n == 240) (n = 497) Political Party Afliliation There was not a significant difference (U = 24084.50, p = .083) between political Party affiliations and support for increasing out-of-pocket costs for Medicare. The null 97 hypothesis “There is no relationship between political party affiliation and support for increasing out-of-pocket costs for Medicare” was not rejected. However, there was a significant relationship (x2 = 66.49, df= 3, p = .000) between political party identification and support for adding prescription drug coverage to Medicare. The null hypothesis of no relationship was rejected. Table 11A shows that 73.5% of Republicans were in favor of the benefit, but 91% of Independents and 95.8% of Democrats agreed that the prescription drug benefit should be added to Medicare. SOSS-18 . . . . . Table 11A Add prescrrptron drug coverage to Medlcare by Polltlcal party . Other Republrcan Independent Political P 4!! Democrat Y s 73.5% 91.0% 100.0% 95.8% e (n = 164) (h = 252) (n = 33) (n = 252) No 26.5% 9.0% 0.0% 4.2% (n=59) (n=25) (n=0) (n=11) Total 100.0% 100.0% 100.0% 100.0% (11 = 223) (n = 277) (n = 33) (n = 263) 12 = 66.49 df= 3 p = .000 Again, political party affiliation was significantly (x2 = 19.96, df= 3, p = .000) related to opinions toward raising taxes to pay for prescription drugs, and the null hypothesis was rejected. As seen in Table 11B, while 85.7% of Democrats were willing to pay higher taxes, only 70.1% of Republicans were in favor. Independents (75.6%) were more likely than Republicans but less likely than Democrats to support higher taxes. 98 SOSS-18 Would you be willing to pay higher taxes to add Rx coverage to Medicare Table 113 by Political party . Other Republican Independent Political Party Democrat Yes 70.1% 75.6% 93. 8% 85.7% (n= 110) (n=183) (n=30) (n=210) No 29.9% 24.4% 6.3% 14.3% (n = 47) (n = 59) (n = 2) (n = 35) Total 100.0% 100.0% 100.0% 100.0% (11 = 157) (n = 242) (n = 32) (n = 245) x2=19.96 df=3 p=.000 The Kruskal-Wallis test also indicated a significant relationship (x2 = 28.24, df= 2, p = .000) between party affiliation and support for raising the age of eligibility for Medicare, and again, the null hypothesis of no relationship was rejected. As Table 11C indicates, the majority of all party members somewhat or strongly disagreed with raising the age of eligibility, however, almost the same percent of Democrats (64. 1%, n = 170) and Republicans (66.5%, n = 151) somewhat or strongly disagreed, while 79% (n= 229) of Independents somewhat or strongly disagreed Independents were less likely than either Republicans or Democrats to somewhat or strongly agree that the age of eligibility should be raised. 99 SOSS-l8 Raise the age of eligibility for Medicare by Political party Table 11C . Other Republican Independent Political Party Democrat Strongly 9.3% 6.9% 28.1% 12.1% Agree (11 = 21) (n = 20) (n = 9) (n = 32) Somewhat 24.2% 14.1% 18.8% 23.8% Agree (11 = 55) (n = 41) (n = 6) (n = 63) Somewhat 33.0% 21.4% 28.1% 16.2% Disagree (n = 75) (n = 62) (n = 9) (n = 42) Strongly 33.5% 57.6% 25.0% 47.9% Disagree (n = 76) (n = 167) (n = 8) (n = 127) Total 100% 100.0% 100.0% 100.0% (n = 227) (n = 290) (n = 32) (n = 265) Income Kendall’s tau-b indicated a significant (t-b = .087, p = .000) correlation between income and support for increasing Medicare’s out-of-pocket costs. However, the relationship was very weak (less than .1) and thus the null hypothesis was not rejected. However, there was a significant difference (U = 23343.00, p = .001) between income level and support for adding prescription drug coverage to Medicare. The mean rank of those giving a "yes” response was 372.13, and the mean rank for those giving a “no” response was 456.72, indicating those with higher incomes were less supportive. Therefore, the null hypothesis “There is no relationship between income and support for adding prescription drug coverage to Medicare” was rejected There was not a significant difference between income level and willingness to pay increased taxes for the prescription drug benefit (U = 30807.00, p = .523). Therefore, the 100 null hypothesis was not rejected. There was not a significant association between income and raising the age of eligibility for Medicare (t-b = -.010, p = .738). Therefore, the null hypothesis “There is no relationship between income and support for raising the age of eligibility for Medicare” was not rejected. Race There were so few Native American, Asian, Pacific Islander, and Other respondents to SOSS-18 (n = 11) that they were not included in the statistical analysis. The sample included 109 Afi'ican Americans and 795 Whites. The Mann-Whitney U statistic indicated a significant (U = 35814.50, p = .025) difference between race and support for increasing or decreasing out-of-pocket costs for Medicare. The mean rank for African Americans was 373.94 and for Whites was 422.05. The null hypothesis of no difference was rejected As Table 12 shows, Afi'ican Americans (35.4%, n = 40) were more likely to support decreasing costs than were Whites (25.7%, n = 187). Whites were somewhat more likely to support increasing costs than were African Americans. SOSS-18 . , Table 12 Increase/decrease Medlcare s out-of-pocket costs by Race Decrease Stay the same Increase Total Whites 25.7% 58.8% 15.5% 100% (n = 187) (n = 429) (n = 113) (n = 729) African 35.4% 53.1% 11.5% 100% Americans (n = 40) (n = 60) (n = 13) (n = 113) There was not a significant (x2 = 3.29, df= 1, p = .070) relationship between race and support for adding the prescription drug benefit to Medicare, and the null hypothesis was 101 not rejected Neither was race significant (x2 = 2.20, df= 1, p = .138) in predicting support for higher taxes to pay for the prescription drug benefit, and the null hypothesis again was not rejected. There also was not a significant (U = 39990.50, p = .104) difference between races and support for raising the age of eligibility for Medicare. Again, the null hypothesis was not rejected Education Level There was not a significant (t-b = .05 1, p = .070) relationship between years of education and support for increasing/decreasing out-of-pocket costs for Medicare. The null hypothesis was accepted Kendall’s tau-b did indicate a significant (‘t-b = -.065, p = .017) correlation between years of education and opinions toward raising the age of eligibility for Medicare, however, the association was very weak (less than .1). Thus, the null hypothesis “There is no relationship between education level and support for raising the age of eligibility for Medicare” was not rejected. An independent samples t-test was used to compare the mean years of education of the two groups (those that answered “yes’ and those that answered “no”) and their responses to the questions “Do you think that Medicare coverage should be expanded to include prescription