How valid is the "F ratio"?

Question

I am using stepwise standard least squares in a paper that I hope to publish as a peer-reviewed article. I am concerned that the "F ratio" listed in the parameter table will not be adequate for publication, has there been any precedent for or against its use?

unknown · Answer

I am not aware of criticisms of the F-ratio in this situation. However, the broad topic of stepwise variable selection in regression has been subject to large amounts of criticism, and very few statisticians that I know of actually approve of stepwise procedures.