The sequencing of social movements Debra C Minkoff American Sociological Review; Oct 1997; 62, 5; ABI/INFORM Global pg. 779

THE SEQUENCING OF SOCIAL MOVEMENTS*

Debra C. Minkoff Yale University

Conventional accounts of protest cycles posit a demonstration effect-suc­ cessful protests incite other constituencies to activism. I offer an alternative theon' that builds on population ecology models of organizational behavior. I argue that the expansion of social movement organizations, or organiza­ tional density, is also an essential component of protest cycles. Multivariate analyses of the effects of civil rights protest and organizational growth on feminist protest and organizational foundings between 1955 and J 985 dem­ onstrate that organizational density promotes the diffusion of protest. Pro­ test also engenders activism by others, but only under favorable political conditions. This implies that an enduring organizational niche and political allies in power are necessary for protest to spread be..... ond single movements and create protest opportunities for other challengers.

n the 1960s and early 1970s, the politics to test the validity of general arguments re­ I and culture of collective action in the garding the structure of protest cycles. I United States experienced dramatic shifts. I examine the dominant model of protest Groups historically shut out of the political cycles developed by Tarrow (1991, 1994) to arena gained access to an extent previously provide a more precise specification that unknown (Oberschall 1978; Za1d 1988). The builds on organizational models of social received wisdom-both popular and aca­ movement development (McCarthy and demic-is that the U.S. civil rights move­ Zald 1977; Minkoff 1994, 1995). I argue ment was critical in shaping the trajectories that protest cycles are the visible manifesta­ of other movements for social change be­ tion of the interaction between organiza­ cause it initiated a widespread "cycle of pro­ tional trajectories and protest-event trajecto­ test" (Tarrow 1991, 1994) that created oppor­ ries. In contrast, Tarrow's perspective em­ tunities for activism by other constituencies, phasizes the generative role of a protest dy­ such as women, progressive students, and namic without adequately considering the lesbians and gay men (McAdam 1995). Al­ role of organizational dynamics. To sub­ though this is an empirically plausible inter­ sume both protest processes and organiza­ pretation of events, there have been no sys­ tional processes under the singular rubric of tematic efforts to model this relationship or "protest cycles" masks the underlying causal dynamic. 2 Tarrow (1991, 1994) ar- * Direct all correspondence to Debra C. Minkoff, Department of Sociology, Yale Univer­ I Koopmans (1993) may be an exception, al­ sity. P.O. Box 208265, New Haven. CT 06520 though his analysis is essentially descriptive and (debra.minkoff@yale. edu). An early version of aggregates protest across a range of groups with­ this paper was presented at the 1995 annual meet­ out considering intermovement influences. ing of the American Sociological Association, 2 Following Tarrow (1994), a cycle of protest Washington. D.C. Steven Andrews, Bob is "a phase of heightened collective conflict and Edwards, Robert Jenkins, David Meyer and contention across the social system that includes: Sidney Tarrow provided invaluable comments. a rapid diffusion of collective action from more Thanks also [0 the ASR reviewers and Deputy mobilized to less mobilized sectors: a quickened Editor Charles Tilly for their extremely useful pace of innovation in the forms of contention: comments. Craig Jenkins, Rachel Rosenfeld, new or transformed collective action frames: a Kathryn Ward, and Anne Costmn generously combtnation of organized and unorganized par­ made their data available for this analysis. ticipation: and sequences of intensified interac-

American Sociological Review, 1997, Vol. 62 (October:779-799) 779

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 780 AMERICAN SOCIOLOGICAL REVIEW

gues that a "demonstration effect" is associ­ portant. Political space is initially opened up ated with successful or effective collective by organizational expansion, but later be­ action: Protest by early risers encourages comes overcrowded and less open to new en­ protest by other participants because early trants. Organizational density is thus critical protest signals the potential vulnerability of throughout the protest cycle-initially open­ elites to challenges. Therefore, it is the in­ ing opportunities for protest and organiza­ formation provided by successful protest tion-building, but later closing these oppor­ that drives the development of broad-based tunities. cycles (Conell and Cohn 1995). In contrast I examine contrasting specifications of the to Tarrow, I argue that trajectories of protest protest cycle using an event-history analysis cycles are jointly determined by increases in of the influence of civil rights protest and or­ the rates of protest and increases in the den­ ganizational density on feminist protest and sity of social movement organizations.' organization-building between 1955 and Increases in organizational density accel­ 1985. These two movements are commonly erate the diffusion of activism across mul­ described as "initiator" and "spin-off' move­ tiple constituencies through a transfer of in­ ments (McAdam \995). The data suggest formation and the construction of a niche or that at the height of a protest cycle a sus­ resource infrastructure. Just as protest activ­ tained sequencing of social movement devel­ ity broadcasts the effectiveness of these tac­ opment is linked to organizational expansion. tics, increases in organizational density dem­ Precedent-setting movements like the civil onstrate movement success in sustaining or­ rights movement open up opportunities via ganizations. This demonstration effect em­ organization-building and protest that allow phasizes organizational processes over pro­ other constituencies to create movements of test dynamics. Increases in organizational their own. The dominant conceptualization density also promote the expansion of a of protest cycles as predominantly cognitive niche-a resource base-for social move­ therefore obscures the important organiza­ ment activity that provides an enduring in­ tional processes undergirding the opportunity frastructure for the diffusion of protest when for broad-based protest. political conditions are favorable. Inter­ organizational competition, however, may THE STRUCTURE OF PROTEST limit future activism. This alternative model CYCLES suggests that it is the organizational compo­ nent of the opportunity structure that is im- The defining characteristic of a protest cycle is the rapid and widespread diffusion of pro­ test behavior across a wide variety of groups. tion between challengers and authorities which The basic argument is that changes in politi­ can end in reform. repression and sometimes cal conditions position "early risers" to take revolution" (p. 153). advantage of the new opportunities for activ­ 3 I define a social movement as a collective ef­ ism (Tarrow 1991, 1994). If these early ris­ fort (0 change the social structure that uses extra­ ers are successful, a demonstration effect en­ institutional methods at least some of the time (Burstein, Einwohner, and Hollander 1995). So­ courages protest by other groups because cial movement organizations (SMOs) are formal success signals the vulnerability and respon­ organizations that attempt to Implement move­ siveness of elites. Successful protest also ment goals (McCarthy and Zald 1977). Organi­ broadcasts that certai n forms of collective ;:ational density refers to the number of organIza­ action are viable and that the interests of tions active in ~ movement or in the sector gener­ elites are at risk (McAdam, Tarrow, and Tilly ally. Protest refers to the use of noninstitu­ 1996). Movements that arise later are "easy tionalized means to influence authorities and riders," who effectively take "advantage of a elites (Jenkins 1987). Protest events include ral­ political opportunity structure that others lies, marches, sit-ins, boycotts, and various forms struggled to open up" (Tarrow Ac­ of civil disobedience. I use the terms protest, ac­ 1991: 84). tivism, and insurgency to represent such extra-or­ cordingly, the information transmitted by ganizational collective action. The emphasis is on successful insurgencies becomes the key in­ the actions, not on the organizations, groups. or dicator of political opportunity for emerging networks that initiate them. movements and drives the development of

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. SEQUENCING OF SOCIAL MOVEMENTS 781

broad-based protest cycles.4 Prior traditions for broad diffusion.' According to McAdam, of organization provide networks of support movement organizations and associational and diffusion during the movement's emer­ networks are the conduits for cultural diffu­ gent phase (Tarrow 1995), but movement or­ sion, especially early in the protest cycle ganizations do not directly affect the inter­ when existing ties among activists provide movement diffusion of activism. direct points of contact for the spread of At the peak of the protest wave, competi­ movement ideas and political socialization. tion for members and supporters leads social Most important, activist networks facilitate movement organizations (SMOs) to escalate processes of social construction that encour­ protest. New organizations emerge and con­ age the "attribution of similarity" necessary ventional interest groups become active in for new constituencies to identify with the the social movement arena. The creation of initiator movement enough to mimic its ef­ new SMOs produces interorganizational forts (McAdam 1995:233). Direct ties also competition as groups try to outdo each other make other tactical and organizational inno­ by adopting radical forms of action in com­ vations available to a wider group of poten­ petition for public support and attention tial activists, but the information function is (Meyer 1993, Meyer and Imig 1993). This central. Later in the protest cycle, McAdam "competitive spiral" is the dynamic behind suggests that the need for direct ties is re­ the protest wave, encouraging tactical inno­ placed by the general availability of the new vation and expanding "the definition of what "organizing template," which explains the constitutes a challenge" (Tarrow 1991:54). widening "ecological scope" (Snow and Increasing radicalization ultimately discour­ Benford 1992) characteristic of protest ages continued participation by most sup­ cycles. Network theories do not explain, porters, although it may also prompt limited however, how protest cycles contract or de­ concessions from elites as they try to defuse cline. the protest wave. McAdam (1995) argues that cultural diffu­ DENSITY-DEPENDENT PROCESSES sion and adaptation drive protest cycles. Ini­ AND PROTEST CYCLES tiator movements, rather than conferring some sort of political leverage, provide a cul­ One way to understand the relationship be­ tural or cognitive signal to others. Initiator tween organizational processes and the de­ movements are the source of new cultural velopment of protest cycles is suggested by forms-of insurgent consciousness, cogni­ density-dependence theory, a part of popula­ tive liberation, injustice frames-and spin­ tion ecology theories of organizational be­ off movements are the adopters and avenues havior (Hannan and Carroll 1992; Hannan and Freeman 1989), Density-dependence theory specifies a curvilinear relationship be­ 4 Note that the protest-cycle model emphasizes tween the number of active organizations successful protest. Presumably, protest that is (density) and organizational founding rates quickly repressed or gains little publicity signals and failure rates (Hannan and Freeman 1989; a less favorable environment for extra-institu­ Hannan and Carroll 1992). A rise in density tional tactics. Conell and Cohn (1995) look at strike behavior in France between 1890 and 1935. establishes a favorable environment for positing that strikes stimulate further strikes group formation by opening up an organiza­ through consCIousness-raising. by defining occa­ tional niche-what Aldrich and Marsden sions for action. and by signalling the relative (1988) define as a resource combination suf­ power of authorities and strikers. Successful ficient to support an organizational form. strikes tend to increase strike rates. but under fa­ High rates of organizational formation and vorable political conditions workers may discount the information provided by unsuccessful strikes. Concll and Cohn' s analysis emphasizes the diffu­ 5 "Initiator movements" are what Tarrow sion of protest (strikes) across a constituency (1994) calls "early risers"; "spin-off movements" (workers) that occupies a relatively similar struc­ are "easy riders." I use each pair of terms inter­ tural position. whereas the interesting aspect of changeably and introduce the term "later en­ Tarrow's protest-cycle theory is the diffusion of trants" to refer to movements that develop later protest behavior across distinct social groups. in the protest cycle.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 782 AMERICAN SOCIOLOGICAL REVIEW

