Problem Identification and Hypothesis Generation in
Preliminary Analytical Review: An Investigation of Experience Effects

Analytical review has been characterized as a diagnostic, sequential and iterative process consisting of four stages -- mental representation, hypothesis generation, information search and hypothesis evaluation [Koonce, forthcoming]. In the mental representation stage (hereafter referred to as "problem identification"), the auditor determines that an unusual fluctuation exists in an account balance. The auditor then develops an hypothesis to explain the unexpected fluctuation. This hypothesis guides the information search for additional data to be used in the final hypothesis evaluation stage. If the hypothesis is determined not to explain the unexpected fluctuation, the auditor generates another hypothesis, and the process begins again. Thus, each unexpected account fluctuation is a problem that must be identified, and the cause of that fluctuation is the problem solution. Previous audit research has primarily focused on the hypothesis generation and hypothesis evaluation stages in isolation (e.g., Church and Schneider 1990; Heiman 1990; Koonce 1992; Libby 1985; Libby and Frederick 1990). However, since analytical review is a sequential process [Blocher and Cooper 1988; Koonce 1993], successful hypothesis generation first requires successful problem identification. This study builds upon these previous studies and contributes to our understanding of the overall analytical review process by investigating the less frequently examined problem identification stage.

Auditors' ability to identify material financial statement errors through analytical procedures is examined using a computerized process-tracing methodology. Since cognitive psychology and auditing research have both demonstrated experience-related differences in problem solving approaches and results [see, e.g., Chi, Glaser and Rees 1982; Bonner 1990; Biggs, Mock and Watkins 1988], four experience measures, one general and three task-specific, are employed to assess auditors' problem identification and hypothesis generation abilities. The results indicate that only one experience measure, formal training in analytical review, explained differences in problem identification ability, i.e. auditors with training in the use of analytical procedures were better at identifying errors than auditors with no such training. This suggests that formal training in the proper use of analytical procedures can improve auditors' ability to detect errors in financial statements. Auditors with more training were also more likely to identify errors in accounts that they had recently found to contain errors in an actual audit. Contrary to expectation, no experience effects were found for auditors' hypothesis generation ability. Such a finding suggests that problem identification skills and hypothesis generation skills may develop separately. The study also fails to replicate the accessibility effects in hypothesis generation reported by Libby [1985] and Libby and Frederick [1990].

This study explicitly examines the link between the identification of a material unexpected account fluctuation and the development of an explanation for that error. Additionally, the experimental materials provided to the auditors create a richer environment in which to study analytical review than many previous studies.

Return to Charles E. Davis' vita

Return to Charles E. Davis' Home Page

Send E-mail to Charles_Davis@baylor.edu.