Table 2.
The Predictive Validity of ERASOR, J-SOAP-II and YLS/CMI for Sexual and Nonsexual Recidivism
| Sexual recidivism | Nonsexual recidivism | |||
|---|---|---|---|---|
| Risk assessment measure | AUC (SE) | 95% CI | AUC (SE) | 95% CI |
| ERASOR (N = 104) | ||||
| Total score | .74* (.07)a | [.61, .88] | .66* (.06)a | [.54, .78] |
| Overall clinical rating | .83**(.07)a | [.70, .96] | .69**(.06)a | [.58, .80] |
| Sexual interests, attitudes, & behaviors | .64 (.10) | [.45, .83] | .67* (.06) | [.54, .79] |
| Historical sexual assaults | .81** (.06)a | [.70, .92] | .54 (.06) | [.42, .67] |
| Psychosocial functioning | .49 (.08) | [.33, .65] | .58 (.07) | [.44, .71] |
| Family/environmental functioning | .49 (.11) | [.28, .70] | .66* (.06)a | [.54, .77] |
| Treatment | .55 (.11) | [.33, .78] | .51 (.07) | [.38, .63] |
| J-SOAP-II (N = 102) | ||||
| Total score | .51 (.09) | [.33, .69] | .79*** (.06)a | [.68, .89] |
| Sexual drive/ preoccupation | .72* (.08)a | [.56, .89] | .52 (.07) | [.39, .66] |
| Impulsive/antisocial behavior | .37 (.10) | [.18, .56] | .71** (.06)a | [.60, .82] |
| Intervention | .41 (.10) | [.22, .60] | .79***(.05)a | [.68, .89] |
| Community stability/ adjustment | .55 (.11) | [.34, 76] | .69** (.06)a | [.58, .81] |
| YLS/CMI (N = 104) | ||||
| Total score | .29 (.08) | [.15, .44] | .65* (.06)a | [.53, .76] |
| Offenses/disposition | .43 (.10) | [.24, .61] | .51 (.07) | [.38, .64] |
| Family circumstances & parenting | .37 (.12) | [.13, .60] | .54 (.06) | [.42, .66] |
| Education/employment | .31 (.08) | [.15, .47] | .66* (.06) | [.54, .79] |
| Peer relations | .31 (.10) | [.12, .50] | .62 (.06) | [.50, .73] |
| Substance abuse | .50 (.11) | [.29, .71] | .48 (.06) | [.36, .61] |
| Lesisure/recreation | .56 (.10) | [.35, .76] | .57 (.06) | [.44, .69] |
| Personality/behavior | .24 (.08) | [.09, .40] | .63* (.06) | [.52, .75] |
| Attitudes/orientation | .36 (.10) | [.16, .55] | .66* (.06) | [.54, .79] |
a.
Denotes that the risk assessment measure significantly predicted recidivistic outcome even after Benjamini and Hochberg False Discovery Rate (FDR) corrections.
*
p < .05. ** p < .01 ***p < .001.