survival affect the survival chances of indi­ of SMOs establish a resource and institu­ vidual organizations that locate within the tional space for protest by the early risers. niche and adopt the dominant model of orga­ This organizational infrastructure provides nization. Survival demonstrates to elites that co-optable networks for mobilization (Free­ new forms of organizing are legitimate man 1973) and bloc recruitment (Oberschall (Hannan and Freeman 1989; Yuchtman and 1973), supplies activists prepared to orga­ Seashore 1967). In this way, organizational nize related efforts (McAdam 1995; Meyer density promotes and shapes new organiza­ and Whittier 1994), and makes material re­ tional activity. As a sector becomes estab­ sources available for organizational start­ lished, however, increases in density encour­ ups and ongoing projects (Walker 1983). In­ age interorganizational competition for re­ creased density also legitimizes participat­ sources, limiting further access to resources ing social movements, bringing in more and depressing new organizational activity third-party support (financial and political) (Hannan and Carroll 1992). and improving their "bargaining positions" The few studies that relate to social move­ (Hannan and Carroll 1992; Minkoff 1994; ment and interest organizations indicate that Yuchtman and Seashore 1967). rates of new group formation are tied to the Increased organizational density also growing density of social and political orga­ weakens the protest cycle by increasing nizations. For example, Walker (1983) found interorganizational competition (Koopmans that the explosion of citizen's groups in the 1993; Tarrow 1991). Having encouraged oth­ early 1970s was related to start-up support ers to enter the social movement arena, ini­ from existing groups that had already gained tiator movements may have to compete with ground in the interest-group sector. Minkoff more groups for resources, public attention, (1995) demonstrated that density-dependence and political support. New constituencies and encouraged the founding of national women's their organizations may later replace early and racial/ethnic organizations, both in the risers. Alternatively, initiator movements, as aggregate and across constituency bound­ niche-builders, may be protected by estab­ aries. For example, civil rights organizational lished streams of support and greater institu­ density promoted the founding of Hispanic­ tional legitimacy. That is, established groups American and Asian-American organizations may have a competitive advantage with re­ between 1955 and 1985. These findings cor­ spect to resource procurement and survival roborate Olzak and West's (1991) results ex­ (Stinchcombe 1965). Rather than a general plaining the founding of ethnic newspapers set of opportunities that increases activism during an earlier historical period. Although across a variety of constituencies, then, there a recent study of interest-group populations may be a more discrete, competitive sequenc­ in six states found no significant relationship ing of social movements.6 between density and lobbying-group entry rates, it noted that the density of interest or­ ganizations usually was correlated with high 6 One criticism of organizational ecology is that it cannot incorporate the insights of political op­ rates of organizational failure or exit from the portunity models that argue that the environment political scene (Gray and Lowery 1995). comprises actors whose actions affect organiza­ Applying the idea of density-dependence tional formation and protest cycles. This requires to protest cycles, I suggest that the influence an analysis of the actions of actors in the envi­ of early risers on subsequent social move­ ronment rather than an examination of aggregate ments is primarily organizational: The results of such actions and how they influence growing density of early-riser organizations subsequent behavior of a movement. Also, the ef­ encourages other constituencies to form or­ fects of countermovement mobilization and re­ ganizations. In turn, opportunities for pro­ pression may be captured in the observed rate of protest so that these effects are built into the as­ test increase because the organizational in­ sociated demonstration effect. Koopmans (1993) frastructure is less vulnerable to changes in criticizes ecological models on other grounds, political conditions (Minkoff 1994). The so­ suggesting that the "growth of SMOs is more a called demonstration effect has a more tan­ product than a cause of protest wa\e" (p. 41). His gible form: The survival of precedent-set­ descriptive time-series data are inadequate to test ting organizations and the growing density this claim.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. SEQUENCING OF SOCIAL MOVEMENTS 783

THE SEQUENCING OF Broad macro-demographic, economic, and SOCIAL MOVEMENTS: political trends were also significant for THE CIVIL RIGHTS MOVEMENT women's mobilization in the following de­ AND THE WOMEN'S MOVEMENT cade-for example, the increase in the labor force participation of White women, chang­ The prototypical example of intermovement ing family patterns, and the growing number diffusion of protest is the relationship be­ of White women and middle-class women tween the U.S. civil rights movement and the with post-secondary degrees (Ferree and contemporary feminist movement: "If there Hess 1985; Klein 1984). Klein (1984) em­ was ever a movement that seemed dependent phasizes the growing feminist consciousness on gates of opportunity that were opened by among women and their increased electoral others, it was the American women's move­ potential. Costain (1992) emphasizes the fa­ ment of the 1960s" (Tarrow 1994: 180).7 How cilitative role of the federal government and did the civil rights movement "open the shifts in the electoral balance of power that gates" for the feminist movement? Was the encouraged political parties to compete for expansion of the cycle related to an informa­ women's votes and support reform tion effect based on successful protests, or legislation. was it the result of an organizational effect McAdam (1995) also suggests that the gen­ linked to an expanding niche for both protest eral opportunities provided by the cycle of and organization? protest initiated earlier by Blacks were im­ The first stirring of the contemporary portant for the emergence of the contempo­ women's movement coincided with the most rary women's movement. There is no ques­ active phase of the civil rights insurgency in tion that women were operating in a legal and the mid-1960s (Ferree and Hess 1985; political environment they shared with civil McAdam 1982). The civil rights movement rights groups. Both constituencies were in­ emerged in the mid-1950s as a result of cluded in key legislation (e.g., Title VII of changes in the political opportunity structure the Civil Rights Act of 1964, dealing with that encouraged the growth of indigenous employment), and both groups pressured the southern organizations (e.g., the NAACP, federal government for legislative solutions Black churches, historically Black colleges to problems of inequality (Burstein 1985). and universities) and provided leverage for The sequencing of the civil rights move­ African Americans in the electoral arena. By ment and the new feminist movement took a the beginning of the 1960s, these trends pro­ variety of direct and indirect forms. The im­ duced a political consciousness and opened petus for much early feminist activism was prospects for protest and the formation of the sexism experienced by women in the civil new organizations (McAdam 1982; Morris rights and student movements of the 1960s 1984). (Evans 1979). Their deep frustration with the contradictions between progressive politics and practice led them to organize indepen­ 7 The roots of the contemporary feminist move­ dently against male-dominated institutions. ment in the civil rights movement have been dis­ Feminists from this younger, women's lib­ cussed extcnsi vely (Evans 1979; Gitlin 1987; eration branch of the movement formed McAdam 1988). In a recent article, Meyer and Whittier (1994) also document the tactical and smaller, less formal, groups that pursued cultural "spillover effects" between the U.S. more radical action, provided resources di­ women's movement and the peace movement, rectly to women, and engaged in conscious­ and Whitticr (1995) suggests links between the ness-raising activities (Ferree and Hess 1985; feminist movement and the peace, environmen­ Freeman 1973; Klein 1984). tal, and lesbian and gay movements of the 1980s. The older branch of the movement was ini­ McAdam (1988) argues that thc student movc­ tiated in 1966 following the Third National ment grew out of the experiences of White activ­ Conference of Commissions on the Status of ists in Freedom Summer. Staggenborg (1991) Women. Conference officials would not al­ notes the indebtedness of the pro-choice move­ ment to both the women's movemcnt and the mo­ low delegates from state commissions to put bilization potential provided by the height of the forward a resolution demanding that the 1960s cycle of protest. Equal Employment Opportunity Commission

