Civic Engagement as a Retirement Role for Aging Americans

Transcription

1 The Gerontologist Vol. 48, No. 3, Copyright 2008 by The Gerontological Society of America Civic Engagement as a Retirement Role for Aging Americans Brian Kaskie, PhD, 1 Sara Imhof, PhD, 1 Joseph Cavanaugh, PhD, 1 and Kennith Culp, PhD 1 Purpose: Public attention directed toward the civic engagement of retired Americans has increased considerably. The purpose of this research was to define civic engagement as a retirement role and differentiate individuals who met this role definition from other retirees. Design and Methods: Retirees who met our definition of civic engagement were identified from a sample of 683 retired older adults living in a rural Midwestern state. Using a multinomial logistic regression analysis, we contrasted this group of engaged retirees to three other groups of retirees: (a) those who were neither working nor volunteering, (b) those who had returned to work in part-time or seasonal occupations, and (c) those who volunteered for fewer than 5 hours each week. Results: The analyses indicated that individuals assigned to the group of engaged retirees were similar across 24 variables. This group of engaged retirees differed from the other groups by gender, education level, and health status; retirement motives and planning efforts; primary retirement activities; and attitudes about volunteering and returning to work. Implications: This research supported the contention that civic engagement could be defined as a formal retirement role, as engaged retirees differ significantly from those who volunteer less, work in noncivic roles, or do neither. Further, we resolved that associating the act of volunteering, in itself, with civic engagement may no This project was supported by a contract awarded to the University of Iowa Center on Aging from Iowa Workforce Development from January 1, 2004 to June 30, We thank Jane Barto, Mary Ubinas, and Paulette Day from Iowa Workforce Development for supporting this research study. We acknowledge the support provided by the University of Iowa Center on Aging and the Center for Health Policy and Research. We also express our appreciation to a number of students who assisted in this effort: Patricia Bonica, Natalie Gerstner, Daniel Gregory, Sara Hill, Scott Kallemeyn, Carly Runestad, Sanjana Raghavan, Nikki Stacjar, and Lora Washington. Address correspondence to Brian Kaskie, PhD, Department of Health Management and Policy, College of Public Health, University of Iowa, 200 Hawkins Drive, E206GH, Iowa City, IA briankaskie@uiowa.edu 1 College of Public Health, University of Iowa, Iowa City. longer be appropriate for retired older adults. The definition of civic engagement as a retirement role should also include those individuals who return to work in organizations that pursue specific civic activities. Key Words: Civic engagement, Older adults, Retirement roles, Volunteering Public attention directed toward the civic engagement of aging Americans has increased considerably over the past 5 years. Civic engagement was a featured topic at the 2005 White House Conference on Aging, and national organizations such as the American Society on Aging (2006), the Gerontological Society of America (2004), and the National Council on Aging (2006) have made civic engagement a programmatic priority. Most recently, the reauthorization of the Older Americans Act provided new authority to the Administration on Aging to develop and implement programs that facilitate the civic engagement of older Americans (Administration on Aging, 2007). One reason for this attention rests on the assumption that retired older adults are an untapped resource, and increasing the civic engagement of retired Americans will correspond with increasing social capital (Galston & Lopez, 2006; Harlow-Rosentraub, Wilson, & Steele, 2006; Sander & Putnam, 2006). Another reason for such ample interest with civic engagement has to do with the potential effect on individual health. Several researchers have linked engagement with healthy, successful aging and have suggested that aging persons who continue to work, find a second career, volunteer, or become involved in local affairs maintain better physical and mental health as they grow older (Glass, Mendez de Leon, Marotolli, & Berkman, 1999; Herzog & House, 1991; Morris & Caro, 1995; Seeman & Crimmins, 2001). In this article, we consider how civic engagement could be defined as a retirement role for aging Americans, analyze survey responses provided by 683 retirees 368 The Gerontologist

2 from a rural Midwestern state, and differentiate those persons who met our definition of civic engagement. What is Civic Engagement? Putnam (1995) suggested that civic engagement is defined by an individual s interest with the improvement of community programs, contemplation of public affairs, and knowledge of political elections. He also defined civic engagement as discrete activities and behaviors. In this case, civic engagement means being involved with community and political affairs helping a frail neighbor with grocery shopping, cleaning up the local park, voting, writing letters to public officials, and contributing to election campaigns. Other definitions of civic engagement only expand upon these notions. For example, the Pew Charitable Trust (2006) suggested that civic engagement consists of taking an interest with issues of public concern and participating in activities such as joining a neighborhood association or attending a community concert. The lack of a more precise definition of civic engagement has been criticized (Martinson & Minkler, 2006; Norris, 1996; Theiss-Morse & Hibbing, 2005). Does participating in a public hearing to support the use of public lands for commercial purposes, including your own, constitute civic engagement? Should voting for the party ticket once every 4 years during the Presidential elections be considered alongside voting in local primaries and board elections every year? The current definition of civic engagement does not provide clear answers to such questions, and the lack of a more precise definition is especially problematic when one is