| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
RESEARCH ARTICLE |
a University of Vermont, Burlington
b Oregon State University, Corvallis
 |
Abstract |
|---|
 TOP Abstract Relevance of Membership in...Inconsistencies in Findings on...Factors Associated With...MethodsResultsDiscussionReferences |
|---|
Methods. Data are drawn from 12 nationally representative surveys covering the period 1974–1994 and including items on membership in 16 types of associations. Age is categorized into 14 groups ranging from 18–24 through 85+. Patterns of age differences in number of association memberships are examined before and after controlling for several variables related both to age and to memberships.
Results. At the bivariate level, number of memberships increases through ages 40–44 and then declines fairly steadily. When controls for compositional differences are introduced, the curvilinear pattern of age differences is replaced by one showing an increasing number of memberships through ages 55–59, a slight decrease in the 60–64 age group, and then generally stable membership levels through the 85+ age group.
Discussion. The lower levels of voluntary association memberships of older persons consistently identified in bivariate analyses appear to be largely due to differences in compositional characteristics between the older and other age groups.
VOLUNTARY association participation has long been important in supplying societal goods and services and bringing satisfaction and meaning to individuals. Although a wealth of information about association participation has accumulated, questions remain unanswered. Are Americans continuing to be joiners as de Tocqueville asserted 170 years ago (Curtis, Grabb, and Baer 1992
; Hausknecht 1962; Lipset 1985)? Does the pattern of voluntary participation show a curvilinear relationship by age as is often claimed? As Rotolo 1999 noted, there is little consensus. For example, Putnam 1995 asserted that participation in voluntary associations has declined precipitously in the last few decades. Putnam is also representative of those who asserted that middle-aged people are the most likely to participate in voluntary associations and that the rates of participation for both their older and younger counterparts are lower (for a review, see Smith 1994). It is the latter question that is the particular focus of this article.
Forecast after forecast (Chambre 1993; Cnaan and Cwikel 1992; Harootyan and Vorek 1994; Warburton, Le Brocque, and Rosenman 1998) has suggested that voluntary participation will increase in importance among older persons as resources shrink and retirement lengthens. The literature also contends that voluntary participation ebbs and flows over the life course. Within gerontology, numerous hypotheses for these shifts have been offered, often building on continuity theory (Monk 1995), activity theory (Chambre 1984; Herzog and Morgan 1993; Stevens 1993), or life-course models (Babchuk and Booth 1973; Chambre 1993). Missing are attempts to disentangle ramifications implicit in various models from compositional characteristics of diverse cohorts as they move through life.
In his original commentary, de Toqueville asserted that analogous patterns exist for people of all ages, dispositions, and circumstances. Yet there is ample evidence suggesting formal participation in voluntary associations is typical of younger, more affluent, and better-educated groups (Curtis et al. 1992; Fischer and Schaffer 1993; Smith 1994). In terms of our interests here, two overlapping patterns have been described. First, there is a contention that older persons' membership in voluntary associations tapers off, and that they are less likely to belong to or be affiliated with voluntary associations than their younger counterparts (Chambre 1984, Chambre 1993; Dye, Goodman, Roth, Bley, and Jensen 1973; Monk 1995). At the same time, the relationship between age and voluntary association participation has been described as curvilinear, reflecting time and resources available to people of differing ages (Cutler 1976; Herzog and Morgan 1993).
Several explanations have been proferred to account for differential rates of voluntary association participation. Generally speaking, one stream is grounded in sociodemographic characteristics, those attributes usually categorized within the array of human capital variables. Despite fluctuations and some variation, there is consensus that the correlation between these factors and voluntary membership is direct and linear. A second stream revolves around a life-course perspective which suggests that differing levels of voluntary association affiliation reflect emerging priorities and role substitutions undertaken in order to maintain self-concepts (Chambre 1993; Gecas and Burke 1995; Herzog and House 1991; Moen 1996; Musick, Herzog, and House 1999). Although a life-course perspective is inherently attractive, speculation without first controlling for compositional changes in key variables may lead to spurious conclusions and will not shed light on the shifting patterns observed among diverse age groups. Few studies of developmental shifts in voluntary association membership or participation have explicitly addressed the role played by shifting compositional factors and how they may mitigate the appearance of developmental or age changes. Our goal in this analysis is to address exactly that point.
