Predictive Validity
Morton (2003) coded the files of 78 adolescent males who had participated in a previous investigation of the efficacy of treatment (Worling & Curwen, 2000). Although many of the files contained only basic demographic data, several of the participants had participated in comprehensive assessments as a result of sexual offenses. Recidivism data (criminal charges) were available from a nation-wide database over an average follow-up period of 68 months (SD=22.0). Although the ERASOR Total score was not found to be predictive of sexual recidivism in this study (AUC=.59; 95% CI .43-.75), the ERASOR was significantly predictive of violent (including sexual) reoffending (AUC=.65; 95% CI .53-.78). Morton also reported that a modified score, based on 9 ERASOR items, was significantly predictive of sexual recidivism (AUC=.74; 95% CI .58-.90).
Skowron (2004) obtained recidivism data (criminal charges) from a nation-wide database over an average follow-up period of 47.3 months (SD = 41.1). Of the 110 adolescent males with a history of sexual offenses, the recidivism rates were as follows: any reoffense (69%), nonsexual violent offense (17%), and sexual reoffense (35%). It was found that the sum of ERASOR risk factors rated “Present” was significantly predictive of any reoffense (AUC=.67; SE=.04), any nonsexual violent offense (AUC=.64; SE=.04), and any sexual reoffense (AUC=.71; SE=.05).
Worling (2004) investigated the ability of the ERASOR to discriminate those adolescent males who had offended sexually and were detected for the first time (“nonrepeaters”) from those adolescent males who offended sexually, were detected, and then reoffended sexually (“repeaters”). Based on clinical ratings completed by clinicians following comprehensive assessments for 136 adolescents, both the overall clinical rating of risk (AUC = .66; 95% CI .55-.76) and the Total score (AUC = .72; 95% CI .62-.80) were significantly predictive of repeater status.
Hersant (2006) examined ERASOR ratings for 91 adolescent males following clinical assessments made at 3 residential treatment programs. It was found that both the Total score (AUC=.66) and the clinical judgment risk ratings (AUC=.67) could significantly differentiate those adolescents known to have offended—been detected—then reoffended (“repeaters”) from those who were being detected for the first time (“nonrepeaters”). Hersant (2006) reported d values; AUC values are based on transformations provided by Rice and Harris, 2005)
McCoy (2007) collected ERASOR ratings made by students using archival file information. With a sample of 128 adolescent males who had offended sexually, and an unspecified follow-up interval, it was found that the ERASOR Total score was not predictive of sexual recidivism (AUC=.50; 95% CI not specified).
Viljoen et al. (2009) examined recidivism data for 193 adolescent males who participated in residential treatment as a result of past sexual offenses. ERASOR ratings were completed following a review of extensive clinical files, and the mean follow-up period was 7.2 years. Although several popular risk-assessment tools were examined (also see Viljoen et al., 2008), the clinical risk rating based on the ERASOR was the only score that was moderately predictive of risk of sexual recidivism (AUC=.64; 95% CI .49-.79).
Rajlic & Gretton (2010) examined ERASOR ratings that were made by undergraduate and graduate students using archival file material for a sample of 268 adolescent males involved in a provincial treatment system. With a mean follow-up period of approximately 6.5 years, they found that the ERASOR clinical judgment ratings (AUC=.67; 95% CI .55-.78) and the ERASOR Total Score (AUC=.71; 95% CI .62-.80) were significantly predictive of sexual reoffending.
Nelson (2011) employed 4 graduate students to code archival files for 93 adolescent males who had been committed to a state juvenile justice agency. Using state records and a mean follow-up period of 6.3 years, it was found that neither the ERASOR Total score (AUC=.48; 95% CI .33-.63) nor the clinical judgment ratings (AUC=.49; 95% CI .28-.69) were predictive of sexual recidivism.
Worling, Bookalam, and Litteljohn (2012) conducted a prospective study of the validity of the ERASOR. They examined ERASOR ratings made by 22 different clinicians immediately following comprehensive clinical evaluations. Ratings for 191 adolescent males from 5 agencies were examined. Recidivism data (criminal charges) were subsequently collected, and two mean follow-up time periods were examined: 3.66 years (SD=2.08) and 1.4 years (SD=.71). Over the shorter follow-up interval, the clinical judgment ratings (AUC=.82; 95% CI .69-.96), the Total score (AUC=.93; 95% CI .86-.99), and the sum of risk factors rated “Present” (AUC=.90; 95% CI .81-.99) were significantly predictive of subsequent sexual recidivism. Using the average follow-up period of 3.66 years, the clinical judgment ratings were not significantly predictive (AUC=.61; 95% CI .48-.74); however, both the Total score (AUC=.72; 95% CI .61-.83) and the sum of risk factors rated “Present” (AUC=.73; 95% CI .63-.84) were significantly predictive of subsequent sexual recidivism over this time period.
