Who volunteers? Constructing a hybrid theory - Wiley Online Library

International Journal of Nonprofit and Voluntary Sector Marketing Int. J. Nonprofit Volunt. Sect. Mark. 16: 298–310 (2011) Published online in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/nvsm.429

Who volunteers? Constructing a hybrid theory Christopher Einolf1 and Susan M. Chambré2 1

School of Public Service, DePaul University, USA Baruch College, City University of New York, USA

2

This paper describes three major theoretical perspectives in research on volunteering: social theories that stress the importance of context, roles, and integration; individual characteristic theories that emphasize values, traits, and motivations; and resource theories that focus on skills and free time. It unites research from multiple disciplines into a single hybrid model, performs a preliminary test of the model on a nationally representative US dataset, and concludes with recommendations for scholars and practitioners. Using the 1995 Midlife in the US dataset, we operationalized concepts from each theoretical category and found that variables measuring each perspective played a substantial and independent role in predicting volunteering. Our hybrid model, which includes significant variables from each theory, offers some directions for recruitment and retention by showing how social roles and networks can constrain or encourage volunteering at different stages of the life course. As social roles and networks are both highly predictive and easily observed, volunteer managers can use them to recruit and retain volunteers. Copyright © 2011 John Wiley & Sons, Ltd.

Introduction Volunteering is a distinctive, significant, and widespread social practice. Surveys reveal that between one quarter and one half of American adults engage in volunteer work over the course of a year, with estimates varying according to differences in survey methodology and how questions define volunteering (Brown, 1999; Rooney et al., 2004). Scholarship on volunteering, much of it summarized in a recent book by Musick and Wilson (2008), is voluminous, but most studies feature a limited range of variables and study volunteering from the perspective of a single academic discipline. This paper unites research from multiple disciplines into a single hybrid model, performs a preliminary test of the model on a nationally *Correspondence to: Christopher Einolf, School of Public Service, DePaul University, Chicago, IL, USA. E-mail: [email protected]

Copyright © 2011 John Wiley & Sons, Ltd.

representative US dataset, and concludes with recommendations for scholars and practitioners.

Review of the literature When David Horton Smith (1994) surveyed research on volunteering and voluntary association membership in the early 1990s, he observed that there had been little theorizing. Although there are significant challenges to creating a unified theory of volunteering (Hustinx et al., 2010), it is fair to say that this situation has changed and that there are numerous theories of volunteering, most of them middle range. We classify them into three theoretical categories (Table 1): social theories that study roles, context, and networks; individual characteristic theories that study traits, values, and motivations; and resource theories that study skills and free time. These theoretical classifications loosely match the disciplines of sociology, psychology, and economics, Int. J. Nonprofit Volunt. Sect. Mark., November 2011 DOI: 10.1002/nvsm

Who volunteers? Constructing a hybrid theory

299

Table 1. Summary of three theoretical categories Name Social factors Context

Theory

Empirical evidence

Volunteers mobilize in response to emergencies and salient social issues and are affected by neighborhoods and cultures of generosity

March of Dimes volunteers 19th century Friendly Visitors AIDS volunteers Variations by state, province, or county Nonprofit density increases volunteering Racial and ethnic diversity decrease voluntary association membership Volunteers have higher levels of engagement in numerous social, voluntary, and leisure activities Levels of social capital are the strongest direct predictors of volunteering Volunteering is a spillover of other social roles: age and lifecycle effects Contradictory findings regarding volunteering and other role obligations Personality: agreeableness, agency, empathic concern, extraversion, conscientiousness, low neuroticism Values: altruism, obligation, religiosity, generative concern Salient role identity Family socialization Education, skills, and free time

Integration

Volunteers are more involved in numerous social and organizational activities

Roles

Other social roles are pathways into and constraints on volunteering

Individual characteristics

Motivations, values, and personality characteristics influence volunteering

Resources

Individuals with higher resources are more likely to volunteer and more likely to be recruited to volunteer

but not exactly, as one can find these theoretical types across disciplinary boundaries. Our categories resemble Wilson and Musick’s (1997a) categories of human, social, and cultural capital, but we further divide social factors into context, integration, and roles.

Social theories

Social context theories study the effect of external events and regional factors, social integration theories examine interpersonal networks and contacts, and social roles examine how volunteering is an expected activity that goes with a particular social status. Social context

Social context theories focus on the role of external events, such as epidemics and natural disasters, on changing patterns of volunteerism. Perhaps the most extensive research on this type of volunteering are studies of AIDS volunteers, which suggest that individuals infected with HIV and their friends, families, and lovers were the Copyright © 2011 John Wiley & Sons, Ltd.

earliest volunteers (Chambré, 1991a, 1991b; Omoto and Snyder, 2002). Once the sense of urgency decreased and more funding became available, organizations relied more upon paid staff than volunteers (Chambré, 2006). Regional factors, corresponding to the neighborhood, state, province, and country where a person lives, also influence volunteering (Anheier and Salamon, 1999; Mellor et al., 2009; Musick and Wilson, 2008; Rotolo and Wilson, 2011). Although there might be regional differences in philanthropy, empirical research on this issue has yielded contradictory findings (Caputo, 2009; Kim and Hong, 1998). Similarly, the impact of neighborhood effects is unclear (Perkins et al., 1996; Wilson, 2000). Although neighborhoods with a relatively higher nonprofit density might have higher participation (Sampson et al., 2005), voluntary association membership tends to be of shorter duration in dense organizational niches (McPherson and Rotolo, 1996). Racially homogeneous neighborhoods have higher rates of volunteering (Portney and Berry, 1997), which is consistent with Putnam’s (2007) observation that racial and ethnic heterogeneity reduce Int. J. Nonprofit Volunt. Sect. Mark., November 2011 DOI: 10.1002/nvsm

300

levels of social capital. Surprisingly, there is only slightly lower participation in central cities (44%) than in suburbs and rural areas (49%) (Musick and Wilson, 2008).

