This paper presents a general model for conceptualizing and testing the interactions of individual differences and subject-matter characteristics with instructional methods. The model postulates certain ways of classifying the variables of interest in such investigations and of conceptualizing the cause-and-effect relationships among those classes of variables. Implications for both research and theory-construction are discussed, and a resolution of the controversy over the optimal breadth of instructional theories is proposed. The paper also describes research methodology that includes aspects of research design and statistical analysis that facilitate the investigation of those interactions. A variation of analysis of covariance is discussed that allows a continuous variable (the covariate) to be analyzed as a factor, complete with its interactions with all the terms in the basic statistical model. (Author)