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 784 AMERICAN SOCIOLOGICAL REVIEW

guarantee that the sex provision of Title VII stitutions as the military and the Catholic of the 1964 Civil Rights Act be enforced to Church. At the national level, women's move­ the same extent as the race provision. A ment organizations, many of which were group of women at the conference met and rooted in the protest cycle precipitated by the formed the National Organization of Women civil rights movement, currently outnumber (NOW): "[S]everal angry women agreed that civil rights groups and Black organizations the time had come to organize a group that (Minkoff 1995). could lobby for women in the same way that the National Association for the Advance­ ment of Colored People (NAACP) worked on RESEARCH QUESTIONS AND HYPOTHESES behalf of blacks" (Ferree and Hess 1985 :54). Costain (1992) documents the rapid develop­ This brief account of the joint development ment and coalition activities of groups such of the civil rights movement and the feminist as NOW, the Women's Equity Action movement suggests the significant influence League, and the National Women's Political of the protest cycle initiated by the civil Caucus, all of which adopted bureaucratic rights movement. To what extent can the de­ models of organization and pursued moder­ velopment of African-American activism be ate goals and tactics. credited with the subsequent increase of The civil rights movement was clearly a women's activism in the late 1960s and early reference point for the initiation of both 1970s? Was this a protest effect or an organi­ strands of the women's movement. The con­ zation effect? Did the growth of the women's temporary women's movement also owed movement playa role in the decline of the much of its resources and institutional access civil rights movement that originally opened to traditional White women's associations, opportunities for it? Figure I summarizes the where many leaders of the older strand predicted influences across the civil rights gained leadership skills (Evans 1979; Gelb and feminist movements. and Paley 1982; Klein 1984; Taylor 1989). According to the protest dynamic model, In addition, in the early 1970s traditional protest by early risers has a positive influ­ White women's groups like the League of ence on protest events and organizational for­ Women Voters, the American Association of mation by others. Black protest events are University Women (AAUW), and the Gen­ the key independent variable. Thus, I expect: eral Federation of Women's Clubs joined with newly formed women's rights groups to H( Increases in civil rights movement activ­ promote passage of the Equal Rights Amend­ ity promoted protest by feminists and the ment (Gelb and Paley 1982). formation of women's organizations. Late in the 1970s, the women's movement moved away from local social movement ac­ The organization dynamic model posits tivity and established a dense and diverse that the growing density of civil rights orga­ field of national organizations (Costain nizations, not the protest events themselves, 1992). This national organizational system is promoted women's activism and organiza­ made up of mass membership associations tional activity. Increased density also in­ like NOW; specialized groups, including liti­ creases interorganizational competition for gation and research groups like the Center for resources, which limits the amount of new Women Policy Studies; single-issue groups activity that can be supported over the course like the National Abortion Rights Action of the protest cycle. This suggests an alter­ League; traditional women's organizations native hypothesis: like the AAUW; and an electoral campaign sector comprising political action committees H2: Increases in African-American organiza­ sponsored by large women's organizations, tional density promoted increases in along with campaign-oriented groups like the feminist activism and organizational ac­ Women's Campaign Fund -(Spalter-Roth and tivity; later in the cycle, this density ef­ Schreiber 1995). Katzenstein (1990) also fect was reversed as a result of inter­ shows that women have been mobilizing "un­ organizational competition (i.e., the den­ obtrusively" within such male-dominated in- sity function is concave).

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. SEQUENCING OF SOCIAL MOVEMENTS 785

Direction Model Effect of: Effect on: of Effect

Protest Dynamic Protest events by early risers Protest events by others Model + Black protest events Feminist protest events and feminist organizational foundings

Organization Organizational density of Protest events by others Dynamic Model early rise rs n

Black organizational density Feminist protest events and feminist organizational foundings

Competitive Protest events by later entrants Protest events by early risers Sequencing - Model Feminist protest events Black protest events and and feminist organizations Black organizational foundings

Figure 1. Characteristics of Models of the Feminist Movement and the Civil Rights Movement

The density of African-American organiza­ count data (Barron 1992; Hannan 1991; tions is the independent variable in this for­ Hannan and Freeman 1989; Olzak and mulation; feminist protest events and organi­ Shanahan 1996). The baseline Poisson model zational founding rates are the dependent takes the form: variables. Pr(Y = )'i) = e-LiL/i / )'i! , I also argue that intermovement competi­ i tion leads to a decline in the protest cycle. where )" is the number of organizations The development of the feminist movement formed or protest events occurring in year i. may have constrained subsequent Black pro­ This is a one-parameter distribution with the test as these later entrants competed with the mean and variance of Yi equal to Li. To incor­ initiator movements for resources, public at­ porate exogenous variables Xi] (j = 1, ... K), tention, and political support. This leads to a including a constant, the parameter Li is third prediction: specified to be

H3a:Increases in the number of feminist or­ Li = exp (X, B). ganizations and protest events decreased The exponential function ensures that the Black activism as the protest cycle de­ founding rate is a nonnegative integer veloped. (Cameron and Trivedi 1986). A competing hypothesis is that civil rights In the Poisson formulation. the assumption and African-American groups, as initiators, that the conditional mean and variance of Yi occupied a protected position in the niche. given Xi are equal fails to account for over­ This suggests: dispersion (when the variance exceeds the mean) and can result in a downward bias of H3b: Increases in the number of feminist or­ the standard errors for coefficients for the ganizations and protest events had no ef­ exogenous variables (Barron 1992; Cameron fect on Black activism or organizational and Trivedi 1986; Hannan 1991). A common activity over the course of the protest approach to this problem has been to esti· cycle. mate the event count using negative binomial regression, which is a generalization of the Poisson model. Choice of model is based on METHODS standard tests of fit for nested models. In the I model the organizational founding rate and analyses that follow, I present negative bino­ the protest rate using Poisson regression mial or Poisson estimates, depending on analysis, which is appropriate for yearly which model fits the data best. The event-

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 786 AMERICAN SOCIOLOGICAL REVIEW

count models are estimated using Limdep 5.1 (Kriesi 1996) and voluntary associations face (Greene 1988). unique problems of resource acquisition (Knoke 1989). Short-lived and radical groups may be underrepresented in the Encyclope­ DATA dia, but these groups may be visible in the The data come from a database of national media (which is one source of the directory's women's, African-American, Asian-Ameri­ information). Some groups may understate can, and Hispanic-American organizations in the institutional challenge they pose to the existence between 1955 and 1985. This data­ status quo. However, because the directory is base was constructed from the first 23 edi­ intended as a resource for potential members tions of the Encyclopedia of Associations, there are incentives to provide relatively ac­ Volume I, National Organizations (Gale Re­ curate, if somewhat benign, self-descrip­ search Company, 1955-1988). This directory tions. Despite a tendency for the Encyclope­ provides descriptive information on the strat­ dia to list more mainstream groups, the gen­ egies, objectives, and membership of a broad eral contours of African-American and range of national associations, interest women's social movement and interest orga­ groups, and social movement organizations. nizational populations are adequately repre­ I coded each edition of the Encyclopedia sented. Minkoff (1995) describes additional separately to generate an over-time census of strengths and limits to data from the Ency­ the women's and racial/ethnic organizational clopedia of Associations for developing sector that tracked entries, exits, and changes event-history and time-series data. in organizational attributes (Minkoff 1995, app. A). Information was collected on 402 national MEASURES women's organizations. Women's organiza­ Organizational Foundings tions are defined as those groups stating that their primary objective is women's rights or The year of founding is reported by the orga­ the status of women in U.S. society. Fifty of nization. Based on this information, I con­ these were organizations of women of color structed yearly counts of the number of new that emphasize gender issues. Information organizations for each constituency. Organi­ was also collected on 342 African-American zations for which the founding date was national organizations, including general missing are excluded when constructing civil rights organizations as well as groups yearly measures. This includes 18 (5 percent) devoted to improving the status of African of the 342 civil rights organizations and 38 Americans in U.S. society. The selected or­ (9 percent) of the 402 women's groups, lead­ ganizations include mainstream advocacy ing to a slightly higher underestimation of and service organizations like NOW, the the founding rate for women. Most women's NAACP, and the National Urban League, as organizations with missing founding dates well as protest and cultural groups organized enter the database after 1974, but civil rights around more activist agendas like the Black organizations lacking this information are Panthers and Women Make Movies. This spread fairly evenly over the 1955-1985 pe­ broad definition of the two "social movement riod. Because the formation of women's industries"-all organizations mobilized groups peaks after 1970, the excluded cases around preferences for change (McCarthy should not make a significant difference in and Zald 1977)-captures the range of alter­ the pattern of results. 8 natives used by marginalized groups to pur­ sue reform and social change. The database is limited to membership or­ K An alternative to excluding cases with miss­ ing data would be to use year of first mention in ganizations. It excludes staff-run nonprofit the Encyclopedia. This. however, misspecifies the advocacy groups and service agencies, along concept of organizational founding because it as­ with government bodies that may promote sumes that organizations are not active until they women's and civil rights issues. Direct par­ are listed in the directory. On average, it can take ticipation of constituents is a distinguishing three years for new organizations to be located feature of social movement organizations and listed in the Encyclopedia (Minkoff 1995).

