The Reduction of Uncertainty and Troubleshooting Performance

Abstract

Two problem-solving experiments were conducted to determine the effects of prior knowledge of problem intermittency on human troubleshooting performance. In-termittency was defined as the rate at which a network component or output, when tested, gave an incorrect report of its status (i.e., reported that it was working when it actually was not). Participants' task was to find faulty components in a set of 18 computer-based problems that caused a network to malfunction. In the first experiment, half the participants were told whether a problem was intermittent or not before they began troubleshooting procedures. In the second experiment, half the participants were given specific rates of intermittency for the problems. It was found that providing participants with only general "intermitten"" or "nonintermtttent" prior information did not affect troubleshooiing performance. Participants given specific problem intermittency rates did, however, perform better on the troubleshooting problems than did participants who were not given this information.