BACKGROUND: For decades, science has relied on an arbitrary p-value of 0.05 to make a distinction between a statistically significant study result, or not. However, there has been more and more debate about abandoning statistical significance testing in its current form, leading to a recent call by three statisticians to abandon statistical significance (Amrhein V, Greenland S, McShane B,. Retire statistical significance. Nature 2019;567:305-7). The call has been signed by more than 800 researchers, many of which from our society, clearly showing the urgency scientists feel to get rid of the current situation.
OBJECTIVES: To identify and understand misconceptions of statistical significance testing. To understand the role of significance testing in decision making on drug approvals from the regulatory perspective and decision making on scientific papers from the perspective of medical journals. To provide alternative ways of interpreting study findings for decision making.
DESCRIPTION: To set the stage, Kenneth Rothman will speak (20 min) about misconceptions related to statistical significance. Using examples of published studies, the audience will get an understanding of why statistical significance testing is misleading. In addition, alternatives for significance testing will be discussed. The second speaker, Hans Hillege will discuss (20 min) the role of significance in regulatory decision making.Lastly, John Fletcher, research editor of BMJ, will speak (20 min) about significance testing from the perspective of a major medical journal. A panel discussion will be moderated by Olaf Klungel (30 min).
