The Dynamics of Protest Diffusion: Movement Organizations, Social Networks, and News Media in the i960 Sit-ins
Kenneth T Andrews Michael Biggs University of North Carolina University of Oxford
The wave of sit-ins that swept through theAmerican South in the spring of 1960 transformed the struggle for racial equality. This episode is widely cited in the literature on over social movements, but the debate its explanation remains unresolved?partly on because previous research has relied case studies of a few large cities. The authors use event-history analysis to trace the diffusion of sit-ins throughout the South and to compare cities where sit-ins occurred with the majority of cities where they did not. They assess the relative importance of three channels of diffusion: movement organizations,
social networks, and news media. The authors find that movement organizations played an was important role in orchestrating protest; what mattered a cadre of activists rather mass than membership. There is little evidence that social networks acted as a channel news for diffusion among cities. By contrast, media were crucial for conveying
information about protests elsewhere. In addition, the authors demonstrate that sit-ins were most likely to occur where there were many college students, where adults in the
black community had greater resources and autonomy, and where political opportunities
were more favorable.
wave of sit-ins that swept through the formed the struggle for racial equality. The American South in the spring of 1960 trans Sociological investigation began within months of the first protest (Laue [1966] 1989; Oppenheimer 1963; Searles andWilliams 1962; Wehr 1960), and the sit-ins have become an Please direct correspondence to Kenneth Andrews, case in the literature on social move Department of Sociology, CB #3210, University of exemplary ments McAdam North Carolina, Chapel Hill, NC 27599 (Killian 1984; 1982, 1983; ([email protected]) or Michael Biggs, Department of McAdam and Sewell 2001; Morris 1981,1984; Sociology, University of Oxford, 0X1 3UQ, United Oberschall 1973, 1989; Piven and Cloward The Kingdom ([email protected]). 1977; Polletta 1998). Despite the amount of authors are listed in each con alphabetical order; research devoted to the sit-ins, there is no con tributed equally.We thankNeil Brown, Doug Grbic, sensus on why they occurred. Some sociologists Aidan Smith, and Jessica Stannard-Friehl for research (e.g., Morris 1981,1984) argue thatmovement assistance, and James Alt for providing an electron the Southern Christian ic copy of the Matthews and Prothro dataset for organizations, especially the cru Southern counties. We also received financial sup Leadership Conference (SCLC), played port from the Center for the Study of theAmerican cial role of coordinating and mobilizing protest. South at UNC. The authors thank Takwing Chan, Others (e.g., Killian 1984; Oberschall 1989) Peter Charles Larry Griffin, Hedstrom, Kurzman, contend that the wave was spontaneous, with Oberschall, Tarrow, the ASR editor Anthony Sidney college students in various cities adopting this and reviewers for their comments and suggestions. novel form of protest because they were inspired Presentations at Harvard University, New York by the actions of students elsewhere. Previous University, University of Oxford, Russell Sage research has not the cities Foundation, and the American Sociological compared swept up in the wave with the more numerous cities that Association's annual meeting all helped to sharpen the argument. remained untouched.
American Sociological Review, 2006, Vol. 71 (October752-777) DYNAMICSOF PROTESTDIFFUSION 753
In this article, we provide a systematic and gate a protest wave (see also Koopmans and comprehensive investigation of the diffusion of Olzak 2004). sit-ins throughout the South in the spring of We begin by sketching the course of events 1960.We assess the relative importance of three and reviewing scholarship on the sit-ins. From channels of diffusion?movement organiza these studies and the wider literature on social tions, social networks, and news media?and movements and collective action, we develop identify the characteristics of a city that made hypotheses to explain the diffusion of protest. outline protest more likely. Event-history analysis is We the method and data for the event used to predict the onset of sit-in campaigns in history analysis and then present the results. To the we the ten weeks following the initial event in complement results, analyze the pattern of and scrutinize Greensboro, North Carolina. This day-by-day newspaper coverage qualita tive evidence on the of diffusion. The analysis encompasses 334 cities. In addition, we process conclusion draws for future scrutinize qualitative evidence on the process of implications research. diffusion and analyze coverage in four Southern newspapers. Our findings offer a new and complex expla THE SIT-INSOF i960 nation for the diffusion of sit-ins. Movement With hindsight, we tend to perceive a civil rights organizations did play a role in orchestrating the movement emerging after World War II and sit-ins, though the role was not as significant as building inexorably to a crescendo in the mid some have argued. The organization with the 1960s. This conventional narrative conceals dis greatest impact was actually the Congress of continuities and critical moments of rapid Racial Equality (CORE) rather than SCLC. mobilization (Carson 1986; McAdam and Membership of the National Association for Sewell 2001). In fact, activists faced a bleak sit theAdvancement of Colored People (NAACP) uation at the end of the 1950s. The NAACP had no discernible effect. There is surprisingly had won a landmark inBrown, but little evidence that social networks acted as a legal ruling the main effect was to provoke massive resist channel for diffusion among cities. By contrast, ance by white Southerners to school news media were crucial. Protesters recalled desegre gation (Klarman 1994). The organization itself first about sit-ins in other cities from learning was outlawed inAlabama. The achievements of or newspaper, radio, television. Event-history SCLC?which had emerged from the victorious analysis demonstrates that protest tended to fol bus boycott inMontgomery, Alabama?were low the newspaper circulation network. In addi also disappointing. Itsmajor campaign for voter tion, the impact of sit-ins in other cities registration, the "Crusade for Citizenship," diminished with distance just as the likelihood floundered. In 1958 Ella Baker, overseeing the of news coverage diminished with distance. campaign, feared that "we are losing the initia Along with these on diffusion, we findings tive in the civil rights struggle" (Fairclough confirm that college students led the and protest, 2001:240). This changed dramatically in the that was more to occur in cities protest likely spring of 1960. where adults in the black had community greater On February 1, four freshmen from North resources and and where autonomy political Carolina Agricultural and Technical (A&T) were more favorable. opportunities College in Greensboro occupied the lunch These have lessons for findings important counter ofWoolworth's after being refused serv the literature on social movements. Attention has ice. The protest was repeated, with increasing focused on the factors recently determining numbers of students, on the following days which events are the media protest reported by (Chafe 1980;Wolff 1970). This form of protest, and Oliver and (e.g., Maney 2001; Myers soon known as a "sit-in," was not new. It can be Caniglia 2004; Smith et al. 2001). Few studies, traced back to 1943, when a handful of CORE examine whether of however, reports protest activists occupied a Chicago restaurant that can further acts of inspire protest elsewhere. refused service to blacks (Meier and Rudwick By combining quantitative analysis of protest 1973). In the 1950s there were several sit-ins in events and of newspaper coverage, we demon the South, though these were tentative tests strate how the media can inadvertently propa rather than concerted campaigns, and they did 754 AMERICAN SOCIOLOGICALREVIEW
not spread to other cities (Morris 1984). A wave except Mississippi. In each city, protest typically of sit-ins in 1958 began inWichita, Kansas, began with students occupying seats at down and spread south to Oklahoma City and three town lunch counters of "five and dime" stores, other cities in Oklahoma (Eick 2002; Graves disrupting business; this often led to con These sit-ins were initiated 1981; Luper 1979). frontation with the police or hostile whites. Sit local NAACP Youth Councils and were con by ins were accompanied by established forms of ducted students from school. primarily by high protest such as picketing, boycotts, and demon Protest was sustained over and it many months, strations. Although these campaigns extended succeeded in several down eventually forcing over many months, the spread of sit-ins to new town lunch counters to serve blacks. Although cities declined by the end of the spring, leaving NAACP's national had leadership initially most cities untouched. opposed the sit-ins, the organization subse The consequences of this protest wave can quently highlighted these victories at its annu hardly be overstated. The sit-ins mobilized tens al convention and in its magazine in 1959. of thousands of blacks (and hundreds of whites) However, there is surprisingly little evidence and created a new movement organization, the linking these previous events with Greensboro. Student Nonviolent Coordinating Committee The four freshmen who started the protest dis (SNCC). Inmany cities, the sit-ins forced white claimed any knowledge of what had happened leaders to negotiate, and eventually broke the in Oklahoma (Warren 1965:360). taboo interracial dining. Nashville was By contrast, the sit-ins inGreensboro inspired against the first to succumb, in 1960. blacks in other cities to adopt this form of major city May While civil advocates had been protest. After a week, sit-ins began elsewhere in rights relying on and build North Carolina; soon the wave of protest surged primarily litigation organization such victories elevated as into other states. Figure 1 traces the course of ing, tangible protest the central of themovement in the this wave in the South (including Maryland, strategy years to was not Kentucky, and West Virginia along with the follow. The impact of the sit-ins only states of the former Confederacy). Within two confined to African Americans; SNCC also months of the initial event in Greensboro, sit proved influential for the white New Left and ins had been staged in every Southern state other social movements (Carson 1981). As
5-| r-|
H Other States 4- North Carolina
M ti CO 9 i3~ n n r
Figure 1. Sit-ins in the American South, February 1 to April 14, 1960 DYNAMICSOF PROTESTDIFFUSION 755
Morris puts it, the sit-ins were the "origins of a has been undertaken to identify what differen decade of disruption" (1984:195). tiated cities with sit-ins from those without.2 The How to explain these events?especially how second problem is the contradictory biases of to explain the rapid diffusion of protest?has qualitative evidence?contemporary testimo been vigorously debated by scholars of social ny and retrospective interviews?on the role of movements. The debate has as one been framed movement organizations. On one hand, black of versus stu spontaneity organization. College protesters faced powerful incentives to empha dents themselves used a narrative of spontane size local initiative to avoid charges of being led which from adult ity, signified "independence astray by "outside agitators" (Killian 1984:783). local and action leadership, urgency, initiative, In Charlotte, North Carolina, for example, local moral This by imperative" (Polletta 1998:138). leaders disavowed any contact with groups like shared observers narrative, by contemporary CORE (Oppenheimer 1963:177, 180; Polletta Constable Lomax was (e.g., 1960; 1960,1962), 1998). On the other hand, when it became appar in developed subsequent sociological analyses ent that the sit-ins had reinvigorated themove Piven and Cloward 1977). Morris (e.g., (1981, ment against racial oppression, leaders had 1984) challenged the notion that the sit-ins powerful incentives to magnify their own role developed spontaneously, arguing instead that in organizing protest. These biases exacerbate they "grew out of pre-existing institutions and the difficulties of interpreting qualitative evi organizational forms" (1981:744). The most dence (see also Carson 1986). important of these institutions was the network Previous research has considered diffusion, of activist churches linked together by SCLC; showing that sit-ins tended to occur earlier in "sit-ins were largely organized at themovement places closer toGreensboro (Laue [1966] 1989; churches, rather than on the campuses" (Morris McAdam 1982; Morris 1981; Oppenheimer 1981:757). Morris' account was in turn chal 1963; Orum 1972). This research, however, lengedby Killian (1984) andOberschall (1989), overlooks the majority of cities that remained who argued that it overestimated the role of untouched sit-ins. It also fails to established leaders and underestimated conflict by separate geo distance from social for between them and student activists. This debate graphical differences; example, the Upper South was much less repres is not only about what happened in 1960; it sive than the South. We overcome these also has crucial theoretical implications. Deep an model of dif Spontaneity fits the older theory of collective problems by using event-history fusion and Tuma This method has behavior and is also compatible with rational (Strang 1993). been used to and choice theory (Oberschall 1989). By contrast, analyze strikes, riots, campus an on and Cohn emphasis preexisting organization dove protest (Conell 1995; Myers 1997, and Olivier Soule tails with solidarity/mobilization theories, which 2000; Olzak, Beasley, 2003; as well as continue to dominate scholarship on social 1997,1999) organizational founding and Voss movements (see also McAdam 1982). (Conell 1990; Hedstrom 1994; The debate over how to explain the sit-ins has Hedstrom, Sandell, and Stern 2000) and repres sive yet to be resolved. Moreover, previous research violence (Beck and Tolnay 1990; Tolnay, has been limited by substantial methodological Beck, and Deane 1996). Event-history analysis problems. The first problem is selection on the allows us to analyze simultaneously where sit dependent variable. Research has focused entire ins occurred, when they occurred, and how they on ly cities with sit-ins?and predominantly on spread from one city to another. This analysis a few large cities like Nashville and Tallahassee. can be buttressed by scrutinizing qualitative By contrast, the majority of cities where no evidence on the process of diffusion. By com occurred are protest ignored.1 No comparison bining these methods, we hope to explain why