considering civic engagement among retired older adults (Wilson & Rymph, 2006). Given that most Social Security and Medicare recipients already meet some civic engagement criteria, the definition is all inclusive; this renders it inadequate. Civic Engagement as a Formal Role for Retired Older Adults Kaskie and Gerstner (2004) fielded surveys from a sample of 254 retired Californians who provided answers to several questions, including these: What are your attitudes about civic engagement?, What sort of activities do you consider to be a form of civic engagement?, and Which community organizations provide opportunities for civic engagement? Their results indicated that retired older adults defined civic engagement with the same broad concepts as Putnam civic engagement consists of a range of attitudes and behaviors. However, they also found that retirees who considered themselves engaged were volunteering as well as working. Among the survey respondents, almost 95% defined their volunteer service as a form of civic engagement. Among the retired survey respondents who had returned to work or had defined themselves as retired but continued to work in bridge jobs, more than 70% indicated they worked because they wanted to keep active, be engaged with other people, and make a contribution to their local community. Fewer than 30% of these retirees indicated they returned to work for financial reasons alone. Further, the sample of retired older adults defined civic engagement by a certain amount of commitment that goes beyond attitudes and discrete behaviors such as writing a letter to a public official or volunteering in a soup kitchen only one time each year. Kaskie and Gerstner (2004) concluded that, when applied to the retired older adult population, civic engagement should be defined as a role that involves voluntary or paid participation in an activity that occurs within an organization that has a direct impact on the local community. In their analytic sample, civic engagement involved a variety of forms such as direct caregiving, teaching, and providing supportive services, and these civic engagement experiences involved a commitment of at least 1 day each week and occurred within many settings such as faith-based organizations, schools, health care programs, social service agencies, and other nonprofit community programs. Defining civic engagement as a retirement role is consistent with the literature in regard to how population aging has affected life-course trajectories (Moen & Fields, 2002; Uhlenberg, 1992). Sheppard and Philibert (1972) were among the first researchers to recognize that retirement represented a transition in which adulthood roles such as career and family raising give way to another set of roles including volunteering and community service. Civic engagement, as defined here, is distinguished from volunteering in that it also can include paid work experiences and requires a greater commitment than most volunteer opportunities such as those provided through voluntary associations (Cutler & Hendricks, 2000). Moreover, the notion that civic engagement constitutes a retirement role was supported by Boggs, Rocco, and Spangler (1995), who suggested that individuals pass through different stages of civic engagement as they age; by retirement, civic engagement could easily constitute a role rather than discrete attitudes or behaviors. To support this position, they recounted how an individual s sense of civic engagement usually develops slowly over the life course, in small steps, such as voting at age 18 or spending an afternoon volunteering in a local nursing home. Through young and middle adulthood, individuals often become more attentive to community matters, and they acquire skills and resources that allow them to make more substantive contributions to organizations such as the Parent Teacher Association. Boggs, Rocco, and Spangler suggested that, by the time many individuals retire, they have Vol. 48, No. 3,

3 acquired a sufficient amount of knowledge, skill, and time that would allow them to make a greater commitment, voluntary or paid, to an organization that has a direct impact on the local community. Assuming that civic engagement does constitute a retirement role, empirical research is needed to establish which individuals occupy this role and how they differ from other retirees who volunteer, return to work, or do neither. Such information can serve at least two purposes. First, a more precise definition of civic engagement may help policy makers and program administrators target their efforts more effectively (AARP, 2003; Martinson & Minkler, 2006; Sheppard & Philibert, 1972; U.S. Government Accountability Office, 2005). Second, being able to define and differentiate individuals who occupy this role may help researchers conduct a more rigorous evaluation of the relationships among civic engagement, social capital, and individual health outcomes (Wilson & Rymph, 2006). Previous Research To our knowledge, there have been few, if any, empirical evaluations of civic engagement as a retirement role (defined here as a volunteer or work role to be occupied for at least 1 day a week within an organization that has a direct impact on the local community). Instead, the majority of research has focused on varying levels and types of volunteerism and work participation among retired older persons. Our review of this research literature pointed to several variables that distinguished retired volunteers and workers from those retirees who were less involved or not at all involved in these activities. For example, Blau (1994) and Dye, Goodman, Roth, Bley, and Jensen (1973) differentiated retirees who volunteered or returned to work by individual characteristics such as age, education, gender, health, and income. Fischer, Meuller, & Cooper (1991) found significant differences in the amount of time older adults spent as volunteers, with many spending fewer than 5 hours each week. Other research indicated that volunteers and working retirees differ in terms of motivation, planning efforts, selective optimization, family role conflicts, personal flexibility, caregiving responsibilities, and the