The Alpha War

Edouard Machery (University of Pittsburgh)

Abstract: Psychology, epidemiology, and more than a few other sciences are in a crisis: Many published results, including classic, textbook ones, appear not to replicate, and whole empirical literatures are now under suspicion. Despite undeniable conservatism, proposals to improve psychology are put forward, discussed in journals and on blogs, and sometimes implemented by audacious editors. The present talk discusses one of these proposals: Cut the conventional significance level by an order of magnitude (Benjamin et al., 2018). This modest, practical proposal has been widely criticized, and its prospects remain unclear. This talk defends this proposal against these criticisms and highlights its virtues.