1 2 Exceptionally, Oppenheimer (1963) discussed A survey of black students in 1962 (Matthews and where black studentse decid Prothro Lawrenceville, Virginia, 1975) enables systematic comparison ed not to sit he this as due to in; explained the town's between those who participated in sit-ins and those small size, isolation, economically vulnerable black who did not (Biggs forthcoming), but does not reveal and population, lack of organizations. the dynamics of diffusion. 756 AMERICAN SOCIOLOGICALREVIEW
sit-ins multiplied so rapidly across the South The central point of contention in the litera after February 1. ture is the role of movement organizations. Consistent with resource mobilization and polit THEORYAND HYPOTHESES ical process theories, Morris (1981) and McAdam (1982) argue that organization was a For we draw from potential explanations, widely precondition of protest. The presence of amove the literature on collective The wave of protest. ment organization in a city provided a cadre of sit-ins in the of 1960 resembled other spring activists who could orchestrate protest. Local where confrontational has episodes protest activists could be directly inspired by protest and observers used spread rapidly. Participants occurring elsewhere. Alternatively, an organi such as metaphors "fever," "contagion," and zation's leaders could coordinate or encourage as have in other "grass fire," just they protest local activists to act. In 1960, NAACP was by waves Polletta (Oppenheimer 1963; 1998; far the strongest movement organization, with Southern Council Walzer Regional 1960; 1960). branch membership comprising an impressive There was a of feedback clearly process positive 1.5 percent of the urban black population in the blacks in one initiat (Biggs 2003, 2005): city South. NAACP also organized separate College ed sit-ins because others elsewhere, beginning Chapters and Youth Councils. SCLC, by con with had done so. For the vast Greensboro, trast, did not recruit individual members; itwas of the sit-in was a novel majority protesters, really a "meso-level network" (Hedstrom et al. form of that denied protest (and eating places 2000) of activist ministers. CORE had few service to blacks were a novel itwas not target); Chapters in the South, but it had the greatest of the of contention part existing repertoire experience with the sit-in tactic (CORE 1960). In the initial (Tilly 1977,1995;Traugott 1993). This diversity of organizations means that we months there was no that the sit-ins would proof can ask not only whether movement organiza achieve success. But as the sit-ins ultimately tions orchestrated protest but also which type of the fact that so were spread, many protesters organization was most significant. Some schol of success blacks in other cities hopeful inspired ars argue that bureaucratic membership organ to initiate sit-ins. "For the this Negroes, rapid izations like NAACP, despite?or because did a deal to enhance its of are sanc growth great hope of?their apparent power, reluctant to recalled William Peace a victory," (1962:101), tion disruptive protest (Piven and Cloward student at Shaw Raleigh's University. 1977). Morris (1981,1984) argues that SCLC, In this abstract characterization of the outline, along with NAACP Youth Councils, played the should not be controversial. The process ques most important role in orchestrating sit-ins. tion is how exactly it occurred. What were the 1:Protest was orchestrated by move most "channels" of diffusion: move Hypothesis important ment or news organizations. ment organizations, social networks, media? The term "channel" ismost appropriate Social networks provide another channel of dif for themedia, which could provide a conduit for fusion. As protest occurred in one locality, peo information about protest events elsewhere. ple there could inform and encourage their Organizations, by contrast, could actively coor acquaintances elsewhere, who in turn could be dinate protest and exhort their members to take inspired to initiate protest. This channel was part. With this qualification, we believe that suggested by Baker: "A sister who had a broth the trichotomy covers themajor potential expla er in school in another town, her town had nations for diffusion. Networks and organiza already sat in. She might call and ask, why tions are often treated together under the heading doesn't his school sit in?" (Cantarow 1980:83). of "relational" diffusion (see McAdam and McAdam even postulates a "well-developed Rucht 1993), but it is worth distinguishing communication network linking the southern between the social networks of ordinary life black college campuses into a loosely integrat (relatives, friends, coworkers) and the networks ed institutional network" (1982:138; see also created intentionally by movement activists to Tarrow and McAdam 2005). One particular further their goals. The trichotomy of organi network has been identified as significant, the zations, networks, and media provides our first intercollegiate athletic circuit. Walzer (1960) three explanatory hypotheses. suggested that sit-ins were propagated by A&T's DYNAMICSOF PROTESTDIFFUSION 757 basketball team as it played other colleges in incorporates insights from recent scholarship on North Carolina in the first two weeks of the diffusion of protest. Advocates of organi February (see also Oberschall 1989). In fact, the zation would clearly expect strong empirical team's games preceded sit-ins only inDurham, support for Hypothesis 1; those who character and Oppenheimer's (1963) subsequent inter ize the sit-ins as spontaneous would expect that views with protesters found no evidence to sup hypothesis to be rejected, and presumably would port this notion. Nevertheless, intercollegiate expect empirical support for Hypotheses 2 and athletics is worth considering as a proxy for 3. We will evaluate these hypotheses using social networks linking students in different event-history analysis. Hypothesis 1 is readily cities. These affiliations were imbued with rival tested by measuring the presence ormembership ry, which helped motivate protest. A sociology of movement organizations in each city. professor in Tallahassee observed that "an ele Hypothesis 2 implies that the impact of sit-ins ment of competition between Negro colleges has in other cities will diminish with distance, as become a part of the civil rights struggle... and personal acquaintances are more likely to live no college or university wants to be left behind" nearby. Hypothesis 3 has the same implication (Smith 1961:228; see also Laue 1966:82; Wehr so long as potential protesters depend on news 1960:25). media located in their own or nearby cities, which are most to events Hypothesis 2: Protest was inspired by infor likely report occurring close to home. This indeed was borne out the mation about protest occurring elsewhere, by survey of black students: newspaper readers conveyed by social networks. were most likely to read a newspaper published The media an alternative channel of provide in their own city, and then a newspaper pub diffusion. As was in the protest reported news, lished somewhere else in the South (Matthews this could others elsewhere to initiate inspire andProthro 1975). A few recent studies demonstrate that protest. Because Hypothesis 2 (social networks) and the news media can have a effect in significant Hypothesis 3 (news media) both imply that the and Olzak propagating protest (Koopmans impact of sit-ins will diminish with distance, For the 2004; Myers 2000).3 sit-ins, contempo ingenuity is required to distinguish between accounts themedia rary emphasize (Laue 1966; them. As a particular test of Hypothesis 2, we Wehr From Oppenheimer 1963:61-62; 1960). will use intercollegiate athletic associations as a of black stu representative survey college a proxy for social networks. There were seven dents in the South in 1962 and (Matthews associations; affiliation depended on region and Prothro we know that almost all 1975), (93 per college status. Colleges in the same associa read a more or less cent) newspaper regularly, tion fielded teams that regularly traveled to while a listened to large majority (83 percent) each other's campus and that were also con radio and two-thirds watched television more or nected by collegial rivalry. We can test whether less regularly (see also Ward 2004).4 the diffusion of sit-ins tended to follow these Hypothesis 3: Protest was inspired by infor intercollegiate links. As a particular test of mation about protest occurring elsewhere, Hypothesis 3, we will use newspaper circulation. News are to conveyed by news reports. media especially likely report events in their home city; these reports are then circu These three reframe the debate over hypotheses lated or broadcast to a wider area. This creates organization versus spontaneity in a way that what Myers (2000) calls an "asymmetric net work," because news tends to flow from larger cities where news organizations are headquar 3 Roscigno and Danaher (2000) argue that radio tered to more peripheral or smaller cities.5 We stations influenced textile strikes during the Great can test whether the diffusion of sit-ins tended Depression, because they supported itinerant musi cians and broadcast messages from the president; news reports, however, are not part of their argu ment. 5 Myers (2000) also suggests another specification, 4 These are to figures restricted students attending the "responsive network." We have tested this spec = colleges located in cities in our dataset (n 218). ification, and find that it has no effect. 758 AMERICAN SOCIOLOGICALREVIEW
