I Volunteer,Therefore I am? Factors Affecting Volunteer Role Identity

Abstract

Volunteer role identity has long been of interest to social scientists seeking to understand volunteer commitment and the psychological consequences of volunteering. The study reported here tests the theory that predicts that people identify more strongly with the volunteer role as compensation for the absence of other productive roles. Using a sample (n = 572) of Dutch volunteers over the age of 50, we find a strong association between age and volunteer role identity. For older volunteers, the volunteer role is a more important part of who they are. We find that retirement plays an important role in this. The retirement effect, in turn, is accounted for by the extra time retirees invest in the role, signaling a compensation strategy. We find a similar substitution effect for the unemployed/disabled, but not for widowhood. The study makes a contribution by situating the explanation of volunteer role identity within a life-course framework.

Keywords

role substitution role identity aging life course volunteering

Introduction

Not all volunteers are alike when it comes to what the role means to them. Some regard their volunteer work as a very important part of their lives. It is so central to their self-perception it virtually defines who they are. They would agree with the man who told Teske (1997, p. 121) that by volunteering he is being true to his own self and “following through on something that is really important” to him. This man could hardly imagine volunteering “just for fun,” to make friends, to help find a job, or simply to fulfill his membership obligation to a voluntary association.

At the other end of the spectrum are volunteers for whom the role is more ephemeral. They regard their volunteer work as not all that relevant to their sense of self. Volunteering is more of a pastime for them: “it’s just something I do or enjoy and not who or what I am” (Thoits, 2013, p. 11). Their sense of self would remain intact even if they ceased volunteering (Gronlund, 2011, p. 6).

This variation in strength of identification with the volunteer role has become an important topic in volunteerism research because two volunteers could perform exactly the same role (e.g., leading a Boy Scout troop) but differ in how strongly they identify with it. What causes this variation? Most attempts to answer this question have focused on the proximate determinants of volunteer role identity (VRI). For example, volunteers might be more inclined to identify with the role if their efforts are recognized and rewarded by others. Such determinants are important, but focusing too narrowly on them obscures the possibility that other roles and role identities might influence VRI. Every volunteer is simultaneously performing multiple roles—as a parent, a neighbor, a co-worker, or a church member. Role theory predicts that these other roles influence how strongly a person identifies with the volunteer role. The influence of other roles has long been recognized in research on the rate of volunteering (i.e., the proportion of people who volunteer) and sometimes in the research on the intensity of volunteering (i.e., the number of hours volunteered) but as far as we are aware role theory has not been used to explain variations in VRI.

In this study we examine VRI among a large sample of volunteers in the Netherlands. We focus on the influence of two situations that have attracted much attention from volunteerism researchers—retirement and widowhood. In both cases theories have been tested that predict change in the rate and intensity of volunteering as people retire or experience the loss of a spouse. But no study has asked whether these major life-course changes have any effect on VRI strength. Using role substitution or compensation theory, we hypothesize that in both cases VRI will be stronger.

Theory and Hypotheses

Volunteer Role Identity

Roles are an important source of self-identity. As all people perform a number of roles, their self (or personal) identity is composed of identities based on more than one role—as well as identities drawn from other sources, such as gender, age, or ethnicity (Stets & Burke, 2005). Thus a priest might shun certain leisure time pursuits even when not “on the job” if his self-identity draws heavily on his role identity as a priest.

The degree to which a person identifies with a role has been referred to as variation in the “salience” of the role (Thoits, 2012). In fact, this term has two meanings. The first is the importance of the role to the individual. It signifies the degree to which the role is viewed as self-descriptive. The second dimension is the ranking given to the role identity in relation to other identities (Thoits, 2013). The issue in this case is the centrality of the role (Charng, Piliavin, & Callero, 1988). Thoits (2013, p. 2) refers to the former as “identity importance” and to the latter as “identity invocation.” It is quite conceivable that a role identity could be important to an individual but also peripheral—or vice versa. For example, a middle-aged woman might identify strongly with the volunteer role insofar as she welcomes it as a self-description (that is the kind of person I am) but not rank it very highly in importance to her personal identity compared with her role identity as an employer, mother, wife, or daughter. In the study reported here, the respondents were not asked to rank the volunteer role in relation to other roles. Instead they were asked how important the volunteer role is to their sense of self. Although Thoits (2012) uses the term “salience” to describe a composite measure consisting of both dimensions (plus a measure of commitment), in this study the term salience refers to the importance of the identity.

Explaining Variation in VRI

Most explanations of variation in VRI strength have focused on proximal factors such as the immediate rewards of performing the role or the validation of the role by significant others (Marta, Manzi, Pozzi, & Vignoles, 2014). These proximal factors also include years of experience as a volunteer; time spent volunteering; training for the volunteer role; expressions of gratitude for volunteering; positive interactions with other volunteers, clients, and staff; and level of satisfaction with volunteer work, belief in the importance of volunteering, and of the issues for which the volunteer work is performed (Callero, Howard, & Piliavin, 1987; Charng et al., 1988; Finkelstein, Penner, & Brannick, 2005; Grube & Piliavin, 2000; Güntert & Wehner, 2015; Moen, Erickson, & Dempster-McClain, 2000; Piliavin, Grube, & Callero, 2002).