drugs?” and “Would you be willing to pay higher taxes so that all elderly people would have prescription drug coverage under Medicare?” There was a significant difference (t = -l.97, df= 903, p = .050) between those who responded “yes” to adding prescription drug coverage to Medicare and those who responded “no,” with those answering “no” having more years of education. Therefore, the null hypothesis was rejected. However, there was not a significant difference in responses regarding willingness to pay higher taxes to finance the drug benefit (t = 1.30, df= 769, p = .196). Thus, the null hypothesis “There is no relationship between 102 education level and support for paying higher taxes for prescription drug coverage for Medicare” was not rejected. Table 13 summarizes means and standard deviations for each group by responses and questions. SOSS-18 Years of education: Means and Standard Deviations Table 13 Add prescription drug coverage to Medicare N M. .82 Yes 802 13.90 2.37 No 103 14.38 2.28 Willing to pay higher taxes Yes 599 13.89 2.31 No 172 13.71 2.64 Summag Each null hypothesis and its status is as follows: SOSS-3 Null Hypothesis Hol. There is no relationship between age of respondent and support for programs for older adults. H02. There is no relationship between sex of respondent and support for programs for older adults. H03. There is no relationship between political party affiliation and support for programs for older adults. H04. There is no relationship between income and support for programs for older adults. H05 . There is no relationship between status as a caregiver and support for programs for older adults. 103 Status Do not reject Reject Reject Do not reject Do not reject H06. There is no relationship between race and support for programs for older adults. H07. There is no relationship between educational level and support for programs for older adults. H08. There is no relationship between those who believe older adults should live with adult children and support for programs for older adults. SOSS-15 H09. There is no relationship between age of respondent and support for programs for older adults. H010. There is no relationship between sex of respondent and support for programs for older adults. H011. There is no relationship between political party affiliation and support for programs for older adults. H012. There is no relationship between income and support for programs for older adults. H013. There is no relationship between status as a caregiver and support for programs for older adults. H014. There is no relationship between race and support for programs for older adults. H015. There is no relationship between educational level and support for programs for older adults. SOSS-l8 H016. There is no relationship between age and support for increasing out-of-pocket costs for Medicare. H017. There is no relationship between sex of respondent and support for increasing out-of-pocket costs for Medicare. H018. There is no relationship between political party affiliation and support for increasing out-of-pocket costs for Medicare. 104 Reject Do not reject Do not reject Reject Reject Reject Do not reject Reject Do not reject Do not reject Do not reject Reject Do not reject H019. There is no relationship between income and support for increasing out-of-pocket costs for Medicare. H020. There is no relationship between race and support for increasing out-of-pocket costs for Medicare. H021. There is no relationship between years of education and support for increasing out-of-pocket costs for Medicare. H022. There is no relationship between age and support for adding prescription drug coverage to Medicare. H023. There is no relationship between sex of respondent and support for adding prescription drug coverage to Medicare. H024. There is no relationship between political party affiliation and support for adding prescription drug coverage to Medicare. H025. There is no relationship between income and support for adding prescription drug coverage to Medicare. H026. There is no a relationship between race and support for adding prescription drug coverage to Medicare. H027 . There is no relationship between years of education and support for adding prescription drug coverage to Medicare. H028. There is no relationship between age and support for paying higher taxes for prescription drug coverage for Medicare. H029. There is no relationship between sex of respondent and support for paying higher taxes for prescription drug coverage for Medicare. H030. There is no relationship between political party affiliation and support for paying higher taxes for prescription drug coverage for Medicare. H031. There is no relationship between income and support for paying higher taxes for prescription drug coverage for Medicare. H032. There is no relationship between race and support for paying higher taxes for prescription drug coverage for Medicare. 105 Do not reject Reject Do not reject Do not reject Reject Reject Reject Do not reject Reject Do not reject Do not reject Reject Do not reject Do not reject H033. There is no relationship between years of education and support for paying higher taxes for prescription drug coverage for Medicare. Do not reject H034. There is no relationship between age and support for raising the age of eligibility for Medicare. Reject H035. There is no relationship between sex of respondent and support for raising the age of eligibility for Medicare. Reject H036. There is no relationship between political party affiliation and support for raising the age of eligibility for Medicare. Reject H037. There is no relationship between income and support for raising the age of eligibility for Medicare. Do not reject H03 8. There is no relationship between race and support for raising the age of eligibility for Medicare. Do not reject H039. There is no relationship between years of education and support for raising the age of eligibility for Medicare. Do not reject The preceding chapter included the findings from statistical analysis of the three SOSS data sets (SOSS-3, SOSS-15, and SOSS-18). The results from SOSS-3 and SOSS- 15 were compared and contrasted on demographic variables and their relationship to responses to the statement “too many tax dollars are going to support older people.” This statement is central to the generational equity debate. SOSS-18 posed a different set of questions regarding expanding Medicare to cover a prescription drug benefit and raising taxes to pay for this benefit. It also asked for respondents’ attitudes on increasing the age of eligibility for Medicare and increasing out-of-pocket costs for Medicare. Issues surrounding costs and benefits of Medicare, a health-care program designed for older adults, also are at the heart of the generational equity debate. These survey items also were analyzed with respect to demographic variables in order to gain a clearer picture of support or lack thereof for expanding or 106 decreasing Medicare benefits. The final chapter, Chapter 5, will discuss SOSS-3, SOSS-15, and SOSS-18 findings further as well their implications. Suggestions for further research also will be discussed. 