If a relationship has practical significance, does it guarantee that statistical significance will be achieved in every study that examines it?

1 Answer
Jun 26, 2018

No. And the converse is also true (and more common).

Explanation:

This is a variation on the lack of connection between causality and correlation. The missing value is the risk perception. A relationship with a practical significance (risk/benefit) as perceived by the individual may not demonstrate "statistical significance" based on the chosen risk factors or confidence intervals.

More often, people make the mistake of assuming that "statistical significance" (often a very minor mathematical difference) implies both causality (remember to look at that confidence interval again) and practical significance. Statistics can (and should) guide decisions, but they are not the only valid basis for decisions.