Variability of Accounting Restatement Measurement in Audit Quality Research

In this research note, we examine the application of three methods used to operationalize the financial restatement variable in empirical archival accounting literature. Research exploring financial reporting and audit quality often employs financial restatements as a proxy for low audit quality. A review of recent articles reveals three distinct variations in operationalizing the restatement variable: announcement date, first occurrence and all occurrences. Using a typical audit quality model and restatement data from 2003-2014 we find statistically distinct results from each of the three measurement choices evaluated. The results suggest an incorrect choice may subject research findings to reductions in explanatory power, risk of type 1 and type 2 errors, and altered direction (sign) of coefficients. Suggestions for applying the findings to select an appropriate measurement approach for a given research question are discussed