JOHN WILSON [*]
MARC MUSICK [**]
To most people, a "volunteer" is someone who contributes time to helping others with no expectation of pay or other material benefit to herself. However, this does not mean that volunteer work is of no consequence for the volunteer. Indeed, it is widely believed that helping others is as beneficial for the donor as it is for the recipient. "Research studies show that most people do in fact hold the belief that helping others is a good way to gain fulfillment for yourself."  In this article, we review some of the research on the supposed benefits of volunteering and describe briefly some of the results of our own work in this area. We first examine the contribution volunteering is thought to make to a society's social capital, its supply of the generalized trust and norms of reciprocity that make democratic politics possible. Are volunteers more civic minded and more likely to take an active role in political life? Next, we examine the possible link between volunteering and "leading the good life." Are vol unteers less likely to engage in anti-social behavior? We then consider the contribution volunteering might make to both physical and mental health. Is there any evidence to suggest that volunteering can make people healthier or contribute positive feelings of well-being? Finally, we examine the contribution volunteering makes to occupational achievement. Is there any empirical evidence to support the notion that volunteering is either a direct path to good jobs or indirectly provides the self-confidence and skills needed to secure good jobs or to do well in the jobs we have?
For a number of reasons, the possibility that volunteering is useful for building and maintaining civil society--a sphere of activity where people feel free to organize groups, engage in public debate, and in which norms of mutual respect and toleration protect the voices of majority and minority alike--has recently been undergoing fresh scrutiny. Part of this renewed interest stems from the collapse of the Soviet Union and the task faced by newly democratizing countries trying to build, or re-build, the infrastructure of participatory politics. How do people acquire the skills and aptitudes necessary for the give-and-take of democratic government? Some of this renewed interest also stems from concerns expressed in established democracies that fewer and fewer people are taking the time to vote, run for local office, or support political organizations with their time and money. Ever since de Tocqueville's mid-nineteenth-century analysis of American democracy, it has been assumed that a healthy voluntary secto r is vital to the survival of democratic politics.  De Tocqueville believed that voluntary associations were essential intermediary bodies between the mass of individuals and their institutions of government.  Active membership in a voluntary association created the generalized trust--a trust that extends beyond the boundaries of kinship and friendship--on which democratic political life depends. If the habit of "joining" were allowed to die, so too would democracy.
Robert Putnam, who found substantial variation in the performance of Italian governmental institutions across different regions, has recently revived this idea.  His explanation for this variation centered on the concept of social capital. Putnam defined social capital as "features of social organization, such as trust, norms, and networks, that can improve the efficiency of society by facilitating coordinated actions."  His theory, redolent of de Tocqueville, is that active membership in voluntary associations generates the trust necessary for people to organize effectively and act collectively. The associations need not be political to have this beneficial effect, although those that were organized "horizontally," where members could easily participate in running the organization, would be more likely to produce it.
The relevance of this research for a contemporary study of the effects of volunteering is clear. Much volunteering consists of active participation in the voluntary associations that Putnam describes. To be sure, the correspondence is less than perfect. Active participation in a voluntary association might simply mean that the individual is an avid consumer of the association's collective goods without helping to produce them. However, Putnam's theory would be more applicable to the organizational volunteer. Recent research by Sidney Verba et al., using a representative sample of the U.S. population, supports this argument.  It is those "joiners" who use their time in the organization to develop civic skills, such as the ability to organize a meeting, who are most likely to be active in the wider political sphere. Organizing monthly committee meetings at the local Kiwanis chapter, setting up a food pantry at church, or chairing community charity drives are all activities that develop the civic skills that increase one's sense of political efficacy and competence. David Knoke reaches the same conclusion from data supplied by the General Social Survey special 1987 Module on Social Participation.  He finds that being highly active inside a voluntary organization--serving on a committee, serving as an officer, or attending conferences and workshops--is positively related to being active in local politics.  These volunteer activities, more than simple membership, help explain why age, education, income, and being black are positively related to political activity levels. Knoke attributes the positive effect of organizational volunteering to an increase in the likelihood of having political discussions with other association members.  Thus, while "active participation" need not mean volunteering in an altruistic sense of the term, it certainly should be considered a form of volunteering if, as the above examples indicate, it means "taking your turn" at administering in the association to which you belong. If this kind of activity should therefore be considered volunteering, there is good reason to believe it has a positive effect on political behavior.
Having argued that active participation in voluntary associations should, in most instances, be treated as a form of volunteering, we would not want to extend this argument to every kind of participation in every kind of association. Doing volunteer work for an organization need not encourage people to become involved in the wider world of politics. For example, Putnam believes that' the strong link between civic engagement and social trust explains the contribution volunteering makes to political participation.  However, there is no consensus as to whether volunteering actually does increase people's trust in others or in public institutions. Dietland Stolie's recent survey of the members of voluntary associations in Germany and Sweden finds no evidence of a linear relationship between the time individuals spend in an association and their level of interpersonal trust.  Instead, it appears that joining an association initially boosts trust, but the effect subsides for members who stay in the group fo r long periods. Stolle hypothesizes that when members have strong bonds to the other members of their association, measured by their trust in the other members, their trust will not generalize to others.  Therefore, it appears that when people are loosely bound to an association, their trust in their fellow members generalizes, but when they are tightly bound, they are more likely to trust only their fellow members.
Stolle and Thomas Rochon analyze data from three nationally representative surveys conducted in the United States in 1983, 1984, and 1986, data from the Swedish Citizen's Survey conducted in 1987, data from the German Allbus survey conducted in 1991, and data from the World Values Survey conducted in 1983 and 1990. They aggregate memberships in forty-three types of organizations into seven associational sectors.  Members of political, economic, and community groups are heavily involved in political and community participation.  Cultural, personal interest, and community organizations are particularly likely to have members scoring high in generalized trust and in reciprocity with neighbors.  Members of cultural, personal interest, and economic groups, but not political groups, are especially high in political trust and efficacy.  In short, the capacity of voluntary associations to produce social capital--for example, generalized trust--is contingent on the mission of the association and the k ind of volunteer work it expects its members to do. Taken together, these studies suggest caution when generalizing about the "benefits" of volunteering as far as democratic action is concerned. If an abundant supply of social capital is believed to be a necessary condition for democratic politics, then volunteering can certainly help supply it, but not all kinds of volunteering do it equally well.
Failure to distinguish different kinds of volunteering might help explain why surveys do not consistently show a link between volunteering and trust. Brown uses data from the 1995 Independent Sector Survey of Giving and Volunteering to show that volunteers score higher on trust measures than do non-volunteers.  It cannot be determined from these cross-sectional data whether trusting people are more likely to volunteer or whether volunteering makes people more trusting. Nor does Brown include an analysis of the impact of trust on political activities. On the other hand, another major survey of volunteering finds no relationship between volunteering and trust.  Peter Hall reports the same null finding in his analysis of voluntary association memberships in Britain.  Although membership in formal associations shows a statistically strong relationship to political activism, it is not strongly correlated with levels of political trust. It is not clear whether Hall's null finding would hold if he had m easured organizational volunteering rather than membership. As argued above, this does not rule out the possibility that trust can link volunteering to civic-minded behavior, because not all volunteering is the same. Some volunteer activities might actually inhibit civic action because they are inward looking, building trust with fellow volunteers but not with other people or public institutions. There is also a strong likelihood that some kinds of volunteer work, especially that which is contributed to political groups and advocacy groups such as those concerned with environmental issues, actually contribute to people's lack of trust in other peoples, elected officials. The less you trust other people, the more motivated you are to be active yourself.
