When ‘significant’ is not significant

Rachel Kennedy, John Scriven and Magda Nenycz-Thiel

Ehrenberg-Bass Institute for Marketing Science, University of South Australia

Introduction

In terms of statistics, a ‘significant’ result means that we are confident at some level of certainty determined by the analyst, e.g. 95%, that the finding really did (or did not) occur in the population being measured, and that it was not just a quirk of the particular sample we happen to have drawn. Such understanding of the word ‘significant’ is very different from the lay term of significance as something meaningful, sufficiently great or important and worthy of attention.

Learning new meanings for commonly used words is part of the challenge of learning statistics.