cognitive and physical demands that they are able to assume (Hendricks & Cutler, 2004; Maestas, 2004; Okun & Schultz, 2003). Elshaug and Metzer (2001) identified significant personality differences between retired older volunteers and retired persons who returned to work. Taken together, these studies suggested that there may be several individual-level variables that differentiate retirees who assume a civic engagement role from those who do not, and there also is reason to suspect that individuals who meet our definition of civic engagement may differ in ways besides whether they serve as volunteers or have returned to work. Research Objectives In this study we examined a sample of 683 retired older adults from a rural Midwestern state and determined how many of these individuals met our criteria for civic engagement. Next, we tested for differences between the engaged retirees who volunteered for more than 5 hours a week and the engaged retirees who took paid jobs in organizations that were tied to civic engagement. Finally, we compared these retirees (both engaged volunteers and engaged workers) with other retirees who neither volunteered nor worked, returned to work in roles that were not considered to be forms of civic engagement, or volunteered for fewer than 5 hours each week. Methods The Iowa Workforce Seniors Study (IWSS) was conducted between August 2004 and March 2005 under contract from the Iowa Workforce Development Agency. The study fielded a written survey concerning retirement, volunteering, and work from a sample population consisting of 1,800 randomly selected registered voters between the ages of 55 and 75 years from three Iowa counties: one county was defined as rural, the second was urban, and the third was a mix of cities and smaller towns. The IWSS received survey responses from 1,170 of the 1,800 individuals (65%) and then separated the survey respondents into two groups: those who reported that they were still working and not retired (n = 487) and those who indicated they were retired (n = 683). The cross-sectional analysis of individuals who were still working and not yet retired has been published elsewhere (Kaskie, Imhof, Cavanaugh, & Culp, 2005). Given our interest in studying civic engagement as a retirement role, in this study we focused on the individuals who identified themselves as retired. Participants The 683 individual IWSS survey respondents who identified themselves as retired constituted the analytic sample for this research. The average age of these retired older adults was 68.1 years (SD=4.9; range = 56 75). Of the participants, 47.0% percent were male and 53% were female; 76.2% were married, 16.6% lived alone, 95.8% were Caucasian, and 74.4% lived in a town or city; 5.1% relied on public transportation or someone who drove for them; 80.2% rated their health as good or excellent; 19.6% completed a college degree; and 66.3% had an average annual income of more than $20,000 (the 2005 federal poverty line for a two-person household was $12,830). Although the sample population did not include the 11.7% of older Iowans who were not registered to vote, the characteristics of the analytic sample were comparable with those of the older adult 370 The Gerontologist

4 population at large living in the three counties (U.S. Bureau of the Census, 2000). Survey Data A 38-question survey was mailed to the IWSS study population. This survey requested data on retirement, volunteering, and work status, as well as individual characteristics such as age, gender, educational level, and self-rated health. Persons who indicated they were retired also provided their reasons for retirement, retirement planning efforts, retirement activities, and attitudes about volunteer and work opportunities in retirement. Some of the questions included in the IWSS survey were taken from the American s Changing Lives Survey, the Current Population Survey, and the Health and Retirement Survey. Other questions were developed specifically for the IWSS study and had not been included in previous survey research. For example, the IWSS collected more detailed information about motives for retirement, retirement activities, and attitudes about volunteering and returning to work. The survey was piloted on a group of volunteers at a local community senior center and reviewed by the staff directors from the Iowa Workforce Development offices located in the three counties. Following the Dillman (2000) technique of survey administration, the IWSS researchers mailed the IWSS surveys to the sample population in October of 2004; a reminder card was sent to individuals who had not completed the survey in November; and a second survey was mailed in December. Research assistants cross-checked all entered survey data for accuracy. Missing items constituted less than 5% of any single variable response and were treated as random. A copy of the survey can be found in the report prepared by Kaskie and colleagues (2005). Measures Status of Retired Iowans. On the basis of their survey responses, we assigned the retired individuals to one of five mutually exclusive groups: (a) engaged volunteers, defined as retired persons who volunteered at least 5 hours (1 day) each week; (b) engaged workers, defined as retired individuals who returned to work in an organization or job role that was associated with education, a faith-based activity, health care, public service, social service, or some other civic enterprise; (c) not engaged, defined as retired individuals who reported that they were neither volunteering nor working; (d) workers, defined as persons who were retired but had returned to work in an industry or job role that was not associated with education, a faith-based activity, health care, public service, social service, or some other civic enterprise; and (e) volunteers, defined as retired individuals who volunteered for fewer than 5 hours (1 day) each week. We assigned those retired individuals who indicated they were both volunteering and working on the basis of the amount of time involved in each activity. For example, a retired individual who reported working 2 days a week at a local shopping mall and volunteering 