Interactions, Configurations, and Nonadditive Models

a consideration of interaction as a construct, to a delineation of its relations to other statistical and psychological constructs, and to a discussion of certain theoretical implications of the phenomena to which the term refers. Later speakers on the program will be discussing interactions within an analysis-of-variance context-pointing to various types of interactions which may occur, to how they affect the interpretation of experimental data, to transformations which, in certain cases, will remove them, and to other matters primarily statistical in nature. To provide a framework for their remarks, I should like to delve a bit into what we might call the &dquo;psychological significance&dquo; of interactions, and into what the finding of interaction effects implies, in general, for psychological theory and research. To clear the ground for this undertaking, let us first define what