De Tocqueville and Putnam draw our attention to the possibility of a connection between widespread volunteering and a thriving civil society. It is but a small step from this idea to the notion that instilling appropriate democratic values and inculcating good democratic habits--for example, voting--might be fostered by getting young people involved in volunteer work early in their lives. Do young people who volunteer exhibit the kinds of attitudes and values that might conduce to active participation in politics later in life? Is there any connection between adolescent volunteering and political behavior in middle age?
One way to answer these questions in quasi-experimental form is to measure people's attitudes before and after their volunteer experience. Stephen Hamilton and Mickey Feuzel administered a "Social and Personal Responsibility Scale" to adolescents aged between eleven and seventeen (n=44) before and after their involvement in either community improvement projects or childcare tasks (for example, summer camp).  The scale measures the extent to which the respondent expresses responsible attitudes toward social welfare and personal duty.  The volunteers recorded a slight gain in attitudes toward society's obligation to meet the needs of others, but there were no significant gains in their own sense of personal duty to help meet the needs of others.  Because the study did not include a control group of non-volunteers, it is difficult to conclude that these changes were the result of volunteering rather than general maturation.
Adolescents also have been the target of researchers interested in finding ways to encourage participation in politics in those European countries undergoing the transition to democracy. Getting youth interested in volunteer work has been promoted as a way of teaching new attitudes and behavior. A 1995 survey of twelve- to eighteen-year-olds in three stable and four transitional democracies provides data on this subject.  Five hundred students were sampled from each country.  The authors were interested in "civic commitment," defined as the importance adolescents attached to public interest as a personal life goal.  The most consistent and robust finding was that a family ethic of social responsibility--for example, children reporting that their parents stressed the need to be attentive to the needs of others--has a positive effect on civic commitment.  The findings on the impact of volunteering were more uneven. Adolescents' engagement in volunteer work was significantly related to boys' civ ic commitment in the United States, Sweden, and the Czech Republic, and to girls' civic commitment in the Czech Republic, Bulgaria, and Russia.  The authors provided no explanation for the interactions between gender and country. Nor is it possible to attribute causality in this cross-sectional study: Even in those countries where some relationship was found, it remains an open question whether adolescents are civic minded because they have volunteered or they volunteer because they are civic minded. One implication of this research is that role modeling by parents and what they teach their children about their social obligations help account both for a higher rate of volunteering and more positive attitudes about one's obligations to one's community. Other studies of the effect of volunteering on political attitudes should be aware of the selection bias induced by failure to control for these family effects.
Data from Monitoring the Future, a periodic survey of high school seniors in the United States, suggest that volunteering can, promote not only favorable attitudes but also good citizenship practices.  Students who volunteered were more likely to be engaged in a variety of conventional political behaviors, such as voting and working on a political campaign, than those who did not volunteer.  Once again, these are cross-sectional data, and it is not possible to conclude from these correlations alone that volunteering causes subsequent political behavior. It is also far from certain that volunteering is unique in providing these benefits. Almost any kind of organizational participation while in high school could have the same consequences. 
A study of the impact of volunteering on subsequent political behavior reported by Margaret Conway et al. does use longitudinal data, although the measurement of volunteering is less than satisfactory for our purposes.  The authors used the National Longitudinal Study of the High School Class of 1972 to examine the impact of 1972 activities on 1974 and 1976 beliefs and practices. To measure high school activities, they used a question asking about membership and level of activity in a variety of student organizations.  They found that the more the students volunteered in 1972, the more likely they were to espouse democratic beliefs in 1974 (for example, value attached to voting).  Not surprisingly, they found that participation in student government or media-related groups was more strongly related to positive beliefs than participation in athletic clubs or expressive clubs.  The researchers then used 1972 volunteering and 1974 beliefs to predict community participation--a measure that includ ed volunteering--in 1976. They found a small positive effect of 1974 beliefs but a stronger effect of 1972 volunteering.  Interestingly, all types of groups, except vocational, contributed to community participation in 1976, including expressive and athletic clubs. Because these analyses control for a number of variables that might be producing spurious effects (for example, socioeconomic status), the findings provide convincing evidence that early volunteering increases the likelihood that young people will become active members of their political community when they graduate from college. One implication of these findings is that the social circumstances that inhibit children's participation in high school extra-curricular activities might also eventually inhibit their involvement in the political process and the volunteer labor force. Thus, the socioeconomic status of a child can have a direct effect on whether he or she will be later engaged in civic life. It also has an indirect effect because it hel ps to determine whether he or she participates in high school extra-curricular activities. The same could be said for race. One reason African-Americans are less involved in the political process is that they are discouraged from getting involved as a volunteer in their high school years. 
Studies of the long-term effect of early volunteer work on citizenship behavior are rare. Fendrich contacted alumni of Florida A & M and Florida State Universities at intervals of ten and twenty-five years after graduation.  Students who had been active in volunteer organizations and social movements while in college belonged to more civic organizations and were more politically engaged than non-activists a decade later.  Twenty-five years later, however, these differences had disappeared, at least for the primarily black alumni of Florida A & M University, although they persisted among the mostly white alumni of Florida State.  In another study, Douglas McAdam found that volunteers for the Mississippi Freedom Summer civil rights campaign remained more politically active in their adult life than those who applied to join but did not participate.  Both studies exhibit considerable ingenuity in data gathering and analysis. Their longitudinal design instills confidence in the argument that early volunteering causes later political participation. However, neither study can entirely rule out the selection bias created by the likelihood that young people who choose to engage in high risk activism have also been socialized by their parents to be civic-minded and active citizens, leaving open the possibility that the link between volunteering and later political activity is spurious.
It is part of folk-wisdom that volunteering "builds character" and helps "keep kids out of trouble." Volunteering is believed to foster interpersonal trust, toleration and empathy for others, and respect for the common good. This, in turn, makes us less likely to engage in socially pathological behavior, such as vandalism, and less likely to prey on other people and engage in self-destructive behavior, or so the reasoning goes. Empirical research on this topic is not very abundant, and its results are somewhat inconsistent, leaving more questions open than answered.
Using data from the National Longitudinal Survey of Youth, Daniel Hart and Robert Atkins found that adolescents involved in voluntary service were less likely to be involved in five of seventeen "problem behaviors" in the same year.  However, cross-sectional analyses such as these do not indicate whether more socially conforming students are more likely to volunteer or whether volunteering makes students more likely to conform. Longitudinal data from the Michigan Study of Adolescent Life Transitions Survey suggest that volunteer work might help prevent delinquency among younger people.  In this survey, the mostly white tenth graders were re-surveyed two years later. Students who had undertaken volunteer work in the interim were less likely to have engaged in risky behavior such as skipping school and using drugs.  On the other hand, involvement in social clubs had no effect on risky behavior, while participation in team sports increased the likelihood. 
Joseph Allen et al. compared one group of high school students who were participating in Teen Outreach with a control group who did not participate. 
Teen Outreach is a school-based program that involves young people in volunteer services in their communities, linking the work to classroom-based, curriculum-guided group discussions. The authors found that "Teen Outreach students significantly decreased their levels of [getting pregnant, failing a course, being suspended from school] relative to the comparison group from program entry to program exit [a year later]."  They caution that unmeasured characteristics might bias the results because students self-select into the Teen Outreach program.  Nevertheless, these findings are robust in analyses controlling for students' grade level, gender, race, entry problem behaviors, parents' level of education, and household composition. It is worth noting that the number of hours volunteered made no difference to the likelihood of engaging in problem behaviors and that the only volunteer work that had this beneficial effect was work students selected and found enjoyable, and which challenged them to think ab out future goals and taught them new skills.  The lesson from this longitudinal study is clear. Volunteer activities must be carefully designed and monitored if they are to have their desired consequence. Programs geared toward encouraging teenagers to volunteer in the expectation this might keep them out of trouble will not work if this encouragement seems to limit the voluntary nature of the work.
