Design of an inquiry support system for managerial problem diagnosis

Problem diagnosis, the process of discovering and constructing the causal structure of a problem, is one of the most critical aspects of decision making. Management information system (MIS) research, as a discipline concerned with improving managerial decision making, must provide adequate support for this critical and difficult decision process.

The purpose of this research was to formulate a conceptual framework for developing problem diagnosis support systems (PDSS), design PDSS based on the framework, and validate the design based on prototype system implementation. The conceptual framework identified essential elements of PDSS, delineated logical steps in PDSS development, specified the individuals' roles, and discussed major issues to be addressed for PDSS development. The framework indicated that PDSS must combine user-driven and system-driven approaches to support semi-structured or unstructured problem diagnosis process. Thus, PDSS must have both cognitive and normative elements. Furthermore, PDSS must reduce, or at least attempt to reduce, the errors, biases, and uncertainties in conceptualizing a problem. The framework indicated that PDSS must structure complex causal relationships, support the entire problem diagnosis process, and maintain extensive interactions with the users.

This research identified various activities inherent in, and essential for, successful problem diagnosis. Also identified are a variety of views necessary to support the problem diagnosis activities. This research integrated structural, statistical, and rule-based modeling for problem diagnosis support. The support functions are designed from both cognitive and normative perspective. This research incorporated the inquiry guaranteeing concepts into PDSS design.

The framework and the conceptual design of PDSS provided a sound basis for developing more powerful, comprehensive, and cooperative PDSS. The prototype system, tested with two different diagnostic problems, demonstrated that it can overcome the limitations of previous PDSS by integrating structural, statistical, and rule-based modeling approaches, supporting the entire problem diagnosis process, and achieving a more cooperative inquiry. The key concepts synthesized in this research include the cooperative system concept, multiple causal modeling approaches, the process-oriented decision support approach, and the inquiry "guaranteeing" concepts.