Are Civilly Detained and Committed Sexually Violent Persons Released After Age 60 Low Risk?

Abstract

Older age is a well-established protective factor against sexual recidivism. However, research on the effects of aging in individuals considered for civil commitment as sexually violent persons (SVPs) is limited. The current study investigated 53 such individuals released after age 60: 15 were released following dismissal of the civil commitment petition, and 38 were discharged following civil commitment. Recidivism outcomes, including details of the release environment, were examined and the groups were compared on recidivism risk and age-related factors. The sexual recidivism rate across both groups was 7.5% over an average follow-up of about 9 years. The rate of sexually violent offenses was 3.8%. Results suggest that age-related protective effects apply to older SVPs. However, the groups differed in total time institutionalized and release environments. Findings highlight the potential impact of release type and environments on observed recidivism rates. Risk management implications for older individuals released from SVP programs are considered.

Keywords

sexually violent persons civil commitment recidivism risk age supervision

Aging is associated with decreased criminal behavior, and sexual offending is no exception (Farrington, 1986; Hanson, 2002). The mitigating effect of aging for sexual recidivism has been studied for decades (e.g., Barbaree & Blanchard, 2008; Hanson & Bussiere, 1998; Thornton, 2006). In a large multisite study, individuals with a history of sexual offending released over age 60 had a recidivism rate of just 3.8%, whereas the overall rate across all age bands was 17.5% (Hanson, 2002). Consequently, the relevance of an individual’s age at release has been emphasized in actuarial assessments of sexual reoffense risk. In the original Static-99 (Hanson & Thornton, 1999), the age item consisted of two age categories: individuals released prior to age 25 and those released after. Subsequent research found this dichotomization overestimated recidivism risk for individuals released after age 60. Age-related scoring criteria were revised for the Static-99R, substantially improving recidivism estimates associated with release after age 60, even after controlling for other static risk factors (Helmus et al., 2012).

Roughly, 85% of individuals released after their 60th birthday fell into the low risk Static-99R category in the Helmus et al. study. This is relevant because aging individuals from highly select samples, such as those considered for civil commitment as sexually violent persons (SVPs), may manifest higher risk for reoffending than individuals representative of routine normative samples. Potential differences between older individuals with more recent histories of sexual offending (due to release after serving their index offense sentences) and those released at older ages following lengthier periods of institutionalization (due to extended confinement under civil commitment laws) have not been examined. Protective effects of advanced age at the time of release may, or may not, apply to older individuals identified as SVPs warranting extended confinement under civil commitment statutes.

SVP Civil Commitment

In the 1990s, many states enacted modern civil commitment statutes to manage and treat individuals at a high risk to reoffend due to predisposing mental disorders (Meloy et al., 2007). Individuals committed under such statutes are referred to as Sexually Violent Persons/Predators. As of 2019, nearly 5,000 individuals were civilly committed as SVPs across 16 programs that responded to an annual survey of SVP programs (Schneider et al., 2019). Variations in structure and statutory criteria exist among programs, but civil commitment as an SVP typically requires three criteria to be met: (a) history of a qualifying sexual offense, (b) diagnosis of a mental disorder that predisposes the individual to offend sexually, and (c) a high likelihood of reoffending sexually. In Wisconsin, the statutory threshold is “more likely than not to reoffend” and is operationally defined as a reoffense risk greater than 50%. Multistage procedures are used to screen and manage individuals identified as possible SVPs. Initial processes cast a wide net, considering all individuals with qualifying offenses. Predisposing mental disorders and reoffense risks are evaluated at subsequent stages, narrowing the number identified as potentially meeting commitment criteria (Sandler & Freeman, 2017). Typically less than 3% of individuals initially screened are civilly committed (e.g., Boccaccini et al., 2009).

