The specificity of phallometric testing for pedophilia has been calculated using sex offenders against adult women. Does the offender's actual number of prior sexual contacts with women affect such estimates? To answer this, the authors' studied 82 male sex offenders against adult women, 172 offenders against unrelated children, and 70 offenders against their own biological children or stepchildren. Phallometric testing included visual and auditory depictions of prepubescent, pubescent, and adult males and females. The results for offenders against women showed that those who had had sexual contact with the greatest number of women (consenting or nonconsenting) had the lowest probability of being diagnosed as pedophilic. Specificity, calculated for those who had sexual contact with the most women and thus the most evidence of attraction to them, was 96%. Sensitivity, calculated analogously for men with the most offenses against children, was 61%.