Data from the National Opinion Research Center's General Social Surveys (GSS) were brought to bear on an examination of membership in voluntary associations by age. We looked at gender, perceived health, marital status, presence of children, race, and personal and social resources known to affect propensity to join voluntary associations. Due to the nature of our data, our analysis was confined to age differences and to membership in voluntary associations. We were unable to examine individual-level age changes or other measures of types and degree of participation or involvement. Because of the nature of our data, our analysis speaks only to levels of association membership and not to other aspects of voluntary participation, activities that may take place in nonassociational contexts. Despite these limitations, the size of the sample, the capacity to examine patterns of differences through the age group 85+, and the robustness of our analysis provide tangible insights that will help ground subsequent life-course explanations of patterns of voluntary association membership and participation.
|
Relevance of Membership in Voluntary Associations |
|---|
|
TOPAbstractRelevance of Membership in...Inconsistencies in Findings on...Factors Associated With...MethodsResultsDiscussionReferences |
|---|
; Rotolo 1999), participation in voluntary groups is increasingly being "marketed" as a noble, cost-saving endeavor crucial to economic and communal well-being for one or another type of association (Curtis et al. 1992). One recent estimate has older volunteers contributing over $77.2 billion worth of services in 1995 (Independent Sector 1998). If this figure is even approximately correct, the "volunteer marketplace" is an apt portrayal in many respects as various forms of associations find themselves vying for members. In an analysis of church-based associations, Bibby 1987 postulated that sectarian competition has prompted vigorous promotional approaches to joining and to participation among religious groups, as well as in secular organizations. He attributed the competition to the fact that voluntary memberships provide unpaid labor for deserving issues (Knoke 1990).
From an individual's point of view, membership or participation in voluntary associations takes many forms and can mean different things to different people at different points in life. Ordinarily, though, it provides a sense of belonging, bonding, participation, and support (Janoski and Wilson 1995; Putnam 1995; Rotolo 1999). Generally, joining one or another type of voluntary association nurtures a sense of well-being and provides opportunities for compensatory roles, affiliations, recognition, and ego gratification (Fischer and Schaffer 1993; Monk 1995; Musick, Herzog, and House 1999). With the use of a conceptual template drawn from a life-course perspective, it appears that the propensity to join differs by event history and by role obligations—in family life, for example—as these have been demonstrated to be important determinants of association membership (Babchuk and Booth 1973; Knoke and Thompson 1977; Rotolo 1999). Insofar as life satisfaction or a sense of well-being is derived from distinctive activities at diverse points in life (N. E. Cutler 1981–82), a thoughtful age-appropriate life-course perspective can build on closer attention to compositional differences to highlight the significance of membership in voluntary associations across the life span.
|
Inconsistencies in Findings on Patterns and Correlates of Voluntary Association Membership |
|---|
|
TOPAbstractRelevance of Membership in...Inconsistencies in Findings on...Factors Associated With...MethodsResultsDiscussionReferences |
|---|
). Janoski and Wilson 1995 noted that most researchers attribute greater levels of voluntary association affiliation to advantageous socioeconomic factors, with rates anchored in human capital characteristics and changes that occur over the life course. Discriminant analyses routinely suggest that among sociodemographic factors, occupational level, education, income, and being married are positively related to rates of participation, other obligations are negatively associated, and age has a curvilinear relationship, with middle-aged adults being most affiliated. Attitudinal characteristics associated with membership in voluntary groups include altruism, a sense of well-being, and other manifestations of positive affect (Axelrod 1973; Babchuk and Booth 1973; Fischer and Schaffer 1993).
In terms of our specific interest in age differences, the results are inconsistent (Babchuk and Booth 1973; Chambre 1993; Cutler 1977; Monk 1995; Rotolo 1999). Chambre 1984 originally asserted that older people volunteer less than their younger counterparts. She reported that a quarter of the general population, but only one sixth of those older than age 60, volunteered. Looking at the intensity or hours spent volunteering, she claimed that older volunteers spent less time volunteering; as few as 5% reported 100 hours of voluntary activity in the previous year.
