 The persistence of poor methods results partly from incentives that favor them, leading to the natural selection of bad science. Successful labs produce more progeny such that their methods are more often copied and their students are more likely to start labs of their own. This article was authored by Pauli Smaldino and Richard McElrath.