Use of systematic screening procedures and strict statutory criteria identify individuals for SVP commitment who are distinguishable from others with past histories of sexual offenses in notable ways. Consideration for SVP commitment is often based on factors like high actuarial risk scores and indicators of sexual deviance (Boccaccini et al., 2009; Freeman & Sandler, 2012; Levenson, 2004; Lu et al., 2015). Paraphilic disorders are prevalent in SVP populations (Elwood et al., 2010; Levenson, 2004) and may be indicative of entrenched traits that are less amenable to protective aging effects (Diagnostic and Statistical Manual of Mental Disorders [5th ed.; DSM-5]; American Psychiatric Association [APA], 2013; Fazio et al., 2017; Seto, 2012). Compared with other sexual offending samples, individuals committed under SVP statutes have high densities of dynamic risk factors (DRFs) that increase risk for sexual recidivism (Hanson & Thornton, 2012; Kelley, 2018; Olver et al., 2018). Age may be a proxy for factors that underlie risk reduction, such as physical health, libido, or self-control (Doren, 2006). For individuals with exceptionally high levels of DRFs, aging alone (i.e., without evidence of improvement in DRFs) may not translate to meaningful risk reduction. Finally, SVPs often spend lengthy periods civilly committed following their prison sentences, which may be a critical difference from most of the participants in samples used by Helmus et al. (2012). Much of the aging by SVPs will have taken place in institutions rather than in the community. Rice and Harris (2014) found increased odds of sexual recidivism associated with greater time spent aging in prison.

The indefinite nature of SVP civil commitment and relatively few releases from such programs has produced increasing numbers of individuals in older age ranges. Annual program surveys indicate the average client age in these programs has increased from the early 40s in 2008 to around 50 years old at present (Jackson et al., 2008; Schneider et al., 2019). This emphasizes the increasing importance of studying older individuals in SVP programs. Knowing whether they demonstrate typical age-related decline in risk has important implications. If aging SVPs show age-related decline in risk similar to other aging individuals with sexual offense histories, more efficient allocation of resources could be achieved while managing public safety concerns.

Sexual Recidivism Studies in Civil Commitment Settings

Recidivism studies from SVP programs are limited, perhaps due to the low numbers of individuals being released (Wilson et al., 2013). Several studies of individuals considered for SVP commitment report recidivism rates around 10% or less, suggesting this population may be lower risk and less exceptional than previously assumed (Boccaccini et al., 2009; DeClue & Rice, 2016; Duwe, 2014; Mercado et al., 2013; Sandler & Freeman, 2017; Wilson et al., 2013). Two studies found sexual recidivism rates above 20%. Prentky and Lee (2007) found that 25.2% of rapists and 31.9% of child molesters were charged with new sexual offenses, but the study period (1959–1985) was much earlier than other SVP recidivism studies. More recently, 23% of individuals released during early SVP screening stages had felony sexual offense convictions postrelease (Milloy, 2007).

Methodological differences, sample characteristics, and release eras may account for discrepant findings. Lower rates could relate to weaker methods, such as not subtracting time in custody from total follow-up time. Similarly, the nature of the release environment during the follow-up periods may differ across studies. In one study, about half of a sample screened for SVP commitment was released under mandatory supervision akin to parole (Boccaccini et al., 2009). After accounting for risk, individuals discharged without supervision had higher sexually violent recidivism rates than those released to mandatory supervision (odds ratio [OR] = 3.27). Other research has also found postrelease supervision substantially influences observed sexual recidivism rates (Minnesota Department of Corrections, 2007). These studies highlight the importance of postrelease environments; yet, most sexual recidivism studies limit focus simply to whether or not an individual reoffended.

Few SVP studies report recidivism rates specific to individuals released at older ages, such as those 60 years old or older (60+). Rapists released from civil commitment at an older age (n = 6, followed up between 1959 and 1985) had no detected sexual recidivism, whereas two of 12 older individuals identified as child molesters sexually recidivated (Prentky & Lee, 2007). Only 1% of those released aged 60+ in modern SVP cohorts sexually recidivated (zero of four in Milloy, 2007; one of 96 in DeClue & Rice, 2016). Sandler and Freeman (2017) also identified “advanced age” as a protective factor, even after controlling for Static-99R.

It should be noted that individuals in these various samples were released during stages of screening for potential civil commitment, not after a period of civil commitment. Such individuals would likely have spent less time aging in institutions. Civil commitment screening processes typically begin as individuals near their mandatory prison release date. In Wisconsin, persons who may meet criteria are referred to district attorney’s office and the Department of Justice prior to their release dates. Either agency can file a petition with the county circuit court to civilly commit the person as an SVP, prompting a probable cause hearing. If probable cause is found, a commitment trial is held (see https://docs.legis.wisconsin.gov/statutes/statutes/980 for details). Defendants found not to meet commitment criteria are dismissed. As the screening process is initiated prior to parole release dates, many dismissed individuals routinely return to the community under supervision. Individuals meeting SVP commitment criteria are civilly committed for control, care, and treatment by the Department of Health Services until they no longer meet criteria. Discharge processes are initiated by patients petitioning the state for release. If the person’s condition has changed sufficiently, a trial is held. The state has the burden to prove by clear and convincing evidence that the person continues to meet commitment criteria as an SVP. In contrast to those dismissed during screening, criminal sentences often expire for committed SVPs discharged after several years of commitment. As such, discharged individuals may have a lower likelihood of being released under parole supervision.

