Throughout the history of the study of collective behavior, many different explanations have been proposed to explain outbreaks of collective violence. From the early works of LeBon through the modern works of Smelser (1963), Blumer (1971), and the school of new social movement theorists (Mueller, 1992), the first of the two major approaches depending largely on psychological explanation has remained a strong force in explaining collective behavior, even through the temporary dominance of Resource Mobilization Theory (McCarthy and Zald, 1977, Oberschall, 1973). Even resource mobilization theorists though, had to admit to the power psychology in understanding the meaning-making and emotional impetus for collective behavior (McAdam, McCarthy, and Zald, 1988). In examining the works that seem to base themselves firmly in structural factors and avoid individual psychological states, it is important to be reminded that collective behavior would not occur if the structural conditions did not have an influence on the cognitive and emotional functions of the individuals who eventually act collectively. For example, the work of Charles Tilly in explaining collective behavior in Europe relies heavily on economic indicators to make his case (Tilly, Tilly, & Tilly, 1975; Shorter & Tilly, 1974), but the economic conditions mean nothing unless translated into an impetus for action by human beings.

On the other hand, psychological approaches to collective behavior have been roundly criticized for portraying collective actors as crazed mobs who are almost inexplicably out of control (McPhail, 1991). While we now have a somewhat more sophisticated view of the individual in collective action, it worthwhile to remind ourselves when we investigate a volatile phenomenon like rioting, that processes may, on the surface, seem to be completely irrational and irresponsible. However, the behavior actually develops from legitimate reasoning, despite the fact that this reasoning is not apparent to a casual observer.

In this study, I will examine patterns of rioting in the United States from 1961-1968 using mainly structural indices of economic conditions, population flows, and the incidence of rioting in other geographic areas. However, it should be clear that these variables only allow for collective action in that they facilitate a realization that current conditions are unacceptable. Rather than individuals being controlled by these conditions, it is the individual who takes stock of the situation and then determines that rioting behavior is justified and purposeful behavior.

THEORETICAL ISSUES

Contagion of Rioting

Among the psychological explanations of rioting is the notion of the contagion of collective activity. Although explanations of collective action that rest on a crazed madness sweeping crowds have been thoroughly debunked in collective action literature (see McPhail, 1991), there remain indications that riots and other collective activity occur in clusters, both temporal and geographic. In Spilerman’s (1970) work on racial riots, he noted several instances in which a clear contagion effect seems to be present in patterns of rioting. The Kerner Commission report, for example, studied the riots of the summer of 1967 and grouped riots in geographic and temporal clusters which seemed to provide strong evidence for the contagion hypothesis (National Advisory Commission on Civil Disorders, 1968). Other accounts have also suggested the spread of collective behavior that is in some way a result of a contagion-through-imitation process (Olzak, 1987; Pitcher, Hamblin, and Miller, 1978).

However, the contagion effects do not monotonically increase ad infinitum. Rather, the contagion appears to be governed by a process that has accelerating effects on collective behavior initially, and then begins to taper. After the apex of rioting, each additional riot appears to slow the rate of rioting as if the process is becoming distasteful to the participants. This exhaustion effect has been hypothesized to be the result of the tiring and calming of the participants or of the increase in repressive mechanisms as social control agents attempt to get the situation under control and maintain calm (see Oberschall, 1978; Kelly and Isaac, 1984; Koopmans, 1993).

While the existence of the contagion/exhaustion process seems clear enough, modeling the actual rise and fall remains problematic. Scholars are unclear on how long a contagion process remains in effect, and at what point it levels off and the exhaustion process takes over. Olzak (1987) used thirty days as the critical period for the process to cycle but did not examine alternative periods.

