Judging Auditor Negligence: De-Biasing Interventions, Outcome Bias, and Reverse Outcome Bias

Individuals judge audit quality, in part, based on adverse outcome information. Assuming that individuals over-rely on outcomes, prior accounting research attempts to improve their judgments by reducing their reliance on outcome information. Logically, however, individuals could either over-rely on outcomes ("outcome bias") or under-rely on outcomes ("reverse outcome bias"). Peecher and Piercey (2008) provide theory and empirical findings that individuals harshly exhibit outcome bias when the Bayesian probability of negligence is below 40% (e.g., a range that would include frivolous lawsuits), but that individuals also leniently exhibit reverse outcome bias when the Bayesian probability of negligence is above 40% (e.g., above key legal thresholds such as "preponderance of the evidence"). Using Support Theory, we predict and find that, by reducing reliance on outcomes, most interventions from prior literature reduce outcome bias for the lower range of Bayesian probabilities but exacerbate reverse outcome bias for the higher range of Bayesian probabilities. Using Cumulative Prospect Theory, we also design a new intervention that, if implemented early during the evaluators' judgment process, successfully reduces both forms of bias. By doing so, we contribute to the accounting literature on de-biasing auditor negligence judgments and to the accounting literature on outcome effects.