The Social and Character Development (SACD) research program was designed to evaluate the effectiveness of seven elementary-school-based programs developed to promote social and emotional competence, positive behavior, a positive school climate, and academic achievement, and to decrease negative behavior. Procedures undertaken by the SACD Consortium to optimize the outcome measures used in the multiprogram evaluation are described. Preliminary analyses of the reliability and validity of the original scales, largely selected from previous research, suggested that a smaller set of outcome measures with stronger psychometric properties could be derived. The factor structure of these measures was examined using exploratory and confirmatory factor analyses to distill the child outcome measures into a more parsimonious and practicable set of measures for these programs. Support was found for a 5-, 3-, and 10-factor solution for the teacher, primary caregiver, and child reports, respectively, which were stable across three assessment times, robust to different statistical assumptions, and invariant across gender, race/ethnicity, and program site. A multitrait, multimethod analysis confirmed construct convergence across reporters but also indicated significant reporter effects. In addition to the measures' utility in evaluating the effects of the SACD program, the process used and knowledge gained are discussed to offer guidance to others who design and conduct evaluations of school-based programs. These include the importance of using multiple reporters of data, assessing actual performance of a measure even if previously published, and including measurement of altruistic behaviors as a unique feature of children's behavior. (Contains 6 notes, 3 tables, and 1 figure.)