For the past forty years, deterrence theory has played a central role in the evolution of U.S. nuclear strategy and national security policy. Proponents of deterrence point to the historical record to argue the theory “works”. Recent critiques of the theory have emphasized, however, that many of the fundamental questions regarding how deterrence works remain unanswered. For example, how do decision makers decide “how much is enough” to deter aggression? What role do cognitive processes play in deterrent decisions? The research described in this paper was designed to examine such questions. Specifically, the effects of decision frame, salience of intergroup boundaries, and feedback regarding the economic consequences of decisions on deterrent behavior were investigated using a laboratory analogue of a security dilemma. Results from two experiments are presented. The first experiment confirms that decision frame and categorization can adversely affect deterrent allocations. The second experiment suggests that reframing decisions and providing feedback regarding the costs of deterrence might reduce decision makers’ tendency to over-allocate resources for defense. The external validity of these findings are discussed, as well as implications of such research for the study of peace among the nuclear powers.