C. Einolf and S. Chambré

children tend to volunteer less than people with no children or people with school-age children, and people who work multiple jobs or jobs with inflexible hours have less time available to volunteer.

Social integration

Volunteering is a social activity, and people who are highly involved in other social activities also tend to do more volunteering. In contrast, individuals who do not volunteer spend more time watching television, listening to the radio, sitting and thinking, or doing nothing (Chambré, 1987; Putnam, 2000). One reason for this is personality, discussed in the succeeding sections, as extroverts tend to volunteer more (Rossi, 2001). Independent of personality, however, people with many social contacts are more likely to be asked to volunteer (Apinunmahakul and Devlin, 2008; Bekkers, 2005; Okun et al., 2007; Rossi, 2001), particularly if they have friends who volunteer (Wymer, 1999). Wives tend to draw husbands into volunteering, but not vice versa (Rotolo and Wilson, 2007). Social integration can also cause people to act upon external norms, so that people volunteer because they feel that other people value volunteering and expect them to participate (Lee et al., 1999). Volunteering is not only an effect of social integration but also a cause of it, as volunteering can be way to cultivate new social relationships (Clary et al., 1998; Prouteau and Wolff, 2008). Some social networks, such as those found in religious congregations, have a particularly strong impact on volunteering (Becker and Dhingra, 2001; Cnaan et al., 1993; Hodgkinson et al., 1990; Park and Smith, 2000; Wilson and Janoski, 1995). Social roles

“Volunteer” is itself a social role, and other social roles can encourage or discourage volunteering. Volunteering can compensate for a lack of fulfillment in other roles (Staines, 1980) or can replace roles lost with life transitions, such as retirement from full-time paid work (Caro and Bass, 1997). Volunteering can be an expected part of the social role of parent (Choi et al. 2007; Rotolo 2000), professional, or employee (Ross, 1954; Wilson and Musick, 1997b). The same roles, however, can discourage volunteering if they too greatly limit the resource of free time. Parents of very young Copyright © 2011 John Wiley & Sons, Ltd.

Individual characteristics

A great deal of research has focused on the importance of individual characteristics, including personality traits, motivations and values, which predispose people to do volunteer work (Clary et al., 1998; Mowen and Sujan, 2005). Although psychologists distinguish among traits, motivations, and values, these terms all refer to individual characteristics that are stable over time and across situations and which influence behaviors. Personality traits that predict volunteering include resilience, extraversion, self efficacy, and low levels of neuroticism (Carlo et al., 2005; Matsuba et al., 2007; Okun et al., 2007). Value orientations that predict volunteering include general measures of altruistic values (Perry et al., 2008; Sundeen, 1992; Wiehe and Isenhour, 1977) and more specific measures such as generative concern (McAdams and de St. Aubin, 1992; Rossi, 2001), moral obligation (Einolf, 2010; Rossi, 2001; Schwartz, 1977; Schwartz and Fleishman, 1978), and moral extensivity (Einolf, 2010; Oliner and Oliner, 1988). Religious values also predict volunteering, even accounting for the effect of religious participation on social roles and integration (Einolf, 2011; Wuthnow, 1991). Of course, volunteers do not have purely altruistic motives, and numerous studies document the social and psychological rewards that motivate volunteering (Andreoni, 1989; Chambré, 2006; Gidron, 1983; Handy et al., 2000; Handy et al., 2010; Midlarsky, 1991). One of the most commonly used measures of motives is Clary and Snyder’s “Volunteer Functions Inventory” (Clary et al., 1998), which proposes one prosocial values motive and five egoistic ones, including the desire to learn new things, experience personal growth, pursue career goals, strengthen social relationships, and protect oneself from negative feelings. Prosocial traits, motivations, and values encourage volunteering but are also strengthened by it. Lee et al. (1999) operationalized the circular relationship of prosocial motives and behavior with Int. J. Nonprofit Volunt. Sect. Mark., November 2011 DOI: 10.1002/nvsm


their theory of “role identity,” which applies to volunteering, charitable giving, and blood donation. A person who volunteers repeatedly may come to think, “I am the kind of person who volunteers,” and eventually, “Volunteering is an important part of who I am.” Role identity helps explain why past volunteering is one of the best predictors of future volunteering (Musick and Wilson, 2008).

Resource theories

Resource theories, which draw from economics but are present in other disciplines, specify that individuals are more likely to volunteer if they possess skills and free time (Wilson and Musick, 1997a; Yörük, 2009). Conceptualizing the decision to volunteer in terms of rational choice theory, Musick and Wilson (2008, p. 113) point out that “volunteering is more attractive to the resource-rich than to the resource-poor. If volunteer work demands money, the rich will find it easier to do; if it demands knowledge and ‘civic skills,’ the well educated will be less challenged by it. . . In other words, the resource-rich are more likely to ‘profit’ from doing volunteer work.” Numerous studies point out that people with higher education and occupational levels are more likely to volunteer (Chambré, 1987; Musick and Wilson, 2008). However, these people also earn more, and traditional economic theory would predict that rational actors would donate less time and more money to a cause as the price of an hour of their time increased. Empirical evidence finds the opposite to be true, however, indicating that volunteered time and donated money are not substitutes (Freeman, 1997). Although free time is a scarce resource, its relationship to volunteering is not straightforward. Employed people and people with children are more likely than non-employed and childless people to volunteer, despite the time demands of work and family. Although non-volunteers most often cite a lack of time as the major reason for not being involved in volunteer work (Sundeen et al., 2007), there does not appear to be a tradeoff in time spent working versus time spent volunteering. Once a person does volunteer, however, the amount of time Copyright © 2011 John Wiley & Sons, Ltd.