The Current Study

The primary aim of the study was to examine recidivism outcomes for individuals considered for civil commitment in Wisconsin and released at age 60 or older through either dismissal of an initial petition for civil commitment or discharge from civil commitment. Recidivism was examined using (a) simple descriptive rates, (b) survival analysis accounting for varying follow-up times, and (c) the Expected/Observed (E/O) index based on Static-99R recidivism norms. Findings have implications regarding risk reduction associated with advanced age in highly select SVP samples, particularly those released at older ages and potentially following lengthy periods of institutionalization. If the protective effects of aging do not apply to these samples, higher recidivism rates than those predicted by the Static-99R would be observed.

The second aim was to compare the dismissed and discharged groups. We hypothesized that older individuals with dismissed petitions would be more likely to be released to supervision. We also expected these individuals would have spent less time aging in secure institutions compared with SVPs discharged after a period of civil commitment.

Finally, features of the release environments, specifically the levels of administrative constraint following dismissal or discharge, were analyzed to understand their impact on observed recidivism rates. Such information frequently goes unreported in recidivism studies but could be methodologically important. The potential impact of release environments on recidivism risk should be an important consideration for forensic evaluators and courts tasked with integrating findings from this study, and others, into the larger body of recidivism research.

Method

Participants

The sample consists of individuals aged 60+ at the time of release after having been (a) found to meet probable cause for commitment under Wisconsin’s SVP statute, but the petition was dismissed (“dismissed”); or (b) civilly committed under Wisconsin’s SVP statue, but subsequently discharged after the court determined they no longer met commitment criteria (“discharged”). Consideration for SVP commitment required all participants to have been identified by a state evaluator as having a lifetime risk for sexually violent offenses that exceeded 50%. Between 1994 and 2016, 15 individuals 60 years old or older were dismissed and 38 were discharged. Individuals with release dates less than 3 months prior to study data collection (August 2016) were excluded. Age and race-related variables were coded from institutional records; race was dichotomized as White/non-White for analyses. Table 1 describes the characteristics of both samples individually and combined. Most individuals were White (84.6%), followed by Black (9.6%), Native American (3.8%), and Biracial (1.9%). Participants ranged from 60 to 80 years old at release (M = 67.86, SD = 4.99). The average follow-up time for the combined sample was 8.97 years (SD = 5.73).

Table 1:

Descriptive Characteristics for Dismissed and Discharged Samples

	Discharged (n = 38)			Dismissed (n = 15)			p	Effect size	Overall released (N = 53)
Characteristic	n	M (SD)	%	n	M (SD)	%	p	Effect size	n	M (SD)	%
Race/ethnicity	38			15					53
Non-White			13.2			26.7					17.0
White			86.8			73.3					83.0
Diagnoses^a	37			10					47
Pedophilic disorder			62.2			70.0					63.8
Sexual sadism disorder			5.4			—					4.3
Any paraphilic disorder			89.2			80.0					87.2
Major mental illness			8.1			20.0					10.6
Alcohol or other drug use			40.5			30.0					38.3
Any personality disorder			64.9			30.0					57.4
Consent to treatment	37		75.7	10		40.0			47		68.1
Released to supervision	38		18.4	15		80.0	<.001	0.58	53		35.8
Age at the time of release	38	68.60 (5.05)		15	66.00 (4.45)		.087	0.55	53	67.86 (4.99)
Age at the time of death	13	74.81 (3.85)		8	78.41 (8.02)		.266	0.57	21	76.18 (5.88)
Age earliest adult DOC entry	38	46.32 (8.10)		15	55.40 (8.16)		.001	1.15	53	48.89 (9.04)
Age last in community	38	51.34 (5.94)		15	58.33 (5.35)		<.001	1.24	53	53.32 (6.55)
Years institutionalized pre-SVP	38	9.34 (4.56)		15	7.13 (4.05)		.107	0.51	53	8.72 (4.50)
Total years institutionalized^b	38	19.60 (5.65)		15	9.01 (5.44)		<.001	1.91	53	16.61 (7.34)
Follow-up time (years)	38	6.64 (3.44)		15	14.86 (6.24)		<.001	1.63	53	8.97 (5.73)
Static-99R	38	2.61 (1.87)		12	1.92 (2.39)		.304	0.32	50	2.44 (2.00)
PCL-R total	35	19.49 (6.53)		6	16.13 (4.71)		.237	0.59	41	19.00 (6.36)

