Acta Psychologica 80 (1992) 49-65 North-Holland 49 Multi-attribute decision support and complexity: An evaluation and process analysis of aided versus unaided decision making * DaniElle Timmermans a and Ch. Vlek h a zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA fJnit,ersity of Leiden, Leiden, The Netherlands h Uniclersity of Groningen, Groningen, The Netherlands The present study addresses the effectiveness of a computerized decision aid (DECAID, Pitz 1987) based on the Multi-Attribute Utility Model (MAW. The effectiveness of DECAID was investigated for personnel selection problems varying in complexity (number of alternatives and number of attributes). Results show that the decision aid was most successful for problems of medium complexity. DECAID significantly affected the thoroughness of the decision, preference formation, and the perceived difficulty of the decision problem for medium complex problems. This was not the case for highly complex problems. Next, a process analysis is presented of unaided decision making focusing on the selection and evaluation of information. This process analysis of verbal protocols was based on the same decision problems as the first study. Results indicate that problem complexity had a significant impact on the decision process. Increasing complexity resulted in a less thorough evaluation, and selective information processing. Subjects tended to focus on one or two most promising alternatives. Furthermore, evaluation strategies diverged from those used in DECAID in terms of standard of comparison. The paper concludes with implications for the design of decision aids. Introduction Many, often computerized, procedures and tools are available to support decision making. Unfortunately, there is very little systematic empirical research about the effectiveness df decision aids and about * The research reported here was carried out when the first author was at the Department of Social Psychology, University of Groningen. The authors wish to thank John Michon for his advice, and Andre Hoeksma for his assistance in carrying out the experiment. Correspondence to: D. Timmermans, Medical Decision Analysis Unit, Faculty of Medicine, University of Leiden, P.O. Box 9600, Bldg. 43, 2300 RC Leiden, The Netherlands. 0001.6918/92/$05.00 0 1992 - Elsevier Science Publishers B.V. All rights reserved