301

devoted to unpaid work is greater when they work part-time or when they are retired than if they work on a full-time basis (Chambré, 1984, 1987; Musick and Wilson, 2008). This review presents social, psychological, and resource-based theories as three separate theoretical categories and further divides social theories into context, integration, and roles. In practice, however, these categories are closely related and tend to overlap, and many commonly used predictors of volunteering measure concepts from more than one category. For example, some studies describe education as a resource, but fail to consider that education also teaches prosocial values and promotes social roles and integration. Variables from one perspective often correlate with variables from another in a complex relationship where causality can go both ways. A person with a strong sense of religious values might join a congregation that further reinforces those values, and a person with good organizing skills might take a leadership position within a voluntary association that helps them further refine those skills. Engaging in volunteer work is likely to increase the same values, social factors, and skills that led the person to become a volunteer in the first place. Even with the best possible data, untangling all these causal relationships would be a daunting task. In conclusion, prior research has identified five categories of causes of volunteering. Social context studies show that people sometimes begin volunteering as a way of responding to natural disasters and epidemics, and aspects of one’s geographic community can support external norms to volunteer and create more opportunities for participation. Social integration causes volunteering by making it more likely that people will be asked to volunteer and by promoting external norms that make people more likely to say yes. Social roles, such as parent, church member, and professional worker, often involve an expectation of volunteering. Individual psychological characteristics cause people to volunteer in order to satisfy a wide range of desires and motives, including altruistic values and a sense of moral duty. The resource of free time makes it less costly for people to volunteer, and the resource of skills makes it more likely for a person to be asked to volunteer and easier for the person to do so. Int. J. Nonprofit Volunt. Sect. Mark., November 2011 DOI: 10.1002/nvsm

302

Toward a hybrid theory of volunteering

There are two ways to empirically test variables from these theoretical perspectives and construct a hybrid theory. One way is to consider the three categories as alternative explanations, asking which of the three theoretical types can best predict participation in volunteering. A second approach is to view them as complementary, working together to predict participation. This study offers a preliminary analysis of the relationship between these theories and volunteering by operationalizing each theory with several variables and testing the ability of these variables to predict variation in volunteering. It asks which set of variables best predicts volunteering and whether we can combine elements from all three categories to formulate a single predictive model. We also examine how concepts from multiple theoretical perspectives can combine in a single variable. We have three hypotheses. H1: Each theoretical perspective will explain a substantial amount of the variation in volunteering. H2: Each theoretical perspective will explain an independent aspect of the variation in volunteering. H3: Variables that embody concepts from more than one theoretical perspective will have a stronger or weaker relationship with volunteering, depending on whether the effects of each theoretical perspective work the same way or at cross-purposes.

Data and methods To test our hybrid theory, we used data from the 1995 wave of the Midlife in the United States (MIDUS) survey. This survey is random and nationally representative, uses a large sample, studies individuals in a broad age range, and includes variables that measure concepts from all three theoretical categories. The MIDUS study used written surveys and telephone interviews with a nationally representative random-digit dialing sample of 3032 non-institutionalized, English-speaking adults, born Copyright © 2011 John Wiley & Sons, Ltd.


between 1920 and 1970. The estimated overall response rate to the first wave was 60.8%, and the data are weighted to adjust for nonresponse. Full information about the sample, response rate, weighting, and survey design are contained in the MIDUS codebook, available from the MIDUS website at midmac. med.harvard.edu/research.html. Although this survey was conducted 16 years ago, no survey since that time has included as complete a range of measures of correlates of volunteering, particularly in regards to psychological traits and motivations.

Dependent variable: volunteer status

The MIDUS survey asked respondents to write in how many hours they spend volunteering each month. Although the survey does not define volunteering, it prompts respondents with questions about volunteering for health, medical, youth, and educational organizations. About a third of respondents, or 36.8%, stated that they volunteered in a typical month. Among those who did volunteer, the mean hours per month spent volunteering were 14.4, the median hours were 8.0, and the standard deviation was 20.5 (Table 2). The variable was skewed to the right, with a small number of respondents contributing a large amount of volunteer time.