to follow the flow of news, by reconstructing the Minkoff 2004). The civil rights movement has network of newspaper circulation among cities. been the core case around which these ideas Alongside these explanations for the diffusion have been developed (Andrews 2004; Jenkins, of we can protest, examine potential explana Jacobs, and Agnone 2003; McAdam 1982, tions for was more in some why protest likely 1983; Meyer and Minkoff 2004). This line of cities than others. Three further can hypotheses thinking was initially formulated to explain be tested. One hypothesis concerns the role of cross-sectional variation across localities students. The conventional view is that the sit (Eisinger 1973), and political opportunities have ins were the work of students. largely college been used to explain variation among munici This accords with the argument that dense social palities, counties, states, and countries (Amenta, networks within a college campus facilitate Dunleavy, and Bernstein 1994; James 1988; mobilization, especially where students live in Kriesi et al. 1995). The sit-ins were initiated first dormitories (McAdam 1982; Orum 1972; Van in the least repressive areas of the South, as Dyke 1998; Zhao 1998). In addition, students many have observed (e.g., Constable 1960; are rather less constrained by the demands of Pollitt 1960). employment and childcare than adults. "As col 6: Protest was most to occur lege students," remarked one of the Greensboro Hypothesis likely were four, "we have no jobs from which to be fired where political opportunities more by people who don't like to see us assert our favorable. selves" (Dykeman and Stokely 1960:12). Nevertheless, Morris argues that the role of stu dents has been exaggerated: "to understand the RESEARCHDESIGN sit-in movement, one must abandon the assump Dependent Variable and Model tion that it was a collegiate phenomenon" We can test this (1981:757). disagreement. Our analysis investigates cities (rather than Hypothesis 4: Protest was more likely where counties), because sit-ins targeted the down were there many college students. town shopping district. We include 334 urban places with a total population of at least ten Even if college students led the sit-ins, adults in thousand and a black population of at least one the black community may have been important thousand.6 Only a handful of smaller places too. Students could have been more likely to experienced sit-ins. Our analysis begins on protest where they anticipated that support (such February 1, when the first sit-in occurred in as money for bail or attorney's fees in case of Greensboro. It ends on 14, the day before arrest) would be forthcoming from adults. April student activists held a conference at Shaw McAdam (1982) points out that cities were less which led to the creation oppressive than rural plantations. By the same University eventually of SNCC. at this logic, we expect that urban communities varied By terminating analysis point, in the extent to which adults had sufficient we exclude consideration of the outcome of the sit-in itwas still an resources and autonomy to support or partici campaigns. By mid-April, whether the sit-ins would achieve pate in protest. Specifically, blacks in unskilled open question occupations may have been least able to chal their aims of breaking down racial barriers.7 that we can be sure lenge racial oppression (James 1988; Salamon The analytic advantage is suc and Van Evera 1973). The underlying theoreti that protest was inspired by expectations of cal rationale is really the same as for the previ cess rather than actual success. ous hypothesis: amaid toiling for a white family is at the other end of the spectrum from a stu dent at an elite university. 6 on The census provides detailed information the Hypothesis 5: Protest was more likely where the "nonwhite" population; in these states, blacks made adult had autonomy up the vast majority of nonwhites. community greater 7 The was where and resources. only exception Galveston, Texas, sit-ins forced merchants to serve blacks in early to Political opportunities refer the configuration April. InNorth Carolina, some dining establishments of institutions, allies, and opponents that enable removed their stools, forcing everyone to dine while a or constrain protest (McAdam 1995; Meyer and standing, but this was temporary expedient. DYNAMICS OF PROTESTDIFFUSION 759
^ Cities # * Sit Ins None
Figure 2. Map Showing Sit-ins in the American South, February 1 to April 14, 1960
The dependent variable is derived from the ther sit-ins. Take the example of Charlotte: the date of a city's first sit-in. We define this as a first sit-in occurred on February 9, and sit-ins were physical occupation of space from which blacks continued until the 12th when they halted was were excluded, usually a commercial eating at the mayor's request; protest repeated from the 15th to the 18th and on the establishment but occasionally a public facili again 27th; sit-ins and continued ty such as a library.We exclude demonstrations picketing sporadically March and that did not physically occupy segregated spaces through April (Oppenheimer 1963). Our excludes such events as well as cases where themere threat of protest analysis subsequent from consideration for and theoreti sufficed to bring about negotiations. Dates of the pragmatic cal reasons. It would be to first sit-in come from contemporary listings impractical gather evidence on the occurrence of sit-ins after (Laue [1966] 1989:Appendix F; Oppenheimer daily the first event. Moreover, we expect that the 1963:63-64; Southern Regional Council initial sit-in?which demonstrated that some 1960:x/x-xxv). Because they were compiled of the city's blacks were willing to physically from multiple newspaper reports, these are far violate racial segregation?was themost impor more comprehensive in coverage than theNew tant signal for blacks elsewhere. This expecta York Times or any single newspaper.8 In a few tion can be tested (and will be confirmed) by cases, between these sources have discrepancies how the of the initial sit-in been resolved information from NAACP estimating impact by diminished with time. and CORE papers and from local newspapers. We seek to explain why blacks initiated sit Sit-ins occurred in 66 cities; four out of five ins in some cities but not others?and why they cities remained unaffected. As Figure 1 depict protested sooner in some places than others. ed the diffusion of sit-ins through time, Figure This requires event-history analysis. A discrete 2 reveals their diffusion over space. time model is appropriate because many events The first sit-in in a marked the city usually are tied with others (on many days, sit-ins were of a often beginning prolonged campaign, initiated inmultiple cities). The unit of analy and as well as fur involving picketing boycotts sis is the "spell," or city-day. The dependent variable is Yit, a dichotomous variable coded 1 if a sit-in occurred in city / at time t, and 0 oth 8 erwise. The index t at 1 Of the four major Southern newspapers we have begins (February 1) and ends at 74 Each enters the examined, the Charlotte News was the most com (April 14). city on that itwas "at risk" of an prehensive, but it reported only 73 percent of the dataset every day sit-in campaigns that we identify. initial sit-in. Once a sit-in occurred, the city 760 AMERICAN SOCIOLOGICALREVIEW
drops out of the risk set. For example, Charlotte For Hypothesis 4, the variable is the number had its first sit-in on February 9, and so Y is of blacks enrolled in college.9 For Hypothesis = coded 0 for t= 1.. .8 and coded 1 for t 9. For 5, the resources and autonomy of the black cities that remained untouched by sit-ins, Y is community are measured (inversely) by the coded 0 for every day. Because sit-ins did not male unemployment rate and the percentage of occu occur on Sunday, this day is omitted. There are themale labor force relegated to unskilled 18,990 spells. pations?servants and laborers.10 For We are interested in the hazard rate: Hypothesis 6, political opportunities can be dif ferentiated into elite allies, repressive capacity, = and electoral The first 0) power (McAdam 1995). pit=pwb(Yit=l\^h is captured by the presence of the Southern a This is the probability that a sit-in was initi Regional Council, leadership organization that interracial ated in city / at time t, conditional on the fact promoted cooperation (Egerton that sit-ins had not occurred there before t The 1995).11 Repression is measured indirectly in two A of blacks threat model to be estimated is a variant of logistic ways.12 high proportion ened white dominance, and led to greater polit regression: ical and economic repression (Key 1949; Matthews and Prothro 1963). Therefore the per a?+ai Wt + 2p Ai +UmDmit centage of blacks in the county is an important Mft;,)= variable; we allow for a non-monotonic effect by introducing a squared term (an orthogonal where a, (3, and 8 are coefficients to be esti polynomial which eliminates the problem of mated. The characteristics of each city aremeas collinearity).13 There is also a dichotomous vari ured by k cross-sectional (time-invariant) able for states of the Deep South, where repres variables, Over such a short there is Xh period, sion was more extreme. Electoral power is no reason to estimate a separate intercept for measured (inversely) by the existence of a state each time t. It is necessary only to represent the poll tax, which was designed to disenfranchise of events. Sit-ins were more fre weekly rhythm blacks.14 quent on Thursday, Friday, and Saturday, which together are indicated by the dichotomous vari 9 able Wt. The impact of sit-ins elsewhere is cap Separate variables for elite and non-elite institutions tured by m diffusion variables, Dm. (based on the faculty-student ratio) did not reveal sig nificant effects. 10 Variables for median years of schooling, median Explanatory Variables male individual income, and the percentage of the male labor force in and technical did Table 1 lists descriptive statistics and bivariate professional positions not reveal significant effects. correlations for the explanatory variables (see 11 Variables for white students and establish for sources and We college Appendix definitions). begin coun ments owned by Fortune 1000 corporations in the that can be tested by by describing hypotheses ty (following James 1988) did not reveal significant For five cross-sectional variables. Hypothesis 1, effects. measure the extent of movement 12 variables Matthews and Prothro's (1966:166-7) measure of organization. NAACP had branches in 206 violence against blacks between 1955 and 1959 did not not ameas cities. Its strength ismeasured by the number of reveal a significant effect. We have included or branch members. In addition, dichotomous vari ure of repressive action such as arrests violence dur of the obstacles to ables are coded for the presence of a separate ing the sit-in campaigns because valid data across all cities this Youth Council (in 125 cities) or College Chapter collecting throughout Qualitative evidence, however, suggests that 15). Neither SCLC nor CORE had a period. (in only had minimal effect on the of formal structure of local branches. We define a repression spread protest. Most arrests occurred in a few cities and toward the end dichotomous variable for SCLC by coding the of the protestwave (Oppenheimer 1963). of an affiliated or mem 13 from presence organization An orthogonal squared term is the residual ber of the executive board in 34 (located cities), regressing JP onX. a 14 and for CORE by coding the presence of A variable for the percentage of blacks registered Chapter that had applied to affiliate with the to vote (U.S.Commission on Civil Rights 1959) did not national organization (located in 12 cities). reveal a significant effect. C o
?tj?-,-_-_-_-
PairwiseCorrelations Pearson
(14)COREChapter .04.19-.11 -.06.23 .21 .06 .07-.08 -.13 .24 .31 .25 .25 ?
(13)SCLCaffiliate .10.30-.15 -.01.35 .40 .00 .17-.11 .07 .46 .44 .29 .29
-.04-.06.04.21(12)CollegeChapter.28NAACP.43.34-.08.14.29.41-.11-.17
(8)DeepSouth (11)NAACPYouth.36.48Council .03.00.00.03 -.05.37-.06.48 .10-.15 .41-.14 .27 -.21 .04 -.16 .07 .45 .60 -.01-.14(V)-.237.6611.57(10)NAACP.65members.10.35.54-.22-.13
(7)Polltax instate .40.49.06 .00 -.14 -.12 -.03 .00
(9) Black population (logged) Cross-sectionalvariables (for 334 SD_(1) (2)(3) (4) (5) (6)1.10 (7) (8) (9) (10) (11) (12)-.29 .04 .75 .48 .12 .50 -.19 .04 (5) Black % of county cities)_Mean 12.93 -.08 .18 .01 .03 8.74 24.75
(6) Black % of county (squared .00 191.42 -.32 .23 .28 .17 .37 (6) D6: Sit-ins within 3 days .03 .18 .09 .19 -.02 .00 .00 3 (5) D5: Sit-ins within (4) D4:3 Sit-ins before 7 (V)distancebydays 105.63 .02.19 .93 .02 .20 8 .05 .06 -.06 -.05 w RegionalCouncilSouthernPresence(4)of-.14-.05.48.27.44 (7) W: .48 .50 .05 .07 .01 .03 .01 .04 55 athletic orthogonal) by if newspaper (2) D2: Sit-ins within if (V) -52 .21 .89 ** (3) Black college distance days 3.02 1.97 -.24 -.03 (logged)(3) D3: Sit-ins 29.88 22.52 -.15 -.03 3 Table 1. Descriptive and Correlation Matrix Thursday/Friday/Saturdaydays days % 7 7 students before (2) BlackStatistics (1) Dl:Sit-ins (for18,990nTime-varyingspells)variables days7 6.174.03 g 8.45 4.17 -.12 within 8.63(1)50.13Black%unskilled
unemployed
O- 762 AMERICAN SOCIOLOGICALREVIEW