Note. Effect sizes are reported as Cohen’s d except for the released to supervision variable, which is reported as ϕ. DOC = Department of Corrections; SVP = sexually violent person; Static-99R (Helmus et al., 2012); PCL-R = Psychopathy Checklist–Revised (Hare, 2003); DSM-5 = Diagnostic and Statistical Manual of Mental Disorders (5th ed.; American Psychiatric Association, 2013).

Diagnoses that pre-dated DSM-5 were recoded to DSM-5 equivalents. ^bDoes not include juvenile custody.

Measures

Static-99R

The Static-99R (Helmus et al., 2012) is a 10-item actuarial measure that assesses items empirically related to sexual recidivism risk. Static-99R items include age, criminal history (e.g., prior sex offenses), and victim characteristics (stranger, male, or unrelated victims) producing a total score ranging from −3 to 12. These scores fall into relative risk categories of very low (−3 to −2), below average (−1 to 0), average (1 to 3), above average (4 to 5), or well above average (≥6) risk (Hanson et al., 2017). Static-99R scores were completed by licensed psychologists as part of forensic evaluations. Participants released prior to revision of the Static-99 (Hanson & Thornton, 1999) were rescored consistent with Static-99R scoring conventions. Static-99R scores were unavailable for three cases dismissed from other facilities. Static-99R scores (n = 50) ranged from −3 to 5 (M = 2.44, SD = 2.00).

The Hare Psychopathy Checklist–Revised (PCL-R)

The Hare PCL-R (Hare, 1991, 2003) consists of 20 items scored on a 3-point scale: 0 = doesn’t apply, 1 = applies somewhat, and 2 = definitely applies. Total scores range from 0 to 40 and are classified as very low (0–8), low (9–16), moderate (17–24), high (25–32), or very high (33–40) (Hare, 2003). PCL-R scores were available in the records for most participants (n = 41), scored by trained licensed psychologists. PCL-R total scores ranged from 10 to 33.3 (M = 19.00, SD = 6.36).

DSM-5 Diagnoses

Psychiatric diagnoses were coded based on the DSM-5 (APA, 2013) taken from the most recent risk assessments conducted by licensed forensic psychologists or medical records created by a staff psychiatrist prior to participants’ release. Diagnoses that pre-dated DSM-5 were recoded to DSM-5 equivalents. Diagnostic categories included were Pedophilic Disorder, Sexual Sadism Disorder, any paraphilic disorder, Major Mental Illness (i.e., Schizophrenia, Schizoaffective, Bipolar Disorder), any alcohol or drug use disorder, and any personality disorder. All categories were coded as 0 (not present) or 1 (present). Approximately 87% of the sample had some type of paraphilic disorder and 57.4% had a personality disorder diagnosis. Substance use and major mental illness diagnoses were less common (see Table 1).

Outcome Data Collection Procedure

Recidivism was defined as charges identified through publicly accessible circuit court records. Chargeable offenses logged in internal state-level corrections/registry tracking systems were also counted. For example, summaries available through corrections databases and interagency contact included details regarding violations that prompted revocation. The Wisconsin circuit court database was systematically reviewed for all cases. Chronological address histories were available from Wisconsin’s Sex Offender Registry Program (SORP) and used to guide case-specific follow-up of potential charges that might have occurred outside Wisconsin (e.g., name searches of publicly accessible correctional websites and circuit court records from other states where the individual was known to have resided). Recidivism events were coded as “sexual,” “violent,” “miscellaneous,” or “technical” (i.e., failure to register, tampering with a Global Positioning System [GPS] tracking device, or technical violations on probation/parole). Technical violations resulting in parole/probation revocation for individuals with time remaining on their criminal sentence, or who incurred additional parole/probation supervision due to postrelease charges, were identified through internal correctional databases.

