This paper assesses the effectiveness of child safety seat laws in the United States. Over the past 35 years, these laws have steadily increased mandatory child safety seat restraint ages. We exploit state‐year level variation in the age until which children are required to ride in child safety seats to estimate triple difference models using Fatality Analysis Reporting System (FARS) data from 1975 to 2011. Our findings show that increasing the age thresholds is effective in increasing the actual age of children in safety seats. Across the child‐age distribution, restraint rates increase by between 10 and 30 percentage points or by between 50 and 170 percent, in the long run. We also estimate the impact of the child safety seat laws on the likelihood that a child dies in a fatal accident. We find that the laws saved up to 39 children per year. Finally, we find that the laws primarily induce compliant parents to switch from traditional seatbelt use to child safety seat use, with only small effects among parents who do not restrain their children.