107 CHAPTER 5 SUMMARY, CONCLUSIONS AND IMPLICATIONS Summm The purpose of this study was to examine similarities and differences between young, middle-aged, and older adults regarding support for government programs that primarily benefit older adults. Population aging in the United States has raised concerns that conflict between working-age adults and retirees exists because of the increase in the cost of programs for the aged Although there have been claims of conflict between age groups, existing research has found very little evidence of intergenerational conflict within American society. Instead, other population characteristics such as political party identification, gender, socioeconomic status, and race have been found to be more predictive of support or lack thereof for programs for older adults. Others also have insisted that the reason conflict between age groups does not exist on a macro or national level is because solidarity between generations exists on a microlevel, meaning that close relationships within families reduce the possibility of generational conflict. Such solidarity within families is translated into solidarity between age groups within society. Thus, intergenerational bonds are forged within the family mi crosystem and are reflected in support among younger age groups for programs benefiting older persons and vice versa. The present study examines similarities and differences between age groups, sociodemographic variables and measures of solidarity such as provision of caregiving in order to better understand opinions toward programs that benefit older adults. Intergenerational solidarity is a complex construct that describes the parent-child relationship across the life span. Bengtson and Roberts (1991) have characterized 108 solidarity or cohesion as consisting of six elements: association, or the frequency and patterns of interaction between the generations; affection, the type and degree of positive sentiments between family members; consensus, or the extent to which family members agree on values and beliefs; function or resource-sharing, the degree to which family members help each other and exchange financial, physical, or emotional resources; the strength of familism norms which refers to the strength of family role commitments and sense of obligation to each other, in particular, one generation to another", and structure, or the proximity of family members to one another. Walker (1996) argues that the so-called generational equity debate cannot be separated from the caring relationship in the family. Although programs such as Social Security and Medicare are govemment-funded and based on social policies that directly affect families, most governments view the family as a private domain and thus avoid policies that share care between the family and the state. While the state may not intervene directly in the provision of family care, Walker (1996) claims that it intervenes ideologically by reinforcing the expectation that care will be provided by a family member, primarily female, to an aging relative. Thus, the norm of reciprocity, the notion that one generation cares for another and a measure of solidarity, is as much a cultural value as it is a family value. Conclusions One of the ways this study is unique is that it tested support for programs benefiting older persons at two points in time. SOSS-3, conducted during the spring of 1995, and SOSS- ] 5, conducted during the summer of 1998, both asked respondents to agree or disagree with the statement “too many tax dollars are going to support older people.” Data Sets from the two surveys were compared and contrasted to assess the level of 109 support for programs that benefit older adults and to gain an understanding of the effects of sociodemographic variables. A third data set, SOSS-18, conducted during the summer of 1999, asked respondents their views on expanding Medicare to include prescription drug coverage, support for raising taxes to pay for prescription drug coverage, and support for raising age of eligibility for Medicare and Medicare’s out-of-pocket costs. Again, the effects of sociodemographic variables on levels of support were assessed to gain an understanding of attitudes toward programs benefiting older persons. The following is a discussion of the findings by individual variable. Each null hypothesis and its status also are presented. Age H01. There is no relationship between age of respondent and support for programs for older adults. Do not reject H09. There is no relationship between age of respondent and support for programs for older adults. Reject Age of respondent, not a significant predictor of support in the SOSS-3 data, is significant in SOSS-15. In 1995, there is strong support among those aged 18 to 24 for programs benefiting older adults as evidenced by 81% of young adults who somewhat or strongly disagree with the statement “too many tax dollars are going to support older people.” However, three years later, in SOSS-15 support has declined by about 10%. In 1 995, 17.9% of those aged 18 to 24 somewhat or strongly agree with the statement; by I 998 that percentage has increased to 28.8% who somewhat or strongly agree, and it would appear that support among the youngest adults for programs benefiting the aged, while still high, is eroding. In a majority of studies, age of respondent has not been a Significant predictor of support for programs for older adults, and younger respondents 110 have been as supportive as older respondents. The one exception is Rosenbaum and Button’s (1993) study of Floridians where there was some erosion of support among younger groups of respondents. H016. There is no relationship between age and support for increasing out-of-pocket costs for Medicare. Do not reject H022. There is no relationship between age and support for adding prescription drug coverage to Medicare. Do not reject H028. There is no relationship between age and support for paying higher taxes for prescription drug coverage for Medicare. Do not reject H034. There is no relationship between age and support for Reject raising the age of eligibility for Medicare. A year later, when SOSS-18 was conducted, in 1999, age is significant in predicting support for increasing the age of eligibility for Medicare. Those aged 18 to 24 are more likely to strongly agree (11.2%) or somewhat agree (41 .8%) that the age of eligibility for Medicare should be raised. Fifty-three percent of young adults aged 18 to 24 agree the age of eligibility for