1 day a month at a church was designated as a retired worker. Independent Variables. Following our review of the research literature, we developed 24 variables to reflect five factors that differentiated retired volunteers and workers, and we hypothesized that these variables would help differentiate engaged retirees as well. These factors included the following: individual characteristics, reasons for retirement, retirement planning efforts, activities, and attitudes. Individual characteristics: The survey responses provided data pertaining to eight individual characteristics that may relate to group status (i.e., engaged worker, engaged volunteer, not engaged, worker, volunteer). These included age, gender, marital status, race, income, education level, self-reported health status, and physical activity level. Reasons for retirement, retirement planning, and retirement activities: Survey respondents indicated their reasons for retirement by selecting among nine choices (e.g., declining health, being laid off, pursuing personal interests, reaching financial security, care-giving responsibilities). They also reported whether they engaged in any formal retirement planning and what that entailed. Survey respondents identified their primary retirement activities by selecting among nine choices (e.g., pursuing hobbies, traveling, visiting family, watching television). Altogether we constructed 12 measures pertaining to motives, planning, and activities. Attitudes about volunteering and work: Retired survey respondents provided information concerning their attitudes about volunteering and working. We used this information to construct four measures concerning the perceived availability of volunteer opportunities, work opportunities, and the importance of salary and benefits. We also asked individuals if they would return to work if the right opportunity was presented. Analysis The analysis first consisted of compiling descriptive statistics and testing for differences among the five groups. We then tested a multivariable logistic regression model to identify any significant differences between those individuals who were involved in a civic engagement role as volunteers and those who worked in a role or organization that constituted a form of civic engagement. Next, we ran a multivariable multinomial logistic regression model in which engaged retirees served as the reference group to the three other groups of retirees: those who neither worked nor volunteered, those who Vol. 48, No. 3,

5 Table 1. Demographic Characteristics, Reasons for Retirement, Planning, Activities, and Attitudes Characteristic or Reason Description (%) Engaged Volunteer Engaged Worker Inactive Retiree Active Worker Active Volunteer Characteristic Gender*** Female Marital status Married Age M (SD) a 67.6 (5.4) 69.0 (4.2) 68.3 (4.7) 67.0 (4.8) 68.5 (4.8) Race b Caucasian Health status*** Very good excellent Education*** Bachelor degree Annual gross income ($)**.39, Physical activity*** Do not exercise Reasons for retirement Health and disability** Yes Financially secure*** Yes Caregiving obligations* Yes Personal interests*** Yes Encouraged to leave** Yes Retirement planning Talking with friends Yes Self-educated Yes Attitudes about community volunteering and work Adequate work opportunities*** Agree Amount of pay is important* Agree Adequate volunteer opportunities*** Agree Adequate retirement planning resources* Agree Retirement activities Travel*** Yes Watch TV** Yes Hobbies* Yes Visit family* Yes Visit friends* Yes Notes: Pearson s chi-square test is used unless otherwise noted. Retiree numbers are as follows: Engaged volunteer, engaged worker, inactive retiree, active worker, and active volunteer, n = 121, 43, 273, 105, and 141, respectively. a Analysis of variance; b Fisher s Exact Test. *p 0.05; **p 0.01; ***p worked in noncivic jobs, and those who volunteered fewer than 5 hours a week. The regression models took the following form: log½p ij ð1 2 p ij ÞŠ ¼ X i b þ Z i a þ M i v þ c i ; where log[p ij (1 p ij )] represented the odds of retired survey respondents i being assigned to group j. For the explanatory variables associated with ij respondents, X i b reflected the individual characteristics, Z i a were variables pertaining to retirement motives, planning, and activities, M i v were variables reflecting personal attitudes, and c i was the county effect. Results Retired Survey Respondents The survey responses indicated that 17.7% of the retired older adults volunteered for more than 5 hours each week (categorized as engaged volunteers) and 6.3% worked in a job that constituted a form of civic engagement (engaged workers). The responses also indicated that 39.9% of the retired survey respondents neither volunteered nor worked (not engaged), 15.4% worked in an organization or job that was not defined as a form of civic engagement (workers), and 20.6% of the retired older adults volunteered for fewer than 5 hours each week (volunteers). Significant group differences were consistent with previous research and indicated that retirees who neither worked nor volunteered generally were less educated, less financially secure, did not plan for their retirement as much, were less healthy, exercised less frequently, and more often cited health problems and disability as a primary reason for retirement. Summary and test statistics concerning the groups, their reasons for retirement, retirement planning efforts, activities, and attitudes about volunteer and work opportunities are presented in Table 1. Comparisons Between Retirees Who Met the Definition of Civic Engagement The univariable logistic regressions revealed no significant effects on the odds of being a retiree who 372 The Gerontologist