The black population is entered as a control the inverse square root (cf., Hedstrom et al. variable.15 For variables that count the number 2000).19 The resulting sum is highly skewed, and of people, a diminishing marginal effect would so the square root is taken. The formula is thus be expected. To take an example, the difference the following: between 100 and 200 college students should be much greater than the difference between 1000 D2,= and 1100. Therefore these variables are VT||v^T logged?excepting NAACP membership, where Together, Dx and D2 capture the influence of sit the root is taken.16 square ins within seven Diffusion variables the of days. capture impact pre The initiation of sit-ins could also have a vious sit-ins elsewhere. Formal notation helps more enduring impact, especially because the to clarify their construction. They are derived initial event usually inaugurated a continuing from a dichotomous variable coded 1 if sit SjT, campaign of protest. A pair of diffusion vari inswere initiated in cityy in time t, and 0 other ables, D3 and Z)4, captures the influence of ear The initiation of sit-ins in another wise.17 city lier events: would have the greatest in the impact following t-S J days, as it provided new information about the of blacks elsewhere to ^3i*= willingness defy segre pf ^ ^/r gation. We have adopted (after testing alterna tives) a duration of seven days?just like the 1960s riots To J (Myers 1997:97-98,2000:185). D4it=VZZ\%1 T=ly=l facilitate interpretation, there is no time decay Two more elaborate diffusion variables within those seven The first diffusion vari days. require reconstructing how some cities were able counts the number of cities: simply connected to others. Hypothesis 2 is tested using t-\ j networks among black colleges established by affiliation with the same athletic association. T=t-lj=l The Central Intercollegiate Athletic Association, for example, linked various colleges in Hypotheses 2 and 3 imply that the impact of Maryland, Virginia, and North Carolina? events elsewhere will diminish with the prior including Charlotte's Johnson C. Smith distance between the cities, The function dtj.18 University. A sit-in in Charlotte, we hypothesize, al form after alternatives, is adopted, testing should have a particularly great impact on an affiliated city, because students there would be more likely to know someone at Johnson C. 15 Smith University, and moreover because they The black population is more appropriate than would not want to be outdone the rival cam the total population, as few whites participated in the by = A matrix is created with elements coded sit-ins. The two figures are highly correlated (r .91), pus. atj was a in was and so substituting the total population makes no 1 if there college cityy that affil difference to the results. iated with the same athletic association as a 16 Where the number of black college students is college in city /, and 0 otherwise. This matrix = zero, it is transformed as ln(\) 0. This procedure is used to construct the diffusion variable: is forNAACP because a inappropriate membership, t-\ J few cities had only a single member. 17 S is identical to Y, but a distinct symbol helps to clarify the difference between the dependent vari able (what happens here today) and the diffusion variables (what happened in other cities before today). 18 19 A few of the cities are very close to one anoth Myers (2001) pioneers a sophisticated method er, as close as one mile. There is only a single instance of empirically deriving the decay function, dividing cases of sit-ins in one city being followed by sit-ins in (i) the distribution of distances between all another city less than 10 miles away, and that hap with an event by (ii) the distribution of distances over a the inverse between all cases. When entered into the pened week later. Because square models, this function inferior to the root would give excessive weight to very close events, however, decay proved distances less than 10 miles are treated as 10 miles. inverse square root. DYNAMICSOF PROTESTDIFFUSION 763
Hypothesis 3 is tested using newspaper cir whether an observed pattern could have been culation. We focus on newspapers rather than produced by chance alone, or whether that is so broadcast media because the geographical extent unlikely that we are justified in attributing a of readership can be reconstructed with some causal relationship (Fox 1997:12). Standard on precision (rather than relying the transmis errors are estimated without assuming that a sion radius of broadcast station). In addition, observations from the same city on different we can content of news examine the coverage. days are independent.20 The model's overall There are data on the circulation of each daily ability to discriminate between spells with the to where it reached newspaper every county event and spells without is measured by the more than 5 of households. To take a percent area under the Receiver Operating Characteristic concrete the Charlotte Observer example, (ROC) curve, which can range from .5 (no dis and News circu (morning edition) (evening) crimination) to 1 (perfect discrimination).21 For lated to more than a dozen cities inNorth and comparison, a minimal model with only two South Carolina, up to 112 miles away. News of control variables?black population and the day sit-ins in Charlotte should therefore have a par of the week?yields an ROC area of .789. Table ticularly great impact on those cities. These cir 2 reports the results; Model 1 includes all vari culation figures refer to the entire population ables, while Model 2 drops the diffusion vari and not the black community alone. able derived from intercollegiate athletics. Nevertheless, black college students tended to There are mixed results for movement organ read white newspapers (61 percent) more than izations (Hypothesis 1). The membership of black newspapers (39 percent), and so these NAACP had no discernible effect, nor did the general circulation figures should capture the presence of aYouth Council.22 The presence of flow of news within the black community an SCLC affiliate is estimated to have had a (Matthews and Prothro 1975). A matrix is cre was a sizeable effect, though it is not statistically sig ated with elements Cy coded 1 if there nificant at the .05 level. The same holds for the newspaper published in cityy that reached more presence of an NAACP in than 5 percent of the households in city f s coun College Chapter Model 1. Because the variable is correlated ty, and 0 otherwise. This matrix is used to con struct the diffusion variable: with intercollegiate associations, Model 2 pro vides a better estimate.23 The effect increases in t~\ j magnitude and attains statistical significance. CORE had the greatest effect of all. (Note, how ever, that we cannot reject the hypothesis that For both diffusion variables, a duration of three SCLC, CORE, and NAACP College Chapters is days adopted after testing alternatives (includ all had the same effect.) These results suggest seven ing days). that protest was indeed orchestrated by some movement organizations. The mass member RESULTS ship of NAACP, however, does not seem to have been important; nor does SCLC have the promi The event-history models are estimated using nence attributed to it by some scholars. rare-events logistic regression (King and Zeng 2001) because the probability that a sit-in would occur for the first time in a on a particular city 20 = are certain day is very low (66 / 18,990 .0035). Models estimated using the ReLogit pack for Stata For ease of interpretation, the odds ratio (the age (Tomz, King, and Zeng 1999), speci of the fying the cluster option. exponent coefficient) is reported. Because 21 Rare-events is not a likelihood the hazard is so low, the odds ratio convenient logistic regression technique, and so the likelihood is not lymeasures how much an increase of one unit 22 log reported. Expressing NAACP membership as a proportion in the independent variable would multiply the of black population also has no significant effect. This probability of a sit-in?on each for a day, city negative finding holds even if the variables forYouth inwhich sit-ins had not occurred. We also report Council and College Chapter are omitted. (as is tests even 23 = customary) hypothesis though The two variables are modestly correlated (r the data constitute the rather than a population .27), as both indirectly reflect the presence of a col sample. Statistical inference helps us to decide lege. 764 AMERICAN SOCIOLOGICALREVIEW
Table 2. Determinants of the Hazard of a City's Initial Sit-in
_1_2_
Rare events logistic regression Odds SE Oddsp pSE Black unskilled % .950 .021 .02* .948 .022 .02*
Black unemployed % .908 .043 .04* .909 .043 .04* Black college students (logged) 1.884 .252 .00*** 1.915 .251 .00*** Presence of Southern Regional Council 1.401 .551 .39 1.404 .552 .39 Black % of county 1.023 .018 .21 1.022 .019 .22 Black % of county (squared orthogonal) .998 .001 .06 .998 .001 .06 Poll tax in state .522 .166 .04* .523 .166 .04*
Deep South .128 .046 .00*** .129 .046 .00*** Black population (logged) 1.148 .262 .55 1.132 .257 .59 NAACP members (V) -984 .011 .15 .985 .011 .18 NAACP Youth Council 1.386 .513 .38 1.366 .506 .40 NAACP College Chapter 1.997 .760 .07 2.141 .783 .04* SCLC affiliate 1.995 .739 .06 1.938 .710 .07
CORE Chapter 2.915 1.163 .01** 2.883 1.123 .01** Dl: Sit-ins within 7 days .739 .058 .00*** .733 .058 .00*** D2: Sit-ins within 7 days by distance (VxlO) 1.068 .167 .00*** 1.933 .302 .00*** D3: Sit-ins before 7 days .960 .020 .05* .960 .019 .05* D4: Sit-ins before 7 days by distance (VxlO) 1.297 .091 .00*** 1.293 .089 .00*** D5: Sit-ins within 3 days if same athletic affiliation 1.232 .248 .30 D6: Sit-ins within 3 days if newspaper circulates 2.986 1.281 .01* 2.882 1.232 .01* W: Thursday/Friday/Saturday 1.718 .443 .04* 1.72 .44 .03*
ROC area_.945_.944_ = = = error. Note: N 18,990 city-days (excluding Sunday). Odds odds ratio; SE robust standard */?<.05; **/?<.01; ***/?<.001 (two-tailedtests).
The diffusion variables reveal that recent unskilled occupations or were unemployed events in nearby cities had the greatest impact. (Hypothesis 5). Political opportunities were (Dx and D2 cannot be interpreted individually (Hypothesis 6) clearly important. Sit-ins because the occurrence of another sit-in affects were less likely in states that imposed a poll tax, less in the South. The both simultaneously; the same is true for Z)3 and and much likely Deep per of blacks in the to have D4.) Diffusion did not follow athletic intercol centage county appears had a non-monotonic effect neither legiate networks (Hypothesis 2). This variable (though term is at the .05 is therefore dropped from Model 2. By con statistically significant level), with sit-ins most where blacks con trast, newspaper circulation (Hypothesis 3) had being likely stituted about one third of the an important effect. The initiation of a sit-in in population. The of these effects Model the headquarters of a newspaper with wide cir magnitude (using can be the culation had a greater impact on cities within its 2) expressed by considering "typical" with median values on all variables circulation area, in the following three days. city, (the mode for dichotomous The This need not indicate that newspapers variables). typical finding a city has no movement organizations except were more important than broadcast media, as branch of the NAACP. The hazard of an initial the transmission networks for radio and televi sit-in is .0022.24 Establishing a CORE Chapter sion may have been similar in coverage. The finding is compelling evidence, however, for the news importance of media. 24 Hazards are calculated using the ReLogit pack Sit-ins were much more to occur where likely age, in the same manner as Clarify (King, Tomz, there were black students cross-sec many college and Wittenberg 2000). The median for (Hypothesis 4). The size of the black population, tional variables is calculated with each city having no were like the the lat by contrast, had effect. Sit-ins less equal weight (not risk-set, though using to ter makes minimal The median for time ly where many blacks were relegated difference). DYNAMICSOF PROTESTDIFFUSION 765
.02 n-*-1
~ '00 +1 sit-in within 3 in whose circulates here A^ days city newspaper
X. +1 sit-in within 7 days
- +1 sit-in beforeTo*ays:-?- ___^__
baseline - .0 4-1-,-r?,-,-,-,-r?,-,- -,- -,-,-,-,- -,-,-,-,-,- 1000 50 150 250200
Distance from Here (miles)
Figure 3. How the Hazard of a Sit-in Increased When Sit-ins Began in Another City would almost triple the hazard to .0064, an is depicted by the next curve. If the event had NAACP College Chapter would more than dou occurred 10miles away, then the hazard would ble it to .0048, and an affiliate of SCLC would nearly quadruple to .0082; if a hundred miles almost double it to .0043 (though the latter is not away, then the hazard would increase by a third, statistically significant at the .05 level). By to .0029. The median distance between cities comparison, increasing the number of black was about 500 miles: at such far remove, of college students from 16 to 425 (the 90th per course, the hazard would not increase. Consider centile) would increase the hazard of a sit-in finally the enduring impact of earlier events. The almost nine-fold, to .0197. lower curve shows how the hazard would change Figure 3 depicts the estimated hazard of a sit if a sit-in had occurred over seven days earlier in in for this typical city. The diffusion variables one more city (raising the total to 28). The impact begin at the median for cities in the risk-set: six is relatively slight, though still noticeable. It again cities had initiated sit-ins within the last seven diminishes with distance. days, and 27 cities had initiated sit-ins before In sum, then, event-history analysis reveals then. The baseline hazard (.0022) is shown by the why sit-ins occurred in some cities rather than horizontal line. Consider first the impact of recent others, and how protest spread across the South. were events. What would happen if another city initi Sit-ins were most likely where there many ated sit-ins, raising the number of recent events black college students (Hypothesis 4), where the more to seven? If the other city published a newspa black community had relatively autono per that circulates here, then the boldest curve my and resources (Hypothesis 5), and where shows how the hazard would be increased?even political opportunities were less unfavorable 250 miles away, the hazard would triple.25 The (Hypothesis 6). In other words, protest occurred where was least severe. Protest was more usual case, where no newspaper circulates, oppression orchestrated by movement organizations (Hypothesis 1), though what mattered was a variables is calculated from the varying risk-set, cadre of activists rather than a large member except the control variable for Thursday/Friday/ ship.26 Potential protesters were clearly inspired Saturday, which is set to 1. 25 The impact diminishes with distance, of course, because the sit-in has an effect via and as well 26 Dx D2 We have also tested a variant of Hypothesis 1, as via D6. in which the presence of an organization enhanced 766 AMERICAN SOCIOLOGICALREVIEW
