Alternative ways of analyzing data from Aptitude Treatment Interactions were examined over a two-year period. In light of past arguments the author maintains that the questions surrounding aggregation have been badly posed and that the customary methods of analysis were either incorrect or subject to misinterpretation. Therefore, the majority of studies of educational effects--whether classroom experiments, or evaluations of programs or surveys--have collected and analyzed data in ways that conceal more than they reveal. The established methods have generated false conclusions in many studies. Further, the traditional research strategy which pits substantive hypotheses against a null hypothesis and requires statistical significance of effects can rarely be used in educational research. Samples large enough to detect strong, but probabilistic effects are likely to be prohibitively costly. (Author/MV)