Ioannidis: significance rules need to be improved, not abandoned

Stanford's John Ioannidis rejected proposals to abandon dichotomous statements of significance based on p-values and instead called for better and clearer rules to protect data from bias and to address gaps in the ways researchers conduct, analyze and report studies.

Last month, over 800 researchers signed on to a Nature comment which proposed halting the dichotomous use of "statistical significance" (see "Nature Comment: Abandon Dichotomous Statistical Thresholds")...