Independent variables Demographic variables

The MIDUS sample was 43.5% men, 11.2% AfricanAmerican, 1.0% Asian-American, and 2.6% Latino. The average age of the respondents was 45.3 years. We use a quadratic term for age in regression analysis, as prior studies of volunteering find that age has a curvilinear relationship with volunteering, increasing through middle age and declining late in life. Social context

Although the MIDUS survey does not have information about levels of volunteering, social capital, and nonprofit activity in each respondent’s geographical area, the survey does have a measure of neighborhood safety and quality originally designed by Keyes (1998) and Keyes and Shapiro (2004). This scale consists of four agree–disagree statements about Int. J. Nonprofit Volunt. Sect. Mark., November 2011 DOI: 10.1002/nvsm


303

Table 2. Descriptive statistics Variable Dependent variable Volunteering (yes/no) Volunteering (h/month, among those who do volunteer) Demographics Male Black Asian Latino Age Social context Neighborhood safety and quality Prosocial and value orientation Generative concern Obligation to volunteer Obligation overall Subjective religiosity Role identity (lifetime contribution to others) Role identity (expectation of high current volunteering) Resources Occupational prestige Income (in thousands) Social integration Social contribution Social integration Informal socializing with neighbors Multiple categories Education Religious services attendance Religious meeting attendance Voluntary association attendance Part-time employed Full-time employed Retired Married Children aged 0–5 years Children aged 6–12 years Children aged 13–17 years

Range

Mean or % yes

Standard deviation

Cronbach’s alpha

0–1 0–120

36.8% 14.4

n/a 20.5

n/a n/a

0–1 0–1 0–1 0–1 25–74

43.5% 11.2% 1.0% 2.6% 45.3

n/a n/a n/a n/a 13.5

n/a n/a n/a n/a n/a

1–4

3.4

0.6

.68

1–4 0–10 0–10 1–4 1–5 0–1

2.8 6.3 7.2 2.8 3.8 27.5%

0.6 2.6 1.4 0.7 0.9 n/a

.84 n/a .88 .87 n/a n/a

8–90 0–250

36.7

13.6

n/a

3–21 3–21 1–6

15.2 13.8 4.0

3.8 4.3 1.3

.67 .73 n/a

1–12 0–15 0–10 0–15 0–1 0–1 0–1 0–1 0–1 0–1 0–1

6.2 2.3 0.8 2.1 13.5% 60.7% 14.1% 68.2% 19.4% 25.6% 18.2%

2.4 3.1 1.8 3.5 n/a n/a n/a n/a n/a n/a n/a

n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a

feeling safe in one’s neighborhood and trusting one’s neighbors, which correlate highly with one another (Cronbach’s alpha = .68). Social roles

Labor force status, marital status, family status, and participation in religious institutions and voluntary associations bring with them social roles that encourage volunteering and also increase social integration. Social integration

The MIDUS survey uses measures of social integration and social contribution originally designed by Copyright © 2011 John Wiley & Sons, Ltd.

Keyes (1998) and Keyes and Shapiro (2004) and two questions about frequency of social contact with one’s neighbors. The Cronbach’s alpha reliability for social contribution and integration are .67 and .73, respectively. Prosocial and value orientations

MIDUS measures generative concern through a six-item reduction of the Loyola Generativity Scale (McAdams and de St. Aubin, 1992), which has a high level of inter-item reliability (Cronbach’s alpha = .84). MIDUS asks individuals how obligated they would feel “to volunteer time or money to Int. J. Nonprofit Volunt. Sect. Mark., November 2011 DOI: 10.1002/nvsm

304

social causes you support.” We used four questions about how religious and how spiritual the respondents are, and how important religion and spirituality are in their lives, to make a single measure of subjective religiosity (Cronbach’s alpha = .87). We measured prosocial role identity through a question that asked respondents to “rate your contribution to the welfare and well-being of other people” over the course of their lives. We measured volunteer role identity through a question that asked whether the respondent expected to volunteer 15 or more hours per week 10 years from now. Resources

Measures of work skills include education, the prestige of respondents’ occupations, and household income. Labor force status (full-time, part-time, or not working) and a dummy variable for the presence of minor children in the household serve as measures of free time. Education, labor force participation, and the presence of children embody other categories as well and are discussed in the next section. Multiple categories

Nine variables measure concepts related to more than one theoretical category, with four measuring concepts that work together and six measuring concepts that work at cross-purposes. Education is a resource and also facilitates participation in social networks. Attendance at religious services and meetings builds social networks and provide opportunities to take on prosocial roles and also encourage prosocial motives through the internalization of external norms. Three variables related to employment status and three related to children measured causal categories that affected volunteering in conflicting ways. Employment can promote volunteering through the social networks and skills that come with it, but can also discourage it because of the demands paid work places on individuals’ time. The presence of children in the household can encourage volunteering through increased social networks and the adoption of the social role of parent, but can also discourage volunteering through the demands children make on parents’ free time. Copyright © 2011 John Wiley & Sons, Ltd.


Method of analysis

Because the dependent variable (hours volunteered) is not normally distributed, with many zero values, we used logistic regression in our analysis and converted the volunteering measure into a binary variable with 0 = no volunteering and 1 = any volunteering. We also used Tobit regression on the original variables. In reporting the results, we estimate the predictive power of each regression model by using the Nagelkerke calculation of pseudo R-squared (Nagelkerke, 1991). However, one cannot use logistic regression to examine mediation. Accordingly, we report the slopes and significance levels for the Tobit regression. The results of the Tobit and logistic regression equations were similar, and the full logistic results are available upon request. We add the variables in groups, with the first model reporting the results from demographic variables, then demographic and social variables, demographic and individual variables, and demographic and resource variables. Variables that measure concepts from more than one theoretical category are included in each model where they apply. For example, religious services attendance is included as an individual characteristic (religiosity) and a social characteristic (integration and roles), and education is included in all three models. The fifth model incorporates all variables, and the last reports a parsimonious model, in which only statistically significant variables are included. This study compares the relative predictive power of variables drawn from each theoretical category, but readers should not interpret the results as proving one or another category to be superior. Which set of variables has the most predictive power depends on the number of variables in the set, the validity of each variable, and the reliability of each variable. Many of the variables used on MIDUS have been validated in earlier studies, and most have high levels of inter-item reliability, as measured by Cronbach’s alpha (Table 2). Exceptions include the measure of obligation to volunteer and the two measures of role identity, which were newly created for the MIDUS survey. There are nine variables in the model for resources and eight in the model for individual characteristics, but 15 in the model of social factors, so one would expect this model to have a higher predictive power than the other two. Int. J. Nonprofit Volunt. Sect. Mark., November 2011 DOI: 10.1002/nvsm


