"IF YOU DON'T DO IT, NOBODY ELSE WILL": ACTIVE AND TOKEN CONTRIBUTORS TO LOCAL COLLECTIVE ACTION*

PAMELA OLIVER
University of Wisconsin

It is commonly assumed that people participate more in collective action when they believe others will. But local activi~tsoften say: " I did it because nobody else would." Investigation of the differences among 1456 Detroit residents who were nonmembers, token members, or active members (either currently active or past leaders) of their neighborhood associations reveals that active members were significantly more pessimistic than token members about the prospects for neighborhood collective action, a finding explained by recent theoretical work on collective action by Oliver et al. (1984). Other findings are that active members are more highly educated than token members; that past leaders know more people and have higher interest in local problems; and that currently active members have more close ties in the neighborhood, like the neighborhood less, and are less likely to be homeowners. Contrasts between members and nonmembers are similar to those found in previous research.

When many people share an interest in some collective good, there are often wide discrepancies in the extent to which they contribute to obtaining it. Some people do nothing at all, others make only small token contributions such as signing a petition or paying dues, while a third group contributes substantial amounts of time and effort. What explains these differences in the willingness to absorb costs in the provision of local collective goods? Why are some people willing to make some real commitment of their time to a cause, while others give only token support or lip service? Are those who make the larger commitments simply those who are more interested in the collective goal and have more at stake in its provision? Do they have an unusually great faith in the willingness of their fellow citizens to back them in their efforts? Are they people with a lot of spare time on their hands, looking for a way to avoid boredom? Are they powerhungry moguls exploiting community needs as an avenue to their own advancement? There is much less research on the question of explaining active versus token contributions than one might expect. Much of this is because

* Direct all correspondence to: Pamela Oliver, Department of Sociology, University of Wisconsin, Madison, WI 53706. I would like to thank Donald Warren for providing a copy of his data, Forrest Graves and Thomas Flory for helping with the data transfer, Mary Brulla and Mark Nelson for research assistance, and Elizabeth Thomson, Sara McLanahan, Annamette Sprensen, Gerald Marwell, and John Lemke for their helpful comments on previous drafts of this manuscript. Partial funding for this research was provided by the Wisconsin Alumni Research Foundation.

it is assumed that the motivation to leadership is relatively unproblematic, that of course people want to be leaders so that they can have access to power in an organization. But scholars of voluntary associations (as opposed to large bureaucratic organizations) know better. Pearce (1980) compared cooperatives which employed paid staff with those relying solely on volunteers. She observed that in employing organizations, leadership positions were hard to get and were sought after by the participants, while in all-volunteer organizations, leadership positions were easy to get but avoided by the participants. In all-volunteer cooperatives, the leaders absorbed high costs with low compensating rewards. Rich (1980b) studied a variety of neighborhood organizations; the leaders of organizations relying solely on volunteers and voluntary contributions absorbed the high cost of their participation while personally realizing relatively little of the collective goals. In short, the active members of all-volunteer local organizations are frequently underrewarded workhorses who provide collective benefits for their often unappreciative constituencies. Why do they do it? This paper reports the differences among residents of Detroit who were nonmembers, token members, or active members of their neighborhood associations. Consistent with previous research, active members are more highly educated than token members. But an additional predictor has not been recognized in previous research: active members are more pessimistic than token members about the prospects for collective action in their neighborhoods. This perhaps surprising finding is, in fact, quite consistent with what activists often 60 1

American Sociological Review, 1984, Vol. 49 (October:601-610)

AMERICAN SOCIOLOGICAL REVIEW say and with recent theoretical work on collective action. THEORY AND PREDICTIONS The concept of collective action provides a framework for understanding the common dilemmas in a wide variety of situations, and many scholars have treated participation in community organizations as a form of collective actiop (for example, Rich, 1980a,b; O'Brien, 1974; Henig, 1982; Wandersman, 1981; Smith, 1981; Sharp, 1978; Stinson and Stam, 1976). The term "collective action" refers to activities which produce collective or public goods, that is, goods with t h e nonexcludability property that their provision to some members of a group means that they cannot be withheld from others in the group (Olston, 1965: 14). People who live in the same local area share common interests which lead to the existence of collective goods, both environmental, such as floods, wind storms, noxious fumes, or commuter traffic, and social, such as garbage collection, street repair, police patrols, and public schools. Neighborhood organizations are explicitly formed to address these collective goods. Surveys of neighborhood organizations find that they address a wide variety of collective goods, such as housing, general city services, various types of crimes, street safety and traffic problems, recreation, senior citizen needs, education, unemployment, health services, commercial revitalization, redlining, highway construction, drug abuse, planning, tenant issues, and pollution (Green, 1979; National Commission on Neighborhoods, 1979; Oliver, 1980b). From our theoretical understanding of collective action, four general factors can be expected to determine varying levels of involvement in neighborhood organizations. The first two are the basic economic factors of interests and costs. Very simply, we expect that larger contributions will come from people who value neighborhood collective goods more or who experience lower costs from their contributions. A third factor, social ties among group members, is stressed by sociologists such as Fireman and Gamson (1979), Tilly (1978), and Granovetter (1 973). The fourth general factor concerns predictions about others' behavior, specifically pessimism about the prospects for collective action by others. In some situations, notably in "large groups" (Olson, 1965), any individual's contribution is too small to make a noticeable difference in the level of the collective good, so everyone's contribution is irrational no matter what anyone else does. But in "small groups," such as the active members of a community organization, individual contributions do make a noticeable difference (Rich, 1980b) and predictions about others' behavior are relevant. People who believe others will provide the collective good are motivated to ride free; people who do not believe others will provide the collective good are motivated to provide the good themselves or do without. More specific predictions can be developed for each of these four factors in turn, beginning with pessimism about others' behavior.
Pessimism About Others' Collective Action

