Gregory Pettis, The Voter, Voter Turnout and the Context of Local Racial Composition

The story of politics in the American South is a story of a certain context, a certain culture, operating upon individuals to produce certain outcomes. Part of this story must involve voter turnout, with turnout in the South having been quite low historically, especially among blacks. But the voter turnout literature in political science cannot uncover some of the nuances in this voter turnout story very well because largely missing from this literature is how individual factors that motivate turnout are related to the aggregate-level phenomena. In other words, the notion that members of the electorate are functioning within a certain context, and that this context affects individual voters in different ways, has not been adequately explored. While there are individual-level models of voter turnout that stress an individual’s attributes that cause them to vote, such as socioeconomic factors (Verba and Nie 1972) and psychological factors (Conover 1984, Miller et. al. 1981), and there are aggregate-level models that stress attributes of state electorates which cause states to have higher or lower turnout, such as competitive elections (Hill and Leighley 1993) or registration barriers (Wolfinger and Rosenstone 1980), there are few models that explain how context affects the individual level turnout factors. What students of turnout are left with is a description of turnout in which individuals turnout to vote in a vacuum, completely oblivious to their surroundings, and aggregate electorates vote at different rates solely because of the different contexts in which they operate. The composition of the voter’s environment needs to be taken into account.

One of these important, overlooked environmental attributes, and one especially important for the study of the American South, is the racial composition of one’s area. This dissertation will be an exploration of how this context interacts with the individual level factors that motivate voters to the polls. In this dissertation I will test the thesis that racial minorities who live in Congressional districts with a sizable plurality or majority of their own race have more salient racial identities, voting for Congress at higher rates than minorities who live in areas where they are isolated from other members of their own race (holding all other factors equal). Blacks and Latinos who have a more salient racial identity will see politics in more group-centric terms, causing them to participate in politics to a greater degree when there are more of their racial group living in the Congressional district. Whites will be affected differently. In areas in which whites are a bare majority or minority, conservatives will withdraw from politics. Liberals, on the other hand, will not be affected, continuing to participate at the same rates.

Identity salience therefore may affect voter turnout. Salience is the degree to which group membership is cognitively present in influencing one’s self-perceptions and behaviors. The salience of identity depends on how many outgroup members one encounters, as it is the presence of outgroup members that triggers the identity. Individuals who live in racially homogenous areas with members of their own race should have the least identity salience, and as racial heterogeneity increases, so should salience. However, for racial minorities living in an environment in which there are few or no members of an individual's racial group, the salience of the identity should be relatively high because being surrounded by members of other races should make the identity salient. However, this identity should become even more salient as there are more racial minorities present in one’s area. This should be true because lacking the presence of other racial minorities may dilute the salience of the identity by undermining a sense of group cohesiveness (and a shared perspective). As more minorities are present this sense of a shared perspective should increase, as should identity salience (Conover 1999). However, at a certain point the more racial minorities there are, the less salient the identity should be because the necessary triggering factor, the presence of members of an outgroup, would be less and less true.

Context should affect whites differently. When there are many racial minorities in an area Anglos may feel group threat. This group threat may not be economic as much as a challenge to their cultural worldview, as well as their view of their place (a superior one and in control) in society and politics (Tyler and Smith 1998).

Not all Anglos have the same worldview, however. Some identify as political liberals, endorsing greater notions of racial equality, motivated by values of humanitarianism and government action that conservatives do not endorse (Feldman and Zaller 1992). Such whites, while still ambivalent because they endorse individualism (and because they have absorbed some negative racial stereotypes; Fiske 1998), are receptive to messages that question the system justifying ideologies (including the idea that equality of opportunity is a reality). White liberals also have more accepting and positive attitudes toward racial minorities, as reflected by their failure to take conservative issue positions in domains of symbolic politics. This suggests that white liberals should react differently from white conservatives to the racial composition of their area. Because liberals are at minimum accepting of racial minorities, those who live in contexts where there are more minorities should not feel threat from them.

This project is relevant to the American South in two ways. First, because racial diversity is the topic of study, the South will be the empirical focus. Most areas that are majority black are located in the South, and as such the observed behavior involve many Southerners. Second, the theory and results are directly applicable to the South. Theories of group threat, and their empirical tests, can be applied to the South to explain racial conflict.

This project’s hypothesizes will be tested in two ways. First, precinct-level aggregate data will be used to test the hypothesis that greater levels of racial diversity predicts greater turnout (King et. al., Record on American Democracy, Harvard University). This will be done in two ways. First, an analysis will be performed to see if, holding other relevant factors equal, minority turnout is higher in precincts as the percentage of the minority population increases. Second, another analysis will be performed in which precinct-level turnout will be compared from when a precinct was part of a majority white Congressional district to when it was redistricted into a majority-minority district. The second way these hypothesizes will be tested is using Black National Election Study data (Jackson et. al. 1984, Tate 1996) and American National Election Studies data to test the individual-level determinants of this turnout. I expect to have this project, my dissertation, completed by the end of the 2002-2003 academic year.

Center for the Study of the American South
411 Hamilton Hall, CB #9127, UNC-CH
Chapel Hill, NC 27599-9127
call: (919) 962-5665 fax: (919) 962-4433
email: bcall@email.unc.edu