In the last 60 years, the proportion of white Americans expressing anti-black prejudice in face-to-face survey interviews has declined dramatically. To test whether social desirability pressures affect the expression of anti-black prejudice, we analyzed a within-subjects experiment in the 2008 American National Election Study in which white respondents first reported their endorsement of stereotypes of blacks confidentially via audio computer-assisted self-interviewing (ACASI) and weeks or months later orally during second interviews. Shifting to ACASI led to a small but significant increase in negative views of blacks. Unexpectedly, shifting to ACASI also led to a similarly large increase in negative views of whites. Furthermore, the ACASI reports had no more predictive validity than did the oral reports. This evidence suggests that social desirability pressures do not seriously compromise oral reports of racial stereotypes in face-to-face interviews.