It is commonly assumed that people are more willing to participate in collective action if they believe that others will. But if you ask someone why he or she agreed to chair a fundraising drive or be recording secretary of a local organization, acommon answer is: "If I don't do it, nobody else will." That is, activists are often quite pessimistic, believing it unlikely that they will be able to rely on the efforts of their neighbors. Recent work on collective action by Oliver et al. (1984) argues that there is an interaction between beliefs about others' willingness to contribute to collective action and the character of the collective good. Optimism about collective action by others makes a person more willing to contribute when contributions have an accelerating impact on the collective good. But when there are diminishing marginal returns to contributions, pessimism about others' actions, not optimism, makes a person more willing to contribute. Contributions have diminishing marginal return when jobs are relatively finite, or when the earliest contributions have the biggest impact. Keeping an organization's checkbook, arranging to rent games for a school fair, or preparing a newsletter have this property: once the job is being done at all, additional contributions produce smaller (although not zero) increments in the collective good. If the job is being done, there is little marginal payoff for helping out, and free riding is likely. Local activism often exhibits this property of diminishing marginal returns. Thursz (1972) stresses that successful community organizations do not require mass participation, citing Alinsky's claim that participation by 3 percent of a community would ensure success of a community organization. Bolduc (1980) provides a typical case study of a neighborhood in which only a dozen residents participated actively in the neighborhood organization although it was viewed as a legitimate representative body by the majority of residents. Since active members and leaders of local voluntary organizations absorb high costs for low re-

LOCAL COLLECTIVE ACTION wards, it stands to reason that they are more likely to make this sacrifice when they believe no one else will. Private incentives (Olson, 1965; Oliver, 1980a) could minimize the impact of this factor, but it is unusual in voluntary community action for the incentives for participation to be large enough to make a person want to absorb the costs of involvement. The incentives for active participation in community organizations are usually found to be psychological, including social contacts, deference or respect, selfactualization, learning new skills, or feeling a sense of accomplishment (Smith, 1981; Salem, 1978; Rich, 1980b; Sharp, 1978). higher costs than nonmembers, although the difference may not be great. It is difficult to make comparisons across people of the costs of action, but it is possible to draw inferences about such costs by making assumptions about the nature of various actions and the effects of people's life circumstances. Education and income. One of the most well-documented correlations in social science is the positive correlation between socioeconomic status and all forms of political or organizational participation (see, e.g., Verba and Nie, 1972; Smith and Freedman, 1972), including participation in community organizations (see Parkum and Parkum, 1980; Vedlitz and Veblen, 1980; Verba and Nie, 1972; Smith and Freedman, 1972). This finding has been explained in cost-benefit terms by many authors, perhaps most forcefully by O'Brien (1974, 1975). The explanation is grounded in the high costs of participation for the poor. O'Brien argues that poor people are too concerned about survival to have time for leisure activities, that their failure to participate in community organizations is not due t o "apathy" but to an acute case of the free-rider problem in which the costs of participation far outweigh the individual's share of the collective good. Psychological costs are also relevant in this context. Organizational activity usually requires skills that are more common among educated people, such as public speaking, Roberts Rules of Order, understanding technical explanations, or knowing how to call City Hall. This means that the cost of such activities is much higher for less-educated people. These cost considerations should hold true both for distinguishing members from nonmembers, and for distinguishing active members from token members of community organizations. Free time. Free or discretionary time is often posited as an important factor in collective action (McCarthy and Zald, 1973). The stereotype of the community volunteer as a bored housewife or retiree is common. Certainly this is a plausible account in cost-benefit terms, since the opportunity cost of an investment of time is lower for a person with more free time. Since we lack direct measures of free time, we may make plausible inferences about free time from certain demographic characteristics. Other things being equal, people who are employed full time should have less free time than those who a r e not. F r e e time is also doubtlessly negatively related to the number of children one has. As plausible as the free-time account is, especially for women's behavior, and especially

Interest in the Collective Good

A person's level of interest in the collective good should always have a positive effect on participating in collective action, but the strength of this impact may vary. We may distinguish subjective interest, as indicated by statements of concern about neighborhood problems, from objective interest, as indicated by demographic characteristics. Concerning subjective interests, several studies and literature reviews indicate that various attitudinal measures which tap what can be thought of as the person's concern about the collective good are important predictors of participation in community organizations (McKenzie, 1981 ; Uzzell, 1980; Nanetti, 1980; Parkum and Parkum, 1980). However, these studies do not distinguish among levels of participation. One study which concerns leaders specifically (Rich, 1980b) argues that psychological incentives, not concern about the collective good, are the best predictors of the willingness to be a leader in an all-volunteer neighborhood organization. The relevant "objective" interest for neighborhood organizations is being a homeowner. Homeownership is likely to distinguish members from nonmembers, but is less likely to distinguish active from token members. The correlation with membership is high because renters are quite unlikely to belong to neighborhood organizations, but too little variance is left to distinguish active from token members. In theoretical terms, gross-category membership is relevant for defining the population at risk, but not for determining the level of contribution a person is willing to make.

