AAA Spark Meeting of the Regions

The Misuse of Regression-Based x-Scores as Dependent Variables

• In Event: Statistical Methods

Abstract

Researchers often use regression-based x-Scores (e.g., conservatism C-Score, misstatement F-Score) from a stage 1 model as a dependent variable in stage 2. We argue that this x-Score analysis can cause major biases and interpretation problems because (1) x-Score cannot capture new sources of variation, and (2) the estimates often hinge on unacknowledged technical assumptions. Instead, researchers should include the test variables and the relevant controls in stage 1, obviating the need for an x-Score. In replication analyses, major published findings change after we remove the bias caused by the (mis)use of x-Score as a dependent variable.

Authors

• Dmitri Byzalov, Temple University

• Sudipta Basu, Temple University