Modeling Type I Going Concern Reporting Errors.

• In Event: Breakfast and Research & Education Table Forum
  In Roundtable Session: Table 3: Audit Quality 1

Abstract

This study highlights an important feature of modeling Type I going concern reporting errors. Prior studies model determinants of these errors in one of two ways. They either 1) model determinants of the going concern opinion conditional upon subsequent firm viability, or 2) model determinants of subsequent firm viability conditional upon the going concern opinion. I demonstrate that an explanatory variable may exhibit either a positive or negative relationship with Type I errors depending entirely upon this research design choice. I examine 14 explanatory variables from prior literature and document that several are at risk for demonstrating such divergent relationships based on unconditional and conditional correlations. I find that two explanatory variables demonstrate statistically significant divergent results in multivariate analyses based entirely upon the research design decision. Future studies should carefully consider the appropriateness of each method and report sensitivity to their research design choice.

Author

• Nathan R Berglund, Mississippi State University