This review of the literature on aptitude by treatment interaction concentrates on studies which investigated two or more instructional treatments in a quantitative subject area--primarily mathematics--with different populations of students on whom measures of individual differences in aptitude were available. Designed to formulate some tentative recommendations for matching instructional methods to individual differences between students, the review examines 61 studies investigating logical vs scrambled sequences in programmed instruction, linear vs branching programmed instruction, covert vs overt responding in programmed instruction, conventional instruction vs programmed instruction, deductive vs inductive sequencing, mastery vs convential instruction, verbal vs visual/spatial instruction, and conventional instruction vs new math. A brief summary of the results with recommendations is followed by six tables summarizing the major findings of the reports in each of the areas studied. (RAO)