by previous sit-ins, predominantly by what hap Whether newspapers were more likely to pened in neighboring cities. There is no positive report events in nearby cities can be investigat evidence for social networks as a channel of dif ed using event-history analysis. Here the "event" a fusion (Hypothesis 2), but the news media was is news report of another city's sit-in campaign. The variable is 1 if news clearly important (Hypothesis 3). dependent Mnjt, coded paper n mentioned a sit-in campaign in cityy at time and 0 otherwise. A is "at risk" ANALYSIS OF NEWSPAPERCOVERAGE t, campaign of being mentioned on each day after the initial We have seen that the initiation of protest in one sit-in. To take an example, Greensboro enters the city had a greater impact on nearby cities, and risk-set every day from February 2 onward, and each four one for news had its greatest impact within seven days. day yields cases, each on their Hypothesis 3 implies that this spatial and tem paper. Excluding newspapers reporting own there are cases poral pattern reflected the pattern of news cov cities, 11,059 (newspaper We are interested in the hazard rate: erage. In other words, potential protesters were city-day). more to take from sit-ins in likely inspiration = 1 = 1A n qnjt =prob(Mnjt V Sj7 *j) nearby cities because such events were more | to be their local news media. likely reported by The model to be estimated is a variant of logis This can be tested implication by examining tic regression: newspaper of the sit-ins. coverage = We compiled complete coverage from four Po+ PiK +M?, +M?/ + ln( " '' major newspapers: Atlanta Constitution, j-2^) (-1 Charlotte Observer, Houston Chronicle, and Miami Herald. For every day from February 1 are to April 14, we coded whether a city's sit-in where P coefficients to be estimated. Because events in are more campaign was mentioned in each newspaper. larger cities likely to be the is con Consider an article stating that "Concord reported, city's population (PJ) trolled. The distance from the head became the ninth North Carolina city touched newspaper's to the is entered as a by themovement Friday, and students followed quarters city (dnJ) A dichotomous variable the same pattern moving into four Rock Hill, four-degree polynomial. is coded for whether the initial sit-in S.C., business establishments. Demonstrations city's occurred within seven days. have also spread to DeLand, Fla., and Hampton, Results are depicted in Figure 4, which shows Va." (Charlotte Observer 1960:1). The four the probability of a newspaper reporting a city's named cities are coded as being mentioned. sit-in campaign on a given day. As expected, Besides reports of sit-ins, we included reports coverage ismost likely within seven days of the of negotiations, boycotts, and other events relat city's initial sit-in. Most importantly, coverage ed to a city's sit-in campaign. Each newspaper diminishes significantly with distance. This mentioned a substantial fraction?from half to echoes the pattern evident in Figure 3, though the 66 cities with sit-ins dur three-quarters?of the decline with distance is less pronounced in the This was suffi ing period. coverage clearly Figure 4. That sit-in campaigns in nearby cities cient to the extent of the wave. convey were more likely to be reported in the local news media therefore helps to explain why near by events had a greater impact. Faraway events the to events in other cities "susceptibility" (Strang had a lesser impact (in part) because potential and Tuma interaction terms 1993), by creating protesters were less likely to learn about them between the number of recent events (D\) and the from their local news media. This finding com dichotomous variables for organization. None of plements the finding that the diffusion of protest these coefficients is significantly different from zero; tended to follow the circulation of newspapers. moreover, none of the coefficients for organization When a newspaper circulated to other cities, is statistically significant. In other words, the data do in those would more read not allow us to discriminate between the main effect then protesters cities learn about sit-ins in the head and any interaction effect. Testing whether college ily newspaper's would have to on students enhanced susceptibility to events elsewhere quarters; otherwise, they rely same own produces the indeterminate result. coverage in their local media. DYNAMICSOF PROTESTDIFFUSION 767
T3 .6 1-1
f
I'4 ^\
: within 7 of initial sit-in ^***? : more than 7 after initial sit-in ^^^''o^^^^s3 days ^ ' ' ' H ? ?-__J .0 H-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1 1000 50 150 250200 Distance from Newspaper (miles) a Figure 4. The Probability of a Sit-in Campaign Being Reported in Daily Newspaper QUALITATIVEEVIDENCE ON recalled "a very small group" (Powledge DIFFUSION 1991:205). Although this group provided cru cial leadership, Lewis noted that "about ninety We turn now to evidence on the qualitative to ninety-five percent of the people who showed of drawn from process diffusion, contempora up at the church the week before [the first sit neous records and from ... organizational partici in] had no training in nonviolence but they accounts. We consider first pants' retrospective were ready to go" (Viorst 1979:107). In movement and then social net organizations, Winston-Salem, North Carolina, the sit-ins were works and news media. led by a recent graduate of the Teachers' College, who was on the executive board of the local Movement Organizations NAACP branch (Bryan 2000; NAACP 1960; Patrick I960).27 Protesters came from the In what ways did movement organizations NAACP College Chapter?with 270 members orchestrate Our protest? quantitative analysis in 1959, it was the largest in the South (the demonstrates that the conventional measure of median was 64). Tallahassee's first sit-ins were number of mem organizational strength?the conducted by a Chapter of CORE that had been bers?was not important. Instead, what mat established inOctober 1959, thanks to the efforts tered was a cadre of whether associated activists, of two sisters from the Agricultural and with an NAACP or CORE, College Chapter, Mechanical College. It counted just 18 dues SCLC. The of these (perhaps) "presence" organ paying members (Rabby 1999:86). izations indicated a network of people com mitted to the movement against racial 27 oppression. A well-known example is thework Morris (1984: 200) contends that inDurham and shop on nonviolence established by the Winston-Salem, "McKissick, Moore, Carey, and others those studentsfrom Nashville Christian Leadership Council (affil helped organize protests by bringing local colleges to churches, where were trained to con iated with SCLC), led by Reverend James they duct sit-ins? and Lawson. A handful of students from the Following training strategy sessions, city's the students went to the local lunch counters and sat in." institutions of education higher began attending Detailed accounts of events in Winston-Salem and the in the fall of 1959 workshop (Halberstam Durham suggest that student leadership and initiative 1998; Wynn 1991). John Lewis, from the was more crucial (Bryan2000; Dalton 2001; Gallo 1978; American Baptist Theological Seminary, Howard 1983;Greene 2005; Patrick 1960;Sindler 1965). 768 AMERICAN SOCIOLOGICALREVIEW We should not overlook cases where move however. The student leader at Charlotte's ment organizations were less successful or were Johnson C. Smith University, Charles Jones, not necessary. Indeed, ten cities with SCLC refused help from NAACP and publicly rebuked affiliates had no sit-ins throughout this period, CORE'S Carey (NAACP 1960:7; Oppenheimer including New Orleans, Mobile, and Louisville 1963:180; Wehr 1960:29). "We are willing to lis (Gaillard 2004; Fairclough 1995). In New ten towhat people like him [presumably Carey] Orleans, Reverend A. L. Davis led the city's have to say," another leader warned, "but we are Interdenominational Ministerial Alliance and going to keep control of this thing ourselves" belonged to SCLC's executive board. Yet the city (Wehr 1960:28). Even Carey acknowledged that had no sit-ins: students at Dillard University the "'outside agitators' were being pulled along were dissuaded by the threat of expulsion by the kids. Itwas purely spontaneous_[W]e (Fairclough 1995; Rogers 1993). Conversely, were not leaders; we were followers" (Powledge there were six cities where sit-ins occurred 1991:214). Morris (1981, 1984) implies that despite the absence of any movement organi Baker coordinated protest from SCLC's head as zation (including an NAACP branch) such quarters in Atlanta. In fact, her achievement Elizabeth City, North Carolina. was to encourage students already involved in a On balance, of course, the presence of local the sit-ins to organize at the regional level, at the cadre of activists did make sit-ins more likely. Easter conference held at Shaw University. As on own Did these local activists act their ini she recalled, "[T]here was little or no commu tiative, or were they responding to directions nication between those who sat in, say, in from the staff and leaders of their organiza Charlotte, North Carolina, and those who sat in tions? Even before the sit-ins had spread beyond at some other place in Virginia or Alabama. asked Greensboro, the local NAACP president They were motivated by what the North CORE'S national office (inNew York) for help. Carolina four had started, but they were not in two sec CORE immediately dispatched field contact with each other" (Cantarow 1980:83). James to retaries, Gordon Carey and McCain, The activity of movement organizations was Given NAACP's ambivalence aid the protesters. surely vital for sustaining the sit-in campaigns? about confrontational protest, the organization but this should not be confused with the orches was to the also surprisingly quick respond: tration of protest in cities where it had not yet youth secretary, Herbert Wright, traveled from occurred. There are surprisingly few instances York to North where he was New Carolina, of organizers initiating protest in other cities. the state's field Charles assisted by secretary, Moore attempted to broaden the protest beyond Reverend Fred McLean. Within SCLC, North Carolina by calling Lawson inNashville and Reverend Moore Shuttlesworth Douglas on February 10 (Wynn 1991:45).29 Despite were the particularly enthusiastic, although Moore's prodding, Lawson and the head of the organization lacked any equivalent paid organ Nashville Christian Leadership Council were Within two all three move izer.28 weeks, then, cautious, and they tried to persuade students in ment were involved. a organizations actively the workshop to delay protest. The riposte by Inmost outside followed rather cases, support student at American Baptist seminary is well in a than preceded protest given locality. known: "I am sick and tired of waiting!" (Morris McLean that the sit-ins in North reported 1984:206). CORE provides less ambiguous Carolina "were spontaneous, but after they start instances of orchestration from outside. McCain to the NAACP ed the student leaders turned called members in Tallahassee, encouraging for funds advice, legal assistance, necessary them to begin protest on February 13 (Rabby and assistance to on the carry protest" (NAACP 1999:88). The city's first sit-in occurred on that 1960:15). Similarly, protesters could ask CORE'S field secretaries or SCLC's ministers to in nonvio come and provide practical training 29 Some accounts (Branch 1988;Halberstam 1998; lence. Outside help was not always welcome, Morris 1984) place this telephone call earlier in the week, on the basis of retrospective interviews. The in the 10th is attested by a contemporary report 28 and SCLC only hired a field secretary after thewave Tennessean, quoting Lawson (see Sumner 1989) in Peck of sit-ins (Fairclough 1987). by Paul Laprad (reproduced 1962:64). DYNAMICSOF PROTESTDIFFUSION 769 date. A visit by McCain to Orangeburg, South a student atMorehouse College, read about the Carolina, also seems to have orchestrated Greensboro sit-ins in the Atlanta Daily