AbstractAutomobile safety standards adopted because of federal legislation have, according to many researchers, failed to save lives; safer cars, they infer, induce more dangerous driving. Results from the major study suggesting this conclusion are shown, however, to be quite sensitive to reasonable changes in the empirical model. An alternative statistical analysis of death rates supports a very different conclusion: Safety standards have saved tens of thousands of lives during the 1970s. More broadly, the article illustrates some general pitfalls in using statistical results in policy analysis, and how they can be avoided.