These studies have added much to our understanding of VRI but they necessarily push to the background more distal but equally important factors such as the other roles the volunteer is performing. This is where role theory is useful because it provides an important clue as to other reasons why some people identify more strongly with the volunteer role than others. It predicts that VRI salience depends on which other roles the volunteer is performing. According to “role substitution theory,” if a person ceases performing a role that once made a major contribution to the person’s sense of who they are—and connected to this, feelings of self-worth and a sense of mattering (Thoits, 2011)—that person will promote another role to serve this purpose. This idea that by volunteering a person is “compensating” for the loss of a role by moving the volunteer role to a more salient position in their lives guides some of the research on the health benefits of volunteering (Greenfield & Marks, 2004; Pavlova & Silbereisen, 2012). The compensation idea has proven to be very useful to researchers interested in how volunteering varies across the life course because it assumes that role identities change with age as individuals assume and later vacate social roles (Lancee & Radl, 2014; Moen et al., 2000; Nesbit, 2013). The theory does not assume that people consciously strive for compensation by volunteering, simply that the volunteer work functions as a compensation device.

Role substitution theory guides much of the research on changes in the rate or intensity of volunteerism in retirement and widowhood. Volunteer work increases in intensity as it becomes a substitute for the role of the paid worker or the role of the spouse. A brief review of the research on this topic is necessary because time spent volunteering (intensity) is regarded as one determinant of VRI strength. If retirees and widows or widowers contribute more hours this might result in a more salient VRI.

Volunteering and Retirees

Many articles have been published on the topic of the volunteering among older adults, several guided by role substitution or compensation theory. It is certainly plausible that people might take up volunteering once they are no longer working for pay or, if they are already volunteering, to increase the hours they contribute now that they have more free time (Chambre, 1984; Moen & Flood, 2013; Mutchler, Burr, & Caro, 2003). Volunteer work is one way retirees “keep busy” (Griffin & Hesketh, 2008; Oakley & Pratt, 1997). Volunteer work also provides protection from the social exclusion and loneliness that might result from retirement (Jensen, Lamura, & Principi, 2014). In fact, retirement has little effect on the rate of volunteering (Broese Van Groenou & Van Tilburg, 2012; Butricia, Johnson, & Zedlewski, 2009; Erlinghagen, 2010; Smith, 2010; Van den Bogaard, Henkens, & Kalmijn, 2014; Warburton, Le Broque & Rosenman, 1998; Wilson, 2012). The best predictor of whether a retiree is a volunteer is whether he or she volunteered before retirement (Butricia et al., 2009; Erlinghagen, 2010; Smith, 2004). When it comes to retirement and time spent volunteering, however, the facts are very different. Almost uniformly, studies show that retirees volunteer more hours than those who remain fully employed (Chambre, 1984; Choi, 2003; Einolf, 2009; Erlinghagen, 2010; McNamara & Gonzales, 2011; Moen et al., 2000; Narushima 2005; Principi, Warburton, Schippers, & Di Rosa, 2012; Sherman & Shavit, 2012; Van den Bogaard et al., 2014; van Ingen & Dekker, 2011).

The findings on intensity are important because frequent enactment of a role encourages people to develop a self-concept consistent with it (Marta et al., 2014; Thoits, 2013). As Thoits (2012, p. 366) puts it, “It seems theoretically plausible that investing many hours in volunteer work makes the volunteer identity more salient to the person.” A busy volunteer spends more time with others who will validate and support her role as well as with people like herself who are doing volunteer work. Thoits’ own research on volunteers acting as hospital visitors for cardiac patients shows that weekly hours spent visiting have a positive effect on volunteer identity salience.

Volunteering and Widowhood

According to role substitution theory, married people increase their volunteering if their spouse dies. This is either because the bereaved are seeking social contacts or emotional support (Tang, Choi, & Morrow-Howell, 2010) or because they have more free time since they are no longer caring for their spouse (Lancee & Radl, 2014).

Li (2007) finds that widowed individuals volunteer at a higher rate, at least up to 4 years after the loss of the spouse, after which the effect seems to disappear. Butricia et al. (2009), on the other hand, find that volunteers who lose a spouse are more likely to quit volunteering than those who remain married, and Nesbit (2013) finds that widows volunteer at a lower rate and with lower intensity than those who are married. Finally, no effect of widowhood is found in studies conducted in the United States (Donnelly & Hinterlong, 2010), Belgium (Dury et al., 2015), and the Netherlands (Broese Van Groenou & Van Tilburg, 2012). Given the mixed results in these studies, another test of the role substitution or compensation theory seems to be called for, this time using a large sample of volunteers and employing as an outcome variable not only intensity of volunteering but also VRI strength. The death of a spouse takes away the psychological benefits of being married. In response, widows and widowers should not only volunteer more hours but identify more strongly with the volunteer role.

Hypotheses

On the basis of compensation theory and in light of the research findings described above, we test the following hypotheses in this study.

Hypothesis 1: Among volunteers, retirees identify more strongly with the volunteer role than those who are currently employed.

Hypothesis 2: Among volunteers, widows or widowers identify more strongly with the volunteer role than those who are currently married.

Earlier, we pointed out that one of the reasons given for variations in VRI strength is time spent volunteering. A second step in our study is to model the possibility that time spent volunteering mediates the relation between role absence and volunteering.

Hypothesis 3: Time spent volunteering mediates the relation between retirement status and VRI strength.

Hypothesis 4: Time spent volunteering mediates the relation between widowhood and VRI strength.

Method

To test the hypotheses we regress a latent factor (VRI salience) on a number of explanatory variables in five steps. First we look at the relation between age and VRI. This sets the stage for the examination of retirement and the death of spouse: if there is no relation between age and VRI then there is little or no role for retirement or widowhood to play. The second model adds a number of controls. We control for gender because of differences in volunteer rates and intensity between men and women that might be reflected in different VRI, and because men and women might cope with retirement and the loss of a spouse in different ways. We control for education because of class differences in volunteerism and the timing and meaning of retirement. We control for health because of the association between health status and volunteer rates, and intensity and the possibility that healthier adults are more likely to continue working in old age. We control for frequency of church attendance because this might influence how events such as retirement and loss of a spouse are handled and because of its possible association with VRI strength. The third model adds a measure of employment status of which retirement is one component. The fourth model adds a measure of marital status of which widowhood is one component. Finally, we add a variable for time spent volunteering. The fit of these models varied: RMSEA ranged between .05 and .08; CFI between .87 and .90.

