While conjoint analysis has been applied to a wide variety of problems in preference analysis, it has heretofore not been used to analyze similarities data. A method for dealing with this problem is described and illustrated on a small set of empirical data