6 Table 2. Multinomial Logistic Regression Model Characteristic or Reason Inactive vs Engaged AOR (95% CI) Active Worker vs Engaged AOR (95% CI) Active Volunteer vs Engaged AOR (95% CI) Characteristic Gender (female vs male) 0.72 ( ) 0.42 ( )** 1.49 (0.89, 2.49) Health status (very good excellent vs 0.67 ( ) 2.67 ( ) 1.38 (0.60, 3.16) poor fair) Education (>high school vs high school) 0.38 ( )*** 0.31 ( )*** 0.49 (0.29, 0.81)** Physical activity (exercise vs do not 0.13 ( )** 0.25 ( ) 0.39 (0.09, 1.72) exercise) Reasons for retirement (yes vs no) Personal interests 0.63 ( ) 1.29 ( ) 0.83 (0.49, 1.40) Retirement planning (yes vs no) Talking with friends 0.69 ( ) 0.93 ( ) 0.42 (0.21, 0.83)* Self-educated 0.57 ( ) 0.82 ( ) 1.13 (0.64, 1.99) Attitudes about volunteering and work (agree vs disagree) Adequate work opportunities 0.42 ( )** 0.83 ( ) 0.65 (0.32, 1.30) Amount of pay is important 2.74 ( ) 9.51 ( )* 0.87 (0.37, 2.07) Adequate volunteer opportunities 0.17 ( )* 0.14 ( )* 1.04 (0.14, 7.97) Adequate retirement planning resources 1.12 ( ) 1.80 ( ) 1.03 (0.47, 2.25) Retirement activities (yes vs no) Watch TV 2.08 ( )** 0.82 ( ) 2.09 (1.19, 3.69)* Hobbies 0.94 ( ) 0.78 ( ) 1.06 (0.63, 1.79) Notes: AOR = adjusted odds ratio; CI = confidence interval. *p 0.05; **p 0.01; ***p volunteered for more than 5 hours each week compared with those of being a retiree who returned to work in a select occupation or organization. There also were no differences among these individuals across the three counties. The individuals who comprised the engaged retirees group were similar across all variables. Multinomial Comparisons We conducted a multinomial logistic regression to differentiate the group of engaged retirees (i.e., combined engaged workers and engaged volunteers) from the three other groups, which consisted of (a) those who neither worked nor volunteered, (b) those who worked in jobs not considered to be forms of civic engagement, and (c) those retirees who volunteered fewer than 5 hours each week. Engaged Versus Not Engaged. When comparing the engaged retirees with those retirees who neither volunteered nor worked, the regression indicated that not-engaged persons were less likely to have completed a high school education (odds ratio or OR = 0.38; confidence interval or CI = ), were less likely to exercise (OR = 0.13; CI = ), and were less likely to agree that work (OR = 0.42; CI = ) and volunteer (OR = 0.17; CI = ) opportunities were adequate in their community. The not-engaged retirees were more likely to indicate that they watch television as a primary retirement activity (OR = 2.08; CI = ). There were no significant differences across the three counties. The results of this multivariable multinomial regression model are featured in Table 2. Retired Workers. When comparing the group of engaged retirees with the group of individuals who worked in a job that was not defined as a form of civic engagement, the multivariable regression indicated that retired workers were less likely to be women (OR = 0.42; CI = ), were less likely to have completed a high school education (OR = 0.31; CI = ), were more likely to agree that being paid to work was important (OR = 9.51; CI = ), and were less likely to agree that volunteer opportunities were available in their community (OR = 0.14; CI = ). There were no differences across the three counties. The results of this multivariable regression are featured in the second column of Table 2. We also conducted a logistic regression to identify more specific differences between engaged workers and the other retired workers (i.e., we removed engaged volunteers from the group of engaged retirees). In addition to what was already reported, we found that engaged workers were less likely to be in very good or excellent health (OR=0.29; CI= ) and were more likely to indicate that they did not work out of financial need (OR = 3.55; CI = ). Vol. 48, No. 3,

7 Further, we conducted a logistic regression to determine if engaged retirees differed from 149 older workers (older than 65 years of age) from the IWSS sample who indicated that they were not yet retired (i.e., we compared the engaged retirees to an analytic sample of individuals selected from the original IWSS population). This analysis revealed that individuals who had retired but returned to work were more likely to exercise than those who continued to work after the age of 65 (OR = 0.18; CI = ). There were no other significant differences between the two groups. However, this analysis included only 12 variables as data on retirement activities, and we did not collect information on attitudes from individuals who indicated they were not retired. These findings are not presented in Table 2. Retired Volunteers. The multinomial regression featuring individuals who volunteered fewer than 5 hours a week indicated that they were less likely to have completed a high school education (OR = 0.49; CI = ), were less likely to discuss retirement plans with friends (OR = 0.42; CI = ), and were more likely to indicate that they watch television as a primary retirement activity (OR = 2.09; CI = ). There were no significant differences across the three counties. The results of this multivariable regression model are featured in the third column of Table 2. We also conducted a logistic regression to identify more specific differences between engaged volunteers and the other retired volunteers (i.e., we removed workers from the group of engaged retirees). We found that engaged volunteers were more likely to have been encouraged to retire (OR = 3.43; CI = ). This finding is not presented in Table 2. Discussion We found empirical support to define civic engagement as a retirement role. The retired persons who met our definition of civic engagement were similar across 24 variables and differed significantly from retirees who volunteered fewer than 5 hours each week, worked in noncivic jobs, or did neither. We also found that this role already may be occupied by as many as one out of every four retired older adults. This effort to establish a more precise definition may serve several purposes, as the civic engagement phenomenon is not likely to go away any time soon. In particular, organizations are being challenged to fill jobs within emerging industries (e.g., technology) and other industries being depleted by an aging workforce (e.g., education) (Kaskie et al., 2005). Many of these organizations can provide opportunities that are well suited for retirees who want to assume a civic engagement role that has a direct impact on their community (e.g., education, health care, and other nonprofit