305

Table 3. Multivariate Tobit regression results

Variable Demographics Male Black Asian Latino Age Age2 Resources Occupational prestige Income Prosocial value orientation Generative concern Obligation to volunteer Subjective religiosity Role identity (lifetime contribution) Role identity (high future volunteering) Social context Neighborhood safety and quality Social integration Social contribution Social integration Informal socializing with neighbors Multiple categories Education Religious services attendance Religious meeting attendance Voluntary association attendance Part-time employed Full-time employed Retired Married Children 0–6 years Children 7–13 years Children 14–17 years Constant N /sigma R2 Log likelihood

Model 1

Model 2

Model 3

Model 4

Model 5

Model 6

Demographics

Resources

Values

Social measures

Full model

Parsimonious

4.1*** 4.2** 1.4 5.4 0.03 0.01

4.6*** 5.6*** 1.3 4.4 0.19 0

1.4 1.9 1.2 8.2** 0.09 0

4.0*** 1.9 1.5 4.6 0.08 0

0.12** 0.02

0.05 0.004 4.1*** 2.5*** 1.6^ 0.5 9.3***

0.6** 1.0*** 1.3*** 3.8* 2.5^ 6.2**

5.0 3004 22.2 .003 6416.7

0.8 7.9*** 4.5*** 17.5* 3002 21.5 .017 6326.7

2.3* 0.3 0.6 5.7^ 0.08 0

37.4*** 2930 19.2 .051 6038.4

1.9* 2.1*** 1.2 0.1 8.5*** 1.0

0.8

1.2*** 0.2^ 1.2***

0.6*** 0.1 0.8*

0.6** 0.8*** 1.6*** 1.4*** 0.1 3.5** 4.2* 2.7** 0.3 6.8*** 2.6* 46.1*** 2939 19.2 .058 6023.7

0.3 0.7*** 1.3*** 1.2*** 3.1* 3.0* 4.3* 3.4*** 1.1 6.9*** 3.2** 51.8*** 2911 18.1 .076 5867.9

p ≤ .10; *p ≤ .05; **p ≤ .01; ***p ≤ .001.

^

Findings The data support the first hypothesis, that each theoretical perspective will explain a substantial amount of the variation in volunteering. Multivariate regression (Table 3) demonstrates the relative predictive power of the variables available in the MIDUS data set that represent each theoretical Copyright © 2011 John Wiley & Sons, Ltd.

category. Demographic controls alone (model 1) explained very little of the variation in decisions to volunteer (Nagelkerke pseudo R-squared = .013). The pseudo R-squared was higher for resource variables (.136), individual characteristics (.284), and social factors (.349). Combining all three perspectives with the demographic controls (model 5) creates a regression Int. J. Nonprofit Volunt. Sect. Mark., November 2011 DOI: 10.1002/nvsm

306

equation with a pseudo R-squared value of .407. We then removed variables one at a time, starting with the ones that were non-significant in the full model to obtain a parsimonious model. Model 6 contains only eight variables (obligation to volunteer, volunteer role identity, education, religious services attendance, voluntary association attendance, full-time employment, children aged 7–13 years, and children aged 14–17 years), but has a high pseudo R-squared value of .361. The data also support the second hypothesis that different theoretical perspectives work independently. Many of the independent variables correlate with one another, particularly within a single theoretical perspective, but also across perspectives. Space does not permit inclusion of these correlations, but they are available from the authors upon request. Despite their correlation, most of the variables that were statistically significant in the partial models (models 2–4) remain significant in the full model (model 5). The parsimonious model (model 6) includes only eight variables, but these variables come from all three theoretical perspectives. To test the third hypothesis, that variables that measure multiple theoretical perspectives have either a larger or smaller predictive power depending on whether causal relationships work together or against one another, we used bivariate logistic regressions and examined model fit, as measured by the Nagelkerke (1991) pseudo R-squared. Variables that measure multiple positive causes of volunteering, such as education (pseudo R-squared = .072) and religious services attendance (R-squared = .076), were good predictors of volunteering, but there were other variables that drew upon a single perspective that were better predictors. These include felt obligation to volunteer (R-squared = .131), planned future volunteering (.081), and social contribution (.113). On the other hand, variables in which different causes worked at cross-purposes tended to have no net relationship with volunteering. The presence of preschool children had no significant relationship with volunteering, perhaps because the gain in social networks when having children is canceled out by the loss of free time. Similarly, retirees were no more likely than the non-retired to volunteer. Copyright © 2011 John Wiley & Sons, Ltd.


