Workshop on Studying Practitioner expertise: Expert System Representations

Moderator:

Mark a. Mattaini, Assistant Professor, Columbia University School of Social Work, New York, NY.

Research, Practice and Expert Systems

Dr. Scheuerman considered ideas emerging from his exploration of the use of expert systems in child welfare decision-making. Expert systems development requires close cooperation between experienced practitioners and systems developers. With colleagues at the Chapin Hall Center for Children, Dr. Scheuerman developed systems for decisions made while investigating allegations of abuse or neglect -- whether an allegation should be "founded"; the assessment of risk to the child; and the decision to place a child in out-of-home care. Dr. Schuerman observed that decision-making in social welfare is characterized by uncertainty, nonlinearity and the confounding of data-gathering, decision and action. He emphasized that decision decisions do not appear to be b ased on prediction of results of various courses of action and noted that, as Dr. Schon suggests, practice is partly a matter of testing subhypotheses, usually in a nonlinear way. Child welfare practitioners appear to base decisions on the balancing of "i nertia" and "give'em a chance", rather than on prediction. He concluded that to he extent to which expert systems can capture expert decision process, they may be able to manage the problem of behavioral unpredictability.

Dr. Ferns and Ms. Riedel described expert systems as software programs that model and expert's judgement in a specific arena. As the expert practitioner and researcher who collaborated to develop lifenet, a computerized expert system for assessing imminent danger of suicide with runaway and street youths, Dr. Ferns and Ms. Riedel outlines parallels between the development of expert systems and the process of acquiring professional practice knowledge. They noted that skilled practitioners become exp erts through constant synthesis of research-based knowledge and practice wisdom. Expert systems development is a cyclical process in which the knowledge engineer and the practice expert, as a team, refine the system's knowledge base. Dr. ferns and Ms. Rie del emphasized that expert systems development recognizes that the practitioner's role as expert is integral to knowledge acquisition. Expert systems are a vehicle for integrating empirical knowledge with practice wisdom to create professional knowledge a nd facilitate its dissemination throughout the practice community. They assert that the practice-research partnership is a natural arrangement and a prerequisite in developing and evaluating expert systems.