Can computerized decision aids do “damage”? A case for tailoring feedback and task complexity based on task experience

The theory of technology dominance [Arnold, V., and Sutton, S.G., 1998. The Theory of Technology Dominance. Advances in Accounting Behavioral Research. 1: 175–194.] predicts that mismatches between the skills of decision-aid users and the complexity of the tasks the decision aid is designed to assess will result in a lack of reliance on the decision aid and, ultimately, a potential for deskilling. The cognitive science literature on cognitive load suggests that varying the level of feedback provided to decision-aid users might moderate the risks that result from under-reliance. The current study presented participants, who were divided between two levels of skills, with decision aids in which the level of feedback was varied. These aids were used to solve cases of differing complexity. The results show that individuals with a high level of skill in understanding internal controls deskilled when they were assigned to non-complex cases. Further, this study found that with complex cases, a lower level of feedback was helpful and with non-complex cases, a higher level of feedback was helpful.