Discussion and conclusion This paper classified existing theories of volunteering into three categories: social factors, individual characteristics, and resources. Using multivariate regression analysis, we found that variables from each category predicted a substantial amount of the likelihood of volunteering and did so independently of variables from other categories. Combining variables from all three categories with demographic controls created a model with a pseudo R-squared value of .407, and a parsimonious model with only eight variables had a pseudo R-squared value of .361. Variables that measured multiple categories were good predictors of volunteering, but not necessarily better predictors than variables that measured only one category. However, variables that measured the positive effects of one category and negative effects of another, such as the presence of preschool children and retirement from full-time work, had no net relationship with volunteering, as the positive and negative effects canceled one another out. Some limitations in the data made a full test of the predictive power of each category impossible. First, the results of any comparative test of types of variables say more about the number, reliability, and validity of the variables used than about the strength of the underlying causal relationships that they are supposed to measure. For this reason, we only tested whether variables from each category had a substantial and independent relationship with volunteering, not which category had the best ability to predict volunteering. Second, the MIDUS data set only had one variable for social context, which measure respondents’ subjective opinions of neighborhoods, thus conflating respondents’ psychological characteristics and the actual characteristics of their neighborhoods. Future research should use data sets that include measures of volunteer participation, nonprofit density, and social capital in the community where the respondent resides. Despite these limitations in the data, the hybrid theory expands theoretically upon earlier comprehensive theories of volunteering. This paper goes beyond Wilson and Musick’s (1997a) division of the causes of volunteering into human capital, social capital, and cultural capital, by disaggregating social capital into social context, social roles, and Int. J. Nonprofit Volunt. Sect. Mark., November 2011 DOI: 10.1002/nvsm


social integration. By separating theoretical categories from variables, the paper clarifies why variables that simultaneously measure concepts from multiple theoretical categories may have larger or smaller than expected effects on volunteering. The theory presented in this paper suggests several avenues for future research. The first avenue is methodological, as the theory suggests that many of the variables commonly used in surveys actually measure multiple theories and should be disaggregated for better accuracy. For example, Wilson and Musick (1997a) take education as a measure of human capital, but education also affects social roles, social integration, and prosocial values. Future surveys should find a way to separate these three effects. As social integration and social roles should have independent effects on volunteering, future surveys should distinguish between activities that merely bring people into contact with others from activities that also assign roles to individuals where volunteering is expected. Our findings also have practical implications for policy makers and volunteer managers. Resource and social factors such as education, family status, employment status, religious participation, voluntary association membership, and social interaction are easily visible to volunteer managers, whereas psychological characteristics such as volunteer role identity and generative concern are not. Fortunately, these easily visible factors are also good predictors of volunteering. Volunteer managers will not have to read the minds of potential volunteers to look for prosocial values, or read their bank statements to ascertain their income, but can rely instead upon easily observable characteristics and behaviors such as church and voluntary association membership, participation in the labor force, and educational achievement. A second use of this theoretical perspective relates to encouraging volunteer work among retirees. People gain free time when they retire, but lose the social roles and networks that come with employment, so that retirement has little net effect on volunteering. Volunteer managers can concentrate on other types of social connections to recruit retired volunteers and emphasize the meaningful and productive nature of volunteering to create a replacement social role for people who have lost their roles as paid workers. Retirees also lose the social networks that came with employment, and retirees who move to a new home Copyright © 2011 John Wiley & Sons, Ltd.

307

may lose all of their social networks as well as the social context that originally encouraged them to volunteer. Taking advantage of the social context of retirement communities to foster social networks that view volunteering as normative may help to compensate for the loss of networks and context that accompanies retirement. Of particular use to program managers and policy makers may be the division of social factors into social context, roles, and networks, as managers can use all three to promote volunteering. Any type of social contact increases the likelihood of volunteering, as each additional friend or acquaintance is an additional person who may ask an individual to volunteer. However, some types of social contact also include social roles and contexts that encourage volunteering. By adding pro-volunteering roles and contexts to existing social networks, programs may encourage volunteering among people who had not previously considered it. A performing arts company, for example, may encourage regular attendees at performances to change the conception of their own role from passive onlooker to volunteer, by calling them “members” or “patrons” and establishing normative expectations for them. As this idea takes hold, existing patrons may encourage new attendees to also take on the role of volunteer, turning the performing arts organization into a context that encourages volunteering. Finally, volunteers encourage other members from their social network to join, first as audience members and then as patrons and volunteers.

Biographical notes Christopher J. Einolf is an Assistant Professor at DePaul University’s School of Public Service. His research focuses on altruism, volunteering, and charitable giving, and he also does research on torture. His current work on altruism examines how gender and family structure influence volunteering and giving and how married couples make decisions about charitable donations. Susan M. Chambré is a Professor of Sociology at Baruch College, City University of New York. Her publications focus on volunteerism by elders, Jewish philanthropy, and the intersection between civil society and health policy. Her current research focuses on how patient advocacy organizations influenced health policy in response to TB, polio, and HIV. Int. J. Nonprofit Volunt. Sect. Mark., November 2011 DOI: 10.1002/nvsm

