Analytic racecraft: Race-based averages create illusory group differences in perceptions of racism

Research practices used by social scientists to understand and dismantle the psychologicalfoundations that uphold racist hierarchies can backfire when they rely on racecraft. Racecraftideology assumes the reality of race(s), an assumption that shapes study designs and inferences tothe detriment of theoretical and practical goals. I showcase how racecraft manifests in studiesseeking to quantify how perceptions of sociopolitical stimuli differ across racialized perceivers(e.g., black, white, latinx). The typical analysis for quantifying perceptions focuses on comparinggroup averages, which assumes the existence of discrete “races” whose perceptions can besufficiently summarized by averages. Across three studies, I used variance component analyses onracism ratings of anti-immigrant tweets from differently-racialized perceivers (N=1,211) to showthere was much larger disagreement than agreement within race categories, even when there wereaverage differences in perceptions across race categories. This analysis shows how analyticpractices can bolster different assumptions about the nature of race, some which reify the illusionthat race categories are stable cohesive groups. Researchers can improve their analytic inferencesand avoid producing race-reifying caricatures of peoples’ perceptions by adding variance mappingto their toolkits and attending to racialization as a dynamic process - needed improvements withinthe psychological study of race and racism, group-based beliefs, and anti-racist research endeavors.