Twenty years have passed since Sears (1986) alerted social psychologists to the many possible dangers faced by relying on a database composed mostly of students, especially with respect to the generalizability of the theoretical conclusions we come to. With a focus this time on the prejudice literature, this article examines how much has changed in our approach to whom we study. Content analyses show that prejudice researchers who publish in social psychology's major journals continue to rely heavily on student samples. Next, data are presented showing that important differences may exist between student and nonstudent participants in terms of how prejudice-related variables are expressed and used. The article concludes by raising metatheoretical concerns about the continued use of student samples both in the conclusions we arrive at as a science and in the very topics we study in the prejudice literature, with various recommendations suggested for decreasing this trend in relying on such a narrow database.