Problems with experimental reproducibility have been of concern in the scientific community of late, which has been attributed as least in part to our misuse of p-values. Misconceptions of the p-value have also been discussed previously in Nature news features.

Last week the American Statistical Association (ASA) released a statement warning about the misconceptions and misuse of the p-value. For example, they highlight how our obsession with p=0.05 as a cutoff may be rooted in a circular argument: we use p=0.05 for experiments because that is what we are taught, and we are taught p=0.05 because this is how we conduct experiments. However, RA Fisher had never intended the p-value to replace scientific inquiry, but rather to assess whether a hypothesis merited further examination. The ASA statement includes six principles, many of which address the misconceptions and misuse of the p-value.

Summaries of the ASA statement are also available from Nature and the ASA.