Precise timelines of postrelease environments were constructed for each individual to account for time spent in custody or at varying levels of administrative constraint. GPS tracking was not included as a type of constraint because this information was not consistently available for all individuals. Detailed timelines were generated using the chronological address histories documented by SORP for all but one individual (who was not subject to Wisconsin’s SORP requirements). Time gaps of less than 35 days were presumed to have been spent at the registered address identified at the end of the gap. Time gaps between registered addresses of 35 days or more were coded as “location unknown” time. Each registered address was coded as one of the following: private residence, shelter, hotel/motel, trailer/recreational vehicle, assisted living center/nursing home, hospital, halfway house, jail/prison, or homeless. These living arrangements were collapsed into five categories based on the level of oversight and presumed degree of constraint limiting offense opportunities. These categories were as follows: (a) unrestricted, meaning the individual was living in private arrangements without supervision or oversight beyond those required by the registry or GPS monitoring; (b) mildly constrained, which included time living in private arrangements, but under probation/parole supervision; (c) moderately restricted, which included living in halfway houses, hospitals, or assisted living centers; (d) secure, which included time spent in jail, prison, or other secure inpatient facility; and (e) unknown. Any time spent in a secure setting was subtracted from the individual’s total time free. The follow-up period ran from each individual’s release date following dismissal or discharge to August 2016 (or date of death if prior to August 2016). Death records were gathered from SORP or publicly accessible sources such as obituaries. No direct contact with participants was made. This study was approved by the site’s Institutional Review Board.

Analyses

Analyses were conducted in SPSS (Version 25) and Microsoft Excel 2010. Descriptive recidivism outcomes were computed for the overall sample and for the dismissed and discharged groups separately. Outcomes were reported for any violation (any reoffense or technical violation, including revocations), as well as by each recidivism type. Life table (Singer & Willett, 2003) and Kaplan–Meier survival analyses were used to estimate absolute recidivism rates for both sexual offenses and any violations, taking into account differing follow-up periods. Due to the varying lengths of follow-up, only Kaplan–Meier estimates were used to calculate the E/O index (see Hanson, 2017). Independent-sample t tests compared means between the dismissed and discharged samples for numeric variables (e.g., Static-99R, time spent aging in institutions), and Pearson chi-square tests compared group frequencies of nominal variables.

Results

Comparing Dismissed With Discharged

As shown in Table 1, the dismissed group was significantly older than the discharged group, on average, when admitted to adult corrections, t(51) = 3.67, p = .001. The discharged group was significantly younger at the time they were last in the community than the dismissed group, t(51) = 3.96, p < .001. The dismissed group spent significantly fewer years institutionalized as adults than the discharged group, t(51) = −6.21, p < .001. The groups did not significantly differ on age at release, t(51) = 1.74, p = .087; years institutionalized prior to SVP referral, t(51) = −1.64, p = .107; Static-99R, t(48) = 1.04, p = .30; or PCL-R scores, t(39) = 1.20, p = .24. Effect sizes and p values are presented in Table 1.

Overall, 19 (37.3%) of the 53 participants were still under a criminal sentence at the time their SVP petition was dismissed or their commitment discharged, so they would have been released to parole supervision. However, there was a significant association between release type and index sentence status, χ²(1) = 17.73, p < .001. Time remained on the index sentences for 18.4% of the discharged group and 80.0% of the dismissed group. Follow-up time ranged from about 2.5 years to over 21 years (M = 14.86, SD = 6.24) for the dismissed sample, and from 3 months to over 13 years (M = 6.64, SD = 3.44) for the discharged sample.

Recidivism Outcomes

Descriptive Rates of Recidivism and Technical Violations

Twenty-four (45.3%) of the 53 participants had committed a violation of some type following their discharge or dismissal (any reoffense or technical violation including revocations). Four had been charged with a sexual offense, including one from the dismissed group and three from the discharged group. Two of the four sexual charges were for 4th Degree Sexual Assault, which is not considered a sexually violent offense under Wisconsin’s SVP statute. The third involved 1st Degree Child Sex Assault—Sexual Contact With Person Under the Age of 13. The charge in the last case was 1st Degree Criminal Sexual Conduct. Twenty individuals committed nonsexual offenses or technical violations, although two incidents did not result in any custody time. Registry, GPS tampering, and other technical violations were most common (n = 17); violent nonsexual offenses (n = 1) and other miscellaneous offenses were rare (n = 2; misdemeanor hit and run and disorderly conduct). The remaining 29 individuals had no known recidivism or technical violations and were either still living in the community (n = 13) or were deceased (n = 16) at the time of data collection. In total, 39.6% of the released sample was deceased at the time of data collection.