We test the hypotheses using structural equation models (SEMs; Stata 13) estimated using maximum likelihood with missing values (StataCorp, 2013), with volunteer role identity as latent endogenous variable. This imputation procedure ensures that all of the available information is used in the model, although the difference with listwise deletion was only 31 respondents. The conclusions from the imputed and non-imputed data are the same.

Data

The Longitudinal Internet Studies for the Social Sciences (LISS) is a large survey based on a true probability sample of the population of the Netherlands of 16 years of age and older. Data collection is coordinated by CentERdata at Tilburg University (see www.lissdata.nl). Questionnaires are answered online. The survey includes non-Internet users, who were equipped with a computer and Internet access, and who received guidance in using them. This is important, because research has shown that even in the Netherlands, with one of the highest Internet adoption rates in the world, adding non-users to an Internet panel improves the quality of the data (Leenheer & Scherpenzeel, 2013). Surveys lasting 15 to 30 min are conducted monthly and respondents are paid 15 euros per hour to complete the questionnaires. Refreshment samples are drawn to deal with attrition. The representativeness of the panel resembles surveys in the Netherlands which collect data in more traditional ways. Older individuals tend to be somewhat underrepresented but not as much as in online access panels (Knoef & De Vos, 2009).

Once a year (in February and March) respondents are asked about their participation in voluntary organizations. To gather data on volunteering and volunteer identities, we designed a special questionnaire that was administered to those respondents who indicated they volunteered for at least one of 12 types of voluntary organizations. Our follow-up survey took place in May 2011. The response rate was 81% and included volunteers from all kinds of organizations. Since the survey is based on a probability sample of the Dutch population and since all important types of volunteer organizations are covered (including an “other” category), our sub-sample should be reasonably representative of volunteers in the Netherlands. It is hard to assess representativeness directly, as population data on volunteers are not available in the Netherlands (e.g., there are no censuses). When we compare the LISS data with a survey by the Central Bureau of Statistics (Schippers & Conen, 2014), the gender and age distributions are fairly similar. However, the oldest old (75+ years) in our data are more active than they are in the other survey (24% volunteering vs. 16% volunteering). The hours volunteered are also fairly similar, although women in the LISS volunteered slightly fewer hours than in the other data (3.8 vs. 4.3 hr/week).

We excluded two groups from our analyses: (a) those who stated (at the end of the questionnaire) that their answers were invalid because they did not participate (or no longer participated) in the mentioned organization (n = 33) and (b) those who gave the “not applicable” answer to the first set of items (see below; n = 10). This second group of respondents apparently misread the question in the preceding module. For instance, the question explicitly referred to “the past 12 months,” but some respondents indicated they participated despite the fact that the last time they volunteered was several years ago.

In the current study, we focus on volunteers older than 50 years of age. Although the role absences that are central to this study can be found in early stages of the life course, it is highly unusual. Very few people retire permanently before the age of 50, and spousal loss, although experienced by people under 50, does not carry the same implications as it does for older adults. The cut-off of 50 also makes sense because in our data, the age group 56 to 60 is the first to show increased VRI. This means the group of 51 to 55 needs to be included to capture the full effect, and it is the most intuitive reference group. The cut-off of 50 is also used in several well-known surveys on aging, such as the Health and Retirement Study (HRS), the English Longitudinal Study of Aging (ELSA), and the Survey of Health, Aging and Retirement in Europe (SHARE). After applying the selection criteria, the data set comprises 572 respondents.

Variables

Volunteer role identity

This is a five-item scale designed by Grube and Piliavin (2000). The items were as follows: (a) My volunteer work is something I rarely ever think about; (b) I would feel a loss if I were forced to give up volunteering; (c) I really don’t have any clear feelings about volunteer work; (d) For me, being a volunteer means more than just doing volunteer work; and (e) Volunteering is an important part of who I am. Answer categories ranged from (1) strongly disagree to (7) strongly agree. The items formed a reliable scale (α = .82). Table 1 shows the descriptive statistics of the continuous variables in our study. As noted earlier, this scale is a measure of salience, or how important the identity is to the volunteer.

Table 1.

Descriptive Statistics of Continuous Variables.

	M	SD	Minimum-maximum	n
Volunteer role identity^a	5.090	1.313	1–7	572
Age	64.010	7.934	51–88	572
Education	3.855	1.489	1–6	560
Health	3.049	0.728	1–5	555
Church attendance	2.276	1.450	1–6	569
Time spent volunteering	5.268	5.346	0–40	572

Based on the average across the VRI items.

Age

Age is a continuous variable, which ranges between 51 and 88 years (see Table 1). It is part of a questionnaire filled out by the contact person of the household on a monthly basis. Since the relation between age and VRI showed a non-linear functional form, we use age as a categorical variable, with the following groups: 51 to 55, 56 to 60, 61 to 65, 66 to 70, 71 to 75, and 76+.

Employment status and marital status

Respondents were instructed to select their current (primary) occupation from a list of 15 different options (“Please indicate which of the following statements apply to your situation”), including “I perform paid work” and “I am a pensioner.” We used this question to determine who is retired. In 2011, every person aged 65 or older was entitled to a basic state pension in the Netherlands. Few Dutch people remained employed after the age of 65 in 2011 (Arts & Otten, 2013). In addition, “being a pensioner” generally refers to having quit working rather than to receiving a state pension in everyday Dutch language. Therefore, it is very unlikely that respondents who indicated that they were a pensioner were in fact still working for pay.

