Practical and Statistical Significance in Two-Way ANOVAs

In Excerpt 14.36 [not shown here], notice how three F-ratios turned out to be significant, yet two of these were labeled by the authors as being of only minor practical significance. Clearly, these researchers (as well as those who conducted the studies appearing in Excerpts 14.34 and 14.35) were aware of the fact that a single inferential result can turn out to be, at the same time, both significant from a statistical perspective and nonsignificant from a practical perspective. They deserve credit for conducting their data analyses with this important distinction in mind, and even further credit for incorporating this distinction into the written summaries of their investigations. Unfortunately, most researchers formally address only the concept of statistical significance, with the notion of practical significance automatically (and incorrectly) superimposed on each and every result that turns out to be statistically significant. In your reading of research reports, remain vigilant for instances of this unjustified interpretation of results.

(From Chapter 14, pp. 421-422)

Copyright © 2003 Schuyler W. Huck
All rights reserved.