The Allen et al. study uses a pre/post-test design and a comparison group to determine whether the improvements were confined to volunteers.  However, the period covered by the study (one year) is perhaps too brief for the benefits of volunteering to appear. Christopher Uggen and Jennifer Janikula used data from the first eight waves of the Youth Development Study (1988-95) to examine the impact of volunteering on the likelihood of being arrested over a longer period.  They used retrospective self-reports from the years 1991-95 to measure timing of first arrest.  Volunteering was measured when respondents were in the eleventh and twelfth grades. Only three percent of the volunteers were arrested in the four years following high school compared to eleven percent of the non-volunteers.  The risk of arrest for both groups rose until age nineteen and then began to decline, but it rose faster and peaked higher for the non-volunteers.  This difference remained even after volunteering prior to e leventh grade was controlled to reduce selection bias. 
In short, these studies indicate that volunteering can inhibit anti-social behavior, albeit to a modest degree. A number of issues remain to be resolved. It is far from clear why volunteering keeps young people out of trouble. Social control theory would lead us to expect that volunteering reduces delinquency because it exposes young people to informal social controls and supervision.  Differential association, or social learning, theory would lead us to expect that volunteering reduces delinquency because it inhibits social contacts with law violators.  It might also be the case that volunteering teaches young people pro-social values that impose normative constraints on delinquent behavior.
Recent advances in medical sociology have drawn attention to the physical health benefits of social integration and social support.  By helping others, individuals may develop stronger networks that buffer stress and reduce disease risk. Additionally, "the altruistic features of volunteerism might reduce destructive levels of self-absorption."  It is possible that the association between volunteering and health is due to the fact that volunteers have access to more information about the benefits of exercise and preventive medical care. If volunteering does enable people to share in the benefits that accrue to socially-engaged persons, it should be especially evident in the later stages of life, when physical health problems are more likely to occur and when, perhaps, other forms of social integration, such as work and marriage, are no longer available.
A number of recent studies give clear indications that there are physical health benefits to volunteering. Because they all use longitudinal data--and some use mortality as the outcome variable--they permit the conclusion that volunteering is causing good health, rather than that healthy people are volunteering more. Phyllis Moen et al. used two-wave data, 1956 and 1986, to examine the effect of volunteering on the onset of serious illness and functional disability.  Women who volunteered on an intermittent basis between the two measurement intervals scored higher on functional ability in 1986 than those who had not.  Edward Sabin used data on respondents aged seventy years of age or older from the Longitudinal Study of Aging to estimate the effects of volunteering on mortality.  About sixteen percent of the respondents had volunteered in the first wave of the research in 1984.  Respondents in this subgroup were less likely to have died by 1988, controlling for physical health in 1984, than t he majority who did not volunteer.  The positive effect of volunteering was stronger for those elderly who reported being in good health in 1984,  suggesting that volunteer work could help maintain good health but not improve the condition of those whose health is poor.
Richard Rogers tracked participants in the National Health Interview Survey, all of whom were at least fifty-five years of age when the survey began, using mortality data from 1991.  Of those still alive in 1991, 21.5% had been volunteers in 1983; of whose who had died by 1991, twelve percent had been volunteers.  Volunteers were less likely to die than non-volunteers, regardless of level of church attendance, age, marital status, education, or gender, although it should be noted that the prophylactic effect of attending shows, movies, or concerts was somewhat stronger.  Unfortunately, the study contained no data on the length of volunteer experience.
Marc Musick et al. also used longitudinal data to estimate the effect of volunteering on mortality.  They selected a subgroup of respondents to the 1986 Americans' Changing Lives survey who were sixty-five years old or older in the first wave.  They measured how many different volunteer activities each respondent reported in the first wave. Using proportional hazards models, they then estimated the effect of volunteering on the risk of dying by 1994.  Volunteers had a lower mortality hazard than did non-volunteers, even after adjustments for socio-economic status, age, gender, and race.  They found the impact of volunteering to be curvilinear: The group with the lowest risk of dying was the moderate volunteers, those who volunteered for just one group.  They also found interaction effects: Volunteering was most beneficial to those who reported low levels of informal social interaction. 
Oman et al. used data on volunteering from four age groups (fifty-five to sixty-four, sixty-five to seventy-four, seventy-five to eighty-four, and eighty-five and older) to predict mortality five years later. They distinguished highly active, moderately active, and inactive volunteers, using both a measure of the number of organizations for which respondents volunteered and the hours volunteered in an average week. Unlike Musick et al.,  they found mortality rates to be lower for highly involved, but not moderately-involved volunteers.  Similar, but smaller, linear positive effects were found for volunteer hours.  The researchers also discovered that the beneficial effects of volunteering were stronger among volunteers for religious organizations.  Interestingly, Oman et al.'s findings on the moderating effects of social integration also seem to contradict those of the Musick et al. study. While the latter suggests that volunteering substitutes for lack of social interaction in producing heal th benefits,  the Oman et al. study finds that volunteerism is more protective at the higher levels of most of their social integration measures (that is, out of the house seven days a week, living with others, having close friends, and attending religious services),  suggesting a complementary role for volunteering. In the case of participation in organized social activities, however, volunteerism had its most beneficial effect among those reporting fewest activities,  suggesting a substitute role similar to that found by Musick et al. 
These various studies of the effect of volunteering on mortality using different data sets based on good random samples of the general population and employing longitudinal data all point in the same direction--that volunteer work does help people live longer. Of course, these are aggregate patterns and cannot be taken as a prescription for longer life. Researchers are aware of the possibility of significant moderating effects such that the benefits might be stronger for certain groups and for certain kinds of volunteering. Most of the remaining work must address the issue of mediating variables or mechanisms. What are the social factors that link volunteer work to reduced mortality? Among the most important of these are social psychological attributes, such as self-efficacy. That is, volunteering lowers the risk of physical ill-health because it boosts the social psychological factors that healthy people have. This topic is the subject of the next section.
VOLUNTEERING AND MENTAL HEALTH
There is a long sociological tradition linking social integration with the individual's subjective evaluation of his or her personal well-being. Social integration signifies the extent to which an individual is connected to other people. One measure of social integration is the number of social roles--mother, worker, sibling, volunteer--being performed at the same time. Another measure is the number of "ties" or contacts with other people. Panel data analyzed by Peggy Thoits shows quite clearly that the acquisition or loss of social roles affects psychological distress in both men and women.  Because volunteering adds a social role to one's life, it should yield mental health benefits.
Unfortunately, past assessments of the psychological benefits of volunteering are inconclusive, largely because they are based on volunteers' own estimations. For example, S. Mark Pancer and Michael Pratt conclude that Canadian youths gain a great deal from volunteering because "[t]hey report high levels of satisfaction with their activities; they indicate that they have acquired new skills and knowledge; and many have begun to think of themselves as lifelong volunteers."  The researchers overlook the possibility that youth with more negative experiences will have dropped out of the sample and that cognitive dissonance could be distorting the responses of the volunteers. Likewise, they do not consider that "new skills and knowledge" could have been gained by other means, such as simply growing older and gaining life experience. Likewise, Nancy Morrow-Howell et al. were commissioned by OASIS, a U.S. nonprofit agency designed to promote volunteering among the elderly, to find out what kinds of benefits vol unteers thought they had derived from volunteer work.  Nearly eight out of ten reported an increase in well-being, the same proportion reported having more friends and acquaintances, and three-quarters believed they now made more productive use of their time.  However, cross-sectional designs that use participants to self-assess the impact of a volunteer program function as little more than market research for the agency concerned. Without a pre/post-test design and a control group, and without more objective and generalizable outcome measures, little can be learned of the benefits of volunteering from these studies.