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. SEQUENCING OF SOCIAL MOVEMENTS 787

Protest Events Organizational Density

Civil rights movement events are measured Organizational density is measured sepa­ as the total number of civil rights and Black rately for women's organizations and civil movement-initiated events occurring per rights organizations. Density is calculated as year as coded from the New York Times An­ the total number of organizations active at nual Index (Jenkins and Ekert 1986; the end of the prior year, plus the number of McAdam 1982). Feminist protest events re­ new entrants, minus the number of groups fers to the total number of pro-feminist­ that exited (either because they failed, be­ movement initiated events occurring each came inactive, or did not respond to repeated year, also coded from the New York Times In­ requests for updated information). Entry into dex (Rosenfeld and Ward 1996). This in­ the population is defined as the reported year cludes institutionalized insurgency (e.g., of founding, or, if such information is miss­ campaigns to enact legislation, register vot­ ing, when an organization entered the sample ers, petition); boycotts and strikes; mass ac­ between editions of the Encyclopedia. Using tions, such as protests, sit-ins, and demon­ year of first mention for missing data on strations; illegal or violent actions; and founding is defensible in this case, because women's politically oriented cultural no assumptions are made about whether the events.9 When included as control variables, organization was active prior to entering the the measures were lagged one or two years. sample. The density of feminist organiza­ In the analysis of the women's movement, tions may be underestimated before 1974 be­ protest events were lagged two years; in the cause most organizations without founding civil rights analysis, they were lagged one dates entered the database after 1974. A qua­ year. There was a theoretical basis for this dratic density term was used to test for the decision-early in the cycle it takes time for hypothesized curvilinear relationship be­ the demonstration effect to stimulate dis­ tween organizational density and protest cernible increases in subsequent protest events and organizational foundings. The events, whereas the effect is more immedi­ squared density terms are divided by 100 to ate later in the cycle. reduce the number of places after the deci­ mal point in reported coefficients. Density is lagged two years, except when modeling the ~ Data on Black protest events were made avail­ effect of the density of women's organiza­ able by Craig Jenkins; data on women's protest tions on civil rights protests in which case it events were provided by Rachel Rosenfeld and Kathryn Ward. The original coding of both vari­ is lagged one year to capture a more proxi­ ables relied on McAdam's (1982) framework. mate effect. The measure of civil rights activism is from a re­ cently recoded and updated time-series used in Control Variables Jenkins and Ekcrt (1986). There are limitations to using media sources for event data (McCarthy, Control variables provide a baseline model McPhail. and West 1996; Olzak 1989), but for the from which to observe the influence of purpose of indexing visible protest activity. the lagged measures of organizational density New York Times Index is sufficient. In fact, its and number of protest events. These mea­ tendency to report only large national demonstra­ sures are derived from accounts of the emer­ tions (McCarthy et al. 1996) provides a strong gence of the civil rights movement and the test of the demonstration effect in protest cycles. contemporary women's movement (Burstein I use a broad specification of protest events that 1985; Costain 1992; Klein 1984; McAdam combines both conventional and contentious 1982; Rosenfeld and Ward 1991. 1996). events (Tarrow \996). This is consistent with the When analyzing the women's movement, I conceptualization of protest cycles as heightened phases of "organized and unorganized participa­ control for the female labor force participa­ tion," A narrower specification of the model tion rate (lagged one year) (Economic Report might use measures of "genuine" (presumably, of the President 1989:table B-36) and for contentious) protest events. An empirical test of public opinion favoring women's rights. these different conceptualizations is beyond the which was measured as the percentage of scope of this article. adults indicating they would vote for a

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 788 AMERICAN SOCIOLOGICAL REVIEW

woman for President of the United States To measure the availability of funding, I (lagged one year) (Neimi, Mueller, and use a one-year lagged measure of total foun­ Smith 1989). I also include a measure of dation and corporate funding for public in­ movement success-the percentage of Con­ terest organizations (in billions, 1982 con­ gressional bill introductions advancing stant dollars) (American Association of women's status passed by Congress between Fund-Raising Council 1988). As the protest 1955 and 1985 (Costain 1992).10 These mea­ cycle waned in the 1970s and into the 1980s, sures are predicted to have a positive influ­ foundation sponsorship of organizations and ence on number of feminist protest events programs serving women and minorities and new organizational activity. grew as part of a general funding trend asso­ I also include two general measures of the ciated with dramatic increases in SMOs and political and social environment. As a mea­ advocacy organizations (Jenkins 1985, 1987; sure of political opportunities, I include a Walker 1983). variable referencing Democratic Party ad­ When modeling the development of the vantage in Congress, measured by the num­ civil rights movement, I include the same ber of Democrats elected to the House of measures of the political/resource environ­ Representatives minus the number of Re­ ment, along with group-specific variables publicans elected (lagged one year) (Con­ (McAdam 1982; Burstein 1985): African­ gressional Quarterly 1993). This variable American labor force participation (lagged measures the presence of influential allies, one year) (Economic Report of the President who form a key component of the political 1989, table B-36); the percentage of adults opportunity structure (Tarrow 1996). willing to vote for an African American for McAdam (1982) and Costain (1992) both President (lagged one year) (Neimi et al. suggest such political leverage facilitated 1989); and the number of Blacks in Con­ the passage of laws favorable to women and gress, which can be considered a measure of Blacks, promoting opportunities for activ­ movement success (lagged two years) (U.S. ism. For example, federal legislation and Bureau of the Census 1978:154; U.S. Depart­ programs associated with President ment of Commerce 1989:244). These mea­ Kennedy's War on Poverty and Johnson's sures are expected to have a direct positive Great Society served a symbolic function by effect on the number of civil rights protests. indicating a favorable environment, lowered costs of action, and so on that fostered sub­ RESULTS sequent collective action (Burstein 1985; Costain 1992; Jenkins and Ekert 1986; Figure 2 depicts the expansion of the McAdam 1982; Walker 1983). II women's movement since 1955. Between 1968 and 1973, the number of feminist orga­ III The information was originally coded USIng nizations founded is inversely related to the u.s. Statutes at Large and the Congressional number of women's protest events. However, Record (Costain 1992:xviii). [ also examined the foundings and protest events moved in tan­ influence of bill introductions relating to dem through the rest of the decade, with women's issues. but this variable was not a sig­ foundings leading protest events by a year or nificant predictor of either women' s organiza­ tional foundings or protest events. two. A final spike in protest events appears II Tarrow (1991) describes this dynamic: around 1982 as new organizational activity "Within the American cycle of protest, policy in­ continued to decline. This increase in protest novations directed at one group could give rise events reflects the final efforts of women's both to new channels of access and to general­ groups to secure ratification of the Equal ized expectations about elite responsiveness Rights Amendment in spite of newly elected among others, leading to new stages of protest President Reagan's antipathy (Costain 1992). which in turn evoked new policy responses. This Figure 3 compares the trajectories of the seems to have been the relationship between the civil rights movement and the feminist move­ Black movement and other movements for minor­ ment after 1955. There is an obvious se­ ity rights which followed similar strategies and succeeded-at a much lower cost-in achieving quencing of these two major U.S. social many of the gains that blacks had fought for in movements. African-American protest events the 1960s" (p. 96)." predate and dwarf feminist protest events and

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. SEQUENCING OF SOCIAL MOVEMENTS 789

50~------______-, 350 Density of feminist '0 organizations e: 300 cu ~ (I)0 2l 40 :I e: 0 CI)'O >CI) 250 ~ W'O_e: Feminist 2- organizations , :ll g 30 (I) 'Ou. founded ~,.' 200 " ~. c:~ :I 'lijo . ' , . ![ :5~ 20 " 150 E.!::! 0 Cl)e: !C u.cu III .... 01 100 :I ...0'" 0 p:;j' 1l 10 / o·~ E Feminist protest :I :I , 50 0 Z events

o ~~~~-r~'+-~-~'-r~~'r'~+-~-+-+-r~+-~r+-+-r-r+-~ 1955 1960 1965 1970 1975 1980 1985 Year Figure 2. Characteristics of the Feminist Social Movement, 1955 to 1985

organizational foundings. Feminist protest stable population of national African-Ameri­ events and foundings are low until about can organizations. 1965, when Black protests peak. As compe­ tition between Black organizations initiated The Intermovement Diffusion of Activism the decline of Black activism, feminist col­ lective action took off. Nonetheless, a sig­ According to the protest dynamic model, in­ nificant overlap in African-American and creases in civil rights movement events women's organization and mobilization re­ should promote protest and organizational mains after 1968. Clearly, the women's formation by feminists. The organization dy­ movement developed in the context of a namic model predicts that the growing den-