Costs
It seems obvious that active members of local organizations absorb higher costs of action than do token members (Pearce, 1980; Rich, 1980a,b). Token members presumably absorb

AMERICAN SOCIOLOGICAL REVIEW considering recent publicity about the decline in volunteerism as women have entered paid employment, there is some contrary evidence. Several studies have found positive correlations between employment and voluntary community participation among women, especially less-educated women (Flynn and Webb, 1975; Schoenberg, 1980; and several unpublished studies cited by Schoenberg, including Schoenberg and Rosenbaum, 1979; Dabrowski, 1979; and Holmes, 1979). These studies argue that the skills and self-confidence obtained from paid employment are necessary for a woman to feel willing to engage in community participation. Having children was a positive predictor of participation in at least one older study (Wright and Hyman, 1958), although Ahlbrandt and Cunningham (1979) report that it did not predict participation in neighborhood organizations in six Pittsburgh neighborhoods.
Social Ties

1969 under the direction of Donald Warren; more details on data collection and sampling procedures may be found in Warren (1975). Twenty-eight elementary school attendance areas within the central city of Detroit were sampled purposively; sixteen of these were Qver 90 percent black and twelve were over 90 percent white. Individuals were randomly sampled within neighborhoods. The 1456 respondents included in the analysis were of the majority race in their neighborhood and were missing information on less than six of the original variables in the analysis. Cases missing information on a variable were assigned the mean for interval variables and the median for ordinal variables, a procedure which is conservative since it tends to attenuate correlations.
Measures of Dependent Variables

In the substantive literature on neighborhoods, a major theme is the social solidarity or social integration within a neighborhood (see, e.g., Warren and Warren, 1977). It is usually assumed that this factor is an important element in a neighborhood's ability to act collectively in response to some threat, although this assumption is rarely subject to test. Sociologists such as Tilly (1978), Fireman and Gamson (1977), Granovetter (1973, 1982), and Snow et al. (1980) stress the importance of social ties for collective action. Tilly (1978) distinguishes feelings of identity or solidarity from network ties. Granovetter (1973, 1982) and Duff and Liu (1972) distinguish weak ties of acquaintance from strong ties of friendship, arguing that weak ties are important for collective action. Social ties may be thought of as indicators of subjective interest in the neighborhood, as factors influencing the availability of solidary incentives for participation in collective action, or as factors reducing the cost of action by making communication easier. All these theoretical interpretations yield the same empirical prediction, that social ties will generally have a positive effect on collective action. None distinguishes theoretically between active and token contributions. In this paper, the effects of social ties are assessed without attempting to determine the best theoretical interpretation of these effects. METHODS AND PROCEDURES Sample The analysis in this paper is based upon data originally collected in Detroit neighborhoods in

The three dependent dummy variables are types of participation in neighborhood improvement organizations: membership, activism, and leadership. Each reported organizational membership was classified by type of organization; "neighborhood improvement association" was an original response category. The handful of people who were members of more than one neighborhood organization were simply coded as members, as there were too few of them to analyze. For each organizational membership, the respondent was asked "How active have you been in the past three years?'The codes active and not active were entered for each oflhe years 1966, 1967, 1968, and 1969. The variable employed in the present analysis contrasts those who were active in any of the four years with all others. Leadership was indicated by the answer to the question: "Have you ever held an office or position of leadership in any of these groups? If yes, which ones?'1
Measure of Predictions About Others' Collective Action

Respondents were asked to indicate on a fivepoint scale from "certaintly will" to "certainly won't" their assessment of "how ready you think your neighbors would be to help each other in various situations." Two were collective: "If the principal of the local school was doing a very poor job, how much could you count on your neighbors for help in doing something about it?'and "If the city were to announce a project that would hurt this neighThe two former neighborhood organization leaders who were not currently members of neighborhood organizations (Table 1) were dropped from the analysis of leadership.