organizations). Martinson and Minkler (2006) echoed this sentiment and suggested the demand for retired adults will increase as federal and state governments continue to reduce financial support for education, health, and social service programs. Another reason that the civic engagement phenomenon will continue to expand relates to the growing number of persons who are retiring from their primary careers but may not be prepared to retire completely. Individuals who currently are between the ages of 55 and 64 years have comparatively lower amounts of wealth than the current generation of older adults and, thus, may be more likely to remain in the workplace or return to work as a way to maintain income (Cahill, Giandrea, & Quinn, 2006; Hershey & Mowen, 2000). Other retirees recognize that they will be spending more years in retirement and may wish to partake in a civic engagement role as a way to maintain their health and contribute to the social capital of their community (Sander & Putnam, 2006). Hendricks and Cutler (2004) found that volunteering in older age was associated with education levels, health, and other socioeconomic indicators. Given that many of these same variables differentiated the engaged retirees in this study, it is possible to follow on Hendricks and Cutler and argue that the coming generation of retired older Americans may have increased levels of civic engagement because there are a greater number of people with higher levels of education, good to excellent health status, and other characteristics that enable them to engage in civic activities. How Does Civic Engagement Stand Apart as a Retirement Role? Our analyses differentiated retirees who occupied a role defined as a form of civic engagement from those who were less involved in volunteer activities, employed in other jobs, or not involved in either. In particular, when contrasted with retirees who neither volunteered nor worked (i.e., nearly 40% of the survey respondents), the engaged retirees were more educated, in better health, and more physically active. These particular findings were consistent with previous research that found such individual characteristics differentiated volunteers from nonvolunteers and workers from nonworking retirees (e.g., Blau, 1994; Cutler & Hendricks, 2000; Dye et al., 1973; Ruhm, 1990). We also found that the engaged retirees were more likely to be aware of work and volunteer opportunities. This last finding follows on previous research that differentiated retirees by their attitudes and expectations, and it suggests that policies and programs that increase awareness and educate aging workers about postretirement options might facilitate increased engagement in work and 374 The Gerontologist

8 voluntary activities (Brougham & Walsh, 2005; Mermin, Johnson, & Murphy, 2006). In comparing the engaged retirees with retirees working in other jobs or organizations, we found that men and individuals with lower levels of education were more often involved in noncivic jobs or organizations. Significant differences also included attitudes about volunteer opportunities, and retirees who went back to work in noncivic roles placed a higher value on being paid to work. Cahill, Giandrea, and Quinn (2006) reported that some older individuals who continued to work were more likely to need supplemental income, whereas other older adults associated being paid to work with their personal quality of life. Considering our results, some retirees may have been in noncivic jobs because these positions paid more whereas others took noncivic jobs because they assumed that earning a higher salary corresponded with a higher quality of life. We suspected there might be other reasons why retirees returned to work (or continued to work) in noncivic roles and organizations. These individuals may have returned to work to obtain supplemental health care coverage and acquire other noncash benefits that accompany the position; earning an income may have been a secondary motivation (Kaskie et al., 2005). Cahill and colleagues (2006) suggested that as individuals move through the retirement process and take bridge jobs, consulting positions, and part-time and seasonal work opportunities, they often remain in the industry in which they spent their primary career. It may not be that these jobs pay more as much as they are familiar. Moreover, we suspect a certain number of retirees had a different perception of civic engagement and did not make the sort of distinctions that we made in this research. Indeed, retirees who return to work in a local coffee shop or hardware store may assume they are engaged in meaningful ways with members of their local community just as much as someone who works in a school, health care facility, or social service organization. There were two compelling differences between the engaged retirees and those who volunteered for fewer than 5 hours each week. First, retired volunteers were more likely to watch television as a primary retirement activity and less likely to talk with their friends about retirement options. This suggests that retired volunteers may be less interested in occupying more time-consuming civic engagement roles. Martinson and Minkler (2006) cautioned that the demand for retired individuals to fill civic engagement roles may be greater among the policy makers and program administrators who need to address imminent labor force challenges than among the retired individuals themselves. As our results suggest, the majority of retired volunteers spend fewer than 5 hours each week in their service, and they may not want to become any more engaged than that. Second, engaged volunteers more often reported that they were encouraged to leave the workforce. This difference suggests that retired volunteers may not have experienced the same sort of role loss (and need for role substitution; see Hendricks & Cutler, 2004) as those who were encouraged to leave the workforce before they were ready to retire. Szinovacz and Davey (2005) reported that as many as one out of every three older adults were forced to retire and were provided few choices at the end of their careers. They depicted how these