Medicare should be raised This could reflect increased concern with the financial precariousness of Medicare and some self-interest on the part of the youngest age group and the one least likely to benefit in the future from Social Security and Medicare. By increasing the age of eligibility for Medicare, the program might be viable longer. These findings differ from those reported by the Third Millennium (2001) study, which found that a majority (5 5%) of young adult respondents oppose raising the age of eligibility for Medicare to 67, and an even greater majority (75%) oppose raising the age of eligibility to 70 over the next 20 years. There is not a significant correlation between age and increasing out-of-pocket costs 111 for Medicare. Again, this differs fiom Third Millennium findings that support raising the out-of-pocket costs of those older persons who can afford to pay them. Lower income older adults would pay fewer out-of-pocket costs. In addition, age is not significant in predicting support for adding a prescription drug benefit to Medicare, or in predicting support for raising taxes to pay for a prescription drug benefit. The construct of family solidarity, the theoretical basis for this study, would suggest that because intergenerational bonds are formed within the family system and the generations learn a sense of obligation toward one another, solidarity between age groups is reflected in support among younger age groups for programs benefiting older persons and vice versa In fact, Silverstein and Parrott (1997) found that young adults who, as children, had experienced a great deal of contact with their grandparents tended to be more supportive of programs benefiting older persons than those young adults who had not. Young adults with no childhood contact with grandparents tended to see benefits to the elderly as too costly, thought Social Security should be taxed, and that older adults should contribute more to paying for their health care (Silverstein & Parrott, 1997). While none of the SOSS instruments assessed past experience with grandparents, it may be that the eroded support among younger age groups for programs benefiting older persons is an indication of less contact between the generations. Pezzin and Schone (1999) found that marital disruption had a negative effect on family solidarity and were concerned that today’s high rates of divorce and remarriage may alter the family’s role as a support network. It may also be that younger cohorts who have eXper-ienced divorce in their families have experienced less contact with grandparents as a result. While Pezzin and Schone (1999) were concerned with the detrimental effects of divorce on older persons, younger cohorts also are negatively 112 affected by divorce, as is family cohesion. If intergenerational bonds forged within the family microsystem have been disrupted by divorce or other kinds of conflict within the family, support among younger cohorts for programs benefiting older persons may be on the decline. Sex of Respondent H02. There is no relationship between sex of respondent and support for programs for older adults. Reject H010. There is no relationship between sex of respondent and support for programs for older adults. Reject Sex of respondent is significant in predicting support for programs benefiting older adults in both SOSS-3 in 1995 and SOSS-15 in 1999. In fact, support among women increases over the three years. In 1995, 82.4% of women respondents somewhat or strongly disagree with the statement “too many tax dollars are going to support older people.” In 1998, that percentage increases to 87.7%. However, results indicate that support among male respondents decreases by about the same percentage over the three years. H017. There is no relationship between sex of respondent and support for increasing out-of-pocket costs for Medicare. Reject H023. There is no relationship between sex of respondent and support for adding prescription drug coverage to Medicare. Reject H029. There is no relationship between sex of respondent and SUPP?“ for paying higher taxes for prescription drug coverage for Medlcare. Do not reject H035. There is no relationship between sex of respondent and support for raising the age of eligibility for Medicare. Reject 113 SOSS-18 results indicate sex of respondent is significant in predicting support for adding a prescription drug benefit to Medicare, as women are more likely to indicate support Women also are more likely to be against raising the age of eligibility for Medicare. Sex of respondent is also significant in predicting support for increasing out- of-pocket costs for Medicare as female respondents are more likely to support decreased costs and less likely to support increasing them. There is no difference in willingness to pay higher taxes for a prescription drug benefit for Medicare. However, for the most part, women are more supportive than men for programs benefiting older adults. These results are consistent with other studies (Cook & Barrett, 1988; Rosenbaum & Button, 1993; Third Millennium, 2001) that find more support among women than men for programs that benefit older adults. Rosenbaum and Button (1993) hypothesize that stronger support on the part of women for programs for older adults may indicate empathy for an aged population that is predominantly female. However, the theory of family solidarity may be most useful in explaining the consistent support among women for programs benefiting older persons. Women are the primary caregivers in the family. Not only are they more involved with the raising of children as “fathers are engaged with their children about 40% of the time mothers are involv ” (Schaie & Willis, 2001), they are also the primary caregivers to aging parents (Brody, 1981; NAC/AARP, 1997; Schaie & Willis, 1996, 2001; Walker, 1996). In fact, often women at midlife are engaged in caring for both the generation preceding and the generation following their own. Caregiving is an example of functional solidarity and involves an exchange of resources within the family that may be financial, physical, or Psychological, as caregiving often involves emotional as well as physical support. 