A group of statisticians are pushing to abolish the notion of 'statistical significance,' NPR reports.
They instead are asking scientists to cope with having uncertainty. "Uncertainty is present always," Ron Wasserstein, the executive director of the American Statistical Association, tells NPR. "That's part of science. So rather than trying to dance around it, we [should] accept it."
The journal the American Statistician has dedicated its most recent issue to the question. In an editorial, Wasserstein and his co-authors argue that "[s]tatistical significance was never meant to imply scientific importance" and that it has taken on an outsized role in science, with statistical significance affecting what papers get to be published and who gets funding or promoted.
In a separate editorial appearing in Nature, the University of Basel's Valentin Amrhein, his co-authors, and some 800 co-signers say they don't want to ban the use of p values, but change how they are used. Rather, they argue that p values should not be used as the arbiters of whether a finding supports a hypothesis or not.
American Statistician editorial co-author Nicole Lazar from the University of Georgia tells Retraction Watch that what they are proposing has been suggested before. In 2017, the ASA called for researchers to change how they rely on p values and a group of psychologists that year also pushed for the p value threshold to be lowered, while another group said researchers should instead explain whey they chose the cutoff value they did.
"The tone now is different," she adds at Retraction Watch, "perhaps because of the more pervasive sense that what we've always done isn't working, and so the time seemed opportune to renew the call."