Perceived variability is a relatively unexplored prejudice reduction approach. We developed a new intervention and tested it in two pre-registered experiments—in field and online laboratory settings. Across both experiments, we found that the intervention did not reduce prejudice. As expected, higher perceived variability correlated with lower prejudice. However, the correlations were weak and markedly lower in the higher-powered online laboratory trial. Overall, these results indicate that perceived variability may have limited scaling potential due to low adaptability and effectiveness. Exploratory analyses indicated that effects of the intervention did not substantially differ across age, gender, and education.