Human and machine diagnosis of scientific problem-solving abilities



Diagnosis of the problem-solving state of a novice student in science, by an accomplished teacher, is studied in order to build a computer system that will simulate the process. Although such “expert” systems have been successfully developed in medicine (MYCIN, INTERNIST/CADUCEUS), very little has been accomplished in science education, even though there is a reasonably close parallel between expert medical diagnosis of patients with physiological problems and expert instructional diagnosis of students with learning problems. The system described in this paper, DIPS: Diagnosis for Instruction in Problem Solving, involves a new line of research for science educators interested in interdisciplinary efforts and ways in which computer technology might be used to better understand how to improve science learning. The basic architecture of the DIPS system is outlined and explained in terms of instruction and research implications, and the role of such “intelligent” computer systems in science education of the future is considered.