Aiding DSS users in the use of complex OR models

Abstract

The under-utilization of decision support systems (DSS) based on complex OR models is a reality faced by many OR/IS system designers. The research finding that DSS users trade off decision making effort versus decision accuracy has been utilized as a ready-made explanation for this phenomenon. That is, it has been argued that complex DSSs may never get to be used at all due to the increased effort required in using them. We argue, in contrast, that DSSs have been found to elicit a zero-sum effort versus accuracy tradeoff precisely because they are architecturally designed to do so. A modification of DSS architecture designed to deal with this problem is proposed, and a research prototype based upon this architecture is described. Two preliminary laboratory experiments designed to examine the utility of this prototype have indicated that DSS design can significantly impact the level of the effort versus accuracy tradeoff. We conclude, therefore, that the "complexity ceiling" which has heretofore limited the amount of complexity an OR/IS designer can usefully incorporate into DSSs is not inviolate.