After merging some of the initial answer categories, nine remained. The two most important for our current purpose are full-time employed (n = 79 or 14% of the sample) and retired or living on a pre-retirement allowance (n = 258 or 45%). The other options were part-time employed (n = 63), other employment (n = 47), unemployed or disabled (n = 41), student (n = 3), homemaker (n = 62), and other (n = 19). Among the “others” were respondents who indicated that volunteer work was their primary occupation. We do not show the results of the “other” and “student” groups in Table 2, but they were included in the estimations.

Table 2.

SEM of Volunteer Role Identity on Age Groups and Employment Status (Unstandardized Path Coefficients and Standard Errors).

	(1)	(2)	(3)	(4)	(5)
	b/SE	b/SE	b/SE	b/SE	b/SE
Age (ref. = 51-55 years)
56-60 years	0.401** (0.145)	0.353* (0.141)	0.256 (0.145)	0.255 (0.144)	0.265 (0.139)
61-65 years	0.458** (0.144)	0.401** (0.141)	0.182 (0.166)	0.171 (0.166)	0.122 (0.159)
66-70 years	0.667** (0.158)	0.569** (0.157)	0.293 (0.200)	0.280 (0.201)	0.280 (0.193)
71-75 years	0.585** (0.169)	0.442** (0.168)	0.177 (0.211)	0.152 (0.212)	0.250 (0.205)
76+ years	0.395* (0.180)	0.271 (0.179)	0.005 (0.223)	−0.018 (0.224)	0.072 (0.215)
Women		0.089 (0.087)	0.068 (0.096)	0.063 (0.099)	0.197* (0.097)
Education		−0.015 (0.030)	−0.013 (0.030)	−0.014 (0.030)	−0.017 (0.029)
Health		−0.006 (0.059)	0.021 (0.060)	0.024 (0.060)	0.008 (0.058)
Church attendance		0.120** (0.031)	0.129** (0.031)	0.131** (0.031)	0.131** (0.030)
Employment (ref. = full-time)
Part-time work			0.074 (0.178)	0.082 (0.178)	0.066 (0.170)
Paid employment: other			0.330 (0.192)	0.338 (0.192)	0.178 (0.184)
Unemployed/disabled			0.438* (0.208)	0.446* (0.209)	0.215 (0.200)
Homemaker			0.181 (0.202)	0.186 (0.204)	0.063 (0.195)
Retired			0.387* (0.186)	0.395* (0.187)	0.214 (0.179)
Marital status (ref. = married)
Separated/divorced				0.076 (0.143)	−0.011 (0.137)
Widow(er)				0.140 (0.162)	0.094 (0.155)
Never been married				−0.059 (0.161)	−0.077 (0.154)
Time spent volunteering					0.064** (0.009)
R ²	.045	.081	.114	.116	.223
N	572	572	572	572	572

Note. SEM = structural equation model.

Respondents >50 years.

p < .05. **p < .01.

The information on the marital status of respondents is updated on a monthly basis. After combining the categories of divorced and separated, this variable has four categories: Married (n = 422; 74%); separated or divorced (n = 58; 10%); widow or widower (n = 46; 8%); and never married (n = 46; 8%). Compared to the Dutch (50+) population in 2011 (married 66%, divorced 11%, widow(er) 14%, never married 9%; CBS Statline, 2016), this signifies that married individuals are somewhat overrepresented in our sample, which is likely due to the underrepresentation of individuals older than 80 years who are more likely to be widowed.

Time volunteering

The variable time volunteering was measured by asking respondents “How many hours a week on average do you spend on voluntary work for [organization X]?”

Other variables

Health is self-reported, ranging from 1 (poor) to 5 (excellent). Education consists of six categories, ranging from primary school to university. Women is a dummy variable for gender. And finally, Church attendance had six answer categories, ranging from (1) never to (6) every day.

Results

Table 1 presents descriptive statistics for the variables used in the study. With four representing the neutral score, the average score on the VRI items (M = 5.090) was toward the “agree” end of the continuum, although it is clear from the standard deviation (SD = 1.313) that there was considerable variation around the mean. Our sample volunteers ranged in age from 51 to 88, with an average of 64 years. The average educational level was the equivalent of junior college. The average health score (3.049) signals “good health.” The average score on church attendance (2.276) indicates slightly more than a few times a year.

Figure 1 plots the differences in time spent volunteering and VRI salience across volunteers in six age groups, controlled for gender, education, and church attendance. Note that these variables have different scales, making it impossible to conclude that scores on one variable are higher than on the other in any given age group. Both time spent volunteering and VRI show higher scores after 50. Time spent volunteering reaches its peak in the 61 to 65 age group. VRI strength peaks somewhat later, in the 66 to 70 age group.

Figure 1.

VRI and time spent volunteering across age groups (controlled for gender, education, and church attendance).

The first model in Table 2 shows that, using 51 to 55 as the reference category, VRI salience increases across the age-groups, beginning to weaken only among volunteers aged 76 or more. Nevertheless even members of this oldest old group identify more strongly with the volunteer role than volunteers in their early 50s.