individuals were also more likely to need financial assistance and more Medicare and Medicaid benefits. Szinovacz and Davey contended that policy makers and program administrators should develop practicable alternatives so that disabled and displaced individuals can remain engaged in some work or voluntary role rather than being forced to retire completely and draw down public benefits. Our results suggest that some individuals already have taken civic engagement roles as a practicable response to having been forced to retire. Directions for Policy and Program Activity As policy makers and program administrators develop and implement strategies to increase civic engagement among older adults (Administration on Aging, 2007), our findings point toward a number of workable alternatives. For example, given that Liebsen-Hawkins (2005) suggested the most effective way to increase someone s level of civic engagement is to ask them, our results would support the implementation of a public awareness and education campaign about civic engagement roles and how to go about finding one. With several of the nonengaged retirees reporting that they were unaware about volunteer and work opportunities in their community and 70% of retirees listing television viewing as a primary activity, it may prove effective to disseminate such information about civic engagement through targeted media campaigns. This approach appears inherently more practicable than focusing on an individual s level of education, health status, or many of the other nonmodifiable variables found to differentiate the engaged from other retirees. Another way to facilitate civic engagement among retirees might be to offer an appealing menu of competitive noncash benefits. Previous studies have revealed that the utility for many volunteers is about 100 hours of annual service (i.e., fewer than 5 hours per week), and few retirees are willing to serve more that 10 hours per week without some sort of incentive (Morrell-Howell, Hinterlong, Rozario, & Tang, 2003). Perhaps providing noncash benefits may be a practicable way to increase their engagement levels. For example, health care organizations could provide retirees who come back to work greater access to free or discounted health care services. Elementary and secondary schools could provide on-site meal vouchers or establish scholarship funds in which retired workers receive tuition credits to pay for the college Vol. 48, No. 3,

9 tuition of a grandchild. In fact, Martinson and Minkler (2006) recounted that offering such benefits has been an effective strategy for long-running national programs such as Foster Grandparents, Retired Senior Volunteers, and Senior Corps. Future Research Even though we found empirical support for our definition of civic engagement as a retirement role for aging Americans, further research is needed to determine the validity of this construct. One area of investigation should focus on the level of commitment required to fulfill a civic engagement role. Should those who volunteer for fewer than 5 hours a week really be separated from others who volunteer more? When does one go from participating in a civic activity to occupying a civic engagement role? Another area of research should consider the type of engagement pursued by the retired adult. Should the civic contributions made by someone managing a retail store or providing informal caregiving really be distinguished from someone working as an afterschool tutor? Does returning to work as a handy man in a faith-based organization differ from taking the same role in an apartment complex? Although we hesitate to draw any conclusions from our post hoc comparison between engaged retirees and the IWSS sample of older adults who continued to work (never self-identified as retired), this initial finding supports the notion that people who never retire could meet the definition of assuming a civic engagement role in their retired years. Finally, many of the variables that distinguished retirees who were involved in civic engagement could be associated with social class effects (Sheppard & Philibert, 1972). If civic engagement is going to be a role for retired Americans, then efforts must focus on increasing access to this role for all retirees and not just those who have higher levels of education, health, and other enabling characteristics. Notwithstanding this need to further explore what actually constitutes a civic engagement role and who can occupy it, our research puts forth an empirically based definition that sets the stage for researchers to examine how varying types of civic engagement affect social capital and individual health outcomes among retired older adults. Moreover, by identifying variables that further distinguish groups of older retirees, we have identified a need for national surveys to expand their scope and incorporate questions that allow researchers to more effectively differentiate retirees by volunteer and work experiences. Conclusions Our analyses support the notion that engaged retirees are significantly different from those who volunteer less, work in noncivic roles, or do neither. To us, it no longer seems useful to associate the act of volunteering, in itself, with civic engagement among retired older adults. Those who volunteer occasionally should be distinguished from those who take on a formal volunteer role. It also does not seem useful to think that civic engagement roles cannot include retirees who have returned to work in particular jobs or organizations. As the population continues to age and the demand for voluntary and paid labor increases, discussions about the civic engagement of retired Americans will become more common and more important. This research constitutes an initial effort to define civic engagement as a retirement role and distinguish persons who occupy this role from other retirees. References AARP. (2003). Staying ahead of the curve: The AARP Working in Retirement Study. Washington, DC: Author. Administration on Aging. (2007, January). HR 6197: Reauthorization of the Older Americans Act. Available from the Administration on Aging Web site, American Society on Aging. (2006). Strategic