World, protest, even though the city lacked a CORE a black newspaper. He took the article to a fel Chapter; the nearest was 37 miles away (Morris low student, Julian Bond, asking, "Don't you 1984:201; Meier and Rudwick 1973:104).30 think it should happen here?" (Raines 1977:84). Overall, then, movement organizations did The two went on to lead the sit-ins inAtlanta. facilitate protest, but the diffusion of sit-ins Broadcast media were also important. In the depended more on the existence of a cadre of same city, Ruby Doris Smith at Spelman College local activists before February 1 than on the watched the Greensboro sit-ins on television: "I subsequent activities of organizational staff and began to think right away about it happening in leaders. Those activities were crucial for sus Atlanta" (Zinn 1964:17, see also Sitkoff 1993; taining protest once it had begun, of course. Tyson 2004). Jones, the student who led the sit Movement organizations also applied additional in in Charlotte, heard about Greensboro on his pressure on department stores. CORE led a car radio. "All of a sudden," he recalled, "there boycott and protest campaign inNorthern cities, was a handle to getting at this stuff" (Powledge and NAACP belatedly endorsed a general boy 1991:224). Contemporary studies likewise cott inmid-March. Focusing on the initiation of emphasize the importance of news media. protest, however, we see that local activists were According to Oppenheimer (1963:62), "[T]he responding above all to sit-ins occurring else most common source of information is attrib where. uted by students to the radio and newspapers" (see also Wehr 1960:25; Laue 1966:81). The Social Networks and News Media importance of the media is compatible with Lazarsfeld's two-step flow of communication, Our that infor quantitative analysis suggests whereby the mass media influences "opinion mation about elsewhere was protest conveyed by leaders," who then transmit ideas more broad news media rather than social networks. ly through personal contact (Lazarsfeld, that on a meas Although finding depends proxy Berelson, Gaudet 1968; Katz and Lazarsfeld ure for it is reinforced networks, by qualitative 1955). Diffusion among cities corresponded to evidence. Wehr discovered "little if (1960:25) the first step as opinion leaders like King in contact and no real liaison" any among pro Atlanta and Jones inCharlotte learned of protest testers at different cities in North Carolina. elsewhere, while the mobilization of students sit-ins were facilitated Clearly by friendships within each city corresponded to the second students within a and among college, friendships step of communication through personal net at same among students different colleges in the works. There is a lack of city. surprising evidence, The news media did not, of course, intend to however, for social networks as chan acting encourage protest. Some white newspapers were nels for the diffusion of cities. In protest among sympathetic to the goals of the sit-ins, but they accounts and inter contemporary retrospective invariably opposed the tactic?as indeed did views, we have found no of examples protest some black newspapers, including the Atlanta ers informed of sit-ins being by personal Daily World (Walker 1964). What mattered was acquaintances elsewhere. not editorial endorsement but rather information By contrast, several mentioned protesters about protest elsewhere (as the example of King that they first heard of events elsewhere from the illustrates). Information, most basically, made news media. When a in reporter Chattanooga potential protesters aware of this novel tactic. a asked protester where the idea of a sit-in was or orig Awareness rarely enough, else the sit-ins he inated, replied: "[W]e read the papers. We got would have all been initiated after the first our ideas from other people" (Atlanta reports of events inGreensboro. Potential pro Constitution 1960:2). InAtlanta, Lonnie were King, testers naturally cautious about directly confronting the system of racial oppression? risking violence, arrest, and expulsion from 30 The of was In this instance, the positive effect of move college. relevance Greensboro ques ment organization is not captured by our event tionable for those who faced far greater oppres history analysis. sion in the Deep South; students inBirmingham 770 AMERICAN SOCIOLOGICALREVIEW actually waited over six weeks to initiate sit-ins interstate buses setting off another tactical inno despite encouragement from SCLC's vation?the Freedom Ride?which was repli Shuttlesworth (Manis 1999). What was crucial, cated over 60 times in 1961 (Arsenault 2006). we argue, was information about the extent of Following the sit-ins, SCLC also shifted its the wave. Quantitative analysis shows how the strategy to intensive protest, most notably in hazard of a sit-in increased as additional cities Albany in 1961-62 and in Birmingham in 1963. were swept up in the wave. The effect was These campaigns represented the tip of the ice expressed by Peace, a student leader inRaleigh, berg as black communities throughout the South more con on February 9, after the wave had reached three organized local campaigns using cities (and would encompass two more by day's frontational tactics. In 1963 alone the Southern end). "The damned thing is spreading all over Regional Council identified 930 demonstra over twen the place and we haven't made a move yet" tions in 115 cities in the South with (Mitchell 1962:75). Raleigh's sit-ins began on ty thousand arrests (Heacock 1965). In short, the following day. "the sit-ins revitalized all of the major civil to Scrutiny of qualitative evidence, along with rights organizations, led the creation of a was to analysis of newspaper coverage, strengthens SNCC, reinforced tactic that domi our confidence in the event-history analysis nate the movement in the next few years, and once and enhances our understanding of the results. generally set the black struggle inmotion These various sources of evidence provide a again" (McAdam and Sewell 2001:108). to historical clear picture of the diffusion of sit-ins in the Corresponding their importance, 1960 are cited in the lit spring of 1960. Blacks, predominantly college the sit-ins frequently on students, initiated sit-ins because they were erature social movements. We have analyzed one dimension of this wave of the dif inspired by previous sit-ins in other cities. protest, across Our Information about events elsewhere came pri fusion of sit-ins to cities the South. involved of marily from news reports. As they learned about analysis systematic comparison cities in the wave with those that the increasing number of sit-ins, especially in swept up remained and detailed attention to cities nearby, potential protesters became more untouched, the of events. from the event optimistic about the prospect of success?not, timing Findings it should be reiterated, because success was evi history analysis of diffusion have been ampli from sources dent, but simply because so many others were fied by evidence contemporary with a initiating sit-ins. Local activists, primarily stu and retrospective interviews, along sys of news dents and ministers, played an important role in tematic analysis reports. The debate over orchestrating protest. Although the leaders of all "organization" (Morris 1981, versus three movement organizations had a modest 1984) "spontaneity" (Killian 1984; Oberschall can now be resolved. Both impact on the spread of protest, they contributed 1989) sides are correct. On one movement to realizing the potential of the sit-ins by pro partly hand, were viding financial assistance and training once organizations undoubtedly important. They were central in and protest was initiated. coordinating sustaining protest campaigns once they were underway. As for the initiation of the presence of CONCLUSION protest, a CORE Chapter or NAACP College Chapter of a The effect By the middle of April 1960, the movement at least doubled the chance sit-in. was similar against racial oppression had been transformed. for SCLC of magnitude, though Tens of thousands of blacks in the South had this estimate is more uncertain. On the other the sit-ins cannot be attributed to physically challenged segregation; black college hand, simply were Most cities students, previously politically quiescent, formal movement organizations. NAACP now in the vanguard of the movement. with organization (including branches) some occurred in the Nonviolent protest increased in intensity as well did not have sit-ins; sit-ins as extent; students arrested during a sit-in in absence of any organization. The best predictor was of black stu Tallahassee staged the first jail-in (by refusing of protest the number college the main channel of diffusion bail) inMarch. In the following year, a group dents. Moreover, because blacks of CORE activists traveled through the South on was the media. Protest spread DYNAMICSOF PROTESTDIFFUSION 771 were inspired by the actions of others else spark a wave of sit-ins in the South (see also where, which they usually learned about from Oppenheimer 1963:53; Oberschall 1989:43).31 news reports. In this sense, the sit-ins were Our findings also illuminate another question: "spontaneous" rather than organized by the why did thewave of sit-ins begin inGreensboro? leaders of movement organizations. Was this geographical origin purely "contin By concentrating on the rapid diffusion of gent"? To answer this question empirically, we protest in the course of months, we have not can return toModel 2 (in Table 2) and compare addressed the broader question of timing: why the estimated hazards for each city on February did the wave of sit-ins occur in 1960 rather than 1.This procedure effectively "controls for" geo some years before or after? The upsurge in graphical location, in a way that a cross protest in 1960 cannot be predicted by long-term sectional analysis would not.32 Although social and political changes. This can be shown Greensboro does not have the highest estimat using a recent time-series analysis of black ed hazard, it ranks third behind Tallahassee and protest from 1947 to 1998 (Jenkins et al. 2003). Durham. These estimates accord with historical Independent variables include NAACP mem evidence. Activist cadres in those two cities had bership, various measures of grievances, such conducted "test" sit-ins in 1958 or 1959, and as black unemployment, and various measures students inTallahassee were actively preparing of political opportunities, such as the power of for a sit-in campaign at the beginning of 1960 non-Southern Democrats. Replicating this (Killian 1984; Morris 1984). Similar prepara analysis, we find that slightly more protest tions were under way inNashville (Halberstam events are predicted in the late 1950s than in the 1998), which ranks fourth. Estimated hazards early 1950s. Nevertheless, the upsurge in 1960 vary by many orders of magnitude: the hazard is not predicted; the residual for that year is is about two thousand times higher for huge (results available from authors). This fail Tallahassee than for the median city, and two ure to predict 1960 using annual intervals and hundred thousand times higher than for the city aggregate measures of political opportunity and with the lowest hazard. The action of four stu more organization underscores the need for dents inGreensboro on February 1was contin temporally and spatially refined analysis. gent in the sense that it is easy to imagine awave Our findings suggest several explanations of sit-ins starting instead in Tallahassee, for wave for the timing of this of protest. First, example?but not in the sense that sit-ins were movement organizations increased their pres likely to have been initiated anywhere in the ence in the late 1950s: SCLC was established, South. CORE to to the and began spread South, Before turning to the broader implications for were NAACP College Chapters founded. What research on the diffusion of protest, we should was not of mass mattered the growth member acknowledge the particular characteristics of measured Jenkins et al. but the ship (as by 2003) this historical episode. Most important was that of activist cadres. black col expansion Second, racial oppression in the South?enforced by the number of students leges expanded; law as well as norms?created an inescapable increased 35 in the 1950s. The by percent third collective identity. Therefore when blacks chal ismore tentative. as point Newspapers?black lenged white supremacy in one place, blacks well as have been less to white?may likely elsewhere in the South could more readily iden report protest before the mid-1950s. After all, the leaders of theMontgomery bus boycott were ignorant of the bus boycott inBaton Rouge that 31 The nearest city with a sit-in in 1960?Little had occurred just two years before (Meier and Rock, Arkansas?is about 200 miles away from Rudwick 1976:382). The Montgomery cam Oklahoma City. paign (and perhaps events in Little Rock, too) 32 Some cities (like Portsmouth, Virginia) were seems to have made black protest in the South surrounded by other cities where conditions were more our demon to newsworthy. Finally, analysis conducive protest, while others (like Austin, Texas) strates how the of a sit-in elsewhere were impact relatively isolated. By explicitly modeling spa diminished with distance. This helps explain tial diffusion variables, we can disentangle charac a why sit-ins in Oklahoma City in 1958 did not teristics of city and its geographical location. 