LOCAL COLLECTIVE ACTION borhood, and some of the neighbors tried to organize a protest, how would the others feel about joining?" The analyzed variable is an index created by summing these two items.
Measures of Interest in the Collective Good

Membership in the appropriate objective interest groups is indicated by whether the person owns or is buying (as opposed to rents) his or her home. Subjective interest in neighborhood improveinent is measured with a composite index based on six variables. The first variable is the number of "problems" the respondent said . ~ other existed in his or her n e i g h b ~ r h o o dThe five variables are the level of dissatisfaction with five city services (parks and playgrounds, sports and recreation centers, police protection, garbage collection, schools), each having four levels of response from generally satisfied to very dissatisfied. Principal-axis factor analysis with varimax rotation identified only one significant dimension; all other factors were of roughly equal weight and unique to a variable. The analyzed variable is the weighted regression factor score.
Costs Education and income. Education is coded as years of formal education. Income was assessed with the 1960 Census groupings; each category was assigned its midpoint, and the variable was treated as interval. Free time. Respondents employed full time were contrasted with those who have no paid employment or only part-time employment. Respondents were coded for the presence or absence of children under age 18 in the household. Measures of Social Ties

The analyzed variables are three indices tapping important dimensions of social ties: positive affect or liking for the neighborhood; the
These were: racial strife; people not keeping their houses up; unemployment; wild teenagers; nationality or religious conflict; people not knowing how to get along in the big city; pressure to keep up with the Joneses; young children not supervised properly; "it's no use trying" attitude toward solving local neighborhood problems; conflicts between older and younger children; people with strange behavior; fear of street crime; militant pressure groups; police harassment; traffic and noise; lack of police services. Several analytic approaches failed to reveal any interesting subdimensions, factors or clusters within these items, so the count of all yes answers is treated as a single variable.

number of acquaintances one has in the neighborhood; and the extent of one's close ties of friendship or kinship in the neighborhood. To create these indices, eight variables were subjected to factor analysis: (1) The person's liking for the neighborhood, measured with a composite index created by summing four questions. The first asked, "At your present time in life, how close are your neighbors to what you think neighbors should be like?" and was recorded on a fourpoint scale from "very close" to "not close at all." The second asked, "In general, how do you feel about this neighborhood?" with four responses ranging.from good to very poor. The third and fourth used four responses from "like very much" to "dislike very m u c h for "your own block" and "the people living right nearby ." (2) Another measure of positive affect, how close the person's neighbors were to their ideal. (3) An index created by summing the responses to questions on the perceived likelihood that one's neighbors would engage in five kinds of "personal" helping: keeping an eye on children; caring for a child while away for a week; helping you when sick; keep an eye on the house for a month; lend you a few dollars. (4) A measure of the person's perception of being like other people in the local area. Respondents were asked, "In general, would you say you and your neighbors share the same or different views on the following matters: best way to raise children, religious beliefs, attitudes about race problems, political attitudes, goals for children, way to enjoy leisure time, kind of person to have as a friend, how to furnish a house in good taste, how to get ahead in the world." An index was constructed by subtracting the number of "different" answers form the number of "same" answers. (5) The number of " w e a k ties: the number of neighbors the person knows "well enough so that you might spend half an hour or so with them now and then." Responses for separate questions for "on this block" and "in this area but not on this block" were summed. (6) A measure of somewhat stronger ties: how often the person gets together with neighbors "at their home or yours," ranging from never to daily. (7) A measure of the strong ties of friendship: whether at least one of the person's three closest friends is in the neighborhood. (8) A measure of the strong ties of kinship: the proportion of the person's relatives in the Detroit area who live in the immediate neighborhood: all, most, about half, only a few, or none.

AMERICAN SOCIOLOGICAL REVIEW The factor analysis reveals only one "significant" dimension for these variables, a generalized satisfaction with the neighborhood. However, the three-factor solution produced theoretically meaningful dimensions, so the three weighted-regression factor scores were used in the analysis to allow specification of the relationship between social ties and local collective action. The variables which loaded high on the first factor were the liking index, whether the neighbors are ideal, the personal helping index, and the attitude-similarity index; this factor score is called the Liking for Neighborhood Scale. The most important variable in the second factor is the number of people known, with visiting with neighbors and believing the neighbors would give personal help also loading on this factor; this factor score is called the Acquaintances Scale. Having one's closest friends in the neighborhood is the only variable which loads highly on the third factor, with having relatives in the neighborhood having a moderate loading; this factor score is called the Close Ties Scale. "joiners." Other memberships is highly correlated with membership in neighborhood organizations but, among members, not with being active. The only effect of including this variable on the other parameters in any equation reported below is to make income nonsignificant as a predictor of membership.

Mode of Analysis
Since the dependent variables are dichotomies, probit analysis is an appropriate statistical tool. The contrast between members and nonmembers is assessed with the total sample. The contrast between token members and active members is assessed with the subsample who are members of the appropriate organization. This mode of analysis plausibly assumes that people are first selected to membership in an organization and, once members, face further selection for becoming active. RESULTS Table 1 shows the distributions of the dependent variables and their interrelations. Nineteen percent of the respondents were members of neighborhood improvement associations at the time of the survey. Of the members, 22 percent said they had been a leader or officer of such a group, and 27 percent said they had been active in the group in at least one of the past four years. Thirty percent of the past leaders were currently active in their organizations, and 25 percent of those currently active had been leaders. The means and standard deviations for all the independent variables are given in Table 2. Table 3 shows the coefficients for a probit analysis of membership and, among members, of the two kinds of active participation. The findings for membership are reasonably consistent with previous research. Members are significantly more likely than nonmembers to be black, to have higher incomes, to own their homes, to score high on the subjective interest scale, to know many of their neighbors, and to believe their neighbors would respond collectively to a collective problem. The fact that those with relatively few friends or relatives in the neighborhood are more likely to be members is perhaps inconsistent with the theoretical literature but somewhat consistent with McCourt's (1977) findings.' The negative coefMcCourt's (1977) study of about 30 women in a white working-class Chicago neighborhood found that having one's own parents in the neighborhood had no effect on level of participation in the neighborhood organization, while having one's husband's parents in the neighborhood had a strongly negative effect.

