Research literature regarding the aptitude treatment interaction (ATI) model of research is reviewed, and arguments for and against the use of the model are described. The purpose of the model is to allow a combination of desired properties of experimental and correlational methods. It yields disordinal interactions when experimental situations are carefully planned, but may yield ordinal interactions when they are not. Situations in which the model might prove valuable are outlined, and recommendations are made for using the model in reading research. It is emphasized that (1) close attention must be paid to defining experimental manipulations, (2) familiarity with instructional and aptitudinal variables is necessary, and (3) careful analysis of theoretical models of learning before application of the model to a research situation should avoid negative results. Reading research studies in which the ATI model was used are described, and figures and references are included. (MS)