Buz Hunt: THE DIAGNOSER project. cognitive psychology, psychometrics, computing and education.

Abstract

Facet based instruction, developed by Jim Minstrell, is a technique of instruction based on the idea that students enter a class with preconceived ideas on the topic or, if the material is totally new, very rapidly develop such ideas. The ideas usually have a grain of truth, but may not be in accord with current scientific concepts. The teacher proceeds by determining what these preconceived ideas are and tailoring instruction to the students ideas, moving from the original notions to current scientific beliefs. To Cognitive Psychologists, facet based instruction will appear closely related to schema theory, and indeed the jargon of the two approaches is almost interchangeable. Facet based instruction assumes that a teacher knows what the preconceived ideas are, how to identify them, and what to do with them when they are found. These assumptions are often not met, for a great deal of teacher training (including training for university professors) emphasizes how to present t! he desired beliefs clearly, rather than engaging in the dialog anticipated by facet based instruction. Hunt, Minstrell, and their several colleagues over the last 15 years have developed a series of computer programs, collectively referred to as DIAGNOSER, that assist teachers by providing student exercises based on the facet based instruction model, offer indicated instruction, report to teachers the ideas that students seem to have, and provide teachers with suggestions for in-class exercises based on reported student performance.

The current version of DIAGNOSER is a WWW based program. It has been operational for instruction in middle school and high school physics for the past two years. Thousands of students and teachers, world wide, have made some use of the program. Last year we conducted a major evaluation of the effectiveness of DIAGNOSER, by examining the performance of students who had used the DIAGNOSER program on the Washington State Assessment of Student Learning (WASL) examinations. Although we do not regard the WASL examination as the best sort of evaluation of this sort of program, it is an important de facto criterion for the success of teaching methods. Students using DIAGNOSER obtained substantially higher scores on the WASL examinations, showing the efficacy of the program by conventional standards. In addition, data gathered internal to the DIAGNOSER itself provides a new way of looking at the organization of students' concepts, a way more compatible with cognitive psychology than ! the conventional right-wrong scoring pervasive in education. The new way could be applied to many other areas of knowledge than those we have studied, both in and out of science.
.