On Using the <i>t</i>-Ratio as a Diagnostic

Abstract

The t-ratio has not one but two uses in econometrics, which should be carefully distinguished. It is used as a test and also as a diagnostic. I emphasize that the commonly-used estimators are in fact pretest estimators, and argue in favor of an improved (continuous) version of pretesting, called model averaging.

Keywords

• <i>t</i>-statistic
• pretest estimator
• model averaging