700,------,50 /) Z '0 Black protest events (\ e: e: 0 600 cu­ / I Feminist organizations 3 (/l e: 40 g e: CI) I \ founded \ 0 .... 0> ~ / \ .... 0 ._w 500 :', Feminist protest IQ­ -;U; I I events III" .!::! CI) :1(1) e:- p:;j"3 I \ /\ 30 111-­ cue 400 \ -:I 01 a. I I" \ 0" iii' 0"': "I \ . :1- ...:0 ' \ , , 0"t1 o.!!! r, I \ ., ' ,,0 culll I \ I \ '.' 20 0- iDe e:(I) I \ I '" i :I!!l. OQi (1).0 i \.1 :\ ~----.--- Q.(I) ~E DenSity of Black;/"_-----.--- '. 0:1 I organizations' '- , 10 a /'. ''\ ~. o 5iz III C 100 :I ~I ,_ 1"\".. '\ '. I\./~ ___ .~,'/ \/V \ Q. r----, \. J~-- o .,,,. ""'r '.,' ,/ o 1955 1960 1965 1970 1975 1980 1985 Year Figure 3. Characteristics of the Civil Rights Movement and the Feminist Social Movement Com­ pared, 1955 to 1985

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 790 AMERICAN SOCIOLOG!CAL REVIEW

Table 1. Poisson Coefficients from the Regression of the Number of Feminist Protest Events on Se- lected Independent Variables: United States, 1955 to 1985

Independent Variable Modell Model 2 Model 3 Model 4 Female labor force participation (t-l) .231'" .457'" .086 .323'" (.042) (.070) (.044) (.080) Percent vote for woman President (t-I) .100'" .068" -.034 -.045 (.022) (.022) (.025) (.032) Democratic Party advantage in -.006' -.005 -.003 -.006 Congress (t-I) (.002) (.003) (.002) (.003 ) Foundation funding (t-I) .259 .666' -.454 .099 (.341 ) (.332) (.375) (.388) Percent laws passed favoring women .048 -.203 -.311' -.387" (.106) (.118) (.130) (.126) 2 Number of feminist protest events (t-I) .173e- .. , .220e-2 '" .078e-2 ' .148e-2 '" (.02ge-2) (.032e-2) (.032e-2) (.038e- 2) Density of feminist organizations (1-2) -.020'" -.042'" -.008' -.029'" (.004) (.006) (.004) (.007) Number of Black protest events (1-2) -.013'" -.014" (.004) (.004) Black protest events x Democratic .005e-2 .006e-2 ' Party advantage (.003e-2) (.003e-2) Density of Black organizations (t-2) .039'" .030'" (.007) (.007)

Constant -9.125'" -9.812'" -1.443 -2.628 ( 1.776) (1.634) ( 1.599) 2.067 Log-likelihood -74.9 -6.2 -61.9 -52.5 Number of cases 31 31 31 31

2 Note: Numbers in parentheses are standard errors; "e- " indicates that the regression coefficient or stan­ 2 dard error should be multiplied times 10- . 'p < .05 "p < .01 '''p < .001 (two-tailed tests)

sity of civil rights organizations, not the feminist protest events and feminist organi­ number of protests, promoted women's pro­ zational density on subsequent feminist pro­ tests and new organizational activity. As the test events, controlling for demographic density of civil rights organizations reaches changes, political opportunities, and re­ a threshold, interorganizational competition sources. There is a significant positive effect takes over, resulting in fewer new organiza­ of the number of prior protest events on sub­ tions and fewer protests. sequent protest events, as expected from the Exploratory analysis found some initial protest dynamic model. There is, however, a support for the organization dynamic model. significant inverse association between There is a significant .72 zero-order correla­ women's organizational density and number tion between the density of Black organiza­ of feminist protests. (A curvilinear specifica­ tions and the number of feminist protest tion of the organizational density variable events and a .69 association with the number was also examined, but it was not significant.) of women's organizations founded. However, This suggests a trade-off between protest the correlations between number of Black events and organization-building-investi ng protest events and number of feminist protest in new organizations may direct resources events or organizational foundings are not away from protest (piven and Cloward 1977). significant (results available from author). Model I in Table I controls for lagged Model I, the baseline model in Table I, rates of female labor force participation. fa­ specifies the influence of (lagged) number of vorable public opinion towards women,

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. SEQUENCING OF SOCIAL MOVEMENTS 791

Democratic Party control of Congress, po­ creases the number of feminist protest tential funding availability, and passage of events. Thus, high numbers of Black protest laws favoring women. As expected, as events and a Democratic Party advantage in women entered the labor market they devel­ Congress are necessary for protest to diffuse oped a feminist consciousness and protested from the civil rights movement to the femI­ gender inequality (Klein 1985). Public opin­ nist movement. ion in support of women's gains also in­ In addition to the observed protest effect, creases women's protest activity. However, an organizational dynamic is clearly signifi­ during times associated with a more open cant in the diffusion of activism between the political climate (measured by Democratic civil rights movement and the feminist move­ Party advantage) there is a lower likelihood ment. Increased density of African-American that women will engage in protest. At such organizations promoted women's activism: times, traditional modes of reform may seem Each addition to the number of Black orga­ more tenable, lowering women's willingness nizations mUltiplies the number of women's to engage in riskier direct actions.12 protest events by close to 3 percent (com­ Model 2 in Table I adds the number of pared to a less than .05 percent increase from Black protest events to Model I; Model 3 Black protest events during periods of adds B lack organizational density to the Democratic Party strength, and a I-percent baseline model. These models examine each decrease at other times). I also tested curvi­ variable's direct effect and its influence on linear specifications of the Black protest the control variables. Model 4 (the full events and density variables, but adding a model) of Table 1 controls for number of quadratic term did not improve the fit of the Black protest events and Black organiza­ model. There were no other significant inter­ tional density simultaneously. Results for actions between number of protest events, Model 4 show that the (lagged) number of organizational density, and the political en­ Black protest events had a significant influ­ vironment. ence on the number of feminist protest Without qualification, increased density events. Rather than promoting the inter­ of organizations in the civil rights move­ movement diffusion of protest in a straight­ ment provided an opportunity for women's forward manner as Tarrow (1994) argues, in­ protest that was independent of political creases in Black protest events apparently conditions, protest effects, and the interac­ dampened feminist protest events-except tion of the two. The density-dependence hy­ when periods of heightened activism coin­ pothesis is supported, but there is no evi­ cided with openings in the political opportu­ dence of a competitive turn around with re­ nity structure. This interpretation is based on spect to the influence of civil rights organi­ the significant interaction effect between the zations on feminist protest. The protest dy­ number of Black protest events and Demo­ namic hypothesis is also supported, with cratic Party advantage in Congress that posi­ some modification to account for how the tively affects the number of feminist protest political context critically determines events. The negative signs for the number of whether protest will diffuse from initiator to B lack protest events and Democratic Party spin-off movements. In a supportive politi­ advantage in Model 4 indicate that each vari­ cal environment, diffusion is more likely; at able has a negative effect when the other other times, protest by early risers may de­ variable equals O. When the number of Black press the protest activities of other constitu­ protest events approaches 0, political control encies. by allies channels support away from Table 2 presents models predicting the women's protest activity; when the political number of women's organizations founded:it balance of power does not favor the move­ the national level. Modell shows that, again. ment, the number of Black protest events de- female labor force participation rates are positively correlated with the founding of new feminist organizations. However. there 12 I also controlled for a time-trend variable (year) to examine whether these baseline vari­ is a negative effect of periods of Democratic ables mask a trend effect. The time-trend variable political advantage, which contradicts theo­ was not significant in any of the models. retical expectations. Also, as public opinion

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 792 AMERICAN SOCIOLOGICAL REVIEW

Table 2. Poisson Coefficients from the Regressions of the Number of Feminist Organizations Founded on Selected Independent Variables: United States, 1955 to 1985

Independent Variable Model I Model 2 Model 3 Model 4 Female labor force participation (I-I) .334*" .339'*' .243'*' .248'" (.039) (.038) (.054) (.055) Percent vote for woman President (t-I) -.115'*, -.106'" -.112'" -.105'" (.027) (.029) (.024) (.026) Democratic Party advantage in -004' -.004 -.044e-2 .005e-2 Congress (I-I) (.002) (.002) (.247e-2) (.25ge-2) Foundation funding (I-I) .301 .305 -.006 .001 (261 ) (.260) (.291 ) (.292) Percent laws passed favoring women .120 .094 -.015 -.030 (084 ) (.089) (.103) (.105) Number of feminist protest events (1-2) .566e-2 .654e-3 .494e-3 .561 e-3 3 3 3 (351 e- ) (.363e- ) (.340e- ) (.353e 3) Density of feminist organizations (t-2) .039'" .034*" .024' .020 (.008) (010) (.010) (.011) (Density of feminist organizations)2 -.014*" -.O\3'*' -.010'" -.009" (.002) (.002) (.003) (003 ) Number of Black protest events (t-2) -.100e-2 -.076e-2 (.108e-2) (.107e-2) Density of Black organizations (1-2) .020' .019' (.009) (.009)

Constant -2.089 -2.163 -.972 -1.015 (1.570) ( 1.566) ( 1.298) ( 1.311) Log-likelihood -62.7 -62.3 -6.4 -6.1 Number of cases 31 31 31 31

