The sensitivity of phallometric measures with rapists

The sensitivity of phallometric measures with rapists was investigated using individual rape indices of rapists and non-sex offenders from 14 published and unpublished data sets.Sensitivity (the probability of detecting deviant sexual arousal among rapists) was determined by alternately setting thespecificity (the probability of not detecting deviant sexual arousal among non-sex offenders) at 70%, 80% or 90%. Eight data sets showed sensitivity values equal to, or greater than, 60% when specificity was set at 90%. Stimulus sets that contained more brutal and graphic descriptions of rape produced higher sensitivity values. A conservative rape index cut-off score of 1.0 seems optimal in that it yields moderate sensitivity values while minimizing the number of individuals who are falsely classified as presenting deviant sexual arousal. Conditions that may increase the sensitivity of phallometric assessments are presented and discussed.