Direct self-reports outperform implicit bias tests in predicting race-related behavior
Can implicit bias tests predict how people actually behave? A new adversarial collaboration finds that while indirect tests capture a tiny fraction of hidden bias, traditional self-report surveys are vastly superior at predicting race-related decisions.
A recent study in the Journal of Personality and Social Psychology highlights the effectiveness of direct self-reports over implicit bias tests in predicting race-related behaviors. Psychologists commonly assess racial attitudes through direct and indirect measures, with direct measures being traditional surveys where individuals consciously express their feelings, beliefs, or preferences.
Indirect measures, like the Implicit Association Test, aim to capture automatic, less controllable associations. The predictive capability of these indirect tests has been a subject of ongoing debate. While some argue that implicit bias shapes discriminatory behaviors, others suggest that these tests may reflect general cultural knowledge rather than personal prejudice.
A 2026 study revealed that different indirect tests often disagreed, complicating their reliability in predicting discriminatory actions.
Conducted by psychologists Jordan R. Axt, Suzanne Hoogeveen, and Eric Luis Uhlmann, the study aimed to determine the predictive power of indirect measures compared to direct self-reports. The researchers recruited 2,114 White American adults to participate in a two-part online study. In the first session, participants completed tasks measuring discriminatory behavior, such as trust games, ultimatum games, and hiring scenarios.
In a subsequent session, they completed four indirect measures, including the Implicit Association Test and an evaluative priming task. Afterward, they completed five direct surveys assessing their explicit racial attitudes.
The findings indicated that White participants generally exhibited a pro-White and anti-Black bias in indirect measures. However, behavioral tasks revealed a slight pro-Black bias in hiring and monetary decisions. Only about seven percent of participants displayed extreme pro-White scores on indirect measures and showed discriminatory behavior.
The indirect measures correlated with discriminatory outcomes, explaining approximately 2.5 percent of unique variance in behavior beyond what self-report surveys captured. This suggests that indirect measures capture a small but distinct aspect of social decision-making, emphasizing the complex relationship between automatic biases and actual discriminatory actions.
Written by urgent.news from PsyPost's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
