Do voters assess political candidates differently based on their race or ethnicity? A recent meta-analysis, published in Acta Politica, sought to answer this by pooling data from 43 experiments conducted over the past decade. The findings reveal that, generally, voters do not assess racial or ethnic minority candidates differently from their white counterparts. These results challenge the notion that racial bias significantly impacts the electoral prospects of minority candidates.
The study, spearheaded by Sanne van Oosten, a postdoctoral researcher at the University of Oxford specializing in voting behavior and diversity in politics, aimed to shed light on how voters perceive candidates of different racial and ethnic backgrounds. Previous studies had shown mixed results, with some suggesting that racial and ethnic minority candidates face discrimination, while others indicated that such candidates might benefit from positive stereotyping. The researchers sought to clarify these conflicting findings by pooling data from multiple experiments and conducting a comprehensive meta-analysis.
A meta-analysis is a statistical method that combines the results of multiple studies to derive a more comprehensive understanding of a particular research question. It is used to increase the overall sample size and statistical power, thereby providing more robust and generalizable findings than any single study alone. Meta-analyses are particularly useful for resolving inconsistencies across different studies and for identifying patterns and overall effects that might not be apparent in individual research projects.
To compile a robust dataset, the researchers conducted a systematic search of political science journals for candidate experiments published between 2012 and 2022. These experiments involved presenting respondents with profiles of fictional political candidates, with their race or ethnicity being randomly varied. Respondents were then asked about their voting preferences and their evaluations of the candidates.
The final selection included 43 studies, with data from over 305,000 observations. The researchers obtained original datasets from the majority of these studies and standardized the results for consistent analysis. For studies where datasets were unavailable, they used the published results. Most of the studies had been conducted in the United States.
The researchers found that, on average, voters did not significantly differentiate between racial or ethnic minority candidates and white candidates. The overall effect size of 0.235 percentage points in favor of minority candidates was not statistically significant, suggesting that race and ethnicity do not play a major role in voter assessments for most minority groups.
“I had expected major discrimination from voters. These outcomes were very surprising to me,” van Oosten told PsyPost.
When examining specific racial and ethnic groups separately, a more nuanced picture emerged. For instance, Asian candidates in the United States were evaluated slightly more positively than their white counterparts, with an effect size of 0.76 percentage points. Although this effect was statistically significant, it was substantively small, indicating only a marginal preference for Asian candidates.
The findings also highlighted the importance of shared racial or ethnic identity between voters and candidates. Voters tended to evaluate candidates of the same race or ethnicity 7.9 percentage points higher than candidates of different racial or ethnic backgrounds. This significant positive effect underscores the role of in-group favoritism in voter behavior, suggesting that voters feel a stronger connection to and preference for candidates who share their racial or ethnic identity.
“Don’t worry about nominating a woman or person of color, voters actually like that,” van Oosten said.
So why is there an underrepresentation of racial/ethnic minorities in politics? The researchers pointed to supply-side factors as a potential explanation for the underrepresentation of racial and ethnic minorities in politics. Despite the lack of significant voter bias against minority candidates, there might be fewer minority candidates available or selected for political office, which could contribute to their lower representation.
While the study provides comprehensive insights into how race and ethnicity influence voter evaluations, it is important to consider potential caveats. One significant limitation is the possibility of social desirability bias, where participants may underreport their biases to appear more socially acceptable, thereby skewing the results. Additionally, the study primarily measures first impressions of hypothetical candidates rather than actual voting behavior, which may not fully capture the complexities and dynamics of real-world elections.
Nevertheless, the findings challenge the notion that racial bias significantly disadvantages minority candidates in the eyes of voters. Instead, the slight preference for Asian candidates and the significant effect of shared identity suggest a more complex interplay of factors influencing voter behavior.
The study, “Race/Ethnicity in Candidate Experiments: a Meta-Analysis and the Case for Shared Identification,” was authored by Sanne van Oosten, Liza Mügge, and Daphne van der Pas.