Chu, Ng, Fong, & Teoh (2012) compared the predictive validity of the ERASOR, J-SOAP-II (Prentky & Righthand, 2001), and the YLS/CMI (Hoge & Andrews, 2002) for youth in Singapore. The authors coded extensive clinical files for 104 male youth who had offended sexually, and they collected official recidivism data from a national database that spanned an average follow-up period of 4.5 years. It was found that the ERASOR clinical judgment rating (AUC=.83; 95% CI .70-.96) and Total score (AUC=.74; 95% CI .61-.88) were significantly predictive of sexual recidivism. It was further reported that ERASOR clinical judgment ratings were significantly more predictive than the ERASOR Total score, and both ERASOR measures were significantly more predictive of sexual recidivism relative to the other tools that were examined. Chu et al. (in press) also found that the ERASOR clinical judgment was the only variable that could predict time to sexual reoffense.
In a recent meta-analysis of 10 investigations, Viljoen, Mordell, & Beneteau (2012) found that both the clinical judgment rating (AUC=.66; 95% CI .61-.72) and the Total score (AUC=.66; 95% CI .60-.71) from the ERASOR were significantly predictive of sexual reoffending. They also estimated that there would need to be 11 studies with an effect size of 0 added to this meta-analysis to make this moderate effect size a small effect size.
Rojas Mejia (2013) coded the ERASOR and several other risk assessment tools for a sample of 100 males aged 12-19. It was found that, with an average follow-up period of just over 11 years, 8 participants had convictions for subsequent sexual crimes, and 24 had convictions for violent (sexual and nonsexual) crimes. Although the ERASOR Total score (AUC=.64; 95% CI .46-.81) and clinical judgment (AU=.67; 95% CI .50-.83) were not significantly predictive for the 8 youth who reoffended sexually, the ERASOR Total score (AUC=.67; 95% CI .54-.79) was predictive of violent (sexual and nonsexual) recidivism. Using survival analysis, it was also found that those adolescents rated as high risk reoffended sexually at a significantly faster rate than those rated as low risk.
References
Chu, M. C., Ng, Kynaston, Fong, J., & Teoh, J. (2012). Assessing youth who sexually offended: The predictive validity of the ERASOR, J-SOAP-II, and the YLS/CMI in a non-western context. Sexual Abuse: A Journal of Research and Treatment, 24, 153-174.
Hersant, J. L. (2006). Risk assessment of juvenile sex offender reoffense. Unpublished doctoral dissertation. Southern Illinois University Carbondale.
Hoge, R. D., & Andrews, D. A. (2002). Youth Level of Service/Case Management Inventory: User’s manual. Toronto, Canada: Multi-Health Systems.
McCoy, W. K. (2007). Predicting treatment outcome and recidivism among juvenile sex offenders: The utility of the JSOAP-II and ERASOR in an outpatient treatment program. Unpublished doctoral dissertation. Sam Houston State University. Hunstville, Texas.
Morton, K. E. (2003). Psychometric properties of four risk assessment measures with male adolescent sex offenders. Unpublished master’s thesis. Carleton University. Ottawa, Ontario, Canada.
Nelson, R. (2011). Predicting recidivism among juvenile sex offenders: The validity of the ERASOR. Unpublished thesis. Roger Williams Univeristy. Bristol, Rhode Island.
Prentky, R., & Righthand, S. (2001). Juvenile Sex Offender Assessment Protocol—II (J-SOAP-II): Manual. Unpublished document. Available from www.csom.org.
Rajlic, G., & Gretton, H. M. (2010). An examination of two sexual recidivism risk measures in adolescent offenders: The moderating effect of offender type. Criminal Justice and Behavior, 37, 1066-1085.
Rice, M. E., & Harris, G. T. (2005). Comparing effect sizes in follow-up studies: ROC area, Cohen’s d, and r. Law and Human Behavior, 29, 615-620.
Rojas Megia, E. Y. (2013), Violence risk assessment with youth who have sexually offended: A psychometric examination of the Violence Risk Scale: Youth Sexual Offender Version (VRS:YSO). Unpublished doctoral dissertation. University of Saskatchewan. Saskatoon, Saskatchewan, Canada.