114 Caregiving is also an example of normative solidarity, measured by the strength of family role commitment and the degree to which the generations feel responsible for each other. Women internalize the norm of filial obligation from their family experience and from the social environment and feel duty-bound to not only care for their offspring but also their aged parents. Thus, women, perhaps even to a greater degree than men, have recognized the value of the caring relationship within the family and have deeply internalized the norm of reciprocity. This norm, according to Walker (1996) and Foner (2000), is reflected on a societal level by support for programs benefiting older persons. Political Party A flil iation H03. There is no relationship between political party affiliation and support for programs for older adults. Reject H011. There is no relationship between political party affiliation and support for programs for older adults. Reject Political party affiliation also is an important predictor of support for older adults. In particular, Democrats in both the 1995 SOSS-3 round and the 1998 SOSS-15 round are significantly more likely (62.6 % and 57.2%) than Republicans (30.9% and 31%) to strongly disagree with the statement “too many tax dollars are going to support older people.” These findings are consistent with other studies. For example, Cook and Barrett (1988) and Rhodebeck (1993) found Democrats to be more supportive than Republicans of programs benefiting older adults. H01 8. There is no relationship between political party affiliation Md support for increasing out-of-pocket costs for Medicare. Do not reject 115 H024. There is no relationship between political party affiliation and support for adding prescription drug coverage to Medicare. Reject H030. There is no relationship between political party affiliation and support for paying higher taxes for prescription drug coverage for Medicare. Reject H036. There is no relationship between political party affiliation and support for raising the age of eligibility for Medicare. Reject The more recent 1999 SOSS-18 also finds political party affiliation is an important indicator of support for Medicare and expanding Medicare’s benefits. Although there is support among the majority of respondents for a prescription drug benefit for Medicare, regardless of political party affiliation, Democrats (95.8%) are significantly more likely to indicate support than are Republicans (73.5%). The Medicare prescription drug benefit is currently under consideration by the United States Congress and while support appears to be bipartisan, there are clearly differences in proposed programs and the extent to which they benefit older adults. Not surprisingly, SOSS-l8 found that Republicans (67.5%) are significantly less likely than Democrats (84%) to support paying higher taxes for prescription drug coverage for Medicare. Income H04. There is no relationship between income and support for programs for older adults. Do not reject H012. There is no relationship between income and support for programs for older adults. Do not reject Income is not a significant predictor of support for programs benefiting older persons in either SOSS-3 or SOSS-15. This is inconsistent with Rhodebeck’s (1993) finding that 116 persons with lower incomes tended to be more supportive of programs benefiting the aged Others (Cook & Barrett, 1988; Rosenbaum & Button, 1993) have found the significance of income to be inconclusive in predicting support. However, Foner (2000) argues that income is indeed important to the debate and claims it is class rather than age that is at the core of disagreement over programs for older persons. Foner (2000) notes that middle and working class individuals generally indicate strong support for Social Security and Medicare, while those with higher incomes see a tax advantage in cutting back social welfare programs that benefit older adults. Thus, it was expected that income would be significant in predict support for programs benefiting the aged. H019. There is no relationship between income and support for increasing out-of-pocket costs for Medicare. Do not reject H025. There is no relationship between income and support for adding prescription drug coverage to Medicare. Reject H031. There is no relationship between income and support for paying higher taxes for prescription drug coverage for Medicare. Do not reject H037 . There is no relationship between income and support for raising the age of eligibility for Medicare. Do not reject Results from SOSS-18 were consistent with those of SOSS-3 and SOSS-15, as there was not a relationship between income and support for increasing out-of-pocket costs for Medicare or support for raising the age of eligibility for Medicare. The one exception was in support for adding the prescription drug benefit to Medicare. Here income was significant, as those with higher incomes were less supportive. Interestingly, however, there was not a significant relationship between income and willingness to pay increased taxes to fund the drug benefit. Therefore, income appears to be inconclusive in 117 predicting support. Status as Caregiver H05. There is no relationship between status as a caregiver and support for programs for older adults. Do not reject H013. There is no relationship between status as a caregiver and support for programs for older adults. Reject Another way this study is unique is that it examined the relationship of caregiving status to support for programs designed for older persons. Since primarily women are the informal caregivers to older persons, and, as indicated previously, women consistently favor programs benefiting older adults, it was expected that individuals involved in caregiving would be more supportive. Yet caregiving status is not significant for SOSS- 3, but it is significant for SOSS-15, as 52.3% of caregivers as opposed to 38.8% of those not providing care strongly disagree with the statement “too many tax dollars are going to support older people.” Unfortunately, caregiving status is not assessed in the 1999 SOSS-18 round Thus, caregiving status is inconclusive in predicting support for programs benefiting older adults. Race H06. There is no relationship between race and support for programs for older adults. Reject H014. There is no relationship between race and support for programs for older adults. Do not reject Race is significant in predicting support for programs benefiting older adults in 1995 as African Americans are more likely to disagree with the statement “too many tax dollars are going to support older adults.” This result is consistent with Cook and 118 Barrett’s (1988) findings that Blacks were somewhat more supportive than Whites of programs benefiting older persons. However, race is not significant in 1998. H020. There is no relationship between race and support for increasing out-of-pocket costs for Medicare. Reject H026. There is no a relationship between race and support for adding prescription drug coverage to Medicare. Do not reject H032. There is no relationship between race and support for paying higher taxes for prescription drug coverage for Medicare. Do not reject H03 8. There is no relationship between race and support for raising the age of eligibility for Medicare. Do not reject Yet in 1999, race is significant in predicting support for increasing out-of-pocket costs for Medicare, as African Americans are more likely to favor decreasing costs