308

References Andreoni J. 1989. Giving with impure altruism: applications to charity and Ricardian equivalence. Journal of Political Economy 97(6): 1447–1458. Anheier HK, Salamon LM. 1999. Volunteering in crossnational perspective: initial comparisons. Law and Contemporary Problems 62(4): 43–65. Apinunmahakul A, Devlin RA. 2008. Social networks and private philanthropy. Journal of Public Economics 92: 309–328. Becker PE, Dhingra PH. 2001. Religious involvement and volunteering: implications for civil society. Sociology of Religion 62: 315–335. Bekkers R. 2005. Participation in voluntary associations: relations with resources, personality, and political values. Political Psychology 26(3): 439–454. Brown E. 1999. The scope of volunteer activity and public service. Law and Contemporary Problems 62(4): 17–42. Caputo R. 2009. Religious capital and intergenerational transmission of volunteering as correlates of civic engagement. Nonprofit and Voluntary Sector Quarterly 38(6): 983–1002. Carlo G, Okun MA, Knight GP, et al. 2005. The interplay of traits and motives on volunteering: agreeableness, extraversion and prosocial value motivation. Personality and Individual Differences 38: 1293–1305. Caro FG, Bass SA. 1997. Receptivity to volunteering in the immediate postretirement period. Journal of Applied Gerontology 16(4): 427–442. Chambré SM. 1984. Is volunteering a substitute for role loss in old age? An empirical test of activity theory. The Gerontologist 24(3): 292–298. Chambré SM. 1987. Good Deeds in Old Age: Volunteering by the New Leisure Class. Lexington Books: Lexington, MA. Chambré SM. 1991a. The volunteer response to the AIDS epidemic: implications for research on voluntarism. Nonprofit and Voluntary Sector Quarterly 20(3): 267–88. Chambré SM. 1991b. Volunteers as witnesses: perspectives on AIDS volunteers in New York City, 1981–1988. The Social Service Review 63: 531–547. Chambré SM. 2006. Fighting for Our Lives: New York’s AIDS Community and the Politics of Disease. Rutgers University Press: New Brunswick, NJ. Choi NG, Burr JA, Mutchler JE, Caro FG. 2007. Formal and informal volunteer activity and spousal caregiving among older adults. Research on Aging 29(2): 99–124. Copyright © 2011 John Wiley & Sons, Ltd.

C. Einolf and S. Chambré Clary EG, Snyder M, Ridge RD, Copeland J, Stukas AA, Haugen J, Miene P. 1998. Understanding and assessing the motivations of volunteers: a functional approach. Journal of Personality and Social Psychology 74: 1516–30. Cnaan RA, Kasternakis A, Wineberg RJ. 1993. Religious people, religious congregations, and volunteerism in human services: is there a link? Nonprofit and Voluntary Sector Quarterly 22(1): 33–51. Einolf CJ. 2010. Does extensivity form part of the altruistic personality? An empirical test of Oliner and Oliner’s theory. Social Science Research 39: 142–151. Einolf CJ. 2011. The link between religion and helping others: the role of values, ideas, and language. Sociology of Religion on-line advance access. DOI:10.1093/ socrel/srr017 Freeman RB. 1997. Working for nothing: the supply of volunteer labor. Journal of Labor Economics 15(1) pt. 2: S140–S166. Gidron B. 1983. Sources of job satisfaction among service volunteers. Journal of Voluntary Action Research 12(1): 30–35. Handy F, Cnaan R, Brudney JL, Ascoli U, Meijs LCMP, Ranade S. 2000. Public perception of “who is a volunteer”: an examination of the net-cost approach from a cross-cultural perspective. Voluntas 11(1): 45–65. Handy F, Cnaan R, Hustinx L, Kang C, Brudney JL, HaskiLeventhal D, Holmes K, Meijs LCMP, Pessi AP, Ranade S, Yamauchi N, Zrinscak S. 2010. A cross-cultural examination of student volunteering: is it all about resumebuilding? Nonprofit and Voluntary Sector Quarterly 39(3): 498–523. Hodgkinson VA, Weitzman MS, Kirsch AD. 1990. From commitment to action: how religious involvement affects giving and volunteering. In Faith and Philanthropy in America: Exploring the Role of Religion in America’s Voluntary Sector, Wuthnow R, Hodgkinson VA (eds). Jossey-Bass: San Francisco; 93–114. Hustinx L, Cnaan RA, Handy F. 2010. Navigating theories of volunteering: a hybrid map for a complex phenomenon. Journal for the Theory of Social Behavior 40(4): 410–434. Keyes CLM. 1998. Social well-being. Social Psychology Quarterly 61: 121–140. Keyes CLM, Shapiro AD. 2004. Social well-being in the United States: a descriptive epidemiology. In How healthy Are We? A National Study of Well-Being at Midlife, Brim OG, Ryff CD, Kessler RC (eds). University of Chicago Press: Chicago; 350–372. Int. J. Nonprofit Volunt. Sect. Mark., November 2011 DOI: 10.1002/nvsm