The second model adds a number of controls. Frequent churchgoers of all ages have stronger VRI. However, the introduction of these controls, although it weakens the 76+ effect somewhat, leaves the overall age effect intact. Church attendance frequency seems to account for stronger VRI among the oldest old because the coefficient for that group becomes insignificant. The third model in Table 2 tests the first hypothesis of this study: that retired volunteers have stronger VRI than other volunteers. Using full-time employed as the reference category, the positive coefficient for “pensioner” is 0.387* and that for unemployed/disabled is 0.438*, which are not small effects judging from the corresponding standardized (dummy) effects of β = .430 and β = .487 (calculations based on the model-implied variance of the latent VRI factor). The inclusion of employment status in the model renders the age dummies non-significant. The fourth model in Table 2 introduces marital status into the equation with married as the reference category. None of the groups deviate significantly from those who are married. The final model introduces time spent volunteering. The coefficient is 0.064* (standardized effect .349). The more time the volunteer spends on volunteer work the more salient the VRI. The introduction of this variable into the model renders all the work status variables insignificant. In other words, retirement and unemployment/disability strengthen VRI because of greater time investment in the role. These mediation effects can be tested directly. The indirect path from retired through time spent volunteering to VRI is significant (b = .186; SE = .065; p = .004), meaning that because of the greater time investment in the volunteer role by retirees their estimated VRI is .186 points higher compared with that of full-time workers (standardized effect .205). The indirect path from widow(er) through time spent volunteering to VRI is non-significant (b = .048; SE = .052; p = .357).

Finally, while the control variables are not the main focus of this study it is worth noting that women and frequent churchgoers identify more strongly with the volunteer role. The gender coefficient is insignificant until time spent volunteering is included in the model. The reason for this is that women spent less time volunteering compared with men (4.2 hr/week vs. 6.2 hr/week). When this difference is taken into account, women have stronger VRI. The church attendance effect—frequent churchgoers identify more strongly with the volunteer role regardless of how much time they spend volunteering—suggests there are ways other than time spent in the role to strengthen identity and, more generally, that not all identities are antithetical to the volunteer role identity but might instead encourage it.

Discussion

There is little disagreement among gerontologists that volunteering has the potential to provide older people, faced with the loss of roles as they retire and, if married, experience the death of a spouse, a new personal identity in which VRI is more salient (Principi, Warburton, Schippers, & Di Rosa, 2012). Up to this point, however, this theory has been tested only on volunteer rates and intensity, which do not measure identity at all. To our knowledge, the present study is the first to ask whether VRI salience is associated with role absence. By focusing on the experiences of role absence in the later stages of life, we add a life-course perspective to VRI studies. We find confirmation of the fact that, among volunteers, VRI salience varies across age groups, as shown in Figure 1. These variations are unlikely to have been caused by the proximal factors usually adduced to explain VRI.

It is clear from the results of this study that the way in which volunteer role identity is shaped by other roles deserves more attention from researchers in this field. For the first time we show that retirees identify more strongly with the volunteer role. It is worth underlining that retirement in this case is complete and final, as is true of most European societies (Erlinghagen, 2010). Retirees in the Netherlands are unlikely to be either working for pay and or searching for paid employment. In the United States, on the other hand, there is “considerable variability in the transition pathways to full retirement” (Tang & Burr, 2015, p. 1742). Studying the effect of taking one or another of these pathways on volunteering has so far been confined to rates and intensity. For example, one study of U.S. adults aged 70 and over found that retirees who had returned to work part-time were more likely to volunteer and contributed more hours than those who had fully retired (Choi, 2003). The association between partial retirement and VRI salience is unknown. It is interesting to speculate what these different pathways to full retirement might mean for VRI salience. The “role overload” thesis would predict that retirees who are working for pay will volunteer less and perhaps the volunteer role would be less salient to them (Mutchler et al., 2003).

Do these results for retirees validate the compensation theory? We contend that the mediation effect of time spent volunteering does support the theory. Retirees choose to devote more time to volunteer work in search of compensation, consciously or not, for the loss of other productive roles and, as a result, identify more strongly with the volunteer role. However, there might be an alternative explanation which would not support the theory: retirees simply have more leisure time compared with full-time workers which they devote to volunteer work. As a result, they identify more strongly with the volunteer role. This means VRI is simply a result of the “opportunity structure” of free time.

Fortunately, our data allowed us to test one of the implications of the opportunity structure argument. If it were true that retirement is significant for VRI simply because it increases leisure time, then older adults with more free time would have stronger role identities. Our data did not support this argument. The weekly amount of free time registered by our respondents was correlated with neither time spent volunteering (r = .01; p = .68) nor VRI (r = .03; p = .38). This would suggest that the compensation theory is more credible.

In addition, unpaid caring roles might also influence retirees’ level of immersion in the volunteer role. For example, one study found that older adults who were raising a grandchild living with them spent less time volunteering compared with those caring for a grandchild not living with them (Bulanda & Jendrick, 2016). This suggests that VRI salience varies in response to competing family roles. However, we found no relation between marital status and VRI salience. Admittedly, the number of widows/widowers in our sample is small and the data might not have enough statistical power to detect any effects, but it is interesting to speculate why volunteers who had lost a spouse devoted no more time to volunteering than those who were still married. One possibility is that, while volunteering is regarded as a suitable and attractive activity for retirees (who are searching for an activity that makes them feel useful) those who have lost a spouse do not have the same motivation.

There is another possible explanation for the absence of any association between marital status and VRI salience. Previous research has shown that the probability of volunteering is greater among widows and widowers whose spouse died up to 3 years ago (Li, 2007), but there is no difference in volunteer activity between widows and the still married in cases where the widow(er) lost his or her spouse more than 3 years ago. This latter group has had more time to adjust to the loss. It could be that the need for compensation is most pressing in the period immediately following the death of the spouse. Unfortunately, our data did not allow us to test these ideas.

