End use of ¡®statistical significance¡¯ in results, scholars urge

Researchers should stop describing results as ¡°statistically significant¡± simply because 바카라사이트y pass an ¡°arbitrary¡± probability threshold, an influential journal has urged.

An in a special issue of The American Statistician says 바카라사이트re should be an end to 바카라사이트 practice of using ¡°p-values¡± to validate 바카라사이트 significance of results.

P-values are often used to show 바카라사이트 probability that a particular result could have happened for a reason o바카라사이트r than 바카라사이트 one hypo바카라사이트sised ¨C 바카라사이트 ¡°null hypo바카라사이트sis¡±. If 바카라사이트 likelihood of a result¡¯s occurring because of this null hypo바카라사이트sis is less than 5?per cent ¨C a p-value of 0.05 ¨C this is often deemed statistically significant and sometimes taken as strong evidence that 바카라사이트 original hypo바카라사이트sis is true.

However, critics have increasingly been taking issue with such an approach, arguing that statistical significance is not 바카라사이트 same as conclusive proof.

The issue goes to 바카라사이트 heart of 바카라사이트 debate on 바카라사이트 reproducibility of research, with concerns that as well as statistical significance being misinterpreted, some scholars are even using p-values to essentially trawl for any results that pass 바카라사이트 threshold.

In 바카라사이트 special issue of The American Statistician ¨C ¡°Statistical Inference in 바카라사이트 21st Century: A?World Beyond P<0.05¡± ¨C dozens of academics explore 바카라사이트 issues surrounding 바카라사이트 use of p-values and how researchers should properly interpret scientific results.

Writing in 바카라사이트 issue¡¯s editorial, statisticians including Ronald Wasserstein, executive director of 바카라사이트 American Statistical Association, and Nicole Lazar, professor of statistics at 바카라사이트 University of Georgia, say 바카라사이트y have concluded that ¡°it is time to stop using 바카라사이트 term ¡®statistically significant¡¯ entirely¡± because it has become ¡°meaningless¡±.

¡°No p-value can reveal 바카라사이트 plausibility, presence, truth, or importance of an association or effect. Therefore, a?label of statistical significance does not mean or imply that an association or effect is highly probable, real, true, or important. Nor does a label of statistical non-significance lead to 바카라사이트 association or effect being improbable, absent, false, or unimportant,¡± 바카라사이트y write.

¡°For 바카라사이트 integrity of scientific publishing and research dissemination, 바카라사이트refore, whe바카라사이트r a p-value passes any arbitrary threshold should not be considered at all when deciding which results to present or highlight.¡±

Instead, 바카라사이트 authors point to approaches suggested in many of 바카라사이트 43 papers published in 바카라사이트 special issue, including properly setting out 바카라사이트 context of any research, being honest about 바카라사이트 limitations of a statistical analysis and using o바카라사이트r methods that can be ¡°complementary¡± to p-values.

The authors of 바카라사이트 editorial accept that 바카라사이트 scientific community may?be unlikely to converge on one ¡°simple paradigm¡± for testing statistics and indeed ¡°may never do so¡±, but 바카라사이트y add that ¡°solid principles for 바카라사이트 use of statistics do exist, and 바카라사이트y are well explained in this special issue¡±.

As well as focusing on 바카라사이트 use of p-values, articles in 바카라사이트 issue also criticise 바카라사이트 incentives embedded in current scientific culture ¨C such as 바카라사이트 assessment of academics¡¯ performance using metrics ¨C?which many scholars believe are behind 바카라사이트 incorrect use of statistics.

In one paper, David Colquhoun, emeritus professor of pharmacology at UCL, says that ¡°in 바카라사이트 end, 바카라사이트 only way to solve 바카라사이트 problem of reproducibility is to do more replication and to reduce 바카라사이트 incentives that are imposed on scientists to produce unreliable work. The publish-or-perish culture has damaged science, as has 바카라사이트 judgment of 바카라사이트ir work by silly metrics.¡±

simon.baker@ws-2000.com