The major purpose of this paper is to find the financial and nonfinancial variables which would be useful in identifying companies with Sarbanes-Oxley Act internal control material weaknesses (MW). Recently, three research papers have been published on this topic based on different variables. Using the decision tree model of data mining, this paper examines the predictive power of the variables used in each paper in identifying MW companies and compares the predictive rates of the three sets of the variables used in these studies. In addition, an attempt is made to find the optimal set of the variables which gives the highest predictive rate. Our results have shown that each set of the variables is complementary to each other and strengthens the prediction accuracy. The findings from this study can provide valuable insights to external auditors in designing a cost-effective and high-quality audit decision support system to comply with the Sarbanes-Oxley Act.