Cutler 1976 and Knoke and Thompson 1977 reported comparable curvilinear patterns of association membership. Explanations for the pattern among younger respondents remain problematic but may relate, independent of socioeconomic facilitators or inhibitors, to time constraints and the presence of very young children. In commenting on the apparent decline among older respondents, Cutler 1976 noted that it is unclear whether differences in affiliation and participation are a function of age, or of other factors related to family life cycle; human capital variables linked to different cohorts; or to unknown factors (see also Smith 1994). Cutler 1976 pointed out that exactly those variables known to be correlated with lower levels of voluntary association membership and participation are also characteristic of compositional differences between cohorts.
Other investigators reported no decline in membership with age. In another study, Cutler 1977 discovered no apparent age-based declines in a small longitudinal study. He contended that rates remain stable until the very oldest years when declining health or other impediments intervene. As he and others have pointed out, what has been lacking is analysis of representative patterns that control for compositional factors differentiating diverse cohorts (Cutler 1976, Cutler 1977; Herzog, Kahn, Morgan, Jackson, and Antonucci 1989; McPherson and Rotolo 1996).
Cnaan and Cwikel 1992 attributed the patterns commonly identified not to human factors, per se, but to shifts in the societal ethos. They contended that following the 1960s, a period of heightened social involvement, rates of voluntary membership have been relatively flat. Putnam 1995 however, argued that rates are currently declining rather significantly. Further, Cnaan and Cwikel 1992 asserted that voluntary association affiliation rates do not increase with age or as a consequence of retirement or more free time. Chambre 1993, who also referred to voluntarism in the 1960s, argued that the activists of the 1960s have grown older, which accounts for the increase in rates of voluntary participation among older persons. She based her conjecture on results reported in over a dozen studies conducted between 1969 and 1990. She had made a comparable claim in an earlier piece and suggested that participation rates among older persons are a consequence of rates established earlier in life and continued into old age (Chambre 1984, Chambre 1993). At the same time, she commented on a changing social ethos concerning appropriate activities in old age and shifts in the demographics of the elderly. As a cautionary note, she alluded to short-term considerations affecting voluntary participation—specifically, seasonal variations, differences by formal (narrow) versus informal (broad) interpretations of participation, as well as variance stemming from type of organizations analyzed.
Relying on GSS data comparable to what we utilize, McPherson and Rotolo 1996 pointed out that although both education and income yield high, positive correlations with voluntary association membership, not all voluntary groups are solely peopled by the well-educated and wealthy. In addition, little is known about the turnover in association memberships, types of associations with which individuals affiliate, and what induces the apparent curvilinear relationship between age and voluntary affiliations. Is it really a function of age or, as Cutler 1976 implied, might conflating factors be involved (Hausknecht 1962; McPherson and Rotolo 1996; Rotolo 1999)? Moreover, if there is indeed a curvilinear relationship, might age discrimination be involved to the extent that positions for volunteers reflect the logic of the labor market (Kieffer 1986)? Similarly, older persons undoubtedly seek meaningful engagements, and do not simply join groups willy-nilly or volunteer to dispense directions or go for coffee. By bringing a robust, concatenated data set to some of these questions, and by conducting a multivariate analysis among diverse age categories, we seek to shed light on age-related patterns of voluntary association participation in general and membership in particular.
|
Factors Associated With Voluntary Membership |
|---|
|
TOPAbstractRelevance of Membership in...Inconsistencies in Findings on...Factors Associated With...MethodsResultsDiscussionReferences |
|---|
now classic assertion that membership in voluntary associations varies by age, stratum, and social group has fueled considerable debate. He was prophetic in his caveat that rates cannot be examined in a vacuum; they are linked to previous experience, family, work, and consumption cycles, and to compositional characteristics of groups. Key characteristics highlighted by previous research and relevant to this study follow.
Education.
