Mediation analysis is a statistical technique that can be used to identify mechanisms by which intervention programs achieve their effects. This paper presents the results of a mediation analysis of Ngao, an acquired immunodeficiency syndrome (AIDS) education program that was implemented with school children in Grades 6 and 7 in Tanzania in the mid-1990s and evaluated using a controlled, group-randomized trial. The study examined which variables mediated the effect Ngao had in regard to (i) fostering positive attitudes towards people living with AIDS and (ii) decreasing intentions to be sexually active in the near future. Data from students who participated in a baseline and 12-month follow-up survey (n=814) were analyzed. Results indicate that increasing exposure to AIDS information and increasing knowledge about human immunodeficiency virus transmission/prevention were significant mediators of the intervention's effect on alleviating the stigma associated with people living with AIDS. Moreover, encouraging more restrictive social norms about sexual intercourse was a significant mediator of the intervention's effect on decreasing students’ intentions to be sexually active in the near future. Implications for future AIDS education programs for school children in this part of Africa designed to achieve similar goals are discussed.