Stages in problem solving

Abstract

The history of problem solving has, with the exception of the Gestalt school (Kohler, 1925; Wertheimer, 1945; Duncker, 1945), been largely created by those psychologists whose interest lay in the manipulation of one or several independent variables for the purpose of observing how this affected one or more dependent variables. Subjects were studied in strictly controlled laboratory situations, as they solved what were often restricted and unchallenging problems. (Two noteworthy exceptions to this are provided by the work of Moore and Anderson, 1954, and John, 1957. Comprehensive reviews of recent problem-solving research and theory is presented in D.M. Johnson, 1955; Gagne, 1959; Duncan, 1959; G. Davis, 1966 and Kleinmuntz, 1966.) The object of this type of research was to express the performance of the problem solver in terms of number of solutions, number of errors, time to solution and so on. This orientation derives, in part, from a view of problem solving as closely related to learning, so that complex behaviour is perceived as not qualitatively different from simple discrimination learning, both being built up from simple S-R units. This is exemplified by articles by Skinner and by Staats in Kleinmuntz (1966, pp. 225-257; pp. 259-339).