Abstract Kinney and Salamon 1982 used simulated data to compare three investigation rules that auditors might use with regression analysis in analytical procedures. This paper presents a comparison of the Kinney and Salamon findings with results from using actual data. The actual data with its greater variability, regression assumption violations, and trends resulted in a greater percentage of months being investigated. The increase in percentage of months investigated produced more Type I errors and fewer Type II errors; however, the relative efficiency and effectiveness of the decision rules found by Kinney and Salamon were generally confirmed in this study.
Wilson et al. (1989) studied this question.