than are Whites who are more likely to favor increased costs. Again, race is not significant in predicting support for adding the prescription drug benefit to Medicare, in predicting support for higher taxes to pay for the prescription drug benefit, or in predicting support for raising the age of eligibility for Medicare. It would appear, then, that race is not a factor in support for programs benefiting older adults. However, as American society becomes increasingly diverse and resources more scarce, “generational politics might become the norm, combined with class, race, and ideology” (Torres-Gil, 1992, p. 35). Level of Education H07. There is no relationship between educational level and support for programs for older adults. Do not reject 119 H015. There is no relationship between educational level and support for programs for older adults. Do not reject There was not a significant relationship between respondents’ educational level as measured by years of education completed and support for programs benefiting older adults for either SOSS-3 or SOSS-15. H021. There is no relationship between years of education and support for increasing out-of-pocket costs for Medicare. Do not reject H027 . There is no relationship between years of education and support for adding prescription drug coverage to Medicare. Reject H033. There is no relationship between years of education and support for paying higher taxes for prescription drug coverage for Medicare. Do not reject H039. There is no relationship between years of education and support for raising the age of eligibility for Medicare. Do not reject Level of education is significant in predicting support for adding the prescription drug benefit to Medicare with those having more years of education less supportive. However, education level is not significant in predicting support for paying higher taxes for prescription drug coverage, for raising the age of eligibility for Medicare, or for increasing Medicare’s out-of-pocket costs. Thus, level of education is inconclusive in predicting support for programs benefiting older adults. Older Adults Living with Adult Children H08. There is no relationship between those who believe older adults should live with adult children and support for programs for older adults. Do not reject One measure of family solidarity is the belief that older family members who are 120 unable to live independently should live with their adult children (Peek et al., 1998). It was expected that those who share this belief would also be more supportive of programs benefiting older persons. However, there was not a statistically significant relationship between these variables. The question was posed in SOSS-3 but not in SOSS-15 or SOSS-18. Implications for Social W011 The SOSS study results are important to social work because they contribute an understanding of opinions toward the larger debate in American politics over social policy. The generational equity debate is ultimately about who is responsible for maintaining older adults, the federal government or private sector, including individuals being responsible for their own retirement income. Ultimately, the decision affects social policy considerations. Those who believe persons should be responsible for themselves tend to see programs such as Social Security and Medicare as “entitlement” programs, welfare programs to support the aged. Currently it is politically fashionable to be anti-welfare and welfare recipients; when programs for older adults are fiamed as welfare, it becomes more acceptable to consider cutting them. The inconsistencies in support among age groups for programs for older adults bear watching. To date, age has not been predictive of support or lack thereof for programs for older adults. However, as the population of the United States continues to age and as the so-called baby boom generation retires and becomes eligible for Medicare and Social Security benefits, there will be more of a strain on the system. There also may be more tension over the increasing costs of assisting an older and larger age cohort and competition for increasingly scarce resources (Torres-Gil, 1992). In addition to policy issues, social work also is concerned with microlevel issues 121 involving families. Close relationships between the generations contribute to the well- being of both children and adults, are a source of social support, and enrich the lives of all. Bengtson (2001) notes that multi generational families are becoming more common with increasing life expectancy. He argues that multigenerational bonds are replacing nuclear family ties as the source of emotional and social support as well as transfers of time and economic aid (Bengtson, 2001). However, Pezzin and Schone (1999) are concerned that the nation’s high divorce rate ultimately will damage multigenerational bonds at the microlevel. Their study indicated that parents’ marital disruption had a negative effect on intergenerational transfers within the family (Pezzin & Schone, 1999 ). Their concern is that older adults of the firture who have experienced significantly higher rates of divorce and remaniage will not experience the benefits of strong intergenerational ties enjoyed by older adults today. In light of the present study, however, a disruption of intergenerational connections on the microlevel of the family may ultimately affect younger cohorts’ support for programs for older adults. There also is evidence that the reverse is true; macrolevel policies affect intergenerational relationships on a microlevel. In a study of five different nations with social security systems for older adults, Kunemund and Rein (1999) found that the most generous systems that gave the greatest resources to the aged helped to increase family solidarity. Kunemund and Rein (1999) conclude that cuts in spending for older adults would ultimately harm the family. As the nation ages and as more families become multi-generational, relations between age groups as well as within families will continue to be salient issues for social workers and policy makers. Social policy affects families as well as age groups, and the microsystem and the macrosystem are closely related It also may be that solidarity 122 within the family translates into intergenerational solidarity throughout the larger society. mam for F1_rrther Resefl Certainly a subject for firture research would be a continuing study of age groups and their support for programs benefiting older adults. The inconsistencies in support among age groups found in this study may indicate that the youngest groups are beginning to resent the level of resources aimed at the aged As F oner (2000) suggests, past support is not a guarantee of future support and the issues we, as an aging nation, must confront may become divisive enough to create a