Research Design

The interesting questions about Blacks and Jews in America are rooted in their historical relations in different locales. We will therefore collect events data for a 20th (1900 to 1997) century time series of three major U.S. cities, located in different regions of the country: New York City, Chicago, and Los Angeles. Why these three cities? New York City and Chicago were selected (representing a northern and midwestern city respectively) because of the long historic presence of both Blacks and Jews within these cities (quite frequently with close geographic proximity to one another), the high profile clashes that took place there, and the wide variety of political leaders as well as organizations that have acted in the different cities across the political spectrum and over time. Los Angeles was selected because it is often identified as an example of successful Black-Jewish relations (Sonenshein 1992). This longitudinal and comparative analysis will allow us to explore the combination of structuralist “nuts and bolts” in SPOT and process-oriented “nuts and bolts” in CARP.

Why so long a time period? From reading the political-historical literature, it is clear that each group has engaged in different types of behavior (more or less) consistently throughout the century. Sometimes, this behavior comes to the attention of the national media. Other times, however, it does not. The lack of reporting does not indicate a lack of behavior/action. Political-historical sources on each ethnic group and observation of Black and Jewish newspapers (at different points in time) identify that both were very much active. There has simply been no attempt made to index exactly what the different groups have done over time. Ours would thus be the first time series to collect the data and investigate the problem.

Events-data collection is now a common methodology that has been employed in the U.S. for Blacks (McAdam 1982), women (Costain 1992), and various ethnic groups (Olzak 1992; Bond and Bond 1995). The first and the third data sets identified above, while relevant to our investigation, are limited for they do not specifically concentrate on the full range of interactions and because they rely upon sources that might not be sensitive to the actions of ethnic minorities. Consequently, we will collect data from numerous government sources (especially those within the National Archives and Records administration, the Department of Justice, and historical records of the Federal Bureau of Investigation) as well as Black and Jewish archives from organizations within each community (for Blacks, the NAACP, the Urban League, important churches/mosques; for Jews, the American Jewish Committee and Anti-Defamation League of B’nai B’rith).

The cornerstone of our research effort will be the use of Black, Jewish and mainstream (White) newspapers which exist throughout the full time period (for list of sources see Appendix 2; all are available for the time period to be examined). As designed, we will use weekly papers from 1900 to 1997 for both communities – from nationalist/Zionist as well as moderate political orientations (N=10,864 for each group, per city). Ideological orientation has been identified by various political-historical references (e.g., for Blacks, A Reference Guide to African American Publishers and Editors, 1827-1946 [1993], and Black Information Index [1970-annual]; for Jews, Jewish Press [1949] and Index to Jewish Periodicals [1963]; for both Gale Directory of Publications and Broadcast Media (1880 – annual]). For mainstream newspapers, only the center position exists and thus the total number of observations, per city, will equal 5,432.

To contend with the issue of intercoder reliability as well as to better enable examination of the means of communication itself (i.e., the stories themselves), relevant articles within different newspapers (our unit of analysis) will be scanned into a computer (excluding advertisements), saved as a text file, and then examined with software entitled “Non-numerical Unstructured Data Indexing Searching and Theorizing” (or Nudist), created by Qualitative Solutions and Research Pty. Ltd. in Victoria, Australia (WWW: http:/www.sagepub.com). For good discussion on Nudist see Tesch 1990; Richards and Richards 1991; Weitzman and Miles 1995. This indexing system will allow effective and efficient cataloging of information, searching through all of the material scanned in, as well as replication of compiled information. For discussion of article selection see Appendix 3.

Specifically, we will code the following five sets of variables in our model.

Political Opportunity Structure. This is frequently discussed in terms of the closeness of the vote (gauging electoral cleavages), war participation, the general “openness” of the local government (i.e., whether the city government is run by the mayor or more biased towards the city council), some measure of state repression, as well as numerous economic characteristics (employment, income inequality, the value of the dollar). These indicators are directly relevant to hypotheses H.5.1 – system openness, and H.5.4 – economic competition.

While important, we are more influenced by the work of Piven and Cloward (1977) and others (Snow and Benford 1992; Della Porta 1996; Gamson and Meyer 1996; McAdam 1996), who suggests that the objective circumstances of political opportunity are less important then what particular groups believe. If a group perceives an opportunity for action, and most individuals rely upon local elites/opinion leaders for information as well as cues, then framing of opportunities is more important. Framing political opportunity structure will be obtained from newspaper sources.