One unanticipated result of our research is that the unemployed/disabled volunteers identify more strongly with the volunteer role than full-time workers. This association disappears once time spent volunteering is controlled suggesting that, like retirees, unemployed/disabled volunteers identify more strongly with the volunteer role because they spend more time volunteering than full-time workers. Unfortunately the numbers in each category were too small to treat them separately in the models (unemployed = 16, disabled = 25). But it is interesting to speculate about what future research with greater numbers might reveal. The latest research suggests that the transition to unemployment results in a lower rate of volunteering, at least in the case of men (Lancee & Radl, 2014). Losing one’s job does not result in a search for volunteer opportunities. Indeed, it seems to discourage volunteering. One theory is that spells of unemployment do not encourage identification with the volunteer role because the unemployed want to return to full employment (Pavlova & Silbereisen, 2012). But in ancillary analysis we found that unemployed/disabled volunteers contribute more hours than employed volunteers (3.7 hr a week more after controls). Judging by the final model in Table 2, they identify more strongly with the volunteer role as a result. This is also confirmed by a direct test of this mediation effect, which is clearly significant (b = .240; SE = .074; p = .001).

How does the finding that the unemployed/disabled also have higher VRI (as a result of spending more time volunteering) relate to role compensation theory? This depends on whether the role loss is perceived as a short-term problem or a lasting situation. As one reviewer suggested, unemployment in this age group (50+) might well be a form of early retirement. This seems in line with findings on long-term unemployment and reemployment rates among the elderly in the Netherlands (Organisation for Economic Co-operation and Development, 2014). If this is true, the urge to compensate for the absence of the work role should be similar to the case of the retired. This explanation only makes sense among older individuals. When younger volunteers become unemployed they may decide to increase the hours they volunteer, but as it is unlikely that they will perceive this situation as long lasting, it is plausible that other motivations are responsible, such as resume building (Handy et al., 2010).

Limitations

Despite the fact that these results, gathered from a large sample of volunteers and using a standard index for VRI, clearly indicate life-course variations in VRI salience, the analysis has a number of unavoidable limitations.

The fact that we use cross-sectional data has two consequences. First, we cannot be sure that VRI is the consequence of retirement (or unemployment/disablement). Particularly in the case of retirement it could be that a strong VRI motivates people to retire from paid employment so that they can spend more time on volunteer work. It should be noted, however, that recent longitudinal analysis indicates that the timing of retirement is not affected by volunteer status (Lancee & Radl, 2012). This argument for reverse causality is even less plausible in the case of unemployment or disability.

Second, we argue that VRI salience is the consequence of spending more time in the role. This expectation is supported by role identity theory. But it is also possible that VRI salience encourages people to spend more time volunteering or that there is a reciprocal relation between role enactment and role identity (Charng et al., 1988; Finkelstein, 2008; Güntert & Wehner, 2015; Penner & Finkelstein, 1998; Thoits, 2012). This argument assumes that retirement results in a re-evaluation of the volunteer role, a re-consideration of what volunteer work means now that the retiree no longer has a job to go to each day. This is certainly plausible. Unfortunately, our data did not allow us to disentangle the direction of causality empirically.

A third limitation is that, although we confine the study to volunteers aged 50 or more, it is still possible that age and cohort effects need to be separated. Some have argued that the intensity of volunteering among older adults is not due to life-course changes but to their membership in a “long civic generation” of altruistic and socially responsible people (Rotolo & Wilson, 2004). The same argument could be applied to the strength of VRI. Another reason cohort might make a difference (rather than role absences) is recent changes in the social organization and culture of volunteerism. The volunteer role is becoming less programmed and more transitory as episodic volunteering increases in popularity. In addition, a change in attitudes toward volunteering is encouraging younger adults to take a more instrumental approach to the role, using it primarily to further their own interests (Hustinx, 2010). For either reason, volunteering might not call for the same level of identification with the role.

A fourth limitation is that we are unable to explore how VRI strength might be associated with the kind of activity in which the volunteer is engaged. It could be that older adults prefer volunteer work that entails face-to-face contact with clients: “older people become involved in activities which encourage them to assume responsibilities, such as mentoring, tutoring, teaching, recruiting, providing repairs/maintenance, granting personal support, coaching, counselling, training, placing, supervising, and storytelling” (Principi, Warburton, et al., 2012, p. 96). This kind of work makes the volunteer role highly visible. Combined with age-specific desires to use acquired skills and aptitudes and to be useful and relevant, a more personal relation with clients could foster closer identification with the role.

Finally, we focused entirely on salience. But identity centrality is an equally important topic. We do not know whether unemployment, disability, retirement, or widowhood would rearrange the individual’s priority list of role identities.

Conclusions

The aim of this study was to broaden the discussion of VRI by taking into account possible structural influences on salience. Future surveys of volunteers should consider including both VRI scales and information on other social roles to throw more light on how identity varies across the life course and across social groups. With longitudinal data especially, it would be possible to see if the transition to retirement or being laid off from one’s job or needing to leave the labor force because of disability is followed by a change in identity strength.

These findings could be useful in guiding the management of volunteers since turnover is costly and time-consuming. Volunteer organizations may wish to target older volunteers who have recently retired to suggest ways in which they could increase their contributions to the organization on the understanding that the more they volunteer the stronger their identification with the role. And by encouraging their volunteers to identify more strongly with their roles, they benefit by having more committed workers. Among other things, they could provide more opportunities to engage in desirable and rewarding activities immediately to promote role identity and commitment (Grube & Piliavin, 2000); foster stronger ties with other volunteers (Thoits, 2013); and encourage volunteers to display symbols of their volunteer status, such as wearing appropriately-marked t-shirts or displaying emblematic license-plate holders on their automobiles (Finkelstein et al., 2005).