schism between young and old. Another potential area for further research isthe long-term effect of marital disruption on intergenerational solidarity within the family and ultimately throughout society. Pezzin and Schone’s (1999) study raises some w__:_ar_ni_ng flags that should be investigated further in light of the divorce rate in the United States. Torres-Gil (1992) also notes that because more adults are remaining childless, older adults in the future will be without children of their own to care for them and will have to depend on other than family members for support. Thus, the sense of obligation between young and old may diminish (Torres-Gil, 1992). Torres-Gil (1992) also predicts that the old-age lobby will subdivide into diverse groups of older women, more affluent older persons, minorities, and persons with disabilities, and all will compete for public resources. Future research also must do a better job of obtaining samples that are more representative of the increasing diversity of American society and inclusive of the many groups that are a part of this culture in order to obtain their opinions about social programs. Hispanics are one of the fastest growing segments of the population, and yet there were so few in the SOSS samples that statistical analysis of their support for programs benefiting older adults is impossible. 123 Foner (2000) claims that the real equity issue is not one of conflict between age groups but conflict between classes, with those with more financial resources less supportive of government programs benefiting older persons. For the most part, the present study indicates income is not a significant predictor of support, with the exception of support for adding a prescription drug benefit to Medicare. Those with higher incomes tend to be less supportive of the prescription drug benefit. A related issue is that of means-testing benefits. Eligibility for Social Security and Medicare has never depended on the beneficiary’s income. Although it would appear Social Security is financially solvent, the same cannot be said for Medicare (Baker & Wesbrot, 1999), and one proposal for improving its financial viability is requiring beneficiaries to contribute more to cover Medicare’s out-of-pocket costs. While the SOSS-18 data does not indicate a great deal of support for increasing out-of-pocket costs, the question of basing those costs on the beneficiary’s ability to pay was not assessed and should be a question for future research The Third Millennium (2001) survey found that young adult respondents were most supportive of a prescription drug benefit for Medicare based on financial need. Again, this is a question that would be useful in directing future research. Some of the preceding issues may be resolved with efforts to unite the generations. Generations United, 8 national coalition dedicated to increasing public awareness of the common issues faced by all generations, was founded in 1986. The organization is co- chaired by the Child Welfare League of America and the National Council on the Aging and works to foster generational interdependence (V entura-Merkel, 1990). One of their major goals is to unite age-differentiated groups on public policy efforts regarding issues that address human needs across all generations. 124 BIBLIOGRAPHY American Association of Retired Persons. (1995, July). Images of Aging in America. Washington, DC: American Association of Retired Persons. Baker, D., & Weisbrot, M. (1999). Social Securig: The phony crisis. Chicago: University of Chicago Press. Bengtson, V. L. (2001). Beyond the nuclear family: The increasing importance of multigenerational bonds. Journal of Marriage and Family, 93, 1-16. Bengtson, V. L., Marti, G., & Roberts, R. E. L. (1991). Age-group relationships: Generational equity and inequity. In K. Pillemer & K. McCartney (Eds), Parent- child relations throughout life (pp. 252-278). Hillsdale, NJ: Lawrence Erlbaum Associates. Bengtson, V. L. & Roberts, R. E. L. (1991). Intergenerational solidarity in aging families: An example of formal theory construction. 10% of Marriage and Family, 5;, 856-870. Berkowitz, E. D. (1997). The historical development of social security in the United States. In E. R. Kingson & J. H. Schulz (Eds), Soci_al Secgfim in the 21% cengy (pp. 22-3 8). New York: Oxford University Press. Bergquist, W. H., Greenberg, E. M., & Klaum, G. A. (1993). In our fifties: Voice§_9j men and women reinventing glen lives. San Francisco: Jossey-Bass. Binstock, R H (1991). Aging, politics, and public policy. In B. B. Hess & E. W. Markson (Eds), (flowing old in America (4th ed.) (pp. 325-340). New Brunswick, NJ: Transaction Publishers. Brody, E. M (1981). “Women in the middle” and family help to older people. The, Gerontologist, _2_1, pp. 471-480. Burton, L. M, & Dilworth-Anderson, P. (1991). Intergenerational family roles of aged Black Americans. In S. K. Pfeifer & M. B. Sussman (Eds), Families: Intergenerational and generational connections (pp. 311-330). New York: Haworth Press. Cook, F. L., & Barrett, E. J. (1988). Public support for Social Security. Journal 9f Aging Studies, _2_, 339-356. Cohler, B. J. and Altergott, K. (1995). The family of the second half of life: Connecting theories and findings. In R Blieszrrer & V. H Bedford (Eds), Handbook of ggigg and the family (pp. 59-94). Westport, CT: Greenwood Press. 125 Day, C. L. (1993). Public opinion toward costs and benefits of Social Security and Medicare. Research on Agipg, L5,, 279-298. Fink, A. (1995). How tp analyz_e survgy data. Thousand Oaks, CA: Sage Publications. Foner, A. (2000). Age integration or age conflict as society ages? The Gerontolog'st, 4_Q, pp. 272-276. George, L. K (1986). Caregiver burden: Conflict between norms of reciprocity and solidarity. In K. A Pillemer & R. S. Wolf (Eds), Elder abge: Conflict in the family (pp. 67-92). Dover, MA: Auburn House. George, L. K. (1998). Intergenerational transfers: Who gives and who gets? Journal of the American Society of CLU and ChFC, 52 (2). 