Ten years after the American Statistical Association released a statement recommending against the dichotomous interpretation of p-values, what has changed and what still needs to change? I will review the history of the tenacious grip of “statistical significance” on scientific thinking, aiming to clarify the underlying problems and summarise progress towards better practice, while suggesting a connection between dichotomania and a perception of statistical methods as ritualised procedures that give the illusion of answers to poorly specified questions.
Slides from the presentation are available.