2 Note: Numbers in parentheses are standard errors. "e "!"e-3" indicates that the regression coefficient or standard error should be multiplied times 10-211 0-). P < .05 *'p < .0 I '*'p < .00 I (two-tailed tests)

becomes more favorable toward women, the 1971 and 1972, when the density of founding rate declines-the obverse of the women's organizations reached roughly variable's effect on women's protest events. 140. This suggests the importance of understand­ Model 2 in Table 2 adds the number of ing protest events and group formation as Black protest events to Model I; Model 3 distinct, although linked, processes. Neither adds Black organizational density to the foundation funding nor passage of favorable baseline model: and Model 4 is the full legislation appears to influence the founding model. The most visible manifestation of the process. civil rights movement-protest-has no dis­ Increases in the (lagged) number of femi­ cernible impact on the founding of women's nist protest events only approach signifi­ organizations. suggesting that the protest dy­ cance (p < .10, two-tailed test), suggesting namic hypothesis cannot account for new that feminist protest may promote the group formation by spin-off movements. In founding of feminist organizations. In­ contrast, the density of Black organizations creased density of feminist organizations significantly promoted the formation of new significantly encourages further organiza­ women's organizations. without a significant tion-building, although interorganization competitive effect (the quadratic term is not competition diminishes investment in orga­ significant). In support of the density-depen­ nizations as density increases. Specifically, dence hypothesis, growth of the civil rights competition became determinant between sector aided the growth of the women's

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. SEQUENCING OF SOCIAL MOVEMENTS 793

movement. 13 The addition of one organiza­ The Competitive Sequencing of Protest tion to the density of African-American or­ ganizations increased the number of new The civil rights movement had a clear impact women's organizations by 2 percent. The on the development of the contemporary most dramatic growth in Black organiza­ feminist movement. Did feminist's activism tional density occurred in the early 1970s, at have a reciprocal influence on the trajectory the same time that women's organizational of the civil rights movement? The relation­ foundings increased as well (Figure 2). ship between early risers and entrants after To summarize, the development of civil the peak of the protest cycle is under­ rights organizations was a critical factor in theorized. Presumably the entry of new women's group formation between 1955 and groups into the social movement arena shifts 1985, as was the growing density of women's the opportunities for ongoing protest and af­ organizations until the early 1970s. Earlier fects organizational mobilization for all feminist protest events did not lead to any groups. Regarding the decline of the protest appreciable push for new organizational ac­ cycle, Tarrow (1991, 1994) emphasizes com­ tivity (although the high correlation between petition among organizations and actors the number of feminist protest events and within initiator movements rather than com­ other independent variables may be masking petition between initiator movements and the effect). The lack of observed influence of later movements. Koopmans' (1993) account Black protest on women's group formation of the radicalization and institutionalization contradicts expectations derived from Tarrow of protest cycles is similarly limited. How­ (1991, 1994) and suggests that intensified ever, competition among early-riser SMOs civil rights protest activity did not provide an may not be the only factor that precipitates a opening for women's organizing efforts. Al­ movement's decline. As more groups popu­ though organizational density played a gen­ late the social movement arena, competi­ erative role, there was no significant inter­ tion-for resources, public support, and po­ organization competition between the two litical inclusion-intensifies. Later entrants, movements. Rather, the density of Black or­ piggybacking on the successful activism of ganizations provided a continuing opportu­ early risers, "out-compete" them. There may nity for women to choose organization-build­ be less space for multiple movements than is ing over protest. Paradoxically, over time this posited by the protest cycle model. Instead, positive density-dependence across move­ a discrete sequencing of movements may oc­ ments may have indirectly limited women's cur such that one major movement dominates protest activity; the results in Table 1 demon­ at any given time. Once feminist's protests strate that the increased density of feminist and organizations became prominent, Afri­ organizations was correlated with a decline can Americans faced a more restricted social in the number of feminist protest events. movement environment. Figure 3 shows that during the early 1970s 13 Women's labor force participation rates are a significant overlap remained in civil rights highly correlated with several of the other inde­ and women's protest events and organiza­ pendent variables in these models, especially vote tional foundings. Despite the decline in for a woman president, the density of women's organizations, foundation funding, and the den­ B lack protest events by 1968, there was a sity of Black organizations. The correlation with substantial amount of action at the national number of feminist protest events is not as high, level. The density of Black organizations be­ but is also significant (available from author). gan increasing in tandem with the increase The inflation of the standard errors for the in women's activism and organization forma­ women's labor force participation variable as tion. Some reciprocal influence of the controls are added to the baseline model suggests women's movement on the civil rights move­ that collinearity may be masking the significant ment therefore seems plausible. effects of other variables in this model. The in­ Zero-order correlations suggest a signifi­ terpretation of the protest and density effects is cant inverse relationship between the number unlikely to be altered, however. Even at the bi­ variate level, there is no significant association of feminist protest events and feminist orga­ between the number of recent Black protest nizational density and the number of Black events and the number of feminist protest events. protest events (-.60 in each case). There is no

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 794 AMERICAN SOCIOLOGICAL REVIEW

Table 3. Negative Binomial Coefficients from the Regressions of the Number of Black Protest Events on Selected Independent Variables: United States, 1955 to 1985

Independent Variable Model I Model 2 Model 3

Black labor force participation (t- I) .020 .020 .020 (.011 ) (.0 II) (.012) Percent vote for Black President (I-I) .080'" .083'" .084*" (.OIS) (.020) (.020) Number of Blacks in Congress (1-2) -.161 -.159 -.214' (.OSS) (.095) (.090) Democratic Party advantage in Congress (t-I) .007' .007' .OOS' (.003) (.003) (.003) Number of Black protest events (I-I) . I 25e-2 • .12Se-2 • .138e-2 • (.054e-2) (.054e-2) (.056e-2) Density of Black organizations ((-2) -.013 -.015 -.013 (.008) (.011) (.OOS) Number of feminist protest events (1-2) .090e- 1 (.358e- l ) Density of feminist orgamzattons (I-I) .223e- 2 (.297e-2)

Alpha .135' .134' .134' (.056) (.059) (.056) Constant .544 .468 .370 (.604) (.641) (.697) Log-likelihood -158.3 -15S.1 -157.9 Number of cases 31 31 31

Note: Numbers in parentheses are standard errors. "e-I"/"e-2" indicates that the regression coefficient or 1 2 standard error should be multiplied times 10- /10- . 'p < .05 '*p < .01 '*'p < .001 (two-tailed tests)

significant association, however, between nizations has a marginally significant nega­ Black protest events and the founding of na­ tive effect on the number of Black protest tional-level African-American organizations. events (p < .10, two-tailed test). (A curvilin­ These zero-order findings were confirmed ear specification of the density effect did not using negative binomial regression (results improve the model's fit.) Increases in Black available upon request). There may be a com­ Congressional representation-considered a petitive replacement effect in which later en­ measure of movement success-marginally trants to the social movement arena displace decreases the number of protest events early risers with respect to protest activity but (p < .06). Each additional Congressional rep­ not their organization-building. resentative decreases Black protest events by Results from multivariate analyses of 16 percent, suggesting that protest behavior Black protest events and Black organiza­ is inversely related to movement gains. This tional foundings, however, showed no sig­ is the kind of channelling activity that nificant reciprocal effects between later en­ Jenkins and Ekert (1986) describe: Electoral trants and early risers. Model I in Table 3 is advances make the environment less open to the baseline model predicting the number of protest by constituencies that are gaining Black protest events. Model I suggests that standing in the polity (also see Piven and a Democratic Party advantage, favorable Cloward 1977). public opinion toward Blacks, and (lagged) Models 2 and 3, which add controls for number of Black protest events significantly feminist protest events and feminist organi­ promoted subsequent Black protest events. zational density, do not improve the fit of the The growing density of national Black orga- baseline model. Rather, the key factor driv-

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. SEQUENCING OF SOCIAL MOVEMENTS 795