There are a number of reasons why volunteering should have positive effects on mental health. First, volunteering is a form of social activity or social participation. It is well-established that social integration--"the quantity of social ties or relationships" a person has--yields positive mental health effects.  Social isolation--"the sense of not having anyone who is someone to you and not being someone to anyone" --is depressing. Social integration seems to explain why married people tend to report better mental health than do single people and why divorce or widowhood can have deleterious mental health effects.88
Another reason why there might be a connection between volunteering and mental health is that volunteer work is a form of helping behavior. Providing help, even to a generalized other, as in formal volunteering, is a self-validating experience. In certain settings, it can foster trust and intimacy and encourage the provider to anticipate that reciprocal help will be forthcoming when it is needed. 
A final reason why volunteering might increase subjective well-being is that it fosters a belief in the individual that he or she can make a difference and thus enhances personal efficacy. Volunteering provides a sense of control over one's life and one's environment, thereby alleviating depression.  Researchers argue that one reason volunteering lowers the risk of mortality is that it increases perceived self-efficacy, self-esteem, and positive affect, which in turn function as a kind of stress inoculation. 
As is the case with physical health, it is unlikely that the effect of volunteering on mental health is uniform across all populations and at all levels and on all kinds of volunteering. A number of refinements to this broad thesis are therefore necessary.
First, the effect of volunteering on mental health might be non-linear. For example, there may be either few gains to volunteering more hours after a small amount is performed, or possibly even a decline in benefits if the volunteer becomes overburdened. Recall that in Marc Musick et al.'s mortality study, the benefits of volunteering were confined to those who contributed fewer than forty hours each year.  In a study of hospital volunteers aged over sixty-two,  those who worked 500 hours each year scored higher on a "contentment" scale than either those who worked fewer than 500 and those who worked more.  Too much volunteering, it seems, can cause role strain and reduce subjective well-being.
Second, the nature of the volunteer work might moderate the effect of volunteering on mental health. As Morris Okun et al. note, it is unlikely that social activities such as volunteering will make much difference to a general measure of subjective well-being because "well-being" describes a wide range of psychological states that are slow to change.  They find that social activity accounts only for somewhere between one and nine percent of the variance in subjective well-being explained in the studies they review.  They suggest that more attention should be paid to identifying variables that interact with social activity to determine subjective well-being.  Another line of investigation is to distinguish the effects of different types of volunteering. Some volunteering can be highly stressful while other kinds can be restorative. Recall that Oman et al. found the positive effects of volunteering to be strongest for those whose work was church-related.  This makes sense in light of the positi ve effects of religion on health.
Third, the effect of volunteering might be modified by life-cycle stages.  It is an axiom of life course research that roles have different impacts at different stages of the life course. Volunteering is believed to assume an especially important role among the elderly, particularly if they are retired, because it can "inoculate, or protect, [them] . . . from the hazards of retirement, physical decline, and inactivity."  For the retired elderly in particular, volunteering is "serious leisure," a high-investment activity calling for time discipline and even the occasional sacrifice.  Volunteering combines the freedom of a voluntarily chosen pursuit with the commitment of social obligation.  Another reason to expect volunteer work to help the old more than the young is that volunteering is a socially approved role, and its positive effects might be especially beneficial for those who, because of age, have experienced role loss through retirement or widowhood. In a meta-analysis by Judith Whe eler et al. of the effects of volunteering on elderly populations, a significant positive relation between volunteering and life-satisfaction was found, even after adjustments for socioeconomic status and physical health.  The elderly are also the subject of the research reported by Frank Young and Nancy Glasgow.  They use data from a random sample of 629 people aged sixty years or more who were asked about their volunteer activities and asked to rate their own health on a scale of "very poor" to "excellent."  Volunteer activities were not separately analyzed in this research but grouped together with club memberships and participation in local politics as a form of "instrumental" participation.  Religious volunteering was treated separately, as was involvement in more expressive associations such as hobby clubs.  Young & Glasgow found that instrumental participation was positively associated with self-reported health. However, religious volunteering had no effect, and expressive part icipation benefited only women. 
The impact of volunteering on the well-being of the elderly is still unclear. Another study, using a random sample of the U.S. population who were sixty or older at the time of the survey, found that volunteering did not exert a positive effect on the well-being of older people, although providing help informally did.  It is significant that most of the studies described here are cross-sectional, that is, they do not answer the question whether volunteering increases well-being over time.  Finally, because they all deal exclusively with elderly populations, it is not possible to examine the moderating effects of age because no comparison with the effects in younger populations has been made.
A rare study of the moderating effect of race on the connection between volunteering and health is reported by Barbara McIntosh and Nicholas Danigelis.  use the Americans' Changing Lives survey, which over sampled black Americans, to determine whether the effects of volunteering on subjective well-being vary by race. They, too, confine their study to the elderly population, aged sixty and older.  They examine the effect of volunteering on two dependent variables: positive affect (for example, "enjoy life") and negative affect (for example, "feel sad").  They distinguish formal/religious, formal/secular, and informal types of volunteering. Religious volunteering enhances positive affect only for the white sub-sample and reduces negative affect only for black females.  Secular volunteering enhances positive affect only among white males and reduces negative affect only among black males.  Informal volunteering enhances positive affect only among white females and reduces negative affec t only among black males and white females.  The authors speculate that the effect of religious volunteering on black women's negative affect could be attributable to the greater importance they attach to their church activities.  They note that, as far as secular volunteering is concerned, white men feel good and black men feel less bad.  It is possible to hypothesize why elderly men, in particular, might benefit more from secular volunteering as they suffer more than women from role loss at that stage of life, but the race differences are difficult to interpret. Informal volunteering has beneficial effects for white women, regardless of the affect measure, but it is not clear why this gender difference is reversed in the case of blacks, where only men are affected. Although the details of this cross-sectional research are sometimes difficult to interpret, the overall thrust of the findings is that the effect of volunteering on well-being probably depends quite a lot on the type of volunteer w ork being done and the population group doing it.
Further exploration of the possible connection between volunteering and mental health would seem to require six things: (1) it be conducted on a nationally representative sample of the entire adult population; (2) it use longitudinal data to measure mental health both before and after volunteering; (3) it control for possible spurious effects; (4) it test for curvilinear effects to see if too much volunteering has negative consequences; (5) it compare the effects of volunteering on younger and older populations; and (6) it examine interaction effects to see if religious volunteering is more beneficial than volunteering in other areas.
The Americans' Changing Lives data set provides an excellent opportunity to examine these issues.  Collected by the Survey Research Center at the University of Michigan, it is a three-wave survey based on a stratified, multistage, area probability sample of non-institutionalized persons aged twenty-five and over and living in the coterminous United States. African-Americans and persons aged sixty and over were sampled at twice the rate of non-African-Americans and persons under age sixty. Initial face-to-face interviews lasting eighty-six minutes on average were completed in the homes of 3,617 respondents between May and October of 1986. The response rate was seventy percent among sampled households and sixty-eight percent among sampled individuals. Another wave of surveys occurred in 1989  and a third in 1994. 
In our analysis, we examine the effect of volunteering on depression as reported in 1994, using an edited version of the Center for Epidemiological Studies Depression Scale.  The scale has been used in numerous studies to assess levels of depressive symptoms and is known to be both a reliable and valid indicator of that outcome.  The index was scored by taking the standardized mean of the eleven component items; higher scores indicate more depressive symptoms.