Survival Analysis Taking Follow-Up Time Into Account

Time at risk was used for time to event calculations (i.e., adjusted for date of death or time in secure settings). The total follow-up time of 8.97 years decreased to 6.76 when time deceased was subtracted. In addition, subtracting time spent in secure custody resulted in an average street time of 6.19 years (range = 0.08–18.99, SD = 4.65). All sexual recidivism occurred in the first 18 months of release. One participant reoffended in the first 6 months after release, two more recidivated during the next 6 months, and the last individual recidivated in the first 6 months of the second year. The cumulative proportion surviving at the end of the 18-month interval was 0.92. Life table analysis for any type of violation found the highest hazard rates during each of the first three 6-month follow-up intervals (.28, .39, and .18, respectively). Of the 24 participants with some type of violation, 18 happened within the first year and a half after release. The cumulative proportion of individuals surviving without a violation at the end of 3 years was 0.57 (the risk set going into this interval included 27 individuals).

Expected Versus Observed Recidivism

Static-99R scores were missing for three cases. Thus, 50 cases were available for calculation of the E/O index using recidivism estimates based on Routine samples. Two additional cases with Static-99R scores of −3 and −2 were excluded from calibration analyses based on High Risk/Need (HRN) samples, because recidivism norms associated with these scores are not available. Due to the varying length of follow-up time, Kaplan–Meier survival analysis was used to estimate the observed number of recidivists (see Hanson, 2017). Using a fixed follow-up would have further reduced the limited sample size and number of recidivists. The sample size and the number of recidivists are small nonetheless, so E/O calculations are preliminary and should be interpreted with caution.

For calibration based on Routine norms, the estimated proportion surviving at the last recidivism event closest (but prior) to 5 years is 0.914. The last recidivism event occurred at 1.17 years, with 41 cases remaining in the risk set. Therefore, the recidivism proportion is 1 – 0.914, or 8.6%. Multiplying the recidivism proportion by the sample size produces an estimated number of observed recidivists of 4.29. Based on the distribution of Static-99R scores in the sample, the 5-year logistic regression estimates for routine samples predict 3.89 recidivists. Thus, the E/O index based on Routine norms is 0.91, 95% confidence interval (CI) = [0.35, 2.31], indicating that these norms predict about 90% of the observed recidivism rate.

For calibration using HRN sample norms, the estimated proportion surviving at the latest recidivism event closest to, but not exceeding, 5 years is 0.910. The last recidivism event occurred at 1.17 years, with 39 cases remaining in the analysis. Thus, the recidivism proportion is 9.0%, with an estimated number of observed recidivists of 4.31 based on the sample size. HRN sample norms indicate an expected number of 6.59 recidivists. This produces an E/O index of 1.53, 95% CI = [0.60, 3.91], indicating the HRN norms predict roughly 50% more recidivists than observed. Both 95% CIs include 1, indicating no statistically significant difference between the observed and expected rates (p > .05).

Features of the Release Environment

The final aim of the study was to examine features of the release environment. The mean proportion of follow-up time spent under some level of constraint was .30 (Mdn = .11, SD = 0.35). On average, the dismissed group spent a greater proportion of follow-up time under constraint (M = 0.46, Mdn = .36, SD = 0.36) than the discharged group, M = 0.24, Mdn = .04, SD = 0.33; t(51) = −2.17, p = .04, d = 0.64. These are based on a wide range of total follow-up times. For example, two individuals may have each spent 50% of their total follow-up times under constraint, but one individual had only a single year of follow-up while another had 10 years.

Consequently, we aligned the release dates for all participants to show levels of constraint on a monthly basis for each participant’s first 3 years released. Frequencies for descriptive 3-year outcome and constraint patterns are presented in Table 2. Whereas 20 of 38 (52.6%) discharged participants had no constraint, this was the case for only one of 15 (6.7%) dismissed. Comparing 3-year outcomes for those with any constraint, seven of 18 (38.9%) discharged participants were successful as were seven of 14 (50.0%) of those dismissed. However, duration of the constraint varied considerably. Irrespective of constraint, technical or nonsexual violations were relatively common in both discharged (31.6%) and dismissed (40.0%) groups.

