In educational and psychological testing people are often evaluated by their responses to items on a test. The answers are typically graded as right or wrong, or as indicating an underlying quantity (suitability for jobs, knowledge of science, intelligence, etc. ) to some degree. An alternative is to regard different people within the same population as displaying different patterns of thought. Patterns can either be assumed to exist or statistical techniques can be used to discover them. In either case the problem is one of cluster analysis. We will present data from educational and industrial psychology settings demonstrating one clustering technique, based on information theory. The technique closely resembles techniques used in data mining. The implications for multivariate analysis of social science and educational data will be discussed
Patterns of Thought
Tara Madyhastha
Room
209