Modern risk-oriented audit is based on risk assessment. Risk of material misstatement is at the core of the modern risk-oriented audit, thus the allocation of audit resources is determined by the assessed risk level of material misstatement. Whether the risk of material misstatement will be evaluated accurately depends on the methods of evaluation. How best can investors, auditors and regulators detect misstatements? Addressing this question is of critical importance to the efficient functioning of capital markets. Previous research focused mainly on the qualitative assessment through theoretical analysis. There is little research on quantitative evaluation by using the method of empirical research.Based on sorting out the current studies and theoretical analysis, this paper selects the material misstatement and no-material misstatement A shares listed companies in Shanghai and Shenzhen Stock market from 2009to2011as the training sample. Using conditional logistic regression to empirically analyze the correlation between characteristic indicators and material misstatement. Then, on the basis of above result, this paper uses Logistic Regression to establish the assessment model. Also, this paper divides the RMM into three levels under90percent guaranteed and95percent guaranteed-high risk, low risk and moderate risk, which represents different level of RMM. This paper uses the data of2012to test the prediction accuracy of predictive model. At last, this paper analyzes the result and gives some suggestions for the expected model users, simultaneously points out the inadequate points and the prospecting points.To sum up, several conclusions are achieved as followings:(1) the differences of financial indicators between companies those have material misstatement and other companies those do not are significant. The empirical result shows that Debt Asset Ratio, Quick Ratio, Ratio of Operating Cash Flow to Core Operating Revenue and Ratio of Period Expenses to Sales have significant positive correlation with risk of material misstatement; Retained Earnings Per Share, ROA, Cash Flow Ratio and Total Assets Growth Rate have significant negative correlation with risk of material misstatement.(2) The assessment model this paper established has good constructive validity。In the process of sample training, the prediction accuracy rate of overall is87.76%, the accuracy is 89.07%in the recognition of no-material misstatement, the accuracy is82.52%in recognition of those companies which have material misstatement. In the process of predictive training, the prediction accuracy rate of overall is86.67%, the accuracy is86.76%in the recognition of no-material misstatement companies, the accuracy is86.27%in recognition of those companies which have material misstatement. As for predictive sample, the result shows that three levels of RMM have good predictive capability, which can be used for the assessment of risk of material misstatement. |