The AB scale has proven to be a very useful tool in psychotherapy research for a simple and compelling reason: it predicts a psychotherapist's potential effectiveness. However, research with the scale is complicated by the current existence of nine different versions. The purpose of this paper was to clarify this situation by means of a comparative analysis of several of these versions.This comparative analysis was conducted by two methods: 1) a review of the “developmental history” of these nine versions, and 2) an empirical study of the accuracy of several versions of the scale in identifying successful therapists.The historical review of the development of the scales revealed that two of the versions and one of the scoring methods had psychometric shortcomings, leading to the suggestion that they probably should not be utilized in research.The empirical portion of the comparative analysis suggested that, of the remaining seven versions of the test, three were the strongest, in the sense of being most reliable and valid. Since at the present time there are no data permitting any distinctions among these three versions, the best strategy for current research is to use all three and to compare their relative effectiveness.Such a strategy still leaves researchers with a choice of three versions, but this represents considerable gain over the nine possibilities which existed prior to this analysis.