Based on the contagion issue I prose several hypotheses. First, riots occurring in the United States cause an increased probability of rioting in other cities in the United States. Second, contagion effects will tend to peak and then decline within a fairly uniform period of time. Given that Olzak (1987) found significant contagion and decline effects in a 30 day period, I propose that longer periods than 30 days would show little more power in explaining the rioting process given that the process has already been shown to be declining. Therefore, I concentrate on shorter periods, 2 weeks and 1 week, along with 4 weeks to determine which has the most power in modeling the contagion/ exhaustion process. Third, I propose that the contagion effect depends largely on the geographic distance from recently occurring riots. Therefore, the probability of rioting for a given city increases more in response to riots that occur in closer geographic proximity than those that occur farther away. Finally, the contagion process predicts that cities which experience riots are more susceptible to future riots. Therefore, the final hypothesis concerning contagion considers the past history of rioting in each city as a predictor of future rioting.

Competition Models

One important process assumed to underlie conflicts of all types is competition for scarce resources. As two or more groups compete for common resources, they inevitably come into varying levels of conflict. As the demand for the good increases and the supply decreases, conflict between those clamoring for them will increase. Such an economic model of ethnic and racial conflict is evident in works as early as that of Park (1950). Park also felt that assimilation would eventually lead this conflict to subside in favor of a more peaceful accommodation. However, it is clear that this process has not occurred and ethnic conflict has not subsided (see Olzak, 1987; Olzak and Nagel, 1986; Nielsen, 1985). To account for this continued conflict, scholars have developed and tested many models of economic processes to attempt to account for collective conflict (Spilerman, 1971; Spilerman, 1970; Spilerman, 1976; Olzak, 1987; Tilly, Tilly & Tilly, 1975, Shorter & Tilly, 1974).

Some of these models incorporate indices of structural economic conditions and others depend on subjective assessment of one group’s economic condition compared to others’. These relative deprivation arguments have largely been empirically invalidated (see Gurney and Tierney, 1982 for a review), therefore I focus on objectively measured economic conditions. Spilerman’s (1970) analysis essentially found all the economic indicators he used to be insignificant in explaining the patterns of rioting. However, each of his measures was concerned with the position of Blacks either in absolute terms, or relative to whites. I propose that economic conditions that are not specific to Blacks can influence the process of collective violence. In fact, this view is more congruent with competition theory in that general economic contraction, which effects either Blacks, Whites, or both, causes increased competition for scarce goods. In turn, the competition leads to conflict and eventually violence. The hypothesis here then is that economic contraction causes an increase in collective violence and rioting through inter-group competition.

Secondly, following the immigration model used by Olzak (1987), the numbers of Blacks in the particular city should cause an increased likelihood of collective violence. Citing Bonacich, (1972), and Lieberson, (1980), she argues that because migration usually occurs from lower-wage areas to higher-wage areas, an influx of a new group into an established area breeds competition as the new group challenges the older group for coveted higher-paying jobs. Retaining her hypothesis for this study, flows of blacks into a city should lead to economic competition that further encourages collective conflict. Given the structure of competition arguments, it is clear that an interactive effect should exist between immigration and economic contraction. The increase in collective violence caused by in-migration of Blacks will be exacerbated by economic contraction.

In addition to riot rates being driven up by immigration, Spilerman (1971) also showed that the size of minority group members present causes an increase in rioting, likely because the assimilation effects proposed by Park (1950) did not occur, thus maintaining the inter-group competition. This hypothesis will also be retained.

City and Population Characteristics

In several different analyses of ethnic or collective violence, various other variables have been included in the analysis and met with mixed success. For example, Tolnay and Beck (1992), when analyzing lynching incidents, include population density as an exogenous variable because lynchings were often associate with rural, isolated areas. Here the opposite effect is hypothesized because collective action requires that individuals come together for the event (McPhail, 1991). Areas with higher population density will allow easier assembly and rioting as well easier transmission of information about a riot in progress.

Another city characteristic included in this analysis is residential segregation. Following from the above discussion, residential segregation represents an additional density of minority and non-minority populations. When groups are concentrated in homogenous living areas, the above mentioned processes that facilitate riot behavior are intensified. Therefore, increased residential segregation should contribute to the rates of riot behaviors.