We used three measures of volunteering. Volunteering Range is based on the number of voluntary organizations for which the respondents volunteered. We coded two dichotomous variables based on this summed score: volunteered for one organization and volunteered for more than one organization. The reference category was no volunteering. Volunteering Amount is the total number of hours volunteers contributed each year, which we coded into two variables: less than forty volunteering hours and forty or more hours. The reference category was no volunteering. Volunteering Length is a sum of the number of periods for which respondents mentioned volunteering, from which we created three dichotomous variables, each indicating a different number of periods (that is, one, two, or three periods). The reference category for these three variables is no volunteering in any wave. To examine the possibility that religious volunteering is more efficacious than secular, we created three dichotomous variables for each wave: volun teered for a religious organization only; volunteered for a secular organization only; and volunteered for religious and secular organizations. We then summed each variable across the three waves to produce a measure of the number of waves respondents engaged in that type of activity. As was the case for the volunteering length variable above, these three summed scores were separated into another three dichotomous variables that indicated the number of waves in which respondents volunteered. This coding scheme produced nine dichotomous variables, each indicating a specific period of time (that is, one, two, or three periods) in which respondents volunteered a specific way (that is, religious only, secular only, religious and secular). The reference category for all nine variables was no volunteering in any wave.
Table 1 reports the effects of volunteering range and amount, measured in 1986, on the level of depression measured in 1994, net of the level of depression in 1986. None of the volunteer measures have any effect on the levels of depression among adults under sixty-five. However, volunteering lowers depression among the respondents aged sixty-five or older whatever the measure used. A two-tailed t-test confirms that these differences in effect are statistically significant. Two measures of volunteering, range and length, show linear effects. The more respondents volunteer, the less likely are they to be depressed. In the case of volunteer hours, the positive effect is slightly stronger among those who work less than forty hours a year.
Table 2 reports the effect of volunteering for religious organizations compared to not volunteering, and to volunteering for secular or a mixture of secular and religious organizations. The mental health of younger and middle aged adults is unaffected by type of volunteering. Likewise, the amount of time spent volunteering has no impact on mental health. The impact of volunteering on the elderly presents a different story. First, it appears that volunteering exclusively for secular organizations is not so beneficial for them. Secondly, it is very clear that the longer the elderly volunteer for each type of organization, the greater the benefits.
This brief analysis of the Americans' Changing Lives data thus both confirms and adds to previous conclusions about the impact of volunteering on mental health. Given that these are analyses of changes in depression as a result of prior volunteering, they provide strong evidence to support the view that becoming a volunteer can have beneficial consequences on subjective wellbeing. Although our longitudinal design takes care of causal attribution problems and warrants our conclusion that well-being is the result, rather than the cause, of volunteering, the truth is probably more complicated. Causal effects and selection effects can be mutually reinforcing as well as mutually exclusive. Volunteering improves health, but it is also likely that healthier people are more likely to volunteer. Essentially, volunteering keeps healthy volunteers healthy.
The results also confirm that volunteering in a religious context is especially conducive to good mental health. On the other hand, the data provide no support to the arguments that too much volunteering can be stressful or that the beneficial effects observed in many elderly populations can be generalized to the entire adult population.
Much work remains to be done in this area. The first task is to determine whether volunteering is a sufficient condition for well-being. Must one volunteer to derive these benefits or can they accrue from virtually any kind of social activity? We control for social interaction in our analyses in an attempt to isolate the unique effects of volunteer work, but this is only a beginning. Is volunteer work simply one of a large number of forms of helping behavior that could easily be substituted without loss of benefits to the individual? In other words, do volunteers enjoy better mental health because they are helping others or simply because they have more frequent interactions with others? Other forms of helping (for example, caring for others on an informal basis or kinship exchanges of support) need to be measured and analyzed in the same population to see if they have equivalent effects. These other forms of helping might also interact with volunteering to produce different results. For example, one study s hows that volunteering increases depression among women who are also caregivers. 
The second task is to uncover the mechanisms that link volunteering and mental health. Do volunteers enjoy better mental health because their volunteer work boosts their self-esteem and sense of self-efficacy? Do volunteers enjoy better mental health because volunteering encourages better physical health practices that in turn make people feel better about themselves? Does volunteering make people less self-absorbed and therefore enable them to enjoy greater peace of mind?
The third task is to better differentiate among the various volunteer activities. The volunteers in the American Changing Lives data set benefited even more when their volunteering was affiliated with a religious organization. Is this a function of the individual volunteer being more religious? Is it a function of the fact that religious congregations provide more enduring communities of participation? Is it a function of the kind of volunteer work religious organizations tend to do?
The fourth task is to understand the age difference better. Why is it that older people receive more mental health benefits from volunteering than do younger people? Are other work- or family-related causes of depression swamping the effect of volunteering among younger people? Is it a function of the fact that older people might have exercised greater choice over their decision to volunteer and feel more free to leave anytime they feel they are not benefiting from it? Is it a function of the fact that older people tend to volunteer for different kinds of activities?
The fifth task is to examine the relation between satisfaction with volunteering and mental health. Although people can give up volunteer work whenever they become disenchanted with it, there are all kinds of familial, organizational, and societal pressures obliging people to volunteer. This pressure leaves room for much variation in the level of satisfaction with the volunteer work among those who continue to do it, and it is likely that more benefits accrue to those who are the most satisfied with their contribution. There is also the question of the fit between volunteer motivation and volunteer task. People who feel their volunteer work is satisfying their needs or their reasons for volunteering are most likely to reap its mental health benefits.
VOLUNTEERING AND OCCUPATIONAL ACHIEVEMENT
It is widely believed that, for those not currently working, volunteering can be a stepping stone to paid employment. Also, it is widely believed that, for those who currently have a job, being a volunteer advances one's chances of getting ahead. For example, nearly one quarter of the volunteers in the 1995 Independent Sector Survey of Giving and Volunteering included as one of their reasons to "make new contacts that might help my business or However, there is little reliable social science evidence to support the idea that volunteering actually helps people find jobs, or improves the quality of those jobs.
One pathway from volunteering to occupational achievement is through education. Studies conducted by the Higher Education Research Institute at the University of California, Los Angeles shed some light on this issue. Longitudinal data from the Corporation for National Service's Learn and Serve America Higher Education program are used to examine the effects of volunteering while in college. Students who chose to participate in service learning projects were more likely to see their grade point averages increase. Likewise, although all students lowered their expectations about pursuing an advanced degree over the course of their college careers, the decline was less steep for volunteers. 
Besides helping people accumulate educational credentials, volunteering also can play a more direct role in improving occupational opportunities. Mark Granovetter documented the importance of informal methods--personal Contacts--in securing information about jobs.  Among individuals who use such methods to find jobs, those who use weak ties--friends of friends or acquaintances--are more successful in locating satisfying jobs and jobs of high income than are individuals who used strong ties--close friends or relatives.  The strength of weak ties appears to lie in connecting the job searcher to dissimilar individuals, who are more likely to provide non-redundant information and influence.  Subsequent research by Nan Lin et al. shows that, among those who secure jobs through personal contacts, the higher the contact used the better the job they get.  Weak ties, rather than strong ties, tend to lead to high-status contacts.