Table 2:

3-Year Administrative Constraint and Outcome Patterns

Degree of constraint and outcome	Discharged (n = 38)		Dismissed (n = 15)		Overall released (N = 53)
Degree of constraint and outcome	Frequency	%	Frequency	%	Frequency	%
3-Year Constraint × Outcome Patterns
No constraint	20		1		21
Success	16	80.0	1	100.0	17	81.0
Technical/nonsexual violations	1	5.0	0	0.0	1	4.8
Sexual reoffense	3	15.0	0	0.0	3	14.3
Any constraint	18		14		32
Success	7	38.9	7	50.0	14	43.8
Technical/nonsexual violations	11	61.1	6	42.9	17	53.1
Sexual reoffense	0	0.0	1	7.1	1	3.1
3-year descriptive outcomes
Success without constraint	16	42.1	1	6.7	17	32.1
Success under constraint	7	18.4	7	46.7	14	26.4
Technical/nonsexual violations with varying constraint	12	31.6	6	40.0	18	34.0
Sexual reoffense	3	7.9	1	6.7	4	7.5
Total	38	100.0	15	100.0	53	100.0

Note. Constraint refers to some level of administrative oversight or supervision (e.g., parole) or living arrangement presumed to limit offense opportunities (e.g., assisted living center). GPS tracking and Sex Offender Registry requirements in and of themselves are not considered as constraints here. Success refers to no known offense or technical violation resulting in revocations.

Details at the individual case level are presented in Figures 1 (Discharged) and 2 (Dismissed) to show constraint levels by month and the occurrence of significant events (e.g., technical violations, sexual offenses, and death). A month was coded as involving constraint if it applied at any time during the month; if multiple levels of constraint applied within a given month, the most restrictive constraint was coded. The cases are sorted to capture similar patterns of outcome and release environment as presented in Table 2.

Figure 1:

Three-Year Patterns of Constraint and Events for Discharged Cases

Figure 2:

Three-Year Patterns of Constraint and Events for Dismissed Cases

Figure 1 shows that the most common pattern for discharged individuals was “success without constraint” (42.1%) although a few had not accumulated 3 years of follow-up time (i.e., died within 3 years of release or released less than 3 years prior to follow-up date). Another 18.4% were offense free with varying amounts of time marked by mild constraint (e.g., under parole supervision) or moderate constraint (e.g., residing in assisted living facilities). About a third had technical violations or miscellaneous offenses; half of these cases spent a small proportion of release time under secure constraint (e.g., arrested for failure to register but spent only a few days of that month in jail), while others spent substantial time securely constrained. Finally, three participants committed sexual offenses, all of which occurred in the first 12 months of release. None of these three individuals were under constraint at the time of the sexual offense. The 4th degree sexual assault charges for Case 36 were dismissed.

Different patterns emerged for dismissed individuals. Figure 2 shows that during the first 3 years of release, 14 of the 15 dismissed cases had constraint at some point and most spent the majority of this time under constraint. The most common pattern was “success under constraint” (46.7%). A combination of technical violations (and one nonsexual violent offense) with varying constraint was also common (40.0%). One participant was mildly constrained until charged with a sexual offense and spent the remaining follow-up time under secure constraint. Thus, the constraints of release environment constitute a key difference between the discharged and dismissed groups.

Discussion

The current study examined recidivism of individuals considered for SVPs civil commitment in Wisconsin and released at age 60 or older through either (a) a dismissal of their petition for civil commitment or (b) a discharge of their civil commitment, and examined differences between the two groups. All participants were found by the courts to meet probable cause for civil commitment and at least one state psychologist opined they were more likely than not (i.e., >50%) to commit sexually violent offenses in the future. For the dismissed group, this opinion was expressed during the rejected commitment petition, near the time they were released back to the community. Those released through discharge were reexamined annually throughout the duration of their commitment as SVPs, during which time their risk may have declined.

Of the 53 individuals released at age 60 or older, 7.5% sexually recidivated. Two of the four sexual offenses would not constitute a sexually violent offense per Wisconsin statutes. Therefore, the observed sexual recidivism rate corresponding to statutory criteria relating specifically to sexually violent offenses was 3.8%. The sexual recidivism rate found in this study is similar to prior studies of individuals released during SVP civil commitment screening processes (DeClue & Rice, 2016; Duwe, 2014; Mercado et al., 2013; Sandler & Freeman, 2017). It was higher, however, than previously reported rates for SVP samples released at older ages (DeClue & Rice, 2016; Milloy, 2007). Both studies restricted their focus to dismissed individuals, who likely spent less time aging in institutions and may have been released to parole. There may also have been significant differences in Static-99R scores, street time, and duration of postrelease supervision between participants in the two prior studies and the current sample. In contrast, two of 18 older SVPs released in an earlier historical cohort (1959–1985) sexually recidivated (Prentky & Lee, 2007). However, rates of sexual offending have been declining over the last several decades (Finkelhor & Jones, 2012). Regardless, differences in sample characteristics and study methodologies add important context for understanding recidivism findings.