One omission of some analyses is the overall population size of the city in question (Spilerman, 1970, 1971; Olzak, 1987). Approaching the question this way may be a mistake however. Throughout Spilerman’s work, he begins by treating all cities as if they have identical risks for riots and proceeds to show that this assumption is not tenable. However, if one is to legitimately examine the idea that aggregation/urbanization causes rioting, it is not possible to assume that cities, as the unit of analysis, are indeed identical. Doing so confounds the dependent variable with the independent. If aggregation/urbanization does cause an increase in rioting, the resulting incidence of rioting should be greater than the incidence contributed by simple summation of the rates related to the individuals who make up the collective. Suppose each individual contributes a certain finite amount to the rate of rioting. If the process is actually invariant across cities, there will still be rioting in larger cities simply because there are more people there. In other words, the amount of rioting in a city of 100,000 would exceed the amount of rioting in a city of 25,000, but only as a factor of the population ratio. The amount of rioting in four cities of 25,000 should equal the amount in one city of 100,000 if in fact the process is invariant with respect to aggregation. It is therefore necessary to control for the overall size of the population to determine the effects of aggregation that exist beyond the simple aggregation of individual contributions.

Both Spilerman (1973) and Mazur (1973) at point treat the South and the non-South regions as if they are subject to fundamentally different processes. It is necessary to examine this assumption by including variables identifying the cities as South or Non-South. I propose construction of two sets of dummy codes for South and Non-south. In the first set, only states in the deep south are considered southern. In the second set, the definition of southern was expanded to include states contiguous to deep south states. Finally, Spilerman (1970), includes several variables he characterized as "Indicators of Social Disorganization." These variables were intended to characterize the lack of coherence in the daily living experience of city dwellers. Spilerman found two of these items to have a unexpected negative relationship with the number of riots. I will retain these two for analysis, percent change in total population, and percent of housing dilapidated, 1950.

 

DATA AND ANALYTIC STRATEGY

Data Source

The data for this study are derived from two sources. Information about riots for cities in the United States comes from the study conducted by Spilerman (1970a). The cities included are all those in the United States with a population of at least 25,000 in the year 1960 (N=676). Each record includes the dates of all riots that occurred in the city from the beginning of 1961 through the end of 1968. Riots are defined as those disorders involving at least 30 people and resulted in some violence or destruction of property. Information about each riot includes the date, duration in days, and type of riot. Riots were classified into 5 types according to the race of the participants and whether the riot was "spontaneous" or arose in a protest situation (e.g. a civil rights demonstration). The resulting 5 types are, 1) Black aggression, spontaneous, 2) Black aggression, non-spontaneous, 3) White aggression, spontaneous, 4) White aggression, non-spontaneous, and 5) Other, not classified. The data were culled from a number of sources including the Congressional Quarterly Civil Disorder Chronology, the New York Times Index, and Brandeis University’s Riot Data Review (See Spilerman, 1970a for details).

The second data source is the Governmental Units Analysis Data, 1960, collected primarily by Alford and Aiken. This data set contains information about the same set of cities including the economic and political information used as independent variables in this study. The first study, Urban Racial Disorders, 1961-1968 is a complementary subset of the latter data set.

Dependent Variable (Need help here on proper approach)

Of the 676 cities included in this sample, 504 cities experienced no riots in the period under question. Furthermore, the majority of the remaining cities experienced only one riot. Therefore, one reasonable procedure would be to construct the dependent variable as a binary variable indicating whether a riot occurred or not. However, it may be the case that cities which experienced more than one riot are qualitatively different than those which experienced only one. Using the number of riots though, is some what questionable given how drastically the distribution tapers above 3 riots. Another possibility is instead of using the entire time range as the period available for the incidence of a riot, to divide the study into smaller segments, for example one year, and determine a variable which indicates either riots or no riots in those time frames. Finally, some analyses of riot data of this period treat rioting that occurred before the assassination of Martin Luther King as qualitatively distinct from rioting that occurred following the assassination. An informative approach would be to analyze riot data placed into pre-assassination and post-assassination categories.

