AN EXAMINATION OF TASK STRUCTURE, TESTING STRATEGY, AND EFFICIENCY IN AUDITORS' HYPOTHESIS TESTING

Prior research on confirmation bias implied that possession of a confirmatory bias by auditors is an undesirable characteristic and should be corrected. However, it has been suggested in the psychology literature that this phenomenon termed "confirmation bias" can be better understood in terms of a "positive test strategy," and it has been demonstrated that in many cases the positive test strategy is a useful heuristic for determining the truth of a hypothesis.; The main purpose of this study is to examine whether the interaction of the task structure (the relation between the true state of affairs and the hypothesis) and hypothesis testing strategy, when viewed in light of a positive test strategy, affects auditors' efficiency. In addition, other factors (hypothesis source and different costs to perform tests) which may affect auditors' hypothesis testing strategy are also studied. Finally, this study examines the issue of whether consideration of auditors' intermediate hypotheses during testing, rather than only the initial hypothesis, affects results obtained.; Ninety-three practicing senior auditors employed by a Big Six public accounting firm participated in this experiment. Subjects were to select from a list of internal control tests those they wished to perform in order to determine which transaction cycle(s) was likely to contain errors causing an unexplained fluctuation in a hypothetical client's profit margin. Results of the internal control test were provided after each selection, allowing subjects to change their hypothesis.; Results show that the interaction of task structure and testing strategy had a significant effect on auditors' efficiency. No significant effect of hypothesis source or the cost factor on the testing strategy was found. However, significantly different results were obtained when classifying tests selected according to the subjects' initial hypothesis, rather than their most recent hypothesis, as much prior research has done.