Race and Big Government in the Age of Obama
University of Minnesota
Home page: http://www.polisci.umn.edu/people/profile.php?UID=pgoren
Sample size: 4050
Field period: 8/13/2010-11/2/2010
This study tests whether the impact of racial stereotypes on social welfare attitudes grows stronger in the minds of white subjects when an African American politician endorses more government. I use a 12-condition experiment that varies the race of source cues to determine if black sources lead whites to evaluate aid to the poor, welfare, federal vs. private health care, and abortion in racialized terms.
H1: When a hypothetical congressman with an African American sounding name takes a pro-government stance, anti-black stereotypes will be activated in the minds of whites and shape their social welfare opinions to a much greater extent than when a black source cue is absent.
H2: The black source cue will not affect abortion opinion. That is, racialization should be confined to the social welfare domain.
My goal is to see if the presence of black source cues leads whites to think about social welfare policies in race-coded terms. I do so with four issue experiments, each of which has three conditions (12 cells total). For each issue experiment, there is a “no-names” condition, a “white names” condition (e.g. Brad Sullivan v. Greg O'Brien), and a “white and black names” condition (e.g. Brad Sullivan v. Darnell Jackson). The names are attributed to hypothetical congressmen endorsing positions on the four issue scales.
In each condition respondents are presented with a choice between a pro-government option at one end of a 7-pt scale and an anti-government option at the other end. The experiments ask about (1) federal aid to the poor; (2) federal aid to people on welfare; (3) federal v. private health care; and (4) abortion.
Regression analyses revealed that the name manipulations do not systematically condition the relationship between a measure of stereotypical beliefs about the work ethic of African Americans and issue positions in any of the cells. Said otherwise, the black name source cues do not appear to have enhanced race-coded thinking in the minds of white respondents.