On the Misleading Reporting of Regression Coefficients

Abstract

Accounting journals regularly include articles where regression coefficients are estimated and reported. However, the level of accuracy with which the estimated coefficients are reported is often not consistent within a single paper. Furthermore, in almost all cases, the level of accuracy claimed for the estimated coefficients, by re porting to so many significant figures, is far in excess of the level of accuracy that is plausible given the reported t-statistics and hence the implied standard errors of the estimated coefficients.