The effectiveness and efficiency of statistical analytical review in an auditing strategy: A comparison of regression and the X-11 model

The use of analytical review procedures (ARPs) has become increasingly important in recent years. Auditors are using ARPs to achieve a more effective and efficient audit. Much research effort has been directed at determining the most effective and efficient ARP.;Most of the research to date has evaluated ARPs based on the frequency of Type I (falsely signalling a material error) and Type II (failing to signal a material error when one exists) errors. This tells only part of the story, however. Auditors are interested not only in the performance of ARPs but in the performance of an overall audit strategy. The overall performance of an audit strategy depends on the interaction between the ARP and subsequent audit tests.;When testing an audit strategy for balance sheet accounts, efficiency may be defined in terms of the sample size needed for tests of details. Effectiveness may be defined in terms of the risk that the combined audit procedures fail to detect material errors.;Only limited attention has been given to ARPs as part of an overall audit strategy and these studies have tested only regression analysis and income statement items. Research has indicated that the Census X-11 model (a time-series model) may outperform regression as an expectation model. The X-11 model, however, has not previously been tested as part of an overall audit strategy. The purpose of this study was to test the performance of the X-11 model in an audit strategy against the performance of regression analysis to determine which strategy is more effective and more efficient.;A multivariate analysis of variance revealed that the X-11 model was more effective and equally efficient as regression when used as an expectation model. That is, X-11 yielded significantly lower achieved detection risk but the DUS sample size was not significantly different from regression.