Control Variables
Race and gender were controlled in the analysis. There is substantial evidence that blacks participate more in community organizations than whites (Warren, 1975, 1974; Ahlbrandt and Cunningham, 1979; London and Hearn, 1977; Phillips, 1975). In addition, men have been found to be more active in some areas of community participation and women in others (Parkum and Parkum, 1980). Preliminary analyses found no significant effects for age, length of residence in the neighborhood, and marital status, so they are excluded from the reported analyses. People in the broadly defined "middle ages" of 35-60 have often been found to be more active (Parkum and Parkum, 1980), although Edwards (1977) reports finding no age effect on general social involvement, and McPherson and Lockwood (1980) use time-series data to demonstrate that the age difference ariscs because memberships accumulate over time, with younger people having higher rates of membership. There is usually a positive correlation betwen the length of time a person has lived in a local area and involvement in community activities (Bell and Force, 1956; Ross, 1972; McKenzie, 1981; Parkum and Parkum, 1980). There is no reported finding that marital status has any significant effect on community participation. Preliminary analysis also controlled for the number of other organizational memberships to determine whether active members of neighborhood organizations were simply

LOCAL COLLECTIVE ACTION

607

Table 1. Frequencies and Interrelations Among Types of Participation in Neighborhood Organizations Percent of Row G r o w in Column G r o w Number Total Sample Members Leaders Currently Active 1456 278 64 76 Current Member 19% 97% 100% Past Leader 4% 22% 25% Currently Active 5% 27% 30%

ficient for having children is counter to Wright and Hyman's (1958) work, but somewhat consistent with free time as an explanation for involvement. Turning to the contrast between active and token members, only two variables predict both indicators of active contributions: education and beliefs about the neighbors' willingness to act collectively. First, former leaders and current activists are more highly educated than other members. This result is in line with past research, and is probably due to the kinds of skills educated people acquire which make such activities easier (and therefore less costly) for them. Secondly, both former leaders and current activists are less optimistic than other members about the prospects for collective action on the part of their neighbors. Presumably they do not believe they can free ride on their neighbors' efforts. Several factors are different for former leaders than for current activists. Former leaders are more likely than other members to be black, to score high on the subjective interest scale, and to know many of their neighbors. By contrast, current activists are less likely to be homeowners, less likely to like their neighborhood, and more likely to have close social ties with the neighbors they know (although they

Table 2. Descriptive Statistics for Independent Variablesa Variable Race (white = 1, black = 0) Education (in years) Income (in $1000~) Has Children Employed Full Time Sex (female = 1, male = 0) Homeowner Subjective Interest Scale Liking for Neighborhood Scale Acquaintances Scale Close Ties Scale Neighbors' Likelihood of Collective Action Scale (range 2 to 10)
a

Mean .427 10.535 9.121 .593 ,544 ,466 .736 ,002 -.001 -.OOl ,000 8.069

Standard Deviation .495 3.369 4.493 .491 ,498 .499 .441 ,883 ,751 ,651 .708 1.608

Number of Observations = 1456.

know nonsignificantly fewer of them). The negative coefficient on homeownership is difficult to explain theoretically and may be nothing more than a random quirk of these data, but it remains when controls for age and length of residence in the neighborhood are included. Let us review the results for each set of predictors. The various measures of social ties in these data have no consistent effects on active participation, although they do have some meaningful patterns. First, members know more people than nonmembers, and leaders know more people than other memb e r s , consistent with the "weak ties" arguments. Secondly, members have fewer close ties in the neighborhood than nonmembers, but currently active members have m o r e close ties while generally liking the neighborhood less than inactive members. This suggests that current activism arises not from a generalized collective neighborhood spirit, but rather from particularistic ties and, perhaps, even a sense of distance from one's neighbors. Wilson (1973) argues that personalized exchanges are an important mechanism for producing local collective goods, although his description may apply more to acquaintances than close friends. Interests also have mixed effects on active participation. As expected, members are much more likelv than nonmembers to be homeowners anh to score high on the subjective interest scale. However, among members, the only significant positive effect is that of subjective interest on having been a leader. As discussed above, homeowners are surprisingly less likely than others to be currently active as members. We may say that interest in the collective good seems to move people from doing nothing to doing something, but interests do not seem to be critical for moving people from doing less to doing more. There is, however, some indication that people who take on leadership roles are especially interested in local issues. Costs of participation were not measured directly, but we assumed that people with more resources would experience lower opportunity costs than those with fewer resources. Time is the central resource for active participation in neighborhood groups, but indicators of free

