Ditching Statistical Significance?!

Nature (!) has just published an editorial discussing and advocating that statistical significance should be ditched. For me, that’s the stuff of dreams, but I have lived to see it happen! I’m so happy!

Here’s one para from the editorial:

The ‘call for scientists to abandon statistical significance‘, by Valentin Amrhein, Sander Greenland, Blake McShane is the Comment that is also in this week’s Nature. Its title is Scientists rise up against statistical significance, and it is supported by 854 scientists from 52 countries. See a list of these good folks here.

The Comment has been developed over past weeks, with revisions in response to suggestions (and exhortations) from many, including Bob and me. Of course I’d like the final version to be stronger, and to call for an end to any use of p values, or their use only in rare cases when we don’t (yet) have good alternatives. But it is pretty strong, and has much wise advice that would lead to much improved practice if widely followed.

The 854 signatures were collected in a very few days, during which the Comment itself was strictly embargoed. Bob and I were very happy to sign.

The ‘series of related articles‘ mentioned above, and published by the American Statistical Association, has not yet appeared, but should do so any moment. That should also be a massive game changer. Let’s hope.

Sometimes, ‘ditching’ can be great progress!

Geoff