Educational attainment appears to have a universal, positive relationship with membership in volunteer groups (Fischer, Mueller, and Cooper 1991; McPherson and Ranger-Moore 1991; McPherson and Rotolo 1996; Monk 1995; Rotolo 1999; Smith 1994), in terms of both number of memberships in voluntary associations and "intensity" of voluntarism. Whether that connection is because better-educated individuals have more volunteer and membership options or is due to other factors is difficult to say. McPherson and Ranger-Moore 1991 examined education over a 15-year period using GSS data, and concluded that it is a stable, strong predictor of membership in voluntary associations. The Minnesota Study of Volunteering reported a linear, positive link between education and organizational, familial, and individual-based volunteering, including voluntary associations (L. R. Fischer et al. 1991). To identify the relevance of education, we utilized five widely recognized educational intervals in examining the relationship net of the influence of other factors.
Employment status.
Work-related influences on voluntary participation are discussed in terms of occupational ranking or whether an individual is employed full-time or part-time or is retired. As far as the role of employment status is concerned, opinions have been mixed (Harootyan and Vorek 1994). Some researchers asserted that retirees may relish a full exit from paid and volunteer obligations (Chambre 1987, Chambre 1993; Fischer and Schaffer 1993) and therefore rates will decline. There is a countervailing view that because time is limited, participation may increase when other obligations lessen—a sort of substitution hypothesis (Chambre 1984). Then again, being active at work may encourage activity in other realms (Fischer and Schaffer 1993; Warburton, Le Brocque, and Rosenman 1998). Smith 1994 reviewed a number of empirical investigations of voluntary association participation and reported that many found higher rates among full-time employees than among any other group. For a sense of which way the pattern flows, we examined employment status.
Marital status.
Married persons join voluntary associations more than their unmarried counterparts, especially in associational and familial contexts (L. R. Fischer, Mueller, and Cooper 1991; Smith 1994). An exception to the pattern was noted by Warburton and colleagues 1998 in an analysis of volunteering in Australia. They did not find marital status to be predictive. When marital status is found to be related to voluntary memberships, it is not clear whether the differences are due to broader social networks in general or are linked to income and other resources (L. R. Fischer, Mueller, and Cooper 1991). Because there is reasonable consensus that marital status, independent of household size, does affect membership in voluntary associations, we included a four-category measure of this variable.
Gender.
Findings concerning the linkage between gender and voluntary association participation are inconsistent. In a number of empirical investigations, men appeared to have higher participation rates than women (Curtis, Grabb, and Baer 1992). In contrast, Chambre 1984, Chambre 1987 claimed that at least among elderly individuals, more older women participate and volunteer because there are more older women than older men. Babchuk and Booth 1973 commented on the gender-based inconsistencies and hypothesized they might result from greater variability by age among men. Warburton and colleagues 1998 believed the absence of a significant gender difference in their data could be due to the blurring of gendered distinctions. As long ago as the mid-1970s, Cutler 1976 suggested male rates are higher at all ages than female rates. In contrast, Chambre 1984 and others asserted that women are more likely than men to join and to volunteer regardless of age. However, close scrutiny may reveal that women are more likely to hold voluntary membership in religious and other expressive organizations and men are more likely to claim voluntary memberships that are worklike in nature (Babchuk, Peters, Hoyt, and Kayser 1979; Gallagher 1994). Fischer, Rapkin, and Rappaport 1991 and others (Wilson and Musick 1997) urged caution in formulating any generalizations by pointing to the difficulties of separating gender from other factors related to employment experiences and income. We incorporated gender into our multivariate analysis to try to isolate its role, net of other factors.
Presence of dependent children.
The literature has long supported the idea that the presence of dependent children promotes voluntary participation, perhaps because child-centered activities are a major avenue of involvement (Harootyan and Vorek 1994; Smith 1994). Wilson and Musick 1997 reported that number of children present in the home has a significant positive effect on formal voluntary participation. They hypothesized that the presence of children facilitates social involvements that might not otherwise occur (Wilson and Musick 1999). In order to assess such a prospect, we, too, included "presence of dependent children" as a variable.
Age.