To the extent volunteering increases a person's weak ties, it should have positive effects on occupational achievement. To date, no research has been conducted on this topic. However, John Beggs and Jeanne Hurbert use data from the 1981 Metropolitan Employer-Worker Survey to evaluate the effect of voluntary association memberships on the chances of finding a job.  Their study is also confined to those who found out about their job through personal contacts. They find that both gender and type of voluntary organization affect the kind of job a person gets.  For men, membership in a fraternal or service organization decreases the chances of using weak ties to find a job, which in turn reduces the quality of the job found, as measured by occupational prestige.  For women, participation in church-related or recreational organizations increases the probability that the job contact would be female, which has a negative effect on the prestige of the job found, although it does increase the chance tha t the woman found the job for which she was looking. 
As the authors acknowledge, these cross-sectional data cannot rule out the possibility that social ties are the cause of association memberships (for example, men with strong ties join fraternal or service organizations) rather a consequence.  Similarly, the data pertain only to membership in voluntary associations and do not record whether the respondents volunteered in connection with them or volunteered in any other capacity. Nevertheless, this is a promising line of research. To the extent volunteer work is considered a social resource it should yield the same benefits as those provided by simply being a member of an association. Indeed, it would stand to reason that these benefits are more likely to accrue to the active than the inactive member. There is evidence to suggest that volunteers indeed have more social contacts. For example, Robert Wuthnow finds that the number of different volunteer activities in which people are involved is positively correlated with the number of social contacts they have (for example, with neighbors, and local government officials), independent of the hours spent volunteering, level of education, number of group memberships, employment status, neighborhood income, church membership, and number of children. 
Anne Statham and Patricia Rhoton's unpublished paper provides one of the few analyses of the effects of volunteering on occupational achievement.  Using data from the Mature Women's Module of the National Longitudinal Survey, they find that, among women who were working for pay in 1981, those who had reported volunteering in 1974 had higher occupational prestige, net of occupational prestige in 1974.  Interestingly, they find that volunteering in 1974 decreased the chances of a woman being in the labor force in 1981, but had a positive effect on the occupational status of those who were. 
Using panel data from the Young Women's Module of the National Longitudinal Survey, we estimated the effect of early volunteering on subsequent occupational achievement. Women were first asked about their volunteer work in 1973 when their ages ranged from nineteen to twenty-nine.  They were asked again in 1978, 1988, and 1991. During each of these waves and in the intervening waves when questions on volunteering were not asked, respondents were asked to report on their labor-force participation. We use data from 1991 pertaining to occupational prestige to estimate the effect of volunteering, controlling for a number of other variables that influence both occupational prestige and volunteering. Occupational prestige depends chiefly on "human capital," which is measured principally by years of education. Ascriptive characteristics, principally race, also help determine occupational achievement. Because education is also a powerful predictor of volunteering, and whites tend to volunteer at higher rates tha n non-whites, we control for these two variables in our analyses. We also control for marital status and the presence of children in the household because these factors increase volunteering and may be related to participation in the labor force in a way that obscures the relation between volunteering and work.
In the following analysis, we use logistic regression to estimate the effect of earlier volunteering on whether the respondent will be in the labor force. We then use OLS to estimate the effect of volunteering on occupational prestige among those women who have a job. We regress 1991 occupational prestige first on 1973 volunteering, then on 1978 and 1988 volunteering. The control variables are always measured contemporaneously with the measure of volunteering being used. Finally, we regress occupational prestige on the number of waves for which women volunteered. Here the control variables are measured in 1973. For the purposes of this analysis, we restricted the data to those women who were present in all four waves.
Table 3 shows that neither 1973 nor 1978 volunteering has any effect on whether the young women are in the labor force in 1991. The odds-ratio for the effect of 1988 volunteering on 1991 labor force participation, because it is net of labor force participation, in 1988, indicates that women who volunteered in 1988 are more likely to have dropped out of the labor force three years later. Labor force participation in 1991 is unaffected by how many times women had reported volunteering in the past.
The OLS coefficients tell a rather different story. They indicate that all three waves of volunteering have a positive effect on 1991 occupational prestige among those with jobs, net of education and occupational prestige in earlier waves. The earlier waves of volunteering appear to have a stronger effect than the 1988 wave. Multiple waves also increase occupational prestige: the more times a woman reported volunteering, the higher her occupational prestige.
These are quite remarkable findings. Controlling for earlier levels of occupational prestige, we examine the effects of volunteering on the probability of an increase in occupational prestige. The mean level of occupational prestige for all women rises slightly over the eighteen-year period covered by the analysis, but it rises more sharply for the women who volunteer. We also demonstrate that a career in volunteering has a cumulative effect on occupational achievement. The more consistently the woman has volunteered, the more prestigious her job in 1991.
Those who believe that volunteering brings occupational benefits can draw some comfort from these results. Whatever their personal motivation for undertaking volunteer work, these young women seemed to benefit from it. It makes considerable sense that early volunteering makes little or no difference to whether women will be working at all. This is more likely to be determined by whether the woman is married, whether her husband works, how many children she has, and their ages. For those in the labor force, however, earlier volunteering does make a difference. Among the tasks ahead is the job of deciding what mechanisms connect volunteering and occupational achievement. Do women who volunteer obtain access to social capital that, in turn, increases the chances they will be exposed to better job opportunities? Do women who volunteer develop work-related skills thereby improving their qualifications for better jobs? Do women who volunteer gain self-confidence and greater self-esteem that in turn makes them more likely to aspire to and apply for better jobs? Another task is to place this relationship into a richer social context to see if other social forces moderate it. Do all volunteers benefit equally from their work when it comes to occupational achievement? Do some kinds of volunteering benefit women more than others? Does volunteering provide some women with an alternative way of achieving work-related skills and confidence and, therefore, do some women benefit more than others? Does volunteering benefit all aspects of women's work? Do women who volunteer enjoy higher salaries net of occupational position? Do they exercise more authority in the workplace? What about the moderating effects of family statuses? Are these benefits confined to those who can most fully take advantage of them? The presence of women in the labor force is known to be highly dependent on their life-cycle stage. It is lowest at the stage when there are young children in the household. Are women who have no competing child care responsibi lities most likely to benefit from their volunteer work? Women's labor force position is also dependent on their marital status. They are more likely to be working, even if they have small children, if they are single, or are married to a husband who is unemployed or earns little money. Perhaps this moderates the effect of volunteering on work. The effect is likely to be much stronger on women who have elected to be in the labor force and selected themselves into their job than women who are in the labor force simply because they have to make ends meet. Volunteering, in this respect, is an enabling factor but will increase occupational achievement only if other conditions are favorable. In short, family life cycle conditions and constrains the effects of Volunteering--and other economic variables--on occupational achievement.
These findings make some contribution to the current debate about the impact of women's increasing presence in the labor force on the supply of volunteer labor. There seems to be no question that women today spend more time working for pay than they did years ago.  It is also clear that women who increase their work hours tend to reduce their volunteer hours and that the most consistent source of volunteer labor comes from part-time workers.  However, it does not seem that this is causing fewer women to volunteer, and the reason seems to be that rising educational and occupational levels among women are compensating for their extra hours at work. Women today might have less time to volunteer, but they are accumulating other resources, both individual and social, that seem to compensate for this loss to maintain their overall level of volunteering and even increase it.
This review of the research on the effects of volunteering leaves little doubt that there are individual benefits to be derived from doing volunteer work that reach far beyond the volunteer act itself and may linger long after the volunteer role is relinquished. Much work remains to be done in the five areas covered, and there are undoubtedly other categories of benefits that could be identified. In conclusion, it is worthwhile to stand back and reconsider the very notion of the benefits of volunteering to the volunteer.
