What explains the low rate of retractions in political science?

Very few political science articles have been retracted. We argue that this surprisingly low number of retractions does not indicate greater a priori honesty within our profession compared to other disciplines. Instead, we identify three alternative explanations: (1) data access and transparency requirements that deter research misconduct, (2) the dominance of observational data analyses and causal models, in which undetectable tweaks to research designs and model specifications are the fraudulent method of choice for producing desired results, and (3) the paucity of institutionalized checks and balances that help with detecting fraud, such as meta-analyses, replication studies, and robustness tests that are not carefully selected by the authors themselves. We suggest that the low number of retractions is best accounted for by the second and third explanations, which combine to create an environment where highly problematic research practices persist without leading to formal retractions.