AMERICAN SOCIOLOGICAL REVIEW

Table 3. Probit Analysis of Membership and Active Participation in Neighborhood Organizations Whole Sample Membership Standard Estimate Error -0.842* 0.100 0.020 0.014 0.021* 0.011 -0.218* 0.087 0.008 0.100 0.070 0.094 0.564* 0.108 0.177* 0.049 0.095 0.068 0.236* 0.073 -0i153* 0.061 0.108* 0.030 -2.261* 0.293 1205.0 1443 Members of Neighborhood Groups Only Has Been Leader Currently Active Standard Standard Estimate Error Estimate Error

Race (white) Education Income Has Children Full-Time Employment Gender (female) Homeowner Subjective Interest Liking Scale Acquaintances Scale Close Ties Scale Neigh Act Coll Constant G-Square D.F. * p < .05, one-tailed.

time-not being employed full time and not having children-failed to predict such participation. Members are less likely to have children than nonmembers, but neither of these variables is a significant predictor of leadership or current activism. Although these factors are often thought to be more salient for women, tests for interactions with gender found no significant coefficients. Furthermore, there was a nonsignificant tendency for men's activism to be lower when they were employed or had children, but for women's to be higher. Thus, such trends as there are run counter to the conventional wisdom that community activists are women who do not have the demands of jobs or children. Higher-income people have more financial resources, and more highly educated people have cultural resources and skills for organizational participation. Income is a weak predictor of membership, while education strongly distinguishes active from token members. This suggests that the salient resource limiting active participation is the skills involved, rather than money or time. Finally, although members are more optimistic than nonmembers about the prospects for neighborhood collective action, both former leaders and currently active members are more pessimistic than token members about such prospects. As a bit of a check on the generality of this finding, the same analysis was performed for parent-teacher associations; the coefficients for former leadership and current activism in PTAs are not significant, but they are negative. Active participants in community organizations seem definitely to be

more skeptical of their neighbors' cooperativeness than token members. CONCLUSIONS To summarize, two factors consistently distinguish active from token contributors to neighborhood organizations: active members are more highly educated and they are more pessimistic about their neighbors' willingness to make active contributions. Distinguishing among the types of active contributors, leaders are interested in local issues and know many people, while those currently active have negative feelings about their neighbors in general but have their closest friends or relatives in the area. The positive effect of education is well known, but the significance of activists' pessimism about their neighbors has not been previously recognized and merits further discussion. Some of the pessimism effect may be consequence rather than cause. People who get involved in community activities often experience frustration when they try to get others involved. Although the numbers involved are very small, the people in these data who had been leaders in past years but were not active in the current year had more pessimistic beliefs about the prospects for action than other active members. Optimism about collective action may be due to simple naivete: many people do not understand the collective-goods dilemma and are shocked when they try to organize collective action. Experience may teach people about the col-

LOCAL COLLECTIVE ACTION lective dilemma. But whether they are pessimistic from the start, or become so through experience, people who make active contributions have less faith in the collective spirit of their neighbors than people who are only token members. This pattern is easily understood within a general collective-action model. Active contributors make a noticeable difference in the provision of the public good. Furthermore, many local collective goods have the property of diminishing marginal returns, in that early contributions have much greater impact than later contributions of the same size. Under these conditions, rational individuals take account of the likelihood that the collective good will be provided through the efforts of others, and are less likely to contribute the more they believe others will. There is a kind of paradox of community life. People with the greatest sense of collective identity and positive regard for their neighbors may not absorb the costs of community activism because they assume that someone else will take care of the problems. The people who are willing to absorb these costs are often precisely those who have less respect and liking for their neighbors and more of a belief that if they want something done they will have to do it themselves. There is often a real tension between community activists and their communities. Of course, in some neighborhoods community activists are able to mobilize widespread activism whenever it is necessary and there are close ties between leaders and residents. But just as Olson's (1965) work made us realize that collective action is problematic, so we need to recognize that this is an exception to be explained rather than a natural or likely turn of events.
Edwards, John N. 1977 "Age and social involvement." Journal of Voluntary Action Research 6:127-32. Fireman, Bruce and William A. Gamson 1979 "Utilitarian logic in the resource mobilization perspective." Pp. 3-44 in Mayer N. Zald and John D. McCarthy (eds.), The Dynamics of Social Movements. Cambridge, MA: Winthrop. Flynn, John P, and Gene E. Webb 1975 "Women's incentives for community participation in policy issues." Journal of Voluntary Action Research 4:137-46. Granovetter, Mark 1973 "The strength of weak ties." American Journal of Sociology 78: 1360-80. 1982 "The strength of weak ties: a network theory revisited." Pp. 105-30 Peter V. Marsden and Nan Lin (eds.), Social Structure and Network Analysis. Beverly Hills, CA: Sage. Green, Gerson 1979 "Who's organizing the neighborhood? Community organizations: a report on their structure, finance, personnel, issues and strategies." Office of Community AntiCrime Programs, Law Enforcement Assistance Administration. Superintendent of Documents stock no. 027-000-00783-0. Henig, Jeffrey R. 1982 "Neighborhood response to gentrification: conditions of mobilization." Urban Affairs Quarterly 17:343-58. Holmes, Sandra 1979 "Middle-class women, paid work and volunteer work." Unpublished manuscript, Department of Sociology, Washington University. London, Bruce and John Hearn 1977 "Ethnic community theory of black social and political participation: additional support." Social Science Quarterly 57:883-91. McCarthy, John D. and Mayer N. Zald 1973 The Trend of Social Movements in America: Professionalization and Resource Mobilization. Morristown, NJ: General Learning Press. McCourt, Kathleen 1977 Working Class Women and Grass Roots Politics. Bloomington: Indiana University Press. McKenzie, W. M. 1981 "Citizen participation: what motivates it." Australian Journal of Social I s s u e s 16:67-79. McPherson, J. Miller and William G. Lockwood 1980 "The longitudinal study of voluntary association memberships: a multivariate analysis." Journal of Voluntary Action Research 9:74-84. Nanetti, Raffaella Y. 1980 "From the top down: government promoted citizen participation." Journal of Voluntary Action Research 9: 149-64. National Commission on Neighborhoods 1979 People Building Neighborhoods. Final Report. Case Study Appendix. Superintendent of Documents stock no. 052-003-00616-2.