First, there can be little doubt that these benefits are usually unintended consequences of behavior that is motivated not by extrinsic but intrinsic rewards. Indeed, there are justifiable fears that attaching rewards to altruism will undermine motivation and distort values. It is not likely, then, that this information on benefits can be used productively as a recruitment tool or means of mobilizing volunteer effort. Indeed, in the case of volunteering's effect on health it might be unwise because the benefits are contingent on the volunteer being intrinsically motivated.
Second, the language of "consequences" encourages us to think of volunteering as the cause of some subsequent outcome. Where the outcome is risk of mortality, this language might well be appropriate, because there can be no doubting the independence of cause and effect. It is also appropriate where the rewards are extrinsic to the volunteer act, as when volunteering builds occupational skills. Presumably, those rewards could have been obtained from a wide array of activities that do not entail helping others. It is therefore appropriate to investigate the probability that volunteers do indeed have more occupational skills than do non-volunteers. However, the consequences of volunteering for many people are intrinsic to the act itself. Volunteering is how they express their identity, or their values. Intrinsic rewards, those that are embedded in the act itself, are difficult to describe in the language of cause and effect. Self-discovery is not the effect of volunteering; it is volunteering itself.
Third, in some part, the interest in the benefits of volunteering is triggered by a desire to encourage more volunteering. If volunteering can be shown to be good for you, perhaps more people will do it. Too much attention to benefits, however, can distract attention from possible costs. Costs are not simply the lack of benefits. For example, much of the study of the effect of volunteering on citizenship posits an ideal democratic citizenship as the possible benefit. Encourage children to volunteer, it is argued, and they will become model citizens. However, too much focus on these benefits distracts attention from other possible consequences of volunteering, some of which might not be seen as quite so beneficial. Therefore, volunteering--to the extent it focuses on face-to-face amelioration of the personal problems of individuals--might encourage political attitudes that are, in many ways, anti-democratic because these attitudes may attribute social problems such as homelessness to the actions of individual s and not to the larger structural forces which might be causing them, forces which can be tackled only by concerted political action. In short, the language of "benefits" is more value-laden than is the language of consequences.
The analysis of data reported in this article was made possible by a grant from the Nonprofit Sector Research Fund of the Aspen Institute.
(*.) Professor of Sociology, Duke University.
(**.) Assistant Professor of Sociology, University of Texas, Austin.
(1.) See ROBERT WUTHNOW, ACTS OF COMPASSION 87 (1991).
(2.) See Robert Putnam, Bowling Alone: America's Declining Social Capital, 6 J. DEMOCRACY 65 (1995).
(3.) See ALEXIS DE TOCQUEVILLE, DEMOCRACY IN AMERICA 107 (Henry Reeve trans., Phillips Bradley ed., 1945).
(4.) See generally ROBERT PUTNAM, MAKING DEMOCRACY WORK: CIVIC TRADITIONS IN MONDERN ITALY (1993).
(5.) Id. at 167.
(6.) See SIDNEY VERBA ET AL., VOICE AND EQUALITY: CIVIC VOLUNTARISM IN AMERICAN
POLITICS 352 (1995).
(7.) See David Knoke, Networks of Political Action: Toward Theory Construction, 68 SOCIAL FORCES 1041, 1057 (1990).
(8.) See id.
(9.) See id. at 1058.
(10.) See Putnam, supra note 2, at 73.
(11.) See Dietland Stolle, Bowling Together, Bowling Alone: The Development of Generalized Trust in Voluntary Associations, 19 POL. PSYCHOL. 497 (1998).
(12.) See id. at 501.
(13.) See Dietland Stolle & Thomas R. Rochon, Are All Associations Alike? Membership Diversity, Associational Type and the Creation of Social Capital, 42 AM. BEHAVIORAL SCIENTIST 47, 55 (1998).
(14.) See id. at 57.
(15.) See id.
(16.) See id.
(17.) See Eleanor Brown, The Scope of Volunteer Activity and Public Service, 62 LAW & CONTEMP. PROBS. 17 (Autumn 1999).
(18.) See generally ANDREW KOHUT, TRUST AND CITIZEN ENGAGEMENT IN METROPOLITAN PHILADELPHIA (1998).
(19.) See Peter Hall, Social Capital in Britain, 29 BRITISH J. POL. SCI. 417 (1999).
(20.) See Stephen Hamilton & Mickey Fenzel, The Impact of Volunteer Experience on Adolescent Social Development: Evidence of Program Effects, 3 1. ADOLESCENT RES. 65, 70 (1988).
(21.) See id.
(22.) See id. at 72 tbl. 1. The effect was stronger for girls and for those involved in community improvement projects.
(23.) See Constance Flanagan et al., Ties that Bind. Correlates of Adolescents, Civic Commitments in Seven Countries, 54 J. SOC. ISSUES 457-75 (1998). Russia, Bulgaria, the Czech Republic, and Hungary represented the transitional democracies while Australia, Sweden, and the United States represented the more stable political systems.
(24.) See id. at 465.
(25.) See id.
(26.) See id at 469 tbl. 3.
(27.) See id.
(28.) See James Youniss et al., The Role of Community Service in Identity Development: Normative, Unconventional, and Deviant Orientation, 14 J. ADOLESCENT RES. 248 (1999).
(29.) See id. at 258-59.
(30.) See VERBA ET AL, supra note 6, at 425.
(31.) See Margaret Conway et al., Democratic Socialization in Schools, in DEMOCRACY, SOCIALIZATION AND CONFLICTING LOYALTIES IN EAST AND WEST 421-40 (Russell Farnen et at. eds., 1996).
(32.) See id. at 424.
(33.) See id. at 431.
(34.) See id.
(35.) See id. at 432.
(36.) See VERBA ET AL., supra note 6, at 465.
(37.) See JAMES MAX FENDRICH, IDEAL CITIZENS: THE LEGACY OF THE CIVIL RIGHTS MOVEMENT 30-31 (1993).
(38.) See id. at 65-67, 89, 95-98.
(39.) See id. at 78-82, 108-13.
(40.) See DOUGLAS MCADAM, FREEDOM SUMMER 214-15 (1988).
(41.) See Daniel Hart & Robert Atkins, Urban America as a Context for the Development of Moral Identity in Adolescence, 54 J. SOC. ISSUES 513 (1998). Examples of "problem behaviors" include stealing something worth more than $50, using force to get something, attacking someone to inflict serious injury, and breaking into a building. See id. at 521.
(42.) See Jacquelynne Eccies & Bonnie Barber, Student Council, Volunteering, Basketball and Marching Band: What Kind of Extracurricular Activity Matters?, 14 J. ADOLESCENT RES. 10, 13 (1999).
(43.) See id. at 18, 19 tbl. 4, 20 tbl. 5.
(44.) See id.
(45.) See Joseph P. Allen et al., Programmatic Prevention of Adolescent Problem Behaviors: The Role of Autonomy, Relatedness, and Volunteer Service in the Teen Outreach Program, 22 AM. J.
COMMUNITY PSYCH. 617 (1994).
(46.) Id. at 627.
(47.) See id. at 636.
(48.) See id. at 633.
(49.) See id. at 626.
(50.) See Christopher Uggen & Jennifer Janikula, Volunteerism and Arrest in the Transition to Adulthood, 78 SOC. FORCES 331, 336-37 (1999).
(51.) See id.
(52.) See id. at 345 tbl. 2.
(53.) See id. at 346, 347 fig. 3.
(54.) See id. at 349.
(55.) See generally TRAVIS HIRSCHI, THE CAUSES OF DELINQUENCY (1969).
(56.) See generally EDWIN SUTHERLAND, PRINCIPLES OF CRIMINOLOGY (1939).
(57.) See generally James House et al., Social Relationships and Health, 241 SCIENCE 540 (1988).
