This paper describes and evaluates the four approaches to nuclear threat assessment that are most commonly employed by U.S. national security analysts: (1) comparisons of U.S. and Soviet expenditures on nuclear forces; (2) comparisons of static indicators of nuclear capability, such as number of weapons or measures of the destructive power of these weapons; (3) using deterrence theory and battle simulations to predict nuclear war outcomes and evaluate the deterrent capability of U.S. nuclear forces; and (4) analyzing official Soviet military doctrine to make inferences about Soviet goals, images, and plans for nuclear war. The analysis of each approach begins with a brief description of how the approach is implemented. Next the conceptual bases of the approach—both theoretical and empirical—are examined. Then applications of the method are evaluated in terms of how well they deal with the uncertainties inherent in any attempt to assess nuclear capability or threat. Finally, applications of each method are evaluated in terms of how they incorporate subjective judgment, how well they document critical assumptions, and the extent to which they use sensitivity analysis to assess the impact of varying critical assumptions.