The volunteer role identity concept is an extremely useful tool for the study of volunteerism. By focusing on how volunteer work can inform people’s self-perceptions, it takes us closer to the actual experience of the volunteer, adding a layer of subjectivity to our understanding of why and how a person volunteers and providing more detail on the benefits and costs of being a volunteer. In addition, social psychologists believe that role identities provide people with a sense of purpose and meaning in life. Much of the research on volunteering assumes that this sense of purpose and meaning is one reason why volunteers enjoy better mental health. The role identity concept adds another layer to this argument by suggesting that a strong identification with the role will result in a stronger sense of purpose and, more psychological benefits. Future research should explore the mediating role of role identity in the connection between volunteer time and psychological well-being.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Author Biographies

Erik van Ingen is associate professor at the Department of Sociology of VU Amsterdam. His research interests include effects of involvement in voluntary associations, personal relations, political socialization, and social and psychological consequences of Internet use.

John Wilson is professor emeritus of sociology at Duke University and the author of a number of articles on volunteering.

References

Arts

Otten

(2013, January). Over 40 percent of employees 65 years or older at retirement. CBS Web Magazine. Retrieved from http://www.cbs.nl/en-GB/menu/themas/arbeid-sociale-zekerheid/publicaties/artikelen/archief/2013/2013-3769-wm.htm?Languageswitch=on

Bulanda

J. R.

Jendrek

M. P.

(2016). Grandparenting roles and volunteer activity. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 71(1), 129-140.

Butricia

Johnson

Zedlewski

(2009). Volunteer dynamics of older Americans. The Journals of Gerontology. Series B, Psychological Sciences and Social Sciences, 64B, 644-655.

Callero

Howard

Piliavin

(1987). Helping behavior as role behavior: Disclosing social structure and history in the analysis of prosocial action. Social Psychology Quarterly, 50, 247-256.

CBS Statline. (2016). Bevolking; geslacht, leeftijd en burgerlijke staat [Population; gender, age and marital status]. Retrieved from http://statline.cbs.nl/Statweb/publication/?DM=SLNL&;PA=7461bev&D1=-a&D2=0&D3=121–130&D4=61&HDR=T&STB=G2,G1,G3&VW=T

Chambre

(1984). Is volunteering a substitute for role loss in old age? An empirical test of activity theory. The Gerontologist, 24, 292-298.

Charng

Piliavin

Callero

(1988). Role identity and reasoned action in the prediction of repeated behaviors. Social Psychology Quarterly, 51, 303-317.

Choi

(2003). Factors affecting volunteerism among older adults. The Journal of Applied Gerontology, 22, 179-196.

Donnelly

Hinterlong

(2010). Changes in social participation and volunteer activity among recently widowed older adults. The Gerontologist, 50, 158-169.

10.

Dury

De Donder

De Witte

Buffel

Jacquet

Verte

(2015). To volunteer or not: The influence of individual characteristics, resources, and social factors on the likelihood of volunteering by older adults. Nonprofit and Voluntary Sector Quarterly, 44, 1107-1128.

11.

Einolf

C. J.

(2009). Will the boomers volunteering during retirement? Comparting the baby boom, silent, and long civic cohorts. Nonprofit and Voluntary Sector Quarterly, 38(2), 181-199.

12.

Erlinghagen

(2010). Volunteering after retirement: Evidence from German panel data. European Societies, 12, 603-625.

13.

Finkelstein

(2008). Predictors of volunteer time: The changing contributions of motive fulfillment and role identity. Social Behavior and Personality, 36, 1353-1364.

14.

Finkelstein

Penner

Brannick

(2005). Motive, role identity, and prosocial personality as predictors of volunteer activity. Social Behavior and Personality, 33, 403-418.

15.

Greenfield

Marks

(2004). Formal volunteering as a protective factor for older adults’ psychological well-being. The Journals of Gerontology. Series B, Psychological Sciences and Social Sciences, 59, S258-S264.

16.

Griffin

Hesketh

(2008). Post-retirement work: The individual determinants of paid and volunteer work. Journal of Occupational and Organizational Psychology, 81, 101-121.

17.

Gronlund

(2011). Identity and volunteering intertwined: Reflections on the values of young adults. Voluntas, 22, 852-874.

18.

Grube

Piliavin

(2000). Role identity, organizational experiences, and volunteer performance. Personality and Social Psychology Bulletin, 26, 1108-1119.

19.

Güntert

Wehner

(2015). The impact of self-determined motivation on volunteer role identities: A cross-lagged panel study. Personality and Individual Differences, 78, 14-18.

20.

Handy

Cnaan

R. A.

Hustinx

Kang

Brudney

J. L.

Haski-Leventhal

Zrinscak

(2010). A cross-cultural examination of student volunteering: Is it all about resume building? Nonprofit and Voluntary Sector Quarterly, 39(3), 498-523.

21.

Hustinx

(2010). Institutionally individualized volunteering: Toward a late modern reconstruction. Journal of Civil Society, 6, 165-179.

22.

Jensen

Lamura

Principi

(2014). Volunteering in older age: A conceptual and analytical framework. In Principi

Jensen

Lamura

(Eds.), Active ageing: Voluntary work by older people in Europe (pp. 1-23). Bristol, UK: Policy Press Scholarship Online. doi:10.1332/policypress/9781447307204.001.0001

23.

Knoef

De Vos

(2009). The representativeness of LISS, an online probability panel. Retrieved from http://www.lissdata.nl/assets/uploaded/file/paper_knoef_devos_website.pdf

24.

Lancee

Radl

(2012). Social connectedness and the transition from work to retirement. The Journals of Gerontology. Series B, Psychological Sciences and Social Sciences, 67, 481-490.

25.