(58.) Doug Oman et al., Volunteerism and Mortality Among Community-dwelling Elderly, 4 J. HEALTH PSYCH. 301, 303 (1999).
(59.) See Phyllis Moen et al., Successful Aging: A Life Course Perspective on Women's Multiple Roles and Health, 97 AM. J. SOC. 1612, 1628 (1992).
(60.) See id. Duration of volunteering between 1956 and 1986 was not related to better physical health in old age, although it was positively correlated with multiple role holding in 1986. See id.
(61.) See Edward P. Sabin, Social Relationships and Mortality Among the Elderly, 12 J. APP. GERONTOLOGY 44 (1993).
(62.) See id. at 50.
(63.) See id. at 54.
(64.) See id.
(65.) See Richard Rogers, The Effects of Family Composition, Health, and Social Support Linkages on Mortality, 37 J. HEALTH & Soc. BEHAVIOR 326, 328-29 (1996).
(66.) See id. at 331 tbl. 1.
(67.) See id. at 334-34 tbl. 3.
(68.) See Marc A. Musick et al., Volunteering and Mortality among Older Adults: Findings from a National Sample, 54B J. GERONTOLOGY S173 (1999).
(69.) See id. at S174.
(70.) See id.
(71.) See id. at S176 tbl. 1.
(72.) See id. at S176.
(73.) See id. at S177.
(74.) See id., at S176.
(75.) See Oman et al., supra note 58, at 307.
(76.) See id.
(77.) See id. at 308.
(78.) See Musick et al., supra note 68, at S177.
(79.) See Oman et al., supra note 58, at 309.
(80.) See id.
(81.) See Musick et al., supra note 68, at S177.
(82.) See Peggy Thoits, Multiple Identities and Psychological Well-being, 48 AM. SOC. REV. 174, 182 (1983).
(83.) S. Mark Pancer & Michael Pratt Social and Family Determinants of Community Service Involvement in Canadian Youth, in ROOTS OF CIVIC IDENTITY: INTERNATIONAL PERSPECTIVES ON COMMUNITY SERVICE AND ACTIVISM IN YOUTH 32, 38 (Miranda Yates & James Youniss eds., 1999).
(84.) Nancy Morrow-Howell et al, The Perceived Benefits of Participation in Volunteer and Educational Activities, 32 J. GERONTOLOGICAL SOC. WORK 65 (1999).
(85.) See id. at 7l.
(86.) House et al., supra note 57, at 302; see also James House et al., Structures and Processes of Social Support, 14 ANN. REV. SOC. 293, 295 (1988).
(87.) JOHN MIROWSKY & CATHERINE ROSS, SOCIAL CAUSES OF PSYCHOLOGICAL DISTRESS 141 (1989).
(88.) See Debra Umberson et al., The Effect of Social Relationships on Psychological Well-being: Are Men and Women Really So Different?, 61 AM. SOC. REV. 837,839 (1996).
(89.) See Neal Krause et al., Providing Support to Others and Well-being in Later Life, 47 J. GERONTOLOGY P300, P300 (1992).
(90.) See MIROWSKY & ROSS, supra note 87, at 167.
(91.) See Musick et al., supra note 68, at S173; Oman et al., supra note 58, at 311.
(92.) See Musick et al., supra note 68, at S178.
(94.) Ronald Jirovec & Christine Hyduk, The Type of Volunteer Experience and Health among Older Adult Volunteers, 30 J. GERONTOLOGICAL SOC. WORK 29, 36 (1998).
(95.) See Morris Okun et al., The Social Activity/Subjective Well-Being Relation, 6 RES. AGING 45, 56 (1984).
(96.) See id. at 55.
(97.) See id. at 62.
(98.) See Oman et al., supra note 58, at 309.
(99.) See id. at 312.
(100.) LUCY ROSE FISCHER & KAY BANISTER SCHAFFER, OLDER VOLUNTEERS: A GUIDE TO RESEARCH & PRACTICE 9-10 (1993).
(101.) See Roger Mannell, High Investment Activity and Life Satisfaction among Older Adults, in ACTIVITY AND AGING 125, 127 (J. Kelly ed., 1993).
(102.) See id.
(103.) Judith Wheeler et al., The Beneficial Effects of Volunteering for Older Adults and the People They Serve, 47 INT'L J. AGING & HUMAN DEV. 69,78 (1998).
(104.) Frank W. Young & Nancy Glasgow, Voluntary Social Participation and Health, 20 RES. AGING 339 (1998).
(105.) See id. at 347-48.
(106.) See id. at 349.
(107.) See id. at 348-55.
(108.) See id. at 356-57.
(109.) See Krause et al., supra note 89, at P308.
(110.) See generally id; Wheeler et al., supra note 103; Young & Glasgow, supra note 104.
(111.) Barbara McIntosh & Nicholas Danigelis, Race, Gender and the Relevance of Productive Activity for Elders' Affect, 50 J. GERONTOLOGY S229 (1995).
(112.) See id. at S229.
(113.) See id.
(114.) See id. at S233.
(115.) See id.
(116.) See id.
(117.) See id. at S236.
(118.) See id. at S237.
(119.) JAMES S. HOUSE, AMERICANS' CHANGING LIVES: WAVES I AND II, 1986 AND 1989 (1996).
(122.) L.S. Radloff, The CES-D Scale: A Self-Report Depression Scale for Research in the General Population, 1 APPLIED PSYCH. MEASUREMENT 385 (1977).
(123.) See Walter M. Easel, Measuring Depression: The CES-D Scale, in SOCIAL SUPPORT, LIFE EVENTS, AND DEPRESSION 51, 53-55 (Nan Lin et al. eds., 1986).
(124.) See Phyllis Moen et al., Caregiving and Women's Well-being: A Life Course Approach, 36 J. HEALTH & SOC. BEHAVIOR 259,267 (1995).
(125.) VIRGINIA HODGKINSON & MURRAY WEITZMAN, GIVING AND VOLUNTEERING IN THE UNITED STATES (1996).
(126.) See Linda J. Sax & Alexander Astin, The Benefits of Service: Evidence from Undergraduates, 78 EDUCATIONAL REC. 25-32 (1997).
(127.) Mark S. Granovetter, The Strength of Weak Ties, 78 AM. J. SOC. 1360 (1973).
(128.) See id. at 1371-73.
(129.) See id.
(130.) Nan Lin et al., Social Resources and Strength of Ties: Structural Factors in Occupational Status Attainment, 46 AM. SOC. REV. 393,398 (1981).
(131.) John J. Beggs & Jeanne S. Hurlbert, The Social Context of Men's and Women's Job Search Ties: Membership in Voluntary Organizations, Social Resources and Job Search Outcomes, 40 SOCIOLOGICAL PERSPECTIVES 601, 609 (1997).
(132.) See id. at 613-14.
(133.) See id. at 611.
(134.) See id. at 616-17.
(135.) See id. at 605.
(136.) ROBERT WUTHNOW, LOOSE CONNECTIONS: JOINING TOGETHER IN AMERICA'S FRAGMENTED COMMUNITIES 235 (1998).
(137.) Anne Statham & Patricia Rhoton, Mature and Young Women's Volunteer Work, 1974-1981 (1985) (unpublished manuscript, on file with the Center for Human Resource Research, Ohio State University).
(138.) See id. at 35.
(139.) See id.
(141.) See ELIZABETH MARET, WOMEN'S CAREER PATTERNS: INFLUENCES ON WORK STABILITY 7 (1983).
(142.) See Richard Freeman, Working for Nothing: The Supply of Volunteer Labor, 15 J. LABOR ECON. 140-67 (1997).…