The Effect of Self-Selection Bias on the Testing of a Stock Price Reaction to Management's Earnings Forecasts.

This study examines the inferential bias due to the failure to control for self-selection when studying the market's reaction to management earnings forecasts. The analysis is conducted by controlling for self-selection and comparing the results to those obtained when self-selection is not controlled. This comparison suggests that the overall inference of a market reaction to the management forecast issuance does not change. However, the statistical significance declines when self-selection is considered. Since the issuance of a management forecast is an obvious self-selection, the results of this study suggest that self-selection should be considered and evaluated in quasi-experimental studies in accounting and finance. Copyright 1995 by Kluwer Academic Publishers