Predicting Material Weaknesses In Internal Control Systems After The Sarbanes-Oxley Act Using Multiple Criteria Linear Programming And Other Data Mining Approaches

Main Article Content

Wikil Kwak
Susan Eldridge
Yong Shi
Gang Kou


Sarbanes-Oxley, data mining, material weakness, multiple criteria linear programming


Our study proposes a multiple criteria linear programming (MCLP) and other data mining methods to predict material weaknesses in a firm’s internal control system after the Sarbanes-Oxley Act (SOX) using 2003-2004 U.S. data.  The results of the MCLP and other data mining approaches in our prediction study show that the MCLP method performs better overall than the other data mining approaches using financial and other data from the Form 10-K report.  Consistent with prior research, firms that disclosed material weaknesses in their SOX Section 302 disclosures were more complex (based on the existence of foreign currency translations), more often used Big 4 auditors, and had lower operating cash flows-to-total assets ratios than the non-material weakness control firms.  Because of mixed results on several profitability measures and marginal predictive ability for the MCLP and other methods used, more research is needed to identify firm characteristics that help investors, auditors, and others predict material weaknesses.


Download data is not yet available.
Abstract 78 | PDF Downloads 79