Investigations of age differences trace to the post-World War II era, when Myrdal echoed de Toqueville's assertion that Americans tend to join voluntary associations, and continued into the 1950s and 1960s (Mayo 1950; Rose 1960). Wilensky 1961 examined more than a dozen studies reporting age differences in voluntary association participation but noted that few attempted to disentangle age from compositional components characterizing cohorts, such as education, income, and so on. He was among the first to call for statistical controls over intervening factors that might contribute to the apparent age differences. In our investigation we categorized age into 14 intervals and explored its relative importance, net of the effects of related predictors.
Race.
Less is known about the relationship between voluntary association involvement and race or ethnicity than should be the case. Evidence from a 1990 AARP survey suggests African Americans are significantly more likely to volunteer and spend more hours doing it than others (Harootyan and Vorek 1994). The curious aspect of this finding is that it runs counter to what is usually assumed to be the case in light of education and income differences. Despite mixed results, and the need for greater specification concerning types of voluntary participation (Gallagher 1994; Smith 1994; Wilson and Musick 1997), there may be subcultural and institutional factors that abet one another, promoting an ethic of reciprocity and helpfulness that overrides other sociodemographic variables (Harootyan and Vorek 1994). Based on the variables available in the GSS data set, we employed a global White/non-White distinction.
Self-perceived health.
Health status and self-perceived health are correlated with many measures of life satisfaction and with voluntary association membership and participation (K. Fischer et al. 1991). Both actual health status and subjective assessments have been linked with participation (Chambre 1987; Herzog and Morgan 1993; Smith 1994). Warburton and colleagues 1998 found that although subjective health was associated with rates of voluntary activity, actual health was not. In reversing the usual directionality, Young and Glasgow 1998 reported a strong relationship between perceived health, net of all other factors, and participation in what they termed instrumental (service-related) organizations for both men and women, although for women this was the case only in expressive (self-gratification) organizations. We utilized a global question of self-perceived health as a criterion variable.
|
Methods |
|---|
|
TOPAbstractRelevance of Membership in...Inconsistencies in Findings on...Factors Associated With...MethodsResultsDiscussionReferences |
|---|
). Based on the availability of variables used in the analysis, 12 surveys conducted between 1974 and 1994 
were selected for inclusion: 1974 
. Sampling for 1974 and half of the 1975 survey was based on a block quota design; the remaining half of the 1975 survey and all other years used full probability sampling designs. Each sample was representative of the population of English-speaking persons 18 years of age and older living in noninstitutional arrangements in the United States. Changes in content coverage over the history of the GSS resulted in a reduction of the total N for the 12 surveys. Several of the variables used in the analysis (age, gender, race, education, employment status, marital status, and presence of children in the household) are part of the series of permanent items asked of all respondents in each survey. Other items have been included on a rotational basis to allow more topics to be covered. Through 1987, rotating items were included in 2 of 3 surveys. Beginning in 1988, the rotation was switched to one in which selected items were asked of two thirds of the respondents within each survey. Information on both voluntary association memberships and self-perceived health was asked of all respondents in the first six surveys listed above. From 1988 through 1994, however, these items were asked of partially overlapping subsets of two thirds of the samples with the effect that information on both association memberships and health was obtained from only about one third of the respondents in these years.
In addition to the reduction in the total N due to the GSS's rotational system, the number of available cases was further reduced by the exclusion of respondents with missing data on any of the variables used in the analysis. The final Ns, with missing data from all sources excluded, are presented in Table 1 by age and survey year.
|
|
and others, number of association memberships was operationalized as a count of the types of voluntary associations to which respondents belonged (corrected for a coding error for the years 1989–1994 [Helliwell and Putnam 1996]), with scores ranging from 0 to 16. As types rather than actual number were utilized, the estimate is conservative and cannot be taken as a measure of the intensity of voluntary participation.
