In a study of academic attribution patterns, teachers were asked to suggest causes for the successes and failures of students whose backgrounds had been manipulated to provide various combinations of social class (middle, lower), race (White, Asian American), and gender. It was found that teachers cited causes internal to the student more often for success than failure, and teacher-related causes more frequently for failure than success. It was also found that the more similar the student's background to the teacher's, the more likely the teacher was to "count" the student's successes and "discount" the failures. Middle class White successes were more likely to be attributed to internal stable causes than were Asian and lower class successes. White female successes were more likely than failures to be attributed to stable effort causes, while White male and Asian student data did not reveal this difference. Also, teachers took more responsibility for Asian females' failures than successes and for Asian females' failures than Asian males' failures. Overall, student background results provided ambiguous evidence concerning whether prior expectations or outgroup prejudice, or both, influence attributions. Lower class students, a low expectation outgroup, showed the least favorable attribution pattern. Asians, a high expectation outgroup, showed more favorable attributions than lower class students but less favorable patterns, relative to middle class Whites, than prior expectations might have predicted. (CMG)