Download data and study materials from OSF
Principal investigators:
Vesla M. Weaver
Yale University
Email: vesla.weaver@yale.edu
Homepage: http://veslaweaver.wordpress.com/
Jennifer Hochschild
Harvard University
Email: hochschild@gov.harvard.edu
Homepage: http://scholar.harvard.edu/jlhochschild
Sample size: 2138
Field period: 9/17/04-9/24/04
With two main puzzles as its springboard -- the underrepresenatation of minorities in elected office and the disadvantage associated with minorities of darker skin -- this study examines the role of race and skin color on candidate preferences. Subjects were randomly assigned to campaign literature of two opposing candidates, in which the race, skin tone, and issue stance of candidates was varied and then asked about their attitudes toward the contenders. The results demonstrate some surprising asymmetries in the operation of race and skin color and finds that there is a "one-drop rule" when blacks run against whites but a color hierarchy when they run against each other.
1. In-group/Out-group
When a candidate is not from the majority racial group, racial dissimilarity is more determinative of voting behavior than ideological, partisan, or other similarity. Therefore, black candidates should receive lower support from white voters, without regard to issue stances.
2. Skin Color
Because race functions differently for darker members of the out-group, the importance of race effects in candidate preference are dependent on candidate skin color. Racialized voting will be even more pronounced for the darker-black candidate. Therefore, light-skinned black candidates will receive more support than their dark-skinned counterparts.
3. Racial Stereotyping
Even when white voters do not disfavor the outgroup, they are more likely to stereotype them as being less qualified, more liberal, and better/worse on certain issues. It may not be simply the race alone that voters react negatively to but the stereotypes that come out of racial cues, for instance, that black candidates support race-based policies they are opposed to.
a. In relation to the skin color hypothesis above, the darker the candidate, the more he will activate negative stereotypes.
b. Counter-stereotypical candidates will moderate the effects; black candidates will be rewarded when they run on a conservative platform.
Using the Internet instrument of the Time-Sharing Experiments for the Social Sciences, the survey experiment was administered on a nationally representative sample of 2138 non-Hispanic white adults. First, subjects were shown the campaign literature of two opposing candidates and informed that the candidates are running for an open Senate seat in a neighboring state. The race and skin color of the candidates varies across experimental groups using a morphing procedure to control for physical differences.
The campaign ads position the candidates in opposing stances on 5 issues -- economic growth, health care, education, public safety, and the environment. The candidates' skin color, name, and platform (conservative vs. liberal) are randomized, creating a total of 16 possible versions of the experimental stimuli.
Respondents were randomly assigned to one of these four campaigns:
* White vs. White
* White vs. Light skinned Black
* White vs. Dark skinned Black
* Light skinned Black vs. Dark skinned Black
Immediately after subjects read the campaign literature, they were asked a battery of questions about their reactions, including: feeling thermometer rating, queries on which candidate will do a better job in office on various issues (crime, health care, economy, disadvantaged), items tapping into stereotypes (which candidate is more experienced, hardworking, intelligent, trustworthy), their interest in the campaign, concern about the outcome of the election, importance of certain candidate characteristics in making their decision of who to vote for, and perceived ideology of the candidates. Finally, subjects voted for their favored candidate on a simulated ballot.
Cross-cutting findings for blacks by skin color: when black candidates run against each other, the lighter-skinned candidate is undeniably favored (he wins by a 16 point margin, larger than the control and regardless of issue stance) but across treatments the light-skinned candidate is less popular.
Interaction between race, skin color, and issue stance: being conservative and light-skinned is an advantage for the black candidate when he runs against another black, and being conservative and dark-skinned is advantageous when the opponent is white.
Stereotypes, Issues, and Interest in the Election:
* Both the light-skinned and dark-skinned black candidates were viewed as considerably less experienced (but harder-working) than their white opponent.
* The light-skinned black was seen as less intelligent.
* Running against a black candidate hurt the white candidate on trustworthiness.
* Black candidates are perceived as better on helping the disadvantaged and not as adept at improving the economy.
* There was no significant difference across treatment groups in interest in the election or concern with the outcome.
By studying the manner in which race and skin color affects black office-seekers, this analysis has attempted to shed light on why blacks have been relatively unsuccessful at mobilizing the white electorate.
This study suggests the complex relationship between candidate race and skin color:
1) Race is not a "deal breaker" for black candidates: black candidates have some advantages they can exploit --being viewed as more hardworking and trustworthy and better suited to help disadvantaged populations.
2) A large portion of the effect of race on subjects' reactions to candidates is due to skin color, not to race alone.
3) However, the results are contradictory with regard to skin color, such that the light-skinned black candidate is unpopular next to his white opponent and relative to the dark-skinned black condition, but simultaneously favored when he runs directly against the dark-skinned black candidate.
We plan to replicate the study to examine: 1) whether skin color operates similarly for Latino, and Asian candidates; and 2) whether respondents of different races and ethnicities respond skin color variation in similar ways as whites.
The survey also included a racial predispositions question separately to test whether racial affect influences reactions to minority candidates.
Note about the data file
Due to a technical error, party affiliation and ideology were not asked of 801 people in the original survey. As a result, a supplemental survey requested the information from the missing respondents, with a 81% response rate (645 of the 801). Of the remaining 156 people who did not complete it, 87 were no longer on the KN panel as of the time the supplemental survey was fielded; so of those who were eligible for it, 90% completed the supplemental survey. Also, only 17 out of the 645 respondents reported that their party affiliation had changed since the time of the initial survey (3%). The data file contains both the original survey and the supplement, along with a codebook for each.
"Race, Skin Color, and Candidate Preference." Paper presented at the Annual Meeting of the Midwest Political Science Association. Chicago, IL. April 2005.
Weaver, Vesla. 2010. The Electoral Consequences of Skin Color: The "Hidden" Side of Race in Politics. Political Behavior (31 December 2010), pp. 1-34. doi:10.1007/s11109-010-9152-7