REFERENCES Ahlbrandt, Roger S., Jr., and James V. Cunningham 1979 A New Public Policy for Neighborhood Preservation. New York: Praeger. Bell, W. and M. T. Force 1956 "Urban neighborhood types and participation in formal associations." American Sociological Review 21:25-34. Bolduc, Vincent L. 1980 "Representation and legitimacy in neighborhood organizations: a case study." Journal of Voluntary Action Research 9: 165-78. Dabrowski, Irene 1979 "Working-class women, paid work and volunteer work." Ph.D. dissertation, Department of Sociology, Washington University. Duff, Robert W. and William T. Liu 1972 "The strength in weak ties." Public Opinion Quarterly 36:361-66.

AMERICAN SOCIOLOGICAL REVIEW

O'Brien, David J. 1974 "Public goods dilemma and apathy of poor toward neighborhood organization." Social Service Review 48:229-44. 1975 Neighborhood Organization and InterestGroup Processes. Princeton: Princeton University Press. Oliver, Pamela 1980a "Rewards and punishments as selective incentives for collective action: theoretical investigations." American Journal of Sociology 85: 1356-75. 1980b "Summary of data collected at 1979 convention of National Association of Neighborhoods." Unpublished manuscript, Department of Sociology, University of Wisconsin. Oliver, Pamela, Gerald Marwell and Ruy Teixeira 1984 "Interdependence, group heterogeneity, and the production of collective action: a theory of the critical mass, I." Unpublished manuscript. Department of Sociology, University of Wisconsin. Olson, Mancur 1965 The Logic of Collective Action. Cambridge, MA: Harvard University Press. Parkum, Kurt H. and Virginia Cohn Parkum 1980 "Citizen participation in community planning and decision making." Pp. 153-67 in David Horton Smith, Jacqueline Macaulay and Associates (eds.), Participation in Social and Political Activities. San Francisco: Jossey-Bass. Pearce, Jone L. 1980 "Apathy or self interest? The volunteer's avoidance of leadership roles." Journal of Voluntary Action Research 9%-94. Phillips, W. M., Jr. 1975 "Educational policy, community participation, and race." Journal of Negro Education 44:257-67. Rich, Richard C. 1980a "A political-economy approach to the study of neighborhood organizations." American Journal of Political Science 2439-92. 1980b "The dynamics of leadership in neighborhood organizations." Social Science Quarterly 60570-87. Ross, J. 1972 "Toward a reconstruction of voluntary association theory." American Sociological Review 22:315-26. Salem, G. W. 1978 "Maintaining participation in community organizations." Journal of Voluntary Action Research 7: 18-27. Schoenberg, Sandra Perlman 1980 "Some trends in the community participation of women in their neighborhoods." Signs 5:261-68. Schoenberg, Sandra Perlman and Patricia Rosenbaum 1979 "Prerequisites to participation by the poor in an urban neighborhood." Paper presented at annual meeting of American Sociological Association. Sharp, Elaine 1978 "Citizen organizations in policing issues and crime prevention: incentives for participation." Journal of Voluntary Action Research 7:45-58. Smith, Constance and Anne Freedman 1972 Voluntary Associations: Perspectives on the Literature. Cambridge, MA: Harvard University Press. Smith, David Horton 1981 "Altruism, volunteers, and volunteerism." Journal of Voluntary Action Research 10:21-36. Snow, David A., Louis A. Zurcher, Jr., and Sheldon Ekland-Olson 1980 "Social networks and social movements: a micro-structural approach to differential recruitment." American Sociological Review 45:787-801. Stinson, Thomas F. and Jerome M. Stam 1976 "Toward an economic model of volunteerism: the case of participation in local government." Journal of Voluntary Action Research 552-60. Thursz, Daniel 1972 "Community participation: should past be prologue?" American Behavioral Scientist 15:733-48. Till y, Charles 1978 From Mobilization to Revolution. Reading, MA: Addison-Wesley. Uzzell, David L. 1980 "Conflicting explanations of participatory group membership." Journal of Voluntary Action Research 9:203-210. Vedlitz, Arnold and Eric P. Veblen 1980 "Voting and contacting: two forms of political participation in a suburban community." Urban Affairs Quarterly 16:31-48. Verba, Sidney and Norman H. Nie 1972 Participation in America. New York: Harper & Row. Wandersman, Abraham 1981 "A framework of participation in community organizations." Journal of Applied Behavioral Science 17:27-58. Warren, Donald 1. 1974 "The linkage between neighborhood and voluntary association patterns: a comparison of black and white urban populations." Journal of Voluntary Action Research 3:l-17. 1975 Black Neighborhoods: An Assessment of Community Power. Ann Arbor: The University of Michigan Press. Warren, Rachelle B. and Donald I. Warren 1977 The Neighborhood Organizer's Handbook. Notre Dame, IN: The University of Notre Dame Press. Wilson, James Q. 1973 Political Organizations. New York: Basic Books. Wright, C. R. and H. Hyman 1958 "Voluntary association memberships of American adults: evidence from national sample surveys." American Sociological Review 23:284-94.

