Science lies nowadays in the centre of several storms. The better known is the finding of non-reproducibility of many scientific results, which stretches from the medical field (clinic and pre-clinic tests) to study on behaviour (priming research). Although the bad use of statistics is reported to be a patent cause of the reproducibility crisis, its deep reasons are to be sought elsewhere; particularly, in the passage from a regimen of little science - regulated by small communities of researchers - to the current big science - identified by a hypertrophic production of millions of research papers and by the imperative "publish or perish", in a setting dominated by market. While spirited debates (on vaccines, climate change, GMO) unfold in society, scientific articles which are bought or withdrawn are the signal of a deep crisis not only of science, but also of the expert thought. In this background, statistics is the main defendant, charged with using methods which experts themselves are not able to explain in an understandable way (p-test). Is there an escape? Yes, there is. Researchers can either court the power and defend the status quo, or contribute to a deep process of reformation, refusing both a vision of science as a religion and the idea that the problem is the poor scientific knowledge of the lay public.