Skowron, C. (2004, December). Differentiation and predictive factors in adolescent sexual offending. Unpublished doctoral dissertation. Carleton University. Ottawa, Ontario, Canada.
Viljoen, J. L., Elkovitch, N., Scalora, M. J., & Ullman, D. (2009). Assessment of reoffense risk in adolescents who have committed sexual offenses: Predictive validity of the ERASOR, PCL:YV, YLS/CMI, and Static-99. Criminal Justice and Behavior, 36, 981-1000.
Viljoen, J. L., Mordell, S., & Beneteau, J. L. (in press). Prediction of adolescent sexual reoffending: A meta-analysis of the J-SOAP-II, ERASOR, J-SORRAT-II, and Static-99. Law and Human Behavior.
Viljoen, J. L., Scalora, M., Cuadra, L., Bader, S., Chávez, V., Ullman, D., & Lawrence, L. (2008). Assessing risk for violence in adolescents who have sexually offended: A comparison of the J-SOAP-II, J-SORRAT-II, and SAVRY. Criminal Justice and Behavior, 35, 5-23.
Worling, J. R. (2004). The Estimate of Risk of Adolescent Sexual Offense Recidivism (ERASOR): Preliminary Psychometric Data. Sexual Abuse: A Journal of Research and Treatment, 16, 235-254.
Worling, J. R., & Curwen, T. (2001). Estimate of Risk of Adolescent Sexual Offense Recidivism (Version 2.0: The “ERASOR”). In M. C. Calder, Juveniles and children who sexually abuse: Frameworks for assessment (pp. 372-397). Lyme Regis, Dorset, UK: Russell House Publishing.
Worling, J. R., & Curwen, T. (2000). Adolescent sexual offender recidivism: Success of specialized treatment and implications for risk prediction. Child Abuse & Neglect, 24, 965-982.
Worling, J. R., Bookalam, D., & Litteljohn, A. (2012). Prospective validity of the Estimate of Risk of Adolescent Sexual Offense Recidivism (ERASOR). Sexual Abuse: A Journal of Research and Treatment, 24, 203-223.
Morton (2003) coded the files of 78 adolescent males who had participated in a previous investigation of the efficacy of treatment (Worling & Curwen, 2000). Although many of the files contained only basic demographic data, several of the participants had participated in comprehensive assessments as a result of sexual offenses. Recidivism data (criminal charges) were available from a nation-wide database over an average follow-up period of 68 months (SD=22.0). Although the ERASOR Total score was not found to be predictive of sexual recidivism in this study (AUC=.59; 95% CI .43-.75), the ERASOR was significantly predictive of violent (including sexual) reoffending (AUC=.65; 95% CI .53-.78). Morton also reported that a modified score, based on 9 ERASOR items, was significantly predictive of sexual recidivism (AUC=.74; 95% CI .58-.90).
Skowron (2004) obtained recidivism data (criminal charges) from a nation-wide database over an average follow-up period of 47.3 months (SD = 41.1). Of the 110 adolescent males with a history of sexual offenses, the recidivism rates were as follows: any reoffense (69%), nonsexual violent offense (17%), and sexual reoffense (35%). It was found that the sum of ERASOR risk factors rated “Present” was significantly predictive of any reoffense (AUC=.67; SE=.04), any nonsexual violent offense (AUC=.64; SE=.04), and any sexual reoffense (AUC=.71; SE=.05).
Worling (2004) investigated the ability of the ERASOR to discriminate those adolescent males who had offended sexually and were detected for the first time (“nonrepeaters”) from those adolescent males who offended sexually, were detected, and then reoffended sexually (“repeaters”). Based on clinical ratings completed by clinicians following comprehensive assessments for 136 adolescents, both the overall clinical rating of risk (AUC = .66; 95% CI .55-.76) and the Total score (AUC = .72; 95% CI .62-.80) were significantly predictive of repeater status.
Hersant (2006) examined ERASOR ratings for 91 adolescent males following clinical assessments made at 3 residential treatment programs. It was found that both the Total score (AUC=.66) and the clinical judgment risk ratings (AUC=.67) could significantly differentiate those adolescents known to have offended—been detected—then reoffended (“repeaters”) from those who were being detected for the first time (“nonrepeaters”). Hersant (2006) reported d values; AUC values are based on transformations provided by Rice and Harris, 2005)
McCoy (2007) collected ERASOR ratings made by students using archival file information. With a sample of 128 adolescent males who had offended sexually, and an unspecified follow-up interval, it was found that the ERASOR Total score was not predictive of sexual recidivism (AUC=.50; 95% CI not specified).
