“A small P-value is just a warning signal that the current model could have a problem, so we should check our model assumptions” https://t.co/v277HwzlWg
@robustgar Haven't read the paper, but thoroughly disagree with the abstract. I wonder how the authors account for the fact that preregistered work paints a picture that is much less rosy than non-prereg work. That gap alone is worth the name "crisis". @chrisdc77
@robustgar Bravo! I am consistently impressed by the wonderfully written pleas for rational science published in the American Statistitician. I am then subsequently crushed that nobody seems to give a damn.