Light water reactor safety in the United States through the Three Mile Island (TMI) accident is reviewed and safety gains up to 1983 are assessed. The work was undertaken during 1982–1983 to support Alvin Weinberg's “Second Nuclear Era Study.” Safety studies, such as the “Rasmussen Reactor Safety Study,” and accidents, such as TMI, disclosed deficiencies in the design basis accident technique and the lack of attention to human factors, including management. Operational problems also highlighted the fact that reactors grew too quickly in size and number for experience to be factored into designs, operating procedures, and regulations. Actions taken by the U.S. Nuclear Regulatory Commission and by the industry following TMI are reviewed. Analytical and experimental programs on severe accidents are examined. Safety gains resulting from these activities and from improved quality of operation are assessed qualitatively and quantitatively. It was determined that a factor of 3 to 6 for core melt frequencies and of at least 10 for public risk have been achieved since TMI.