This study investigates the relationship between computer self-efficacy, mentoring, and the gender of mentors and students. The decline of women in IT-related programs of U.S. universities has led scholars to suggest that making more female faculty mentors available could raise female students' computer self-efficacy. This could address women's computational reticence and encourage more women to enter and remain in the programs. A survey of students from IT-related programs in five U.S. universities shows that extent of mentoring received, gender of students, and time students spent on computers as teenagers were significant predictors of computer self-efficacy. Extent of mentoring, gender of students, and gender of mentors have significant main effects on computer self-efficacy. Students with male mentors reported significantly higher computer self-efficacy while students with female mentors reported lower computer self-efficacy. Female students with no mentors reported the lowest level of computer self-efficacy. (Contains 6 tables and 1 footnote.)