Who volunteers? Constructing a hybrid theory Kim SY, Hong GS. 1998. Volunteer participation and time commitment by older Americans. Family and Consumer Sciences Research Journal 27(2): 146–166. Lee L, Piliavin JA, Call VRA. 1999. Giving time, blood, and money: differences and similarities. Social Psychology Quarterly 62(3): 276–290. Matsuba MK, Hart D, Atkins R. 2007. Psychological and social–structural influences on commitment to volunteering. Journal of Research in Personality 41: 889–907. McAdams DP, de St. Aubin E. 1992. A theory of generativity and its assessment through self-report, behavioral acts, and narrative themes in autobiography. Journal of Personality and Social Psychology 62: 1003–1015. McPherson JM, Rotolo T. 1996. Testing a dynamic model of social composition: diversity and change in voluntary groups. American Sociological Review 61: 179–202. Mellor D, Hayashi Y, Stokes M, Firth L, Lake L, Staples M, Chambers S, Cummins R. 2009. Volunteering and its relationship with personal and neighborhood wellbeing. Nonprofit and Voluntary Sector Quarterly 38(1): 144–159. Midlarsky E. 1991. Helping as coping. Review of Personality and Social Psychology 12: 238–264. Mowen JC, Sujan H. 2005. Volunteer behavior: a hierarchical model approach for investigating its trait and functional motives antecedents. Journal of Consumer Psychology 15(2): 170–182. Musick MA, Wilson J. 2008. Volunteers: A Social Profile. Indiana University Press: Indianapolis, IN. Nagelkerke NJD. 1991. A note on a general definition of the coefficient of determination. Biometrika 78(3): 691–2. Okun MA, Pugliese J, Rook KS. 2007. Unpacking the relation between extraversion and volunteering in later life: the role of social capital. Personality and Individual Differences 42: 1467–1477. Oliner S, Oliner P. 1988. The Altruistic Personality. Free Press: New York. Omoto AM, Snyder M. 2002. Considerations of community: the context and process of volunteerism. American Behavioral Scientist 45(5): 846–867. Park JZ, Smith C. 2000. “To whom much has been given. . .”: religious capital and community voluntarism among churchgoing Protestants. Journal for the Scientific Study of Religion 39(3): 272–286. Perkins DD, Brown BB, Taylor RB. 1996. The ecology of empowerment: predicting participation in community organizations. Journal of Social Issues 52(1): 85–110. Copyright © 2011 John Wiley & Sons, Ltd.

309 Perry JL, Brudney JL, Coursey D, Littlepage L. 2008. What drives morally committed citizens? A study of the antecedents of public service motivation. Public Administration Review 68(3): 445–458. Portney KE, Berry JE. 1997. Mobilizing minority communities: Social capital and participation in urban neighborhoods. American Behavioral Scientist 40(5): 632–644. Prouteau L, Wolff FC. 2008. On the relational motive for volunteer work. Journal of Economic Psychology 29(3): 314–335. Putnam R. 2007. E pluribus unum: diversity and community in the twenty-first century. Scandanavian Political Studies 30(2): 137–174. Putnam R. 2000. Bowling alone: the collapse and revival of American community. Simon and Schuster: New York. Rooney P, Steinberg K, Schervish PG. 2004. Methodology is destiny: the effect of survey prompts on reported levels of giving and volunteering. Nonprofit and Voluntary Sector Quarterly 33(4): 628–654. Ross AD. 1954. Philanthropic activity and the business career. Social Forces 32(3): 274–280. Rossi AS (ed). 2001. Caring and Doing for Others: Social Responsibility in the Domains of Work, Family, and Community. University of Chicago Press: Chicago. Rotolo T. 2000. A time to join, a time to quit: the influence of life cycle transitions on voluntary association membership. Social Forces 78: 1133–1161. Rotolo T, Wilson J. 2007. The effects of children and employment status on the volunteer work of American women. Nonprofit and Voluntary Sector Quarterly 36: 487–503. Rotolo T, Wilson J. 2011. State-level differences in volunteerism in the United States: research based on demographic, institutional and cultural macro-level theories. Nonprofit and Voluntary Sector Quarterly. DOI:10.1177/0899764011412383 Sampson RJ, McAdam D, MacIndoe H, Weffer-Elizondo S. 2005. Civil society reconsidered: the durable nature and community structure of collective civic action. The American Journal of Sociology 111(3): 673–714. Schwartz SH. 1977. Normative influences on altruism. Advances in Experimental Social Psychology 10: 221–279. Schwartz SH, Fleishman JA. 1978. Personal norms and the mediation of legitimacy effects on helping. Social Psychology 41: 306–315. Smith DH. 1994. Determinants of voluntary association participation and volunteering: a literature review. Int. J. Nonprofit Volunt. Sect. Mark., November 2011 DOI: 10.1002/nvsm

310 Nonprofit and Voluntary Sector Quarterly 23(3): 243–264. Staines GL. 1980. Spillover versus compensation: a review of the literature on the relationship between work and nonwork. Human Relations 33(2): 111–129. Sundeen RA. 1992. Differences in personal goals and attitudes among volunteers. Nonprofit and Voluntary Sector Quarterly 21(1): 271–291. Sundeen RA, Raskoff SA, Garcia MC. 2007. Differences in perceived barriers to volunteering to formal organizations: lack of time versus lack of interest. Nonprofit Management & Leadership 17: 279–300. Wiehe VE, Isenhour L. 1977. Motivation of volunteers. Journal of Social Welfare 4(2–3): 73–80. Wilson J. 2000. Volunteering. Annual Review of Sociology 26: 215–40.

Copyright © 2011 John Wiley & Sons, Ltd.

C. Einolf and S. Chambré Wilson J, Janoski T. 1995. The contribution of religion to volunteer work. Sociology of Religion 56(2): 137–153. Wilson J, Musick M. 1997a. Who Cares? Towards an integrated theory of volunteer work. American Sociological Review 62(5): 694–713. Wilson J, Musick M. 1997b. Work and Volunteering: The Long Arm of The Job. Social Forces 76(1): 251–272. Wuthnow R. 1991. Acts of Compassion: Caring for Others and Helping Ourselves. Princeton University Press: Princeton. Wymer WW. 1999. Understanding volunteer markets: the case of senior volunteers. Journal of Nonprofit and Public Sector Marketing 6(2/3): 1–23. Yörük BK. 2009. How responsive are charitable donors to requests to give? Journal of Public Economics 93(9/10): 1111–1117.

Int. J. Nonprofit Volunt. Sect. Mark., November 2011 DOI: 10.1002/nvsm