Anchoring effects associated with expert system recommendations : an experimental analysis in accounting /
This study investigates the effect of expert system recommendations on the judgment behavior of system users in an auditing task. Specifically, two experiments were conducted to determine whether the provision of such recommendations are associated with anchoring effects among users. In each exper...
| Main Author: | |
|---|---|
| Format: | Thesis Book |
| Language: | English |
| Published: |
[Place of publication not identified] :
[publisher not identified] ;
1996.
|
| Subjects: | |
| Online Access: | http://proxy.library.tamu.edu/login?url=http://proquest.umi.com/pqdweb?did=739325741&sid=1&Fmt=2&clientId=2945&RQT=309&VName=PQD |
| Summary: | This study investigates the effect of expert system recommendations on the judgment behavior of system users in an auditing task. Specifically, two experiments were conducted to determine whether the provision of such recommendations are associated with anchoring effects among users. In each experiment, undergraduate auditing students were assigned to treatment groups where a portion of the students used an expert decision aid which provided recommendations and a portion used a decision aid which did not provide a recommendation. The experimental task was to evaluate a hypothetical audit client and determine whether there was substantial doubt about the company's ability to continue as a going-concern. The results suggest that anchoring effects were evident, in each experiment, for only one of two opposing recommendations. Because anchoring occurred for the recommendation toward which each of the experimental cases was biased, anchoring effects may only occur for recommendations that are consistent with case information. Results of the study also show, however, that subjects exhibited a confirmatory search behavior in weighting the relevancy of data items in their decisions. The implications of this are that improper attention may be given to evidence which is incongruent with a system recommendation, regardless of whether the recommendation is consistent with case information. Finally, evidence in the study suggests that, contrary to expectations, individuals who received recommendations did not exhibit inflated confidence beliefs about either their judgments on the cases or in their ability to perform the task. However, those who received system solutions perceived less effort required for the task and provided fewer comments to support their decision. The implications of less effort and lower confidence among expert system users are discussed. |
|---|---|
| Item Description: | Vita. "Major Subject: Accounting". |
| Physical Description: | xi, 114 leaves : illustrations ; 28 cm. Issued also on microfiche from University Microfilms Inc. |
| Bibliography: | Includes bibliographical references: pages 98-102. |