Predicting hands-on child sexual offenses among possessors of Internet child pornography.

Abstract

The recent Supreme Court ruling in U.S. v. Comstock (2010) upheld the constitutionality of The Adam Walsh Act, which provides for civil commitment of child pornography (CP) offenders certified as sexually dangerous, thereby approving litigation of all such prisoners in the federal system. The two studies reported here sought to address the question: What is the likelihood that an individual convicted of child pornography offenses has a prior history of a hands-on sexual offense involving a child or has a high probability of committing such an offense? Our sample consisted of 349 participants: 113 who committed an Internet sexual offense only and no other known or self-reported hands-on sexual offense, 176 child molesters who reported no Internet sexual offense, and 60 child molesters that reported committing an Internet sexual offense. Study 1 yielded two scales, one reflecting Antisocial Behavior (AB) and one reflecting Internet Preoccupation (IP). Those two scales predicted membership in the combined sample of child molesters with a high degree of accuracy (c = 0.75). Study two revealed that all three groups were discrete with respect to AB and IP. By increasing the IP scale by 1 point, the odds of being an IO rather than a CM increased by 86%. The plotted conditional probabilities increased linearly as values on the AB scale increase, from 0.27 when AB = 0 to 0.84 when AB = 13. Our results are discussed in terms of risk discrimination among possessors of child pornography, relevance of risk to the statutory third prong element of serious difficulty, and the policy implications of the findings reported here.