Viljoen et al. (2009) examined recidivism data for 193 adolescent males who participated in residential treatment as a result of past sexual offenses. ERASOR ratings were completed following a review of extensive clinical files, and the mean follow-up period was 7.2 years. Although several popular risk-assessment tools were examined (also see Viljoen et al., 2008), the clinical risk rating based on the ERASOR was the only score that was moderately predictive of risk of sexual recidivism (AUC=.64; 95% CI .49-.79).
Rajlic & Gretton (2010) examined ERASOR ratings that were made by undergraduate and graduate students using archival file material for a sample of 268 adolescent males involved in a provincial treatment system. With a mean follow-up period of approximately 6.5 years, they found that the ERASOR clinical judgment ratings (AUC=.67; 95% CI .55-.78) and the ERASOR Total Score (AUC=.71; 95% CI .62-.80) were significantly predictive of sexual reoffending.
Nelson (2011) employed 4 graduate students to code archival files for 93 adolescent males who had been committed to a state juvenile justice agency. Using state records and a mean follow-up period of 6.3 years, it was found that neither the ERASOR Total score (AUC=.48; 95% CI .33-.63) nor the clinical judgment ratings (AUC=.49; 95% CI .28-.69) were predictive of sexual recidivism.
Worling, Bookalam, and Litteljohn (2012) conducted a prospective study of the validity of the ERASOR. They examined ERASOR ratings made by 22 different clinicians immediately following comprehensive clinical evaluations. Ratings for 191 adolescent males from 5 agencies were examined. Recidivism data (criminal charges) were subsequently collected, and two mean follow-up time periods were examined: 3.66 years (SD=2.08) and 1.4 years (SD=.71). Over the shorter follow-up interval, the clinical judgment ratings (AUC=.82; 95% CI .69-.96), the Total score (AUC=.93; 95% CI .86-.99), and the sum of risk factors rated “Present” (AUC=.90; 95% CI .81-.99) were significantly predictive of subsequent sexual recidivism. Using the average follow-up period of 3.66 years, the clinical judgment ratings were not significantly predictive (AUC=.61; 95% CI .48-.74); however, both the Total score (AUC=.72; 95% CI .61-.83) and the sum of risk factors rated “Present” (AUC=.73; 95% CI .63-.84) were significantly predictive of subsequent sexual recidivism over this time period.
Chu, Ng, Fong, & Teoh (2012) compared the predictive validity of the ERASOR, J-SOAP-II (Prentky & Righthand, 2001), and the YLS/CMI (Hoge & Andrews, 2002) for youth in Singapore. The authors coded extensive clinical files for 104 male youth who had offended sexually, and they collected official recidivism data from a national database that spanned an average follow-up period of 4.5 years. It was found that the ERASOR clinical judgment rating (AUC=.83; 95% CI .70-.96) and Total score (AUC=.74; 95% CI .61-.88) were significantly predictive of sexual recidivism. It was further reported that ERASOR clinical judgment ratings were significantly more predictive than the ERASOR Total score, and both ERASOR measures were significantly more predictive of sexual recidivism relative to the other tools that were examined. Chu et al. (in press) also found that the ERASOR clinical judgment was the only variable that could predict time to sexual reoffense.
In a recent meta-analysis of 10 investigations, Viljoen, Mordell, & Beneteau (2012) found that both the clinical judgment rating (AUC=.66; 95% CI .61-.72) and the Total score (AUC=.66; 95% CI .60-.71) from the ERASOR were significantly predictive of sexual reoffending. They also estimated that there would need to be 11 studies with an effect size of 0 added to this meta-analysis to make this moderate effect size a small effect size.
Rojas Mejia (2013) coded the ERASOR and several other risk assessment tools for a sample of 100 males aged 12-19. It was found that, with an average follow-up period of just over 11 years, 8 participants had convictions for subsequent sexual crimes, and 24 had convictions for violent (sexual and nonsexual) crimes. Although the ERASOR Total score (AUC=.64; 95% CI .46-.81) and clinical judgment (AU=.67; 95% CI .50-.83) were not significantly predictive for the 8 youth who reoffended sexually, the ERASOR Total score (AUC=.67; 95% CI .54-.79) was predictive of violent (sexual and nonsexual) recidivism. Using survival analysis, it was also found that those adolescents rated as high risk reoffended sexually at a significantly faster rate than those rated as low risk.
