A meta-analysis of 105 studies reporting 152 self-concept interventions in school settings was conducted. The aims of the study were twofold: to explore the construct validity approach to self-concept interventions, and to examine aspects of the administration of the interventions, namely treatment setting, administrator type, administrator training, and implementation standardisation procedures. In support of the multidimensional perspective of self-concept, results of the random effects model analyses suggest that targeting specific self-concept domains when measuring self-concept outcomes lead to higher effect sizes (p less than 0.001). Interestingly, the treatment setting, intervention administrator type, administrator training, and the use of standardisation procedures were not significant moderators. The implications for self-concept intervention administration in school settings are discussed. (Contains 1 footnote and 5 tables.)