Plan Available from the American Society on Aging Web site, gov/plan/ Blau, D. (1994). Labor force dynamics of older men. Econometrica, 62, Boggs, D., Rocco, T., & Spangler, S. (1995). A framework for understanding older adults civic behavior. Educational Gerontology, 91, Brougham, R., & Walsh, D. (2005). Goal expectations and retirement intentions. International Journal of Aging and Human Development, 61, Cahill, K., Giandrea, M., & Quinn, J. (2006). Retirement patterns from career employment. The Gerontologist, 46, Cutler, S., & Hendricks, J. (2000). Age differences in voluntary association memberships: Fact or artifact. Journal of Gerontology: Social Sciences, 55B, S98 S107. Dillman, D. (2000). Mail and internet surveys: The tailored design (2nd ed.). New York: Wiley. Dye, D., Goodman, M., Roth, M., Bley, N., & Jensen, K. (1973). The older adult volunteer compared to the nonvolunteer. The Gerontologist, 13, Elshaug, C., & Metzer, J. (2001). Personality attributes of volunteers and paid workers engaged in similar occupational tasks. The Journal of Social Psychology, 141, Fischer, L., Meuller, D., & Cooper, P. (1991). Older volunteers: A discussion of the Minnesota Senior Study. The Gerontologist, 31, Galston, W., & Lopez, M. (2006). Civic engagement in the United States. In L. Wilson & S. Simson (Eds.), Civic engagement and the baby boomer generation: Research, policy and practice perspectives (pp. 3 20). New York: Haworth Press. Gerontological Society of America. (2006). A brief history. Advancing the Field. Available from the Gerontological Society of America Web site, Glass, T., Mendez de Leon, C., Marottoli, R., & Berkman, L. (1999). Population based study of social and productive activities as predictors of survival among elderly Americans. British Medical Journal, 319, Harlow-Rosentraub, K., Wilson, L., & Steele, J. (2006). Expanding youth service concepts for older adults: AmeriCorps results. In L. Wilson & S. Simson (Eds.), Civic engagement and the baby boomer generation: Research, policy and practice perspectives (pp ). New York: Haworth Press. Hendricks, J., & Cutler, S. (2004). Volunteerism and socioemotional selectivity in later life. Journal of Gerontology: Social Sciences, 59B, S251 S257. Hershey, D., & Mowen, J. (2000). Psychological determinants of financial preparedness for retirement. The Gerontologist, 40, Herzog, R., & House, J. (1991). Productive activities and aging well. Generations, 15, Kaskie, B., & Gerstner, N. (2004). Work and volunteer opportunities for older persons in Marin County: A market analysis. Larkspur, CA: Marin Community Foundation. 376 The Gerontologist

10 Kaskie, B., Imhof, S., Cavanaugh, J., & Culp, K. (2005). The Iowa Workforce Seniors Study. Des Moines, IA: Iowa Workforce Development. Liebsen-Hawkins, B. (2005). Coming of age. Boston, MA: Minuteman Press. Maestas, N. (2004). Back to work: Expectations and realizations of work after retirement. Unpublished manuscript, University of Michigan at Ann Arbor. Martinson, M., & Minkler, M. (2006). Civic engagement and older adults: A critical perspective. The Gerontologist, 46, Mermin, G., Johnson, R., & Murphy, D. (2006). Why do boomers plan to work so long? Washington, DC: The Urban Institute. Moen, P., & Fields, V. (2002). Midcourse in the United States: Does unpaid community participation replace paid work? Ageing International, 27, Morris, R., & Caro, F. (1995). The young-old, productive aging, and public policy. Generations, 19, Morrow-Howell, N., Hinterlong, J., Rozario, P., & Tang, F. (2003). Effects of volunteering on the well-being of older adults. Journal of Gerontology: Social Sciences, 38B, S137 S145. National Council on the Aging. (2006). Mature workers. Available from the National Council on the Aging Web site, Norris, P. (1996). Does television erode social capital? A reply to Putnam. Political Science and Politics, 9, Okun, M. A., & Schultz, A. (2003). Age and motives for volunteering: Testing hypotheses derived from socioemotional selectivity theory. Psychology and Aging, 18, Okun, M. A., Barr, A., & Herzog, A. R. (1998). Motivation to volunteer by older adults: A test of competing measurement models. Psychology and Aging, 13, Pew Charitable Trust. (2006). Supporting civic life. Available from the Pew Charitable Trust Web site, index.cfm?area=3 Putnam, R. (1995). Bowling alone revisited. The Journal of Democracy, 6, 65. Ruhm, C. (1990). Bridge jobs and partial retirement. Journal of Labor Economics, 8, Sander, T., & Putnam, R. (2006). Social capital and civic engagement of individuals over age fifty in the United States. In L. Wilson & S. Simson (Eds.), Civic engagement and the baby boomer generation: Research, policy and practice perspectives (pp ). New York: Haworth Press. Seeman, T., & Crimmins, E. (2001) Social environment effects on health and aging: Integrating epidemiologic and demographic approaches and perspectives. Annals of the New York Academy of Sciences, 954, Sheppard, H., & Philibert, M. (1972). Employment and retirement: Roles and activities. The Gerontologist, 12, Szinovacz, M., & Davey, A. (2005). Predictors of perceptions of involuntary retirement. The Gerontologist, 45, Theiss-Morse, E., & Hibbing, J. (2005). Citizenship and civic engagement. Annual Review of Political Science, 8, Uhlenberg, P. (1992). Population aging and social policy. Annual Review of Sociology, 18, U.S. Bureau of the Census Census of United States population: Washington, DC: Author. U.S. Government Accountability Office. (2005). Older workers: Labor can help employers and employees plan better for the future. Washington, DC: Author. Wilson, L., & Rymph, D. (2006). Research issues in civic engagement: Outcomes of a national agenda setting meeting. In L. Wilson & S. Simson (Eds.), Civic engagement and the baby boomer generation: Research, policy and practice perspectives (pp ). New York: Haworth Press. Zedlewski, S. R., & Schaner, S. G. (2005). Older adults engagement should be recognized and encouraged. In The Retirement Project: Perspectives on productive aging. Washington, DC: Urban Institute. Received March 6, 2007 Accepted June 12, 2006 Decision Editor: William J. McAuley, PhD Vol. 48, No. 3,