http://www.jstor.org

LINKED CITATIONS
- Page 1 of 2 -

You have printed the following article: "If You Don't Do it, Nobody Else Will": Active and Token Contributors to Local Collective Action Pamela Oliver American Sociological Review, Vol. 49, No. 5. (Oct., 1984), pp. 601-610.
Stable URL:
http://links.jstor.org/sici?sici=0003-1224%28198410%2949%3A5%3C601%3A%22YDDIN%3E2.0.CO%3B2-G

This article references the following linked citations. If you are trying to access articles from an off-campus location, you may be required to first logon via your library web site to access JSTOR. Please visit your library's website or contact a librarian to learn about options for remote access to JSTOR.

References
Urban Neighborhood Types and Participation in Formal Associations Wendell Bell; Maryanne T. Force American Sociological Review, Vol. 21, No. 1. (Feb., 1956), pp. 25-34.
Stable URL: http://links.jstor.org/sici?sici=0003-1224%28195602%2921%3A1%3C25%3AUNTAPI%3E2.0.CO%3B2-8

The Strength of Weak Ties Mark S. Granovetter The American Journal of Sociology, Vol. 78, No. 6. (May, 1973), pp. 1360-1380.
Stable URL:
http://links.jstor.org/sici?sici=0002-9602%28197305%2978%3A6%3C1360%3ATSOWT%3E2.0.CO%3B2-E

Rewards and Punishments as Selective Incentives for Collective Action: Theoretical Investigations Pamela Oliver The American Journal of Sociology, Vol. 85, No. 6. (May, 1980), pp. 1356-1375.
Stable URL:
http://links.jstor.org/sici?sici=0002-9602%28198005%2985%3A6%3C1356%3ARAPASI%3E2.0.CO%3B2-M

http://www.jstor.org

LINKED CITATIONS
- Page 2 of 2 -

Educational Policy, Community Participation, and Race W. M. Phillips, Jr. The Journal of Negro Education, Vol. 44, No. 3, The Continuing Crisis of Urban Education. (Summer, 1975), pp. 257-267.
Stable URL:
http://links.jstor.org/sici?sici=0022-2984%28197522%2944%3A3%3C257%3AEPCPAR%3E2.0.CO%3B2-H

A Political-Economy Approach to the Study of Neighborhood Organizations Richard C. Rich American Journal of Political Science, Vol. 24, No. 4. (Nov., 1980), pp. 559-592.
Stable URL:
http://links.jstor.org/sici?sici=0092-5853%28198011%2924%3A4%3C559%3AAPATTS%3E2.0.CO%3B2-2

Some Trends in the Community Participation of Women in Their Neighborhoods Sandra Perlman Schoenberg Signs, Vol. 5, No. 3, Supplement. Women and the American City. (Spring, 1980), pp. S261-S268.
Stable URL:
http://links.jstor.org/sici?sici=0097-9740%28198021%295%3A3%3CS261%3ASTITCP%3E2.0.CO%3B2-R

Social Networks and Social Movements: A Microstructural Approach to Differential Recruitment David A. Snow; Louis A. Zurcher, Jr.; Sheldon Ekland-Olson American Sociological Review, Vol. 45, No. 5. (Oct., 1980), pp. 787-801.
Stable URL:
http://links.jstor.org/sici?sici=0003-1224%28198010%2945%3A5%3C787%3ASNASMA%3E2.0.CO%3B2-Z

Voluntary Association Memberships of American Adults: Evidence from National Sample Surveys Charles R. Wright; Herbert H. Hyman American Sociological Review, Vol. 23, No. 3. (Jun., 1958), pp. 284-294.
Stable URL:
http://links.jstor.org/sici?sici=0003-1224%28195806%2923%3A3%3C284%3AVAMOAA%3E2.0.CO%3B2-Q