Nature Comment: abandon dichotomous statistical thresholds

A trio of authors called for an end to the dichotomous use of "statistical significance" in a comment published Wednesday in Nature, which over 800 international researchers signed.

The authors -- Valentin Amrhein, Sander Greenland and Blake McShane -- said scientists should never conclude whether a result supports a scientific hypothesis based solely on significance thresholds, adding that the conventional p-value threshold of 0.05 is in itself arbitrary...