Specifically, we will identify each community’s perceptions of government (at local, state, and federal levels) – as to whether they are “open” or “weak.” This will be determined by the identification of different phrases within the text: for example, economic crisis (significant unemployment, stock crashes, increasing cost of living), external conflict, moral dilemma, corruption, closeness of electoral contest/election, an assassination, or some other factor that contributes to political-economic instability. Attention will be paid to the ideological orientation of the newspaper, in order to gauge consistencies across the various sources (greater credence to the claim of political opportunity will be given to situations when all three papers within each respective community identifies the same situation and frames it in the same manner). Additionally, where available we will supplement these more “subjective” measures of political opportunity with government sources regarding actual electoral outcomes at different levels as well as various economic characteristics.

2) Mobilizing Structures concerns each ethnic group’s civic, religious and political organizations. Using the Black and Jewish newspapers, we will collect information on various organizational characteristics: names, approximate size, geographic location, funding sources (hypotheses H.3.1 and H.3.2), activities (e.g., health services, education, lobbying, fund raising, protest, legal, electoral, religious/cultural), ideological orientation as portrayed (hypothesis H.5.2 – the effect of nationalistic mobilization). This information will be supplemented with archival sources of different groups as well as the National Archives and Records Administration.

3) Cultural Frames concern group concepts of the major problems confronted by the community, suggested resolutions to these problems, depictions/representations of the other relevant ethnic group and their relationship to the specific problem identified - protagonist, antagonist, or audience (hypothesis H.5.3 – the influence of hostile/inflammatory rhetoric), and the degree to which diversity of opinion within respective communities exists (hypotheses H.4.1 and H.4.2 - intra-community competition). These will be collected from Black and Jewish newspapers. In addition to the numerous factors just identified, we will also note what page the article appeared on as well the size of the article itself as a way of indexing importance.

4) Instances of Conflict (protest behavior or contentious politics directed against the government as well as the other ethnic group) and Cooperation (instances of coordinated activity between the two ethnic groups for a wide variety of purposes) will also be identified. Here, we will note the specific date at which an event took place (start as well as end), the target of activity, the type of action taken, the specific location of the activity (by SMSA code[s] and street names), the groups involved (from the listing generated under mobilizing structures), the stated objective of the action, the degree of violence present, as well as the outcome (e.g., tolerance/inaction from political authorities, negative sanction, accommodation, and the amount of damage/deaths). Additionally, we will note the degree to which both communities are involved: e.g., the numbers of individuals present (small [1-9] to tens of thousands [10,000 or more], and a relative assessment of financial burden.

These indicators will allow us to identify “anti-government acts” (hypotheses H.1.1 and H.1.2 – the mobilization/ countermobilization issue). They will also allow us to identify the temporal sequence of different events employed by each group (hypotheses H.2.1 and H.2.2 – investigating the diffusion argument).

5) These indicators will also allow us to identify Government/Police Behavior directed towards the different ethnic groups. This includes reactive as well as proactive behavior directed towards each ethnic group (tolerance/inaction, negative sanctions [arrests, harassment, raids, investigation or the imposition of laws that restricts political behavior], accommodation [passing of specific legislation relating to particular groups stated grievances or meeting with members of the community to discuss problems/solutions]). These indicators will allow us to identify another aspect of political opportunity structure (hypothesis H.5.1 – relative system openness).

In line with our interests, we will investigate four types of relationships. First, we will explore the effects of political opportunity, mobilizing structures, cultural frames and government/police behavior on Black and Jewish collective action respectively (i.e., its type and its frequency). The type of behavior observed (i.e., different protest events) will be examined with the use of a Multinomial logit model (or polytomous logistic regression model). This enables us to investigate unordered categorical variables, such as different types of protest behavior – without having to force the variables to fit on some continuum of conflict or cooperation (for discussion, see Aldrich and Nelson 1984, 73-77; Greene 1993, chapter 21; Judge et al. 1985, 768-772). Event frequency will be examined with OLS regression or Poisson regression, depending upon the actual distribution of the different variables considered. Second, we will employ the same independent variables (e.g., political opportunity and so forth) to investigate how well the model accounts for Black-Jewish interactions (i.e., event type – inaction, conflict or cooperation; event frequency – the number of events identified; and, event scope – the relative degree of Black-Jewish involvement). In line with the explanation provided above, the first will be examined with a Multinomial logit model, while the second and third will be analyzed with an OLS/Poisson regression model. Third, we will explore the effects of lagged determinants across the different explanatory variables identified above. For this, an ARIMA intervention model will be employed (McCleary and Hay 1980). Finally, we will examine the between instances of cooperation and conflict with a hazard rate model. The examination employed here will allow us to address the issue of what prolongs periods of quiescent, non-conflictual interactions.