28-33. Harootyan, R. A. & Bengtson, V. L. (1994). Intergenerational linkages: The context of the study. In V. L. Bengtson & R. A. Harootyan (Eds), Intergenerational linkages: Hidden connections in American Society (pp. 1-18). New York: Springer. Hareven, T. K. (1996). Introduction. In T. K Hareven (Ed), Aging and generational relations over the life cogrse: Life-course and gross-cult_l_l_ra_l mrspe_c_tives (pp. ix- xxv). New York: Aldine De Gruyter. Haveman, R. (1993). When problems outrun policy. Challgnge, _3_§ (3), 28-41. Hembroff, L. A., & Silver, B. D. (1995). State of the sate survey: Methodological repprt (1995-B). East Lansing, MI: Michigan State University, Institute for Public Policy and Social Research, Survey Research Division. Hembroff, L. A., & Silver, B. D. (1998). State pf the state survey: Methodological repprt ( 1998-B). East Lansing, MI: Michigan State University, Institute for Public Policy and Social Research, Survey Research Division Hembroff, L. A., & Silver, B. D. (1999). State of the Etc survey: Methodological remrt ( 1998-B ). East Lansing, MI: Michigan State University, Institute for Public Policy and Social Research, Survey Research Division. Hobbs, R B., & Damon, B. L. (1996). “65+ in the United States.” Current Population Revels: Special Smdies. Online. US. Bureau of the Census. Available: http:/www.census.gov/prod/1/pop/p23-190.html. 29 Aug. 2001. Kerlinger, F. N. (1986). Foundatiops of behavioral research (3rd ed). Fort Worth: Holt, Rinehart and Winston, Inc. Kronebusch, K., & Schlesinger, M. (1994). Intergenerational transfers. In V. L Bengtson & R. A. Harootyan (Eds), Interggerational linkages: Hidden conngtions in American socie_ty (pp. 112-151). New York: Springer. 126 Kunemund, H and Rein, M. (1999). There is more to receiving than needing: Theoretical arguments and empirical explorations of crowding in and crowding out. Ageing and Sociegy, 12, 93-121. Lawton, L., Silverstein, M, & Bengtson, V. L. (1994). Solidarity between generations in families. In V. L. Bengtson & R A. Harootyan (Eds), Intergeneretionel lipk_ages: Hidden connections in American socieg (pp. 19-42). New York: Springer. Lee, G. R., Netzer, J. K., & Coward, R. T. (1994). Filial responsibility expectations and patterns of intergenerational assistance. Journal of Marriage and Family, 151, 559- 565. Lee, Y. J ., & Aytac, I. A. (1998). Intergenerational financial support among Whites, African Americans, and Latinos. Journal of Marriage and Family, _6_(_)_, 426-441. Logan, J. R. & Spitze, G. D. (1995). Self-interest and altruism in intergenerational relations. Demoggphy, Q, 353-364. Luescher, K. and Pillemer, K. (1998). Intergenerational ambivalence: A new approach to the study of parent-child relations in later life. J oumal of Maniege and Family, 60, 413-425. Marmor, T. R., Cook, F. L., & Scher, S. (1997). Social security politics and the conflict between generations. Are we asking the right questions? In E. R. Kingson & J. H Schulz (Eds), Social Security in the let Cent1_1_ry (pp. 195-207). New York: Oxford University Press. MacManus, S. A. (1996). Young v. old: Generational combat in the let cent_r_ny. Boulder, CO: Westview Press. Minkler, M (1987). The politics of generational equity. Social Poliey, 11(2), 48- 52. Moody, H. R (1994). Aging. Thousand Oaks, CA: Pine Forge Press. Moon, M (1997). Are Social Security benefits too high or too low? In E. R Kingson & J. H. Schulz (Eds), Social Securig in the 21 st Cent_t_lry (pp. 62-75). New York: Oxford University Press. Morris, C. R. (1996). The AARP: Americe’s most mwerful lobm and the clash of generations. New York: Times Books. National Alliance for Caregiving and American Association of Retired Persons. (1997, June). Family Caregiving in the US: Findinge from a national survey. Bethesda, MD: National Alliance for Caregiving. 127 Peek, M. K., Coward, R. T., Peek, C. W., & Lee, G. R. (1998). Are expectations for care related to the receipt of care? An analysis of parent care among disabled elders. The Journals of Gerontology, 53B, $127-$136. Peek, M. K., Coward, R. T., & Peek, C. W. (2000). Race, aging, and care. Reseach on Aging, 2;, 117-142. Pezzin, L. E., & Schone, B. S. (1999). Parental marital disruption and intergenerational transfers: An analysis of lone elderly parents and their children. Demogr_aphy, 3_6, 287-297. Phillips, L. R. (1986). Theoretical explanations of elder abuse: Competing hypotheses and unresolved issues. In K. A Pillemer &. R. S. Wolf (Eds), Elder abuse: Conflict in the family (pp. 197-217). Dover, MA: Auburn House. Ponza, M, Duncan, G. J., Corcoran, M, & Groskind, F. (1988). The guns of autumn? Age differences in support for income transfers to the young and old. Public Opinion Merly, _5_2_, 441-466. Price, M C. (1997). Jp_stice between generations: The growing ppwer of the elderly in America Westport, CT: Praeger Publishers. Reno, V. P. & Friedland, R. B. (1997). Strong support but low confidence. What explains the contradiction? In E. R. Kingson & J. H. Schulz (Eds), Social Secu_rity in the let Cen_tu_ry (pp. 178-194). New York: Oxford University Press. Rhodebeck, L. A. (1993). The politics of greed? Political preferences among the elderly. The Journal of Polities, s5, 342-364). Rogler, L. H, & Cooney, R. S. (1991). Puerto Rican families in New York City: Intergenerational processes. In S. K. Pfeifer & M. B. Sussman (Eds), Families: Intergenerational and generational connectiope (pp. 331-349). New York: Haworth Press. Rosenbaum, W. A., & Button, J. W. (1993). The unquiet future of intergenerational politics. The Gerontofigiet, 3;, 481-490. Schaie, K. W., & Willis, S. L. (1996). Adult development and aging (4th ed). New York: HarperCollins. Schaie, K. W., & Willis, S. L. (2001). Adult development and agpr' g (5th ed). New York: HarperCollins. Silverstein, M., & Parrott, T. M. (1997). Attitudes toward public support of the elderly: Does early involvement with grandparents moderate generational tensions? Research on Aging, _12, 108-132. Third Millennium. (1993). Third Millennium Declaration, [Online]. Available: 128 http://www.thirdrnil.org/publications/declare.html [2001, April 8]. Third Millennium. (2001). Third Millennium Medicare and prescription drugs survey of 500 Americans ages 18-34, [Online]. Available: http://www.thirdmil.org/publications/surveys/O103medicarefinding.html [2001, April 8]. Thompson, L. H., & Upp, M. M. (1997). The social insurance approach and social security. In E. R Kingson & J. H. Schulz (Eds), Social securig in the let cent_u;y (pp. 3-21). New York: Oxford University Press. Torres-Gil, F. M. (1992) The new aging: Politics and change in Americe. NY: Auburn House. Ventura-Merkel, C. (1990). Strategies for change: Builging state and local coalitions on intergenerational issues and mogzgrps. Washington, DC: Generations United Walker, A. (1996). Intergenerational relations and the provision of welfare. In A Walker (Ed), The new generational connect: Intergenerational relations, old age end welfare (pp. 10-36). London: UCL Press. Weisberg, H F ., Krosnick, J. A, & Bowen, B. D. (1996). An introduction to survey research, mlling, and data analysis (3rd ed). Thousand Oaks, CA: Sage. Williams, A, Coupland, J., Folwell, A, & Sparks, L. (1997). Talking about Generation X: Defining them as they define themselves. Journal of Langu_age and Social Psychology, L6, 251-277. Wisensale, S. K (1988). Generational equity and intergenerational policies. m Gerontologist, _28, 773-778. Wong, K, Kitayarna, K. E., & Soldo, B. J. (1999). Ethnic differences in time transfers from adult children to elderly parents. Unobserved heterogeneity across families? Research on Agi_ng, 21, 144-175. 129