Introduction

In Part I of this volume, James Voss presents a discussion of problem representation as it fits into the problem-solving framework. Voss also discusses several problems with the problem-solving model (the need for a metric, lack of application to groups and organizations, defining constraints and assessing their importance, how information is retrieved and used, the difficulty of prediction), which, at their heart, point to the difficulty of building appropriate models without knowing the content of particular problem representations or the processes of reasoning used. To move forward, as Charles Kegley notes, we need to “map the cognitive terrain” and build “dynamic process models.” That is, we need to build carefully representations (models) of the problem representations of individuals whose policy choices we seek to explain and predict. Furthermore, these representations provide one set of inputs for dynamic process models, with other inputs coming from the policy makers’ environment. Representing problem representations provides a basis for solving the “problems of problem solving” and provides meaning to discussions of change by suggesting mechanisms for change. A useful strategy for accomplishing these goals is a combination of careful content analysis and computational modeling, which also offers a reasonable hope of prediction in addition to understanding.