 

References

Alford, Robert & Michael Aiken. Governmental Units Analysis, 1960. (Machine Readable Data File. University of Wisconsin, Data and Program Library Service.

Blumer, Herbert. 1971. Social Problems as Collective Behavior. Social Problems, 18:298-306.

Bonacich, E. 1972. Advance Capitalism and Black/White Relations, American Sociological Review. 41:31-51.

Gurney, Joan Neff & Kathleen Tierney. 1982. Relative Deprivation and Social Movements: A Critical Look at Twenty Years of Theory and Research. Sociological Quarterly, 23:33-47.

Kelly, W. & L. Isaac. 1984. The rise and Fall of Urban racial Violence in the U.S.: 1948-1979, Research in Social Movements Conflict and Change. 7:203-233.

Koopmans, Ruud. 1993. The Dynamics of Protest Waves: West Germany, 1965 to 1989, American Sociological Review. 58(5):637-658.

Lieberson, Stanley. 1980. A Piece of the Pie: Blacks and White Immigrants since 1880, University of California Press, Berkeley.

Mazur, Allan. 1973. The Causes of Black Riots. American Sociological Review, 38:490-493.

McAdam, Doug, John McCarthy & Mayer Zald. 1988. in Neil Smelser, Handbook of Sociology, Newbury Park, CA: Sage Publications.

McCarthy, John & Mayer Zald. 1977. Resource Mobilization and Social Movements: A Partial Theory. American Journal of Sociology, 82(6):121-1241.

McPhail, Clark. 1991. The Myth of the Madding Crowd, New York: DeGruyter.

Mueller, Carol McClurg. 1992. Building Social Movement Theory, in Aldon Morris & Carol McClurg Mueller, Frontiers in Social Movement Theory, New Haven: Yale University Press.

National Advisory Commission on Civil Disorders. 1968. A Report. Washington: Government Printing Office.

Nielsen, F. 1985. Ethnic Solidarity in Modern Societies. American Sociological Review. 45:76-94.

Oberschall, Anthony. 1973. Social Conflict and Social Movements. Englewood Cliffs, NJ. Prentice-Hall.

Oberschall, Anthony. 1978. The Decline of the 1960s Social Movements. Research In Social Movements, Conflict and Change. 1:257-289.

Olzak, Susan. 1987. Causes of Ethnic Protest and Conflict in Urban America, 1877-1889. Social Science Research

Olzak, Susan & J. Nagel 1986. Competitive Ethnic Relations, Orlando: Academic Press.

Park, R. 1950. Race and Culture, Glencoe, IL: Free Press.

Pitcher, B., R. Hamblin & J. Miller. 1978. Diffusion of Collective Violence, American Sociological Review, 43:23-35.

Shorter, E. and Charles Tilly. 1974. Strikes in France: 1830-1968. New York: Cambridge University Press.

Smelser, Neil. 1963. Theory of Collective Behavior. New York: Free Press.

Spilerman, Seymour. 1970. The Causes of Racial Disturbances: A Comparison of Alternative Explanations. American Sociological Review, 35:627-649

Spilerman, Seymour. 1970a. Governmental Units Analysis Data, 1960: Urban Racial Disorders, 1961-1968. (Machine Readable Data File) University of Wisconsin, Data and Program Library Service.

Spilerman, Seymour. 1971. The Causes of Racial Disturbances: Test of an Explanation. American Sociological Review, 36:427-442.

Spilerman, Seymour. 1973. Strategic Considerations in Analyzing the Distribution of Racial Disturbances. American Sociological Review, 38:493-499

Spilerman, Seymour. 1976. Structural Characteristics of Cities and the Severity of Racial Disorders. American Sociological Review, 41:771-793.

Tilly, Charles, Louise Tilly, & Richard Tilly. 1975. The Rebellious Century: 1830-1975. Cambridge: Harvard University Press.

Tolnay, Stewart & E. Beck. 1992. Racial Violence and Black Migration in the American South, 1910 to 1930. American Sociological Review, 57(1):103-116.