Will history repeat itself? Growth mixture modeling of suspected serial sexual offending using forensic DNA evidence

Sexual offenders often commit more than one sexual assault, but there is variability in how many assaults they commit and in what pattern over time. Trajectory modeling studies typically use criminal history records as a data source to model perpetrators' sexual assault convictions, but this may underestimate the scope of offending because so few sexual assaults result in a conviction. We used both criminal history records and forensic DNA evidence from sexual assault kits (SAKs; also termed ‘rape kit’) to identity a sample of n = 392 serial offenders, all of whom were suspected of committing two or more sexual assaults. Using growth mixture models, we identified and validated a four-class model of suspected serial sexual offending spanning ages 16 to 60. The four classes varied in the overall number of sexual assaults committed by each perpetrator and the ages of peak offending. All classes included sexual assaults identified through rape kit testing and through criminal history records. Forensic DNA testing of rape kits can help identify suspected serial sexual offenders. DNA testing plus criminal history searches on identified offenders as standard investigation practice would provide police with a more complete picture of offenders' criminal behavior. •Criminal history records underestimate rates of serial sexual offending.•Forensic DNA testing of sexual assault kits can identify serial sexual offenders.•Used criminal records and forensic DNA to model perpetration trajectories.•Validated a four-class model of sexual offending spanning ages 16 to 60.