ing the decline of Black protest events ap­ ing templates with them (McAdam 1995). pears to be the success of the movement Resource-sharing and information-sharing coupled with competition among civil rights are other direct ways that established organi­ groups. This fits McAdam's (1982) descrip­ zations transmit advantages to new move­ tion of the civil rights movement's trajectory ments and their organizations (Walker 1983). and confirms Tarrow's (1994) suggestion Although the migration of activists, re­ that, at least within a given movement, inter­ sources, and information between move­ organizational competition replaces the pro­ ment-related organizations is central, we also test demonstration effects. The fact that there need to learn more about how affinities or is no effect of the feminist movement on the differences in ideologies, strategic choices, founding of civil rights organizations (results or the resonance of organizational forms af­ available upon request) also confirms that fect collaborative efforts (Clemens 1996). early risers establish a protected place in the Activists must be able to adapt models of social movement arena and face little com­ protest and organization to other movements. petition from later entrants. Processes of legitimacy-building by initia­ tor movements and how legitimacy is trans­ ferred to subsequent movements represent DISCUSSION AND CONCLUSION more indirect mechanisms by which organi­ The development and decline of the 1960s zational effects are transmitted. Organiza­ U.S. protest cycle was contingent on both or­ tional ecology models suggest that early in­ ganizational and protest dynamics. Surges in vestments in organization-building payoff, protests by early risers generate activism by not only because resource flows become rou­ other constituencies. However, my analyses tinized and are available for use by other demonstrate that the diffusion of protest is constituencies, but also because the success­ situational: The recruitment of political al­ ful maintenance and growth of organizations lies and the expectation of elite responsive­ increases their acceptance by powerful third ness encourage protest by a range of groups parties (Hannan and Carroll J 992). Such le­ and initiate the onset of a protest cycle. gitimacy affects the willingness of funders to While Democrats had the political advantage support new constituencies, of authorities to for much of the period studied, periods of tolerate their dissent, and of the media to Republican dominance limited the diffusion broadcast their claims in a favorable light. of protest from the civil rights movement to Initiator movement organizations thus must the feminist movement. alter the readiness of outsiders to support In addition to political circumstances, an movement demands by establishing accep­ enduring organizational niche is essential for tance of the forms that their claims-making protest to spread beyond a single movement activities take. Organizational dynamics play and to create protest opportunities for other a central role in the general legitimation pro­ challengers. Increases in organizational den­ cesses, but few social movement researchers sity encourage resource flows that can sus­ give them serious consideration (Clemens tain ongoing activities even as political op­ 1993; Minkoff 1994). portunities contract. An increase in move­ I also show that to understand protest ment organizations signals favorable politi­ cycles, the processes by which allies are cal conditions to other constituencies and placed in positions of power must be scruti­ also provides resources for activism. In part, nized. Allies represent access and political this process reflects the spread of informa­ leverage, factors that are necessary for keep­ tion and political learning. The process also ing the potential for protest open even as creates resources and institutional space for movements make policy gains. Without ac­ sustained protest and organization-building tive and ongoing intervention by activists and by a diversity of actors. their political sponsors, movement accom­ Direct and indirect mechanisms promote plishments (the passage of legislation, the the organizational effects that I have docu­ election of representatives from the move­ mented. Activists become available for re­ ment) are likely to close down protest oppor­ cruitment and mobilization; they can also tunities for all groups even though such gains cross movement boundaries, taking organiz- are usually partial at best (Piven and Cloward

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 796 AMERICAN SOCIOLOGJCAL REVIEW

1977). In fact, the growth of a national orga­ gay and lesbian movement (D'Emilio 1983), nizational infrastructure may encourage and the lesbian feminist mobilization multimovement coalitions that can position (Whittier 1995). These movements have supporters inside the political system thereby overlapping constituencies. They also devel­ maintaining readiness for future protest. oped in a more restricted political climate, There was less clear support for the view suggesting that the presence of political al­ that the expansion of later movements con­ lies may be less critical as the movement sec­ tributes to intermovement competition and tor becomes more institutionalized. the decline of initiator movements. Later The influence of organizational expansion movements piggyback on the precedent-set­ on the diffusion of protest cycles across ting efforts of early risers without placing movements is understandable in the U.S. limits on them. Thus, early risers can estab­ context. National social movement organiza­ lish and maintain a protected niche for their tions have become increasingly important as organizations even as more groups enter the vehicles for citizen participation in North social movement arena. American and Western European countries This analysis focused on the two most vis­ over the past century (Tarrow 1994; Tilly ible and enduring U.S. social movements of 1984). Nowhere are they more central to the recent decades. However, the civil rights operation of institutional life than in the con­ movement influenced a broad set of social temporary United States. However, different movements, including the New Left and anti­ state structures and political systems may war movements (Gitlin 1987), the peace produce distinctive patterns of interaction movement (Meyer 1992), the ecology move­ between organization and mobilization. Such ment (Epstein 1991), the gay liberation variations will depend on the cultural and in­ movement (D'Emilio 1983), the Chicano stitutional support for, and restrictions on, movement (ami and Winant 1994), and the different types of mobilizing structures political resurgence of the American Indian (McCarthy 1996). movement (Cornell 1988). Was there a simi­ This analysis underscores the continued lar interplay between protest and organiza­ relevance of organizational analYSIS for the tion in the diffusion of activism to these study of social movements. Both movement­ movements? The analysis also leaves open based and organizational dynamics affect the the possibility that initiator movements are development of national social movements. vulnerable to displacement by the new move­ Spontaneous collective action builds on and ments they engender. offers an alternative for new organizational A final research question is, how does the initiatives; the establishment and subsequent pattern of cross effects vary for movements growth of a new niche for organizational ac­ that develop later in the protest cycle. What tivity encourages activism-both for early is the pattern of sequencing between the risers and so-called easy riders. Thus, the or­ women's movement and subsequent mobili­ ganizational system provides opportunities zations, for example? Is diffusion similar to for the expansion of activism. The protest that between the civil rights movement and cycle initiated by the civil rights movement the feminist movement, or does the fact that in the 1960s clearly shows that investments later movements enter into a heavily popu­ in organization were critical both early on lated niche change the intermovement dy­ and as the movement spread from a specific namic? Some argue that density creates the constituency to a wave of insurgency encom­ possibility for new niches once competitive passing newly empowered citizens. forces are released (Olzak 1992; Wilson 1973). Meyer and Whittier (1994) argue that Debra C. Minkoff is Associate Professor of Soci­ the feminist movement influenced the form ology at Yale Unil'ersitv. Her work explores the organizational dynamiCs of soc wi mOI'ements, and development of the peace movement, but bringing macro-organizational theory to bear Oil there may be competitive effects as well. questions of mOl'ement emergence and develop­ Likewise, both mutualistic and competitive ment. She IS all/hor of Organizing for Equality: dynamics may have infused the relationship The Evolution of Women's and Racial-Ethnic Or­ between the women's movement and the pro­ ganizations. 1955-1985 (Rutgers University choice movement (Staggenborg 1991), the Press, 19951 and a number of articles exp/orLng

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. SEQUENCING OF SOCIAL MOVEMENTS 797 the organizational dimensions of the civil rights 1989. Washington, DC: Government Printing andfeminist movements. Her current researchfo­ Office. cuses on the structure and determinants of the Epstein, Barbara. 1991. Political Protest and Cul­ national U.S. social movement sector. tural Revolution. Berkeley. CA: University of California Press. REFERENCES Evans, Sara. 1979. Personal Politics. New York: Vintage Books. Aldrich. Howard and Peter V. Marsden. 1988. Ferree, Myra and Beth Hess. 1985. Controversv "Organizations and Environments." Pp. 361-92 and Coalition. Boston, MA: Twayne. ' in Handbook of Sociology, edited by N. Freeman, Jo. 1973. "The Emergence of the Smelser. Beverly Hills, CA: Sage. Women's Liberation Movement." American American Association of Fund Raising Council. Journal of Sociology 78:792-811. 1988. Giving USA. New York: AAFRC. Gale Research Company. 1955-1988. Encyclope­ Barron, David. 1992. 'The Analysis of Count dia of Associations. Vol. 1, Nalional Organi­ Data: Overdispersion and Autocorrelation." Pp. zations. Eds. I to 23. Detroit, MI: Gale Rc­ 179-220 in Sociological Methodology, edited search. by P. V. Marsden. Cambridge, MA: Blackwell. Gelb, Joyce and Marian Paley. 1982. Women and Burstein, Paul. 1985. Discrimination, Jobs and Public Policy. Princeton, NJ: Princeton Univer­ Politics. Chicago, IL: University of Chicago sity Press. Press. Gitlin, Todd. 1987. The Sixties: Days of Hope, Burstein, Paul, Rachel Einwohner, and Jocelyn Days of Rage. New York: Bantam Books. Hollander. 1995. "The Success of Political Gray, Virginia and David Lowery. 1995. "The Movements: A Bargaining Perspective." Pp. Demography of interest Organization Commu­ 275-95 in The Politics of Social Protest, ed­ nities: Institutions, Associations, and Member­ ited by J. C. Jenkins and B. Klandermans. ship Groups." American Politics Quarterly Minncapolis, MN: University of Minnesota 23:3-32. Press. Greene, William H. 1988. Limdep, Version 5.1. Cameron, A. Colin and Pravin K. Trivedi. 1986. New York: New York University. "Econometric Models Based on Count Data: Hannan, Michael T. 1991. "Theoretical and Comparisons and Applications of Somc Esti­ Methodological Issues in Analysis of Density­ mators and Tests." Journal of Applied Econo­ Dependent Legitimation in Organizational metrics 1:29-53. Evolution." Sociological Methodology 21: 1- Clemens, Elisabeth. 1993. "Women's Groups and 42. the Transformation of U.S. Politics, 1892- Hannan, Michael T. and Glenn R. Carroll. 1992. 1920." American Journal of Sociology 98:755- Dynamics of Organizational Popu/atioll. New 98. York: Oxford University Press. ---. 1996. "Organizational Form as Frame: Hannan Michael T. and John Freeman. 1989. Or­ Collective Identity and Political Strategy in the ganizatIOnal Ecology. Cambridge, MA: American Labor Movement, 1880-1920." Pp. Harvard University Press. 205-26 in Comparative Perspectives on Social Jenkins, J. Craig. 1985. "Foundation Funding of Movements, edited by D. McAdam, 1. D. Progressive Social Movements." Pp. 7-17 in McCarthy and M. N. Zald. New York: Cam­ The Grant Seekers Guide, 2d ed., edited by 1. bridge University Press. Shellow. Nyack, NY: Glenmeade. Conell, Carol and Samuel Cohn. 1995. "Learning ---. 1987. "Nonprofit Organizations and from Other People's Actions: Environmental Policy Advocacy." Pp. 296-318 in The Non­ Variation and Diffusion in French Coal Min­ Profit Sector, edited by W. Powell. New Ha­ ing Strikes, 1890-1935." American Journal of ven, CT: Yale University Press. Sociology 101:366-403. Jenkins, 1. Craig and Craig M. Ekert. 1986. Congressional Quarterly. 1993. Congress and the "Channeling Black Insurgency." American So­ Nation. Vol. 8, 1989-1992. Washington. DC: ciological Review 51 :812-29. Congressional Quarterly. Katzenstein, Mary. 1990. "Feminism within in­ Cornell, Stephen. 1988. Return of the Native. stitutions: Unobtrusive Mobilization in the New York: Oxford University Press. 1980s" Signs 16:27-54. Costain, Anne N. 1992. Inviting Women's Rebel­ Klein, Ethel. 1984. Gender Politics. Cambridge, lion. Baltimore, MD: Johns Hopkins University MA: Harvard University Press. Press. Knoke, David. 1989. "Resource Acquisition and D'Emilio, John. 1983. Sexual Politics, Sexual Allocation in U.S. National Associations." In­ Communities. Chicago, IL: University of Chi­ ternational Social Movement Research 2: 129- cago Press. 54. Economic Report of the President: January 1989. Koopmans. Ruud. 1993. "The Dynamics of Pro-