We employed other variables in the analysis, in addition to gender. To enable us to look at detailed patterns occurring at the older ages, age was categorized into 14 groups as shown in Table 3 . Self-perceived health was based on the following question: "Would you say your own health, in general, is excellent, good, fair, or poor?" Educational attainment was divided into five categories as shown in Table 3 . (Although several studies [e.g., K. Fischer, Rapkin, and Rappaport 1991] noted that income has a direct influence on various forms of voluntary participation, we chose not to include family income as another indicator of socioeconomic status to prevent additional loss of cases from missing data and because of its significant correlation with education 
.) Employment status was categorized as working full-time, working part-time, unemployed, retired, and other. Marital status was characterized as married, widowed, divorced or separated, and never married. A dichotomous variable was constructed to indicate presence of children in the household (through age 18 years), and race was also a dichotomy, comparing Whites and non-Whites.
|
|
Results |
|---|
|
TOPAbstractRelevance of Membership in...Inconsistencies in Findings on...Factors Associated With...MethodsResultsDiscussionReferences |
|---|
nor the age2/year interaction term 
was significant, suggesting that the nature of the relationship between age and number of memberships did not vary over time. In the interests of parsimony and, most important, to have a sufficient number of cases to examine the patterns of membership in some detail at the very oldest ages, we based the remainder of the analysis on the composite of all time points. The data presented in Table 3 show the relationships between age, the other predictors, and number of voluntary association memberships. The bivariate relationship between age and number of memberships, as indicated by the gross mean scores, is statistically significant and consistent with the curvilinear pattern noted in previous studies. Membership rates are low at the younger ages and rise to a peak among those aged 40–44 years. Thereafter, membership levels decline fairly steadily, with a substantial decrease occurring between the age groups 55–59 and 60–64 years. Those aged 75–79 years show a membership rate at about the same level as persons aged 18–24 years, and respondents aged 80–84 and 85+ exhibit the lowest rates of any age group examined.
Our central concern was to address the reasons for the successively lower membership levels among older persons. As noted earlier, there are important compositional differences between age groups and considerable evidence suggesting that voluntary association membership levels are related to these compositional characteristics. For the data analyzed here, there was a clear relationship between age and self-perceived health: older persons were significantly more likely to report their health as poor and they were less likely to describe their health as excellent (
2 = 1190.2, df = 21, p < .001). Similarly, older persons had lower levels of educational attainment and they were less likely to be found at the higher educational levels (2 = 2120.4, df = 28, p < .001). Older persons were less likely to be working full-time and part-time and were more likely to be retired (2 = 5452.7, df = 28, p < .001). The older age groups also had higher percentages of females (2 = 29.6, df = 7, p < .001), respondents who were widowed (2 = 6014.2, df = 21, p < .001), and persons living in households without children (2 = 3057.7, df = 7, p < .001). Finally, there was a small but significant tendency for the group aged 85+ to have a higher percentage of non-Whites (2 = 50.0, df = 7, p < .001).
As the data in Table 3 indicate, mean voluntary association membership levels were in turn related to each of these compositional characteristics. Lower membership levels were significantly associated with poorer self-perceived health, lower educational attainment, being retired, being female, being widowed, residing in a household where children are not present, and being non-White.
Given that the compositional characteristics of older age groups differed in ways that implied lower levels of voluntary association memberships, it was essential to determine the pattern of age differences in membership levels after controlling for the effects of these compositional differences. The net mean scores presented in Table 3 indicate the effects of age differences on association memberships after partialling out effects due to compositional characteristics. These net mean scores clearly indicate that a major source of the lower association memberships observed in the bivariate data among the older age groups stemmed from their compositional characteristics. As is apparent, mean membership levels remained low at the younger ages, but steadily rose through ages 55–59, dropped somewhat in the 60–64 age group, and then remained relatively stable up through ages 85+. In other words, the pronounced curvilinear pattern of age differences in association memberships observed at the bivariate level was replaced by a statistically significant pattern that peaked at a later age and was far more stable at the older ages. Closer scrutiny actually revealed that once compositional differences are taken into account, persons 55–59, 65–69, and 75–79 had the highest net mean levels of affiliation. This latter finding is particularly relevant given what has customarily been reported for the relationship between age and voluntary association membership.
