A group of over 800 scientists have signed their names to an article published in Nature, explaining why statistical significance shouldn’t be relied on so heavily as a measure of the success of an experiment. We asked statistics buff Andrew Steele to explain.