I know what I shouldn’t say, but what should I say? An approach to present results without statistical significance

Abstract

Read online

There has been an important move away from the term “statistical significance” in the scientific and statistical community. The desire to “retire 0.05” is rooted in improving scientific reporting by ensuring that researchers report more information than simply whether a p-value passes a specific threshold. At the same time, the desire to stop using statistical significance has led to uncertainty regarding how to present results. In this paper, a two dimensional classification scheme using the estimated association and p-value/confidence interval is proposed to provide guidance regarding how to report results from studies.

Keywords

• introductory statistics
• reporting of results
• statistical significance
• p-value
• confidence interval