The net effects of the remaining predictors continue to be significant. Although the strength of the relationship was reduced, number of association memberships was clearly related to self-perceived health. Level of educational attainment exerted the strongest effect of all the predictors. Employment status and marital status differences were somewhat reduced in magnitude but, of interest, the net mean membership levels of retirees and widows were higher than the corresponding gross mean levels. The introduction of the controls, in other words, appears to have had the same effect on these age-associated categories as it did for the older age groups. Men and Whites continued to exhibit higher membership levels, although here again the magnitude of the gender and racial differences was diminished after partialling out the effects of the other control variables. Finally, net of the other predictors, having children in the household exerted an even stronger effect on number of voluntary association memberships than was the case prior to the introduction of the controls.
The combined data from the 12 surveys are presented in another way to facilitate visual inspection. Fig. 1 plots the gross and net mean number of voluntary association memberships for each age group as well as the 95% confidence intervals for the gross means. If the array of compositional characteristics were to have no bearing on the relationship between age and average number of association memberships, the gross and net effects would not differ and the plotted lines would converge. If the net mean score for any age group were to be higher than its gross mean score, it would indicate that compositional characteristics act to dampen its membership levels compared to other age groups having characteristics that were more favorable to voluntary association memberships. Conversely, if the net mean score for a particular age group were to be lower than its gross mean score, favorable characteristics would operate to inflate its membership level. The 95% confidence intervals are used to provide a rough approximation of the significance of the changes between the gross and net means for each of the age groups.
|
The 95% confidence intervals are presented under the assumption that whether an age group's net mean score falls within or outside of the interval for its gross mean can be taken as an approximate indication of the significance of the change when compositional differences are controlled. Inspection of the confidence intervals in Fig. 1 provides further support for our conclusions. Specifically, the adjusted means fall within the 95% confidence intervals for three of the age groups: 18–24 
. In these instances, the adjustment process had little effect. For the remaining age groups, the net mean scores lie outside of the gross mean's confidence interval: 25–29 
. For these age groups, controls in place for compositional characteristics appear to have had an appreciable effect.
|
Discussion |
|---|
|
TOPAbstractRelevance of Membership in...Inconsistencies in Findings on...Factors Associated With...MethodsResultsDiscussionReferences |
|---|
To reach this conclusion, we aggregated 12 years of the GSS after having determined that the nature of the relationship between age and number of memberships did not vary over the 20-year period encompassed by the surveys. Utilizing a multivariate strategy, we looked at gross and adjusted mean levels of voluntary association membership across 14 age intervals. With no controls in place, there appears to be a curvilinear relationship with peak levels of participation occurring among respondents in the 40–44 year age interval. If we had looked no further, we might have joined those who have maintained that rates of voluntary association membership first build then decline later in life. On first inspection, we might have concluded that older persons have substantially lower levels of affiliation than all but the initial adult age group (18–24).
Realizing that many of the compositional factors characterizing older cohorts have also been identified as affecting propensities to belong to voluntary associations, we submitted these same data to multivariate procedures in the form of a Multiple Classification Analysis capable of examining the net effects of age independent of other characteristics thought to affect membership rates. As is apparent in Table 3 , beginning with the 25–29 year age interval, there was a more nearly direct, linear increase in voluntary association memberships through ages 55–59 years when compositional differences were controlled, a slight decrease in the 60–64 age group, and generally stable membership levels thereafter through the 85+ age group. These age differences cannot be construed as conclusive, to be sure, but they are certainly at variance with what is often reported.
Fig. 1 shows that age differences and the effects of compositional factors were significant. Had the latter been irrelevant to a group's propensity to claim membership in voluntary associations, the two lines would be identical. As noted, between the ages of 25 and 44 years, an array of human capital variables and sociodemographic factors facilitate greater rates of affiliation than are seen in younger or older age intervals. When persons reach a point somewhere in their 50s, these factors dampen previous patterns, leading to the assumption that older persons are less inclined to join voluntary groups. Yet, when the effects of factors such as male mortality and levels of educational attainment accrued at earlier points in the life cycle are partialled out, thereby in effect comparing different age groups with comparable human capital and sociodemographic attributes, the long-identified curvilinear relationship disappears.