772 AMERICAN SOCIOLOGICALREVIEW tify with their actions. The attribution of simi importance of media, either by tracing the phys larity was even more pronounced among black ical transmission of information (Myers 2000) college students. These were unusually fruitful or by analyzing the information itself conditions for the rapid diffusion of protest. (Koopmans and Olzak 2004). Here we have Many other social movements lack such preex done both (like Roscigno and Danaher 2000), isting cultural conditions for mobilization looking at the geographical circulation of news of (Morris and Braine 2001). Characteristics papers and the content of news reports. The the tactic itself have enhanced the may possi findings of our analysis and other recent stud of diffusion. extended sit-in bility Although ies also raise important questions about long and campaigns required planning coordination, term historical change. Social movements and a sit-in could be initiated a handful of activists by protest waves have been shaped by the mass more than enthusiasm and possessing nothing media, more than we have hitherto appreciated. courage. these characteristics Despite particular Therefore changes in media technologies and of the sit-ins in 1960, our has two analysis institutions have important implications. organiza important implications?regarding Whether new communication technologies?the tion and media?for the literature on social Internet and cell phones?are beginning to sup movements. plant the mass media?newspapers, radio, and Our findings confirm the significance of television?is a question for studies of con movement organizations, but they also high temporary protest waves. light crucial differences among types of organ Most studies of protest focus on the longer ization and specify how organization mattered. development of protest cycles (Koopmans 1993; NAACP was extraordinarily strong, with mass Oliver and Myers 2003; Tarrow 1998). We have membership and considerable resources. CORE, made the case for studying the spread of con by contrast, was weak; in 1959, its income was frontational protest forms over a relatively short only 7 percent of NAACP's (McAdam time where new insights regarding protest 1982:253; Marger 1984:23). Despite the dis period can be revealed (Koopmans 2004; parity, CORE played themore important role in dynamics McAdam, Tarrow, and 2001). This ismoti propagating the sit-ins (see also Ganz 2000). Tilly vated the that of Moreover, NAACP's contribution to the sit-ins by expectation periods rapid came more from the actions of members in diffusion have transformative effects on the of social movements College Chapters (and decisions of local branch development (McAdam and Sewell Further is needed to es to support protest) than from the decisions of 2001). analysis determine whether if how moments the organization's national leadership. and, so, like the of 1960 the of Movement organizations were important inso spring shape trajectories the and far as they incubated local cadres of activists; individual activists; goals, leadership, themobi such activists were ready to take advantage of resources of movement organizations; the protest occurring elsewhere and did not have to lization and strategies of opponents; and wait for orders from above. Future research attainment of movement goals at the community should develop amore nuanced understanding or national level. of the relationship between formal organiza Kenneth T. Andrews is Assistant tion and collective protest. Professor of at the North Carolina at the importance of news Sociology University of By emphasizing move Hill. His research focuses on social media, our findings warn against the tendency Chapel institutions, and social His to assume that relational channels of diffusion? ments, political change. is a Constant recently published book?Freedom here differentiated into formal organization and Struggle: The Mississippi Civil Rights Movement informal networks?are inevitably more impor and ItsLegacy (University of Chicago Press, 2004)? tant than non-relational channels (in a different movement examines the influence of the civil rights see Van den Bulte and Lilien social context, 2001). on electoral politics, school desegregation, and Mobilization on the local level on social depends policies. Other current projects investigate organi of between networks, but the diffusion protest zation and leadership in contemporary movements, on sector in locales may depend information transmitted including a study of the environmental by the media. This should not be surprising. North Carolina and a national study of the Sierra Recent studies have also demonstrated the Club's state and local organizations. DYNAMICSOF PROTESTDIFFUSION 773 Michael Biggs is lecturer in Sociology at the appeared inAmerican Journal of Sociology, Theory on University of Oxford. A companion article the and Society, and Social Science History. He has examines on sit-ins, which the determinants of individ begun research self-inflicted suffering as a means ual participation, will appear in Mobilization. of protest: an essay on self-immolation is included in on Previous research the dynamics of protest, inves Making Sense of SuicideMissions (OxfordUniversity tigating the strike wave in Chicago in 1886, has Press, 2005). APPENDIX: EXPLANATORYVARIABLES Table Al. Independent Variables Variable Name Description Source Black population Nonwhite population, 1960 U.S. Bureau of the Census 1963, table 21 Black unemployed % Unemployed nonwhite males / US. Bureau of the Census 1963, table 77 nonwhite males in civilian labor force, 1960 Black unskilled % Nonwhite males in unskilled US. Bureau of the Census 1963, table 78 occupations (private household workers; service workers; farm laborers excluding unpaid and foremen; other laborers) / non white males in civilian labor force, 1960 NAACP members Average number of members of NAACP Papers, Part 25, Series D, Reel 3: NAACP chapter, 1957 and 1959 Total 1957 Memberships and Freedom Fund Contributions Received from Branches; Part 25, Series D, Reel 24, Total 1959 Memberships and Freedom Fund Contributions Received NAACP Youth Council 1 if city has NAACP Youth NAACP Papers, Part 19, Series D, Reel Chapter, 1958 or 1959 14:Total 1958Youth Membership Received; Youth and Student NAACP College Chapter 1 if city has NAACP College Memberships Received from Region V Chapter, 1958 or 1959 During 1959;Youth and Student Memberships Received from Region VI During 1959; Statement of Virginia Youth Memberships SCLC affiliate 1 if city has SCLC affiliate(s) or is SCLC Papers, Reel 1,Part 2: Affiliates of on the SCLC represented the Southern Christian Leadership Executive Board, February 3, Conference, Inc., February 3, 1960 1960 CORE 1 if has CORE at Chapter city Chapter the CORE Papers, multiple reels; Meier and beginning of 1960 Rudwick 1973:83-92 Black college students Nonwhites enrolled in college, US. Bureau of the Census 1963, table 77 1960 Black % of county Nonwhite population / total popu- U.S. Bureau of the Census 1963, table 21 lation of county, 1960 Presence of Southern 1 if has individual or Regional city organiza- Southern Regional Council Papers, Reel Council tion affiliated with Southern 75: State Organizations, officer lists, Regional Council, 1955 Feb. 4, 1953-Dec. 31, 1967; SRC Affiliated Organizations, 1955 Poll tax in state 1 if state had a tax poll Matthews and Prothro's Southern County Data (courtesy of James Alt); Key 1950; Keyssar 2000 Deep South 1 for Alabama, Georgia, Louisiana, Black and Black 1987 Mississippi, South Carolina 774 AMERICAN SOCIOLOGICALREVIEW Table A2. Matrices for Diffusion Variables Median distance Matrix DescriptionSource between linked cities Distance in miles between i and dis- U.S. 508 miles djj city j; Geological Survey's < tances 10 miles treated as 10 miles Geographic Names Information System, http://mapping.usgs.gov ay 1 if city i has college affiliated with same Chicago Defender, 13Aug. 210 miles (160 city intercollegiate association (Central, pairs)1960 Eastern, Gulf Coast, Midwestern, Southern, South Central, or Southeastern) as college in city j 1 if reached Standard Rate and Data Cy newspaper published in city j 47 miles (423 city pairs) more than 5% of the households in city i's Service 1960 county, for daily (morning or evening) _editions_ REFERENCES Mountain: Women Working for Social Change. Old Westbury, NY: Feminist Press. Kathleen and Amenta, Edwin, Dunleavy, Mary Carson, Claybome. 1981. In Struggle: SNCC and the Bernstein. 1994. "Stolen Thunder? Huey Long's Black Awakening of the 1960s. Cambridge, MA: 'Share Our Wealth,' Political Mediation, and the Harvard University Press. Second New Deal." American Review Sociological -. 1986. "Civil Rights Reform and theBlack 59:678-702. Freedom Struggle." Pp. 19-32 in The Civil Rights Andrews, Kenneth T. 2004. Freedom Is a Constant Movement in America, edited by Charles W. Eagles. Struggle: TheMississippi Civil Rights Movement Jackson, MS, and London, England: University and Its Legacy. Chicago, IL: University of Chicago Press of Mississippi. Press. Chafe, William H. 1980. Civilities and Civil Rights. Arsenault, Raymond. 2006. Freedom Riders: 1961 New York: Oxford University Press. and the Struggle for Racial Justice. New York: Charlotte Observer. 1960. "41 Raleigh Negroes Oxford University Press. Arrested in Protest for Food Service: N.C. Move Atlanta Constitution. 1960. "Cafe Push to Spreads 1. Spreads to 9 Cities." February 13, p. Tennessee." February 20, p. 2. Conell, Carol and Cohn, Samuel. 1995. "Learning Beck, E. M. and Stewart E. Tolnay. 1990. "The Killing from Other People's Actions." American Journal Fields of the South." American Deep Sociological 101:366-403. Review 55:526-39. of Sociology Carol and Kim Voss. 1990. "Formal in Conell, Biggs, Michael. 2003. "Positive Feedback and the Fate of Social Movements." Collective Mobilization: The American Strike Organization American Sociological Review 55:255-69. Wave of 1886." Theory and Society 32:217-54. Constable, John. 1960. "Negro Student Protests -. 2005. "Strikes as Forest Fires: Chicago Challenge N.C. Leaders." New South 15:3-10. and Paris in the Late 19th Century." American CORE. 1960. the Color Line: Non-Violent Journal of Sociology 110:1684-714. Cracking Direct Action Methods of Eliminating Racial -. Forthcoming. "Who Joined the Sit-ins and Discrimination. New York: CORE. Why: Southern Black Students in theEarly 1960s." director. 2001. Mobilization. Dalton, Mary M., [Videorecording] "I'm Not Brother's Leadership and Black, Earl and Merle Black. 1987. Politics and My Keeper: Civil inWinston-Salem, North Carolina." Society in the South. Cambridge, MA: Harvard Rights Winston-Salem, NC: Wake Forest University. University Press. in Wilma and James Stokely. 1960. "Sit Branch, Taylor. 1988. Parting the Waters: America Dykeman, Down Chillun, Sit Down!" The June: the King Years, 1954-63. New York: Touchstone. Progressive 8-12. Bryan, G. McLeod. 2000. Making History: February John. 1995. Now the The 23, 1960, Winston-Salem, North Carolina. Egerton, Speak Against Day: Generation the Civil Movement in Winston-Salem, NC: Wake Forest University and Before Rights NC: of North Winston-Salem State University. the South. Chapel Hill, University Press. Cantarow, Ellen with Susan Gushee O'Malley and Carolina Dissent in Wichita: The Sharon Hartman Strom. 1980. Moving the Eick, Gretchen Cassel. 2002. DYNAMICSOF PROTESTDIFFUSION 775 Civil Rights Movement in theMidwest, 1954-1972. American Protest, 1948-1997." American Journal Urbana and Chicago: University of Illinois Press. of Sociology 109:277-303. Eisinger, Peter K. 1973. "The Conditions of Protest Katz, Elihu and Paul F. Lazarsfeld. 1955. Personal Behavior inAmerican Cities." American Political Influence: The Part Played by People in theFlow Science Review. 67:11-28 of Mass Communications. New York: Free Press. Fairclough, Adam. 1987. To Redeem the Soul of Key, V O., Jr. 1949. Southern Politics in State and America. Athens, GA: University of Georgia Press. Nation. New York: Knopf. -. 1995.Race & Democracy: The Civil Rights Keyssar, Alexander. 2000. The Right to Vote. New Struggle in Louisiana, 1915-1972. Athens, GA: York: Basic Books. University of Georgia Press. Killian, Lewis M. 1984. "Organization, Rationality -. 2001. Better Day Coming. New York: and Spontaneity in the Civil Rights Movement." Penguin. American Sociological Review 49:770-83. Fox, John. 1997'. Applied Regression Analysis, Linear King, Gary, Michael Tomz, and Jason Wittenberg. Models, and Related Methods. Thousand Oaks, 2000. "Making the Most of Statistical Analyses." CA: Sage. American Journal of Political Science 44:347-61. Gaillard, Frye. 2004. Cradle of Freedom: Alabama King, Cary and Langche Zeng. 2001. "Explaining and the Movement that Changed America. Rare Events in International Relations." Tuscaloosa, AL: University of Alabama Press. International Organization 55:693-715. Gallo, Elyse. 