References
Chu, M. C., Ng, Kynaston, Fong, J., & Teoh, J. (2012). Assessing youth who sexually offended: The predictive validity of the ERASOR, J-SOAP-II, and the YLS/CMI in a non-western context. Sexual Abuse: A Journal of Research and Treatment, 24, 153-174.
Hersant, J. L. (2006). Risk assessment of juvenile sex offender reoffense. Unpublished doctoral dissertation. Southern Illinois University Carbondale.
Hoge, R. D., & Andrews, D. A. (2002). Youth Level of Service/Case Management Inventory: User’s manual. Toronto, Canada: Multi-Health Systems.
McCoy, W. K. (2007). Predicting treatment outcome and recidivism among juvenile sex offenders: The utility of the JSOAP-II and ERASOR in an outpatient treatment program. Unpublished doctoral dissertation. Sam Houston State University. Hunstville, Texas.
Morton, K. E. (2003). Psychometric properties of four risk assessment measures with male adolescent sex offenders. Unpublished master’s thesis. Carleton University. Ottawa, Ontario, Canada.
Nelson, R. (2011). Predicting recidivism among juvenile sex offenders: The validity of the ERASOR. Unpublished thesis. Roger Williams Univeristy. Bristol, Rhode Island.
Prentky, R., & Righthand, S. (2001). Juvenile Sex Offender Assessment Protocol—II (J-SOAP-II): Manual. Unpublished document. Available from www.csom.org.
Rajlic, G., & Gretton, H. M. (2010). An examination of two sexual recidivism risk measures in adolescent offenders: The moderating effect of offender type. Criminal Justice and Behavior, 37, 1066-1085.
Rice, M. E., & Harris, G. T. (2005). Comparing effect sizes in follow-up studies: ROC area, Cohen’s d, and r. Law and Human Behavior, 29, 615-620.
Rojas Megia, E. Y. (2013), Violence risk assessment with youth who have sexually offended: A psychometric examination of the Violence Risk Scale: Youth Sexual Offender Version (VRS:YSO). Unpublished doctoral dissertation. University of Saskatchewan. Saskatoon, Saskatchewan, Canada.
Skowron, C. (2004, December). Differentiation and predictive factors in adolescent sexual offending. Unpublished doctoral dissertation. Carleton University. Ottawa, Ontario, Canada.
Viljoen, J. L., Elkovitch, N., Scalora, M. J., & Ullman, D. (2009). Assessment of reoffense risk in adolescents who have committed sexual offenses: Predictive validity of the ERASOR, PCL:YV, YLS/CMI, and Static-99. Criminal Justice and Behavior, 36, 981-1000.
Viljoen, J. L., Mordell, S., & Beneteau, J. L. (in press). Prediction of adolescent sexual reoffending: A meta-analysis of the J-SOAP-II, ERASOR, J-SORRAT-II, and Static-99. Law and Human Behavior.
Viljoen, J. L., Scalora, M., Cuadra, L., Bader, S., Chávez, V., Ullman, D., & Lawrence, L. (2008). Assessing risk for violence in adolescents who have sexually offended: A comparison of the J-SOAP-II, J-SORRAT-II, and SAVRY. Criminal Justice and Behavior, 35, 5-23.
Worling, J. R. (2004). The Estimate of Risk of Adolescent Sexual Offense Recidivism (ERASOR): Preliminary Psychometric Data. Sexual Abuse: A Journal of Research and Treatment, 16, 235-254.
Worling, J. R., & Curwen, T. (2001). Estimate of Risk of Adolescent Sexual Offense Recidivism (Version 2.0: The “ERASOR”). In M. C. Calder, Juveniles and children who sexually abuse: Frameworks for assessment (pp. 372-397). Lyme Regis, Dorset, UK: Russell House Publishing.
Worling, J. R., & Curwen, T. (2000). Adolescent sexual offender recidivism: Success of specialized treatment and implications for risk prediction. Child Abuse & Neglect, 24, 965-982.
Worling, J. R., Bookalam, D., & Litteljohn, A. (2012). Prospective validity of the Estimate of Risk of Adolescent Sexual Offense Recidivism (ERASOR). Sexual Abuse: A Journal of Research and Treatment, 24, 203-223.
