Despite the fact that the roll call voting patterns of legislators are being used more frequently in the political arena, the statistical methods employed by political scientists for analyzing roll calls (e.g., Guttman scaling, factor analysis) are not being used. The main reason is the failure to incorporate in a systematic manner the subjective estimates of the political organizations and individuals who must use the output of such analyses. This paper presents two methods, paired comparison and constant sum, which use judgmental data in assessing roll calls. Using a set of seven defense policy roll calls from the 1976 US Senate, the methods are described and compared using two sets of judges.