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 798 AMERICAN SOCIOLOGICAL REVIEW

test Waves: West Germany. 1965 to 1989." Smith. 1989. Trends in Public 0pll1lOn: A American Sociological Review 58:637-58. Compendium of Survey Data. New York: Kriesi, Hanspeter. 1996. "The Organizational Greenwood Press. Structure of New Social Movements in a Po­ Oberschall. Anthony. 1973. Socia! Conflicts and litical Context." Pp. 152-84 in Comparatil'e Social Movements. Englewood Cliffs. N1: Perspectll'es on Social Movements, edited by Prentice-Hall. D. McAdam. J. D. McCarthy and M. N. Zald. ---. 1978. "The Decline of 1900s Social New York: Cambridge University Press. Movements." Research ill Social Movements, McAdam. Doug. 1982. Political Process and the Conflict and Change 1:257-89. Development oj Black Insurgency. Chicago, IL: Olzak. Susan. 1989. "Analysis of Events in the University of Chicago Press. Study of Collective Action." Allllual Rel,iew oj ---. 1988. Freedom Summer: The Idealists Sociology 15:119-41. Revisited. New York: Oxford University Press. ---. 1992. The Dynamics of Etilllic Competi­ ---. 1995. "'Initiator' and 'Spin-off' Move­ tioll and COllflict. Stanford, CA: Stanford Uni­ ments: Diffusion Processes in Protest Cycles." versity Press. Pp. 217-39 in Repertoires and Cycles oj Col­ Olzak, Susan and Suzanne Shanahan. 1996. lective Action. edited by M. Traugott. Durham. "Deprivation and Race Riots: An Extension of NC: Duke University Press. Spilerman's Analysis." Social Forces 74:931- McAdam. Doug. Sidney Tarrow. and Charles 61. Ti lly. 1996. "To Map Contentious Politics." Olzak. Susan and Elizabeth West. 1991. "Ethnic Mobilization I: 17-34. Conflict and the Rise and Fall of Ethnic News­ McCarthy. 10hn D. 1996. "Constraints and Op­ papers." American SOCIOlogical Review 56: portunities in Adopting, Adapting. and Invent­ 458-74. ing." Pp. 141-51 in Comparative Perspectives Omi. Michael and Howard Winant. 1994. Racial on Social Movements, edited by D. McAdam. Formation ill the United States. 2d ed. New 1. D. McCarthy and M. N. Zald. New York: York: Routledge. Cambridge University Press. Piven. Frances Fox and Richard Cloward. 1977. McCarthy. 10hn D., Clark McPhaiL and 1ackie Poor People's Movements. New York: Pan­ West. 1996. "Images of Protest: Dimensions of theon. Selection Bias in Media Coverage of Washing­ Rosenfeld. Rachel A. and Kathryn B. Ward. ton Demonstrations 1982 and 1991." American 1991. "The Contemporary lJ.S. Women's Sociological Review 61 :478-99. Movement: An Empirical Example of Compe­ McCarthy, 10hn D. and Mayer Zald. 1977. "Re­ tition Theory." Sociological Forum 6:471-500. source Mobilization and Social Movements." ---. 1996. "Evolution of the Contemporary American Journal of Sociology 82: 1212-41. U.S. Women's Movement." Research in Social Meyer, David S. 1993. "Protest Cycles and Po­ Movements, Conflict and Change. 19:51-73. litical Process: American Peace Movements in Snow, David and Robert Benford. 1992. "Master the Nuclear Age." Political Research Quar­ Frames and Cycles of Protest." Pp. 133-55 in terly 46: 451-79. Frontiers in Social Movement Theory, edited 1992. A Winter of Discontent. by A. Morris and C. Mueller. New Haven, CT: New York: Praeger. Yale University Press. Meyer. David S. and Douglas R. Imig. 1993. "Po­ Spalter-Roth, Roberta and Ronnie Schreiber. litical Opportunity and the Rise and Decline of 1995. "Outsider Issues and Insider Tactics: Interest Group Sectors." The Social Science Strategic Tensions in the Women's Policy Net­ Journal 30:253-70. work during the 1980s." Pp. 105-27 in Femi­ Meyer, David S. and Nancy Whittier. 1994. "So­ nist Organizations. edited by M. M. Ferree and cial Movement Spillover: The Effects of the P. Y. Martin. Philadelphia, PA: Temple Uni­ Women's Movement on the Peace Movement" versity Press. Social Problems 41 :277-98. Staggenborg. Suzanne. 1991. The Pro-Choice Minkoff. Debra C. 1994. "From Service Provi­ MOl'ement. New York: Oxford University sIOn to Advocacy: The Shifting Legiumacy of Press. Organizational Forms." Social Forces 72:943- Stinchcombe, Arthur. 1965. "Social Structure and 69. Organizations." Pp. 142-93 in Handbook of ---. 1995. Organizing for Equality: The Evo­ Organizations. edited by 1. March. New York: lution of Women's and Racial-Ethnic Organi­ Rand McNally. zations in America, 1955-1985. New Bruns­ Tarrow, Sidney. 1991. "Struggle, Politics. and wick, N1: Rutgers University Press. Reform: Collective Action, Social Movements, Morris, Aldon. 1984. The Origins of the Ol'il and Cycles of Protest" (Western Societies Pro­ Rights Movement. New York: Free Press. gram, Occasional Paper No. 21, 2d ed.). Cen­ Neimi, Richard G., 10hn Mueller. and Tom W. ter for International Studies. Cornell Univer-

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. SEQUENCING OF SOCIAL MOVEMENTS 799

sity, Ithaca, NY. Michigan Press. ---. 1994. Power in Movement: Social Move­ U.S. Bureau of Census. 1978. Current Populatioll ments, Collective Action and Politics. Ncw Reports (Series P-23, 1'<0. 80). Washington, York: Cambridge University Press. DC: Government Printing Office. ---. 1995. "Cycles of Collective Action: Be­ U.S. Department of Commerce. 1989. Statisllcal tween Moments of Madness and the Reper­ Abstract oj the United States: 1988 Wash1l1g- toire of Contention." Pp. 89-115 in Reper­ ton, DC: Government Printing Office. ~ toires and Cycles oj Collective Action, edited Walker, Jack. 1983. "The Origins and Mainte­ by M. Traugott. Durham, NC: Duke Univer­ nance of Interest Groups In America." Ameri­ sity Press. can Political Sciellce ReView 77:390-406. ---. 1996. "States and Opportunities: The Po­ Whittier. Nancy. 1995. Femitllst Generations: litical Structuring of Social Movements." Pp. The Persistence of rhe Radical Women's Move­ 41-61 in Comparative Perspectives on Social ment. Philadelphia, PA: Temple University MOI'ements, edited by D. McAdam, 1. D. Press. McCarthy and M. N. Zald. New York: Cam­ Wi Ison, James Q. 1973. Po/irical Organizations. bridge University Press. New York: Basic Books. Taylor, Verta. 1989. "Social Movement Continu­ Yuchtman, Ephraim and Stanley E. Seashore. ity: The Women's Movement in Abeyance." 1967. "A System Resource Approach to Orga­ American Sociological Review 54:761-75. nizational Effecti veness." American Sociologi­ Tilly, Charles. 1984. "Social Movements and ~a­ cal Review 32:891-903. tional Politics." Pp. 297-317 in Statemaking Zald, Mayer N. 1988. "The Trajectory of Social and Social Movements, edited by C. Bright and Movements in America." Research in Social S. Harding. Ann Arbor. MI: University of Movements, Conflict and Change 10: [9-41.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.