Scientists like to think of science as self-correcting. ... The peer reviewers who evaluate papers before journals commit to publishing them are much .... Fraud is very likely second to incompetence in generating erroneous results, ... Such headlines are rare, though, because replication is hard and thankless.