were sorted and recommendations were forwarded to suggest possible solutions . The purpose of this study was to investigate the relationships among affirmative action programs, self­ efficacy and academic performance. The study was carried out using 243 undergraduate students of Bahir Dar University who have been attending their study in 2004 E.C The students were selected using stratified random sampling and proportional probability size sampling techniques. The reliability coefficient of the self-efficacy scale was 0.885. The affirmative action programs were used to assess student's involvement frequency through descriptive statistics. The academic performance of students measured as per student's cumulative grade point average of first semester in 2004£. C Correlation coefficient, linear regression and partial correlation statistics methods were employed to analyze the data. Then the result of correlation coefficient indicates that the entire three variables (affirmative action programs, self-efficacy and academic performance) were significantly correlated as group. But among affirmative action programs only language training have no significant correlation with self-efficacy and academic perfor.¥-iance. The. linear regression result showed that while academic performance was . predicted by tutorial and mentoring, self-efficacy was found to be predicted by tutorial only. Statistically, affirmative action programs contributed 13% of variability in self-efficacy and 28% of variability in academic performance. The result of partial correlation in combination with regression of the dependent variable, academic performance (as measured by CGPA) demonstrates the direct and indirect effects of - independeni-variables over the dependent variable through mediating variable of self-efficacy. The direct effect of the dependent variables was by far greater than the indirect effects. Statistically, the direct effect • !2f tutorial on academic performance was 0.345, while its indirect effect was 0.044. The direct effect of mentoring was 0.180, while its indirect effect was 0.021. Jn addition, semi-structured interview questions were conducted for the gender officer and counseling and guidance officer of the university to cross check .. the data gained from students. Finally, depend upon the results of the study, discussion and summaries were sorted and recommendations were forwarded to suggest possible solutions .