Lancee

Radl

(2014). Volunteering over the life course. Social Forces, 93(2), 833-862.

26.

Leenheer

Scherpenzeel

A. C.

(2013). Does it pay off to include non-Internet households in an Internet panel? International Journal of Internet Science, 8, 17-29.

27.

(2007). Recovering from spousal bereavement in later life: Does volunteer participation play a role? The Journals of Gerontology. Series B, Psychological Sciences and Social Sciences, 62, S257-S266.

28.

Marta

Manzi

Pozzi

Vignoles

(2014). Identity and the theory of planned behavior: Predicting maintenance of volunteering after three years. The Journal of Social Psychology, 154, 198-207.

29.

McNamara

Gonzales

(2011). Volunteer transitions among older adults: The role of human, social, and cultural capital in later life. The Journals of Gerontology. Series B, Psychological Sciences and Social Sciences, 66, 490-501.

30.

Moen

Erickson

Dempster-McClain

(2000). Social role identities among older adults in a continuing care retirement community. Research on Aging, 22, 559-579.

31.

Moen

Flood

(2013). Limited engagements? Women’s and men’s work/volunteer time in the encore life course stage. Social Problems, 60, 206-233.

32.

Mutchler

Burr

Caro

(2003). From paid worker to volunteer: Leaving the paid workforce and volunteering in later life. Social Forces, 81, 1267-1293.

33.

Narushima

(2005). “Payback time:” Community volunteering among older adults as a transformative mechanism. Ageing & Society, 25, 567-584.

34.

Nesbit

(2013). The influence of family and household members on individual volunteer choices. Nonprofit and Voluntary Sector Quarterly, 42, 1134-1154.

35.

Oakley

Pratt

(1997). Voluntary work in the lives of post-retirement adults. British Journal of Occupational Therapy, 60, 273-276.

36.

Organisation for Economic Co-operation and Development. (2014). The labour market situation of older workers in the Netherlands. Paris, France: Author.

37.

Pavlova

Silbereisen

(2012). Participation in voluntary organizations and volunteer work as a compensation for the absence of work or partnership? Evidence from two German samples of younger and older adults. The Journals of Gerontology. Series B, Psychological Sciences and Social Sciences, 67, 514-524.

38.

Penner

Finkelstein

(1998). Dispositional and structural determinants of volunteerism. Journal of Personality and Social Psychology, 74, 525-537.

39.

Piliavin

Grube

Callero

(2002). Role as resource for action in public service. Journal of Social Issues, 58, 469-485.

40.

Principi

Warburton

Schippers

Di Rosa

(2012). The role of work status on European older volunteers’ motivation. Research on Aging, 35, 710-735.

41.

Rotolo

Wilson

(2004). What happened to the “long civic generation?” Explaining cohort differences in volunteerism. Social Forces, 82, 1091-1121.

42.

Schippers

Conen

(2014). Older volunteers in the Netherlands. In Principi

Jensen

Lamura

(Eds.), Active ageing: Voluntary work by older people in Europe (pp. 1-23). Bristol, UK: Policy Press Scholarship Online. doi:10.1332/policypress/9781447307204.003.0010

43.

Sherman

Shavit

(2012). How the lifecycle hypothesis explains volunteering during retirement. Ageing & Society, 32, 1360-1381.

44.

Smith

D. B.

(2004). Volunteering in retirement: Perceptions of midlife workers. Nonprofit and Voluntary Sector Quarterly, 33(1), 55-73.

45.

Smith

(2010). Volunteer patterns of mid- and later life American couples. Journal of Sociology & Social Welfare, 37, 131-154.

46.

StataCorp. (2013). Stata user’s guide—Release 13. College Station, TX: Stata Press.

47.

Stets

Burke

(2005). A sociological approach to self and identity. In Leary

Tangney

J. P.

(Eds.), Handbook of self and identity (pp. 128-152). New York, NY: Guilford Press.

48.

Tang

Burr

(2015). Revisiting the pathways to retirement: A latent structure model of the dynamics of transition from work to retirement. Ageing & Society, 35, 1739-1770.

49.

Tang

Choi

Morrow-Howell

(2010). Organizational support and volunteering benefits for older adults. The Gerontologist, 50, 6003-6612.

50.

Teske

(1997). Political activists in America: The identity construction model of political participation. Cambridge, UK: Cambridge University Press.

51.

Thoits

(2011). Mechanisms linking social ties and support to physical and mental health. Journal of Health and Social Behavior, 52, 145-161.

52.

Thoits

(2012). Role-identity salience, purpose and meaning in life, and well-being among volunteers. Social Psychology Quarterly, 75, 360-384.

53.

Thoits

(2013). Volunteer identity salience, role enactment, and well-being: Comparisons of three salience constructs. Social Psychology Quarterly, 76, 373-398.

54.

Van den Bogaard

Henkens

Kalmijn

(2014). So now what? Effects of retirement on civic engagement. Ageing & Society, 34, 1170-1192.

55.

Van Groenou

M. B.

Van Tilburg

(2012). Six-year follow-up on volunteering in later life: A cohort comparison in the Netherlands. European Sociological Review, 28, 1-11.

56.

van Ingen

Dekker

(2011). Changes in the determinants of volunteering: Participation and time investment in the case of the Netherlands between 1975 and 2005. Nonprofit and Voluntary Sector Quarterly, 40, 682-702.

57.

Warburton

Le Broque

Rosenman

(1998). Older people—The reserve army of volunteers? An analysis of volunteerism among older Australians. International Journal of Aging & Human Development, 46, 229-246.

58.

Wilson

(2012). Volunteerism research: A review essay. Nonprofit and Voluntary Sector Quarterly, 41, 176-212.