We analyze a game-theoretic model of crime and crime reporting to study the quality of crime statistics. A citizen potentially engages in illicit behavior; an enforcement agency chooses effort and how to report outcomes. Because of signaling concerns, the agency may misreport. We show that multiple equilibria can exist and characterize when crime is under- or over-reported. Increasing the agency’s costs of data manipulation can backfire, increasing misreporting in crime statistics. When calculating treatment effects of parameter changes, the effect on reported statistics will not equal the effect on true statistics, and the true effect can be under- or overestimated.