Despite a large volume of work on the defense-growth relationship in the United States, previous studies have failed to reach a consensus. Part of the explanation of the mixed results may be that various data sets and model designs have been employed. Thus, using an updated (1948-1996) and longer data set, I test several of the more prominent defense-growth models. This research design will eliminate sampling bias and increase reliability of findings. The results of this research design show that defense sp;ending does not have a significant, direct impact on economic growth in the United States. Instead, nonmilitary government expenditures have significant, negative effects on economic growth.