1978. "The Emergence of Direct Action: Klarman, Michael J. 1994. "How Brown Changed The Early Civil Rights Movement in Durham, Race Relations: The Backlash Thesis." Journal of North Carolina." Manuscript, Chris Howard American History 81:81-118. Papers, Duke University, Durham, NC. Koopmans, Ruud. 1993. "The Dynamics of Protest Ganz, Marshall. 2000. "Resources and Waves: West Germany, 1965-1989." American Resourcefulness: Strategic Capacity in the Sociological Review 58:637-58. -. Unionization of California Agriculture, 2004. "Protest in Time and Space." Pp. 1959-1966." American Journal of Sociology 19-46 in The Blackwell Companion to Social 105:1003-62. Movements, edited by D. A. Snow, S. A. Soule, and Gould, Roger. 1991. "Multiple Networks and H. Kriesi. London, England: Blackwell. Mobilization in the Paris Commune, 1871." Koopmans, Ruud and Susan Olzak. 2004. "Discursive American Sociological Review 56:716-29. Opportunities and the Evolution of Right-Wing R. to Graves, Carl 1981. "The Right be Served: Violence in Germany." American Journal of Oklahoma City's Lunch Counter Sit-ins, Sociology 110:198-230. 1958-1964." Chronicles of Oklahoma 59:152-66. Kriesi, Hanspeter, Ruud Koopmans, Jan Willem Greene, Christina. 2005. Our Separate Ways: Women Duyvendak, and Marco Giugni. 1995. New Social and the Black Freedom Movement in Durham, Movements in Western Europe: A Comparative North Carolina. Chapel Hill, NC: University of Analysis. Minneapolis, MN: University of North Carolina Press. Minnesota Press. Halberstam, David. 1998. The Children. New York: Laue, James H. [1966] 1989. Direct Action and Fawcett. Desegregation, 1960-1962. Brooklyn, NY: Heacock, Ronald T. 1965. Understanding the Negro Carlson. Protest. New York: Pageant. Lazarsfeld, Paul F, Bernard Berelson, and Hazel Peter. 1994. Hedstrom, "Contagious Collectivities: Gaudet. 1968. The People s Choice: How the Voter On the a Spatial Diffusion of Swedish Trade Unions, Makes up His Mind in Presidential Campaign, 1890-1949." American Journal of Sociology 3d ed. New York and London, England: Columbia 99:1157-79. University Press. Rickard and Carlotta Stern. Hedstrom, Peter, Sandell, Lomax, Louis E. 1960. "The Negro Revolt Against 2000. "Mesolevel Networks and the of Diffusion 'The Negro Leaders.'" Harpers, June, pp. 41-48. Social Movements." American Journal -. of 1962. The Negro Revolt. New York: Harper Sociology 106:145-72. and Row. Chris D. on Howard, 1983. "Keep Your Eyes the Luper, Clara. 1979. Behold the Walls. Oklahoma Prize: The Black Struggle for Civic Equality in City, OK: JimWire. North Durham, Carolina, 1954-1963." Thesis, Maney, Gregory M. and Pamela E. Oliver. 2001. of Duke Department History, University, Durham, "Finding Collective Events." Sociological Methods NC. and Research 30:131-69. James, David R. 1988. "The Transformation of the Manis, Andrew M. 1999. A Fire You Can't Put Out: Southern Racial State." American Sociological The Civil Rights Life of Birmingham s Reverend Review 53:191-208. Fred Shuttlesworth. Tuscaloosa, AL: University Jenkins, J. Craig, David Jacobs, and Jon Agnone. of Alabama Press. 2003. "Political Opportunities and African Marger, Martin N. 1984. "Social Movement 776 AMERICAN SOCIOLOGICALREVIEW to Organizations and Response Environmental Movements and Oppositional Consciousness." Pp. Change: The NAACP, 1960-1973." Social 20-37 in Oppositional Consciousness: The Problems 32:16-30. Subjective Roots of Social Protest, edited by Jane Matthews, Donald R. and James W Prothro. 1963. Mansbridge and Aldon Morris. Chicago, IL, and "Political Factors andNegro Voter Registration in London, England: University of Chicago Press. the South." American Political Science Review Myers, Daniel J. 1997. "Racial Rioting in the 1960s: 57:355-67. An Event-History Analysis of Local Conditions." -. 1966. and the New Southern Negroes American Sociological Review 62:94-112. Politics. New York: Harcourt, Brace and World. -. 2000. "The Diffusion of Collective -. 1975. Political Negro Participation Study, Violence: Infectiousness, Susceptibility, and Mass 1961-1962. [computer file] Interuniversity Media Networks." American Journal of Sociology Consortium for Political and Social Research. 106:173-208. 1982. Political Process and the McAdam, Doug. -. 2001. "Modeling Social Diffusion Black 1930-1970. Development of Insurgency, Processes Using Event-History Analysis," IL: of Press. Chicago, University Chicago Discussion paper, Institute of Social and Economic -. 1983. "Tactical Innovation and the Pace of Research and Policy, Columbia University, New Insurgency." American Sociological Review York. 48:735-54. Myers, Daniel J. and Beth Schaefer Caniglia. 2004 -. 1995. Current "Conceptual Origins, "All the Rioting That's Fit to Print: Selection Future Directions." 23-40 in Problems, Pp. Effects inNational Newspaper Coverage of Civil Comparative Perspectives on Social Movements, Disorders, 1968-1969." American Sociological edited D. J. D. and M. N. by McAdam, McCarthy, Review 59:519-43. Zald. New York: Press. Cambridge University NAACP. [I960]. Special Report on the Sitdowns: McAdam, and Dieter Rucht. 1993. "The Cross Doug NAACP Staff Activity in the Sitdowns, Branch National Diffusion of Movement Ideas." Annals of Department. NAACP Papers, Part 19, Section D, the American Political and Social Academy of Reel 19. Sciences 528:56-74. Oberschall, 1973. Social and Social and William H. Jr. 2001 Anthony. Conflict McAdam, Doug Sewell, Movements. NJ: Prentice-Hall. "It's About Time: in the of Englewood Cliffs, Temporality Study -. 1989. "The 1960s Sit-ins: Protest Diffusion Social Movements and Revolutions." 89-125 Pp. and Movement Take-Off." Research in Social in Silence and Voice in the Study of Contentious Movements, Conflict and Change 11:31?53. Politics, edited by R. R. Aminzade, J.A. Goldstone, Oliver, Pamela E. and Daniel Myers. 2003. "The D. E. J. W H. S. McAdam, Perry, Sewell, Jr., Coevolution of Social Movements." Mobilization Tarrow, and C. Tilly. Cambridge, England: 8:1-25. Cambridge University Press. Olzak, Susan, Maya Beasley, and Johan Olivier. McAdam, Doug, Tarrow Sidney, and Tilly Charles. 2003. "The Impact of State Reforms on Protest 2001. The Dynamics of Contention. Cambridge, Against Apartheid in South Africa." Mobilization England: Cambridge University Press. 8:27-50. Meier, August and Elliott Rudwick. 1973. CORE: A Oppenheimer, Martin. 1963. "The Genesis of the Study in the Civil Rights Movement. New York: Southern Negro Student Movement (Sit-in Oxford University Press. of -. Movement)." Ph.D. dissertation, Department 1976. "The Origins of Nonviolent Direct of Action in Afro-American Protest: A Note on Sociology, University Pennsylvania, PA. Historical Discontinuities." Pp. 307-404 inAlong Philadelphia, M. 1972. Black Students in Protest: theColor Line, edited byA. Meier and E. Rudwick. Orum, Anthony A the the Black Student Urbana, IL:University of Illinois Press. Study of Origins of Movement. DC: American Meyer, David and Debra Minkoff. 2004 Washington, Association. "Conceptualizing Political Opportunity." Social Sociological H. 1960. Lunch-Counter Forces 82:1457-92. Patrick, Clarence in North Carolina. Mitchell, Glenford E. 1962. "College Students Take Desegregation Winston-Salem, of Wake Forest Over." Pp. 73-95 in The Angry Black South, edit Department Sociology, College. William III. 1962. "The South Reacts." ed by G. E. Mitchell andW H. Peace. New York: Peace, H., Pp. Corinth. 96-127 in The Angry Black South, edited by G. E. Mitchell and W H. Peace. New York: Corinth. Morris, Aldon D. 1981. "Black Southern Student James. 1962. Freedom Ride. New York: Grove. Sit-in Movement: An Analysis of Internal Peck, Review Frances Fox and Richard Cloward. 1977. Poor Organization." American Sociological Piven, 36:744-67. People sMovements. New York: Pantheon. -. 1984. The Origins of the Civil Rights Polletta, Francesca. 1998. "'It Was Like a Fever...' Movement. New York: Free Press. Narrative and Identity in Social Protest." Social Morris, Aldon D. and Naomi Braine. 2001. "Social Problems 45:137-59. DYNAMICSOF PROTESTDIFFUSION 777 Pollitt, Daniel. 1960. "Dime Store Demonstrations." edited by D. della Porta and S. Tarrow. Boulder, Duke Law Journal 9:315-65. CO: Rowman and Littlefield. at Powledge, Fred. 1991. Free Last? Boston, MA: Tilly, Charles. 1977. "Getting ItTogether inBurgundy, Brown. Little, 1675-1975." Theory and Society 4:479-504. Glenda Alice. 1999. The Pain and the -. Rabby, 1995. Popular Contention in Great Britain, Promise: The Civil in Struggle for Rights 1758-1834. Cambridge, MA, and London, Tallahassee, Florida. Athens, GA: University of England: Harvard University Press. Georgia Press. Tolnay, Stewart E., Glenn Deane, and E. M. Beck. Raines, Howell. 1977'. My Soul Is Rested. New York: 1996. "Vicarious Violence: Spatial Effects on Penguin. Southern Lynchings, 1890-1919." American Rogers, Kim Lacy. 1993. Righteous Lives: Narratives Journal of Sociology 102:788-815. of the New Orleans Civil Rights Movement. New Tomz, Michael, Gary King, and Langche Zeng. 1999. York: New York University Press. Relogit: Rare Events Logistic Regression, Version Roscigno, Vincent J. and William F. Danaher. 2000. 1.1. MA: Harvard "Media and Mobilization." American Sociological Cambridge, University. Review 66:21^8. Traugott, Mark. 1993. "Barricades as Repertoire: Continuities andDiscontinuities in the of Salamon, Lester, and Stephen Van Evera. 1973. "Fear, History Apathy and Discrimination: A Test of Three French Contention." Social Science History 17:309-23. Explanations of Political Participation." American Political Science Review 67:1288-306. Tyson, Timothy. 2004. Blood Done Sign My Name. Searles, Ruth and J. Allen Williams. 1962. "Negro New York: Crown. Students' in Sit-ins." Social College Participation U.S. Bureau of the Census. 1963. Census of Forces 40:215-20. Population, 1960, Vol. 1: Characteristics of the Sindler, Allan P. 1965. Politics and Local Negro Population. Washington, DC: Government Printing Politics inDurham, NC. New York: McGraw-Hill. Office. Sitkoff, Harvard. 1993. The Struggle for Racial U.S. Commission on Civil Rights. 1959. Report of the Equality. New York: Hill andWang. United States Commission of Civil Rights, 1959. Smith, Charles U. 1961. "The Sit-ins and the New Washington, DC: Government Office. Student." Journal Relations Printing Negro oflntergroup Van den Bulte, and L. Lilien. 2001. 2:223-29. Christophe Gary "Medical Innovation Revisited: Social Contagion Smith, Jackie, John D. McCarthy, Clark McPhail, and versus Marketing Effort." American Journal of Boguslaw Augustyn. 2001. "From Protest to 106:1409-35. Agenda Building: Description Bias inMedia Sociology Van Dyke, Nella. 1998. "Hotbeds of Activism." Social Coverage of Protest Events inWashington, DC." Social Forces 79:1397^23. Problems 45:205-19. Milton. 1979. Fire in the in Soule, Sarah A. 1997. "The Student Divestment Viorst, Streets: America Movement in the United States and Tactical the 1960s. New York: Touchstone. Diffusion." Social Forces 75:855-83. Walker, Jack L. 1964. Sit-ins in Atlanta. New York: -. 1999. "The Diffusion of an Unsuccessful McGraw-Hill. Innovation: The Case of the Shantytown Protest Walzer, Michael. 1960. "A Cup of Coffee and a Seat." Tactic." The Annals of the American Academy of Dissent l:\l\-20. Political and Social Science 566:120-31. Ward, Brian. 2004. Radio and the Struggle for Civil Southern Council. 1960. in Regional "Special Report: Rights the South. Gainesville, FL: University of The Student Protest Movement, Winter 1960." Florida Press. SRC-13. Warren, Robert Penn. 1965. Who Speaks for the Standard Rate and Data Service. 1960. SRDS Negro? New York: Vintage. Newspaper Circulation Analysis, 1960 ed. Skokie, Wehr, Paul. 1960. "The Sit-Down Protests." M.A. the IL: Standard Rate and Data Service. sis, Department of Sociology, University of North Strang, David and Nancy Brandon Tuma. 1993. Carolina, Chapel Hill, NC. "Spatial andTemporal Heterogeneity inDiffusion" Wolff, Miles. 1970. Lunch at the Five and Ten. New American Journal of Sociology 99:614-39. York: Stein and Sumner, David E. 1989. "The Local Press and the Day. Linda T. 1991. "The of a New Nashville Student Movement, 1960." Ph.D. dis Wynn, Dawning Day: The Nashville Sit-ins, sertation, College of Communications, University February 13-May 10,1960." Tennessee Historical of Tennessee, Knoxville, TN. Quarterly 50:42-54 1998. Tarrow, Sidney. 1998. Power in Movement. 2d ed. Zhao, Dingxin. "Ecologies of Social Movements." American Cambridge, England: Cambridge University Press. Journal of Sociology Tarrow, Sidney and Doug McAdam. 2005. "Scale 103:1493-529. Shift in Transnational Contention" Pp. 121-150 Zinn, Howard. 1964. SNCC: The New Abolitionists. in Transnational Processes and Social Movements, Boston, MA: Beacon.