One obvious conclusion from these data is that the older age groups are virtually the most involved, at least in terms of number of memberships, when intervening conditions are controlled. A possible consequence of that fact is that upward shifts in educational attainment and improved longevity for men will yield even higher rates of voluntary association membership among future cohorts of older persons. This prospect is particularly relevant to discussions of whether older persons are merely a drain on available resources or also make productive contributions to society through their various affiliations (Hendricks, Hatch, and Cutler 1999). The findings also bear on the contention that there has been a decline in civic participation in the United States (Putnam 1995; Helliwell and Putnam 1996). Using the same data employed in our analysis, Rotolo 1999 demonstrated that voluntary association membership has, in fact, been increasing since 1984. That the characteristics of future cohorts of older persons will likely predispose them to higher levels of voluntary association membership and participation suggests that this recent trend will continue. Given the implications of the work of Musick, Herzog, and House 1999, among others (e.g., Moen, Dempster-McClain, and Williams 1992), that volunteering has a protective effect on mortality and sense of well-being, our findings imply that the situation among older persons who volunteer may not be as dire as some have suggested.
Because our study had only one question pertaining to perceived health status, our findings are surely tentative. Nonetheless, we noted a significant link between how respondents described their health and their net rates of association membership. Also, in light of what some investigators have noted about voluntarism being nested in church or religious affiliations on the one hand and in union membership on the other, we intend to examine these two spheres in subsequent analyses. With the temporally based shifts occurring in each in terms of their prominence on the national scene, it will be meaningful to look at gross and net levels of voluntary participation when church and union memberships are controlled.
Though our focus has been on rates of formal voluntary association memberships, it is important to point out there is more to voluntary participation than meets the eye (Herzog et al. 1989). In an overview of the issues, Monk 1995 pointed out that definitions of voluntary participation range from assisting formal organizations to meet their goals, to playing in adult sports leagues, to helping a neighbor with a chore. K. Fischer and colleagues 1991 offered a threefold typology based on whether volunteering is "formal" or "informal" or regular or occasional, and on how it relates to the nature of the activity itself. If the measure of volunteering is membership in a formal or charitable organization, church-related activities may indeed account for the lion's share of association memberships (K. Fischer et al. 1991). A thoughtful analysis of the issues must also recognize that from a societal or organizational point of view, voluntary participation is no less thorny. As others have noted, social science tends to "measure what we treasure," and therefore unpaid involvement or memberships have traditionally been excluded from definitions of productive activity. For these and other reasons, it is important to isolate exactly what membership in voluntary associations means (Committee on an Aging Society 1986; Herzog et al. 1989).
|
Acknowledgments |
|---|
Received for publication July 17, 1998. Accepted for publication October 4, 1999.
|
References |
|---|
|
TOPAbstractRelevance of Membership in...Inconsistencies in Findings on...Factors Associated With...MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  K. Hank and M. Erlinghagen Dynamics of Volunteering in Older Europeans Gerontologist, August 7, 2009; (2009) gnp122v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Kaskie, S. Imhof, J. Cavanaugh, and K. Culp Civic Engagement as a Retirement Role for Aging Americans Gerontologist, June 1, 2008; 48(3): 368 - 377. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Fengyan Tang Socioeconomic Disparities in Voluntary Organization Involvement Among Older Adults Nonprofit and Voluntary Sector Quarterly, March 1, 2008; 37(1): 57 - 75. [Abstract] [PDF]
|
|
|
|
 |
|
 E. A. Greenfield and N. F. Marks Continuous Participation in Voluntary Groups as a Protective Factor for the Psychological Well-Being of Adults Who Develop Functional Limitations: Evidence From the National Survey of Families and Households J. Gerontol. B. Psychol. Sci. Soc. Sci., January 1, 2007; 62(1): S60 - S68. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Tang What Resources Are Needed for Volunteerism? A Life Course Perspective Journal of Applied Gerontology, November 1, 2006; 25(5): 375 - 390. [Abstract] [PDF]
|
|
|
|
|
|
J. Hendricks and S. J. Cutler The Effects of Membership in Church-Related Associations and Labor Unions on Age Differences in Voluntary Association Affiliations Gerontologist, April 1, 2001; 41(2): 250 - 256. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