The discharged and dismissed groups did not differ in average age at release, but they were distinguishable on other age-related variables. Compared with those dismissed, the discharged group had an earlier entry into the adult corrections system, were younger at the time they were last in the community, and spent more time institutionalized as adults. These differences could reflect more severe or enduring offending patterns within the discharged group; however, by the time they were released, the groups did not differ significantly in levels of actuarial risk.

The time elapsed since these individuals were last in the community is remarkable, especially for those discharged who had aged an average of about 17 years since they were last in the community. In contrast, the change in age from index to release was about 5 years in another study looking at age-related variables and recidivism (Rice & Harris, 2014). Lengthy periods of institutionalization have important implications for risk management. Difficulties with establishing relationships and securing employment present obstacles to desistance from sexual offending by way of informal social controls (Harris, 2014). Non-age-related protective factors, such as social support or stable employment, may be more likely to erode over lengthy periods of institutionalization, or more difficult to establish upon return to the community. Extensive release planning efforts may be required for individuals following SVP commitment. As hypothesized, older individuals whose SVP commitment petitions were dismissed were more commonly released to parole supervision than those committed and subsequently discharged. This affords some transitional support and may have suppressed risk for sexual offending.

Notably, all of the sexual recidivism that occurred in the current study happened within 18 months of release. This is consistent with previous findings that risk for sexual recidivism is highest in the first few years after release (Hanson et al., 2014). Risk diminishes among those remaining in the community offense free, even for initially high-risk individuals. Following 2005 Wisconsin Act 431, virtually all SVPs in Wisconsin are subject to lifetime registration and GPS monitoring, and yet no sexual offenses were detected beyond a year and a half of their release. GPS was not captured as a type of constraint in the current study, but the timing of sexual offenses suggests the enduring nature of such restrictions may be unnecessary. Alternatively, one might argue the ongoing restrictions help deter would-be recidivists, but other research has identified inherent limitations with registration (e.g., Sandler et al., 2008). Results of the current study also suggest that for many older individuals, opportunities to reoffend may be constrained by their living arrangement. For example, older individuals with chronic health concerns residing in assisted care facilities may not require continuous GPS tracking. A more nuanced approach to employing these restrictions might take into account factors such as time offense free and the degree of constraint already imposed by an individual’s living arrangement. Additional research is needed to examine the timing of sexual recidivism for older individuals released from SVP programs.

The current study sample was small and the sexual recidivists too few to draw firm conclusions regarding the calibration of the Static-99R for older individuals released after being detained or committed under an SVP statute. The preliminary findings suggest the Static-99R Routine norms may underestimate risk by about 10%, whereas HRN norms may overestimate risk by about 50%. Observed recidivism rates were not significantly different than those expected, based on the 95% CIs. These results may simply reflect the relatively low statistical power of a follow-up study with only 53 individuals and four sexual recidivists. Notably, materially higher recidivism would have been expected based on the Static-99, which does not account for older age. Our findings do not indicate that SVPs released at older ages are riskier than would be expected based on the Static-99R, but larger studies are needed to confirm whether the typical protective effects of aging apply to these highly select populations. It is also possible that the low recidivism rates observed are due to the truly higher risk SVPs remaining committed. That is, if all of the older SVPs still committed were to be discharged, there might be a materially higher rate of observed sexual recidivism.

Future studies should compare older SVPs who are released with those who remain committed to examine potential differences in risk-relevant variables. Perhaps older individuals who remain under civil commitment have higher actuarial risk scores or striking behavioral indicators presumed to underlie protective aging effects, such as libido or self-control (e.g., Doren, 2006). Such observations might lead forensic evaluators and courts to question whether protective effects of aging apply for these individuals. A case study of an atypical individual who was sexually offending well into his 80s suggested that a plausible explanation for his unyielding risk was his exceptionally good health offsetting the typical protective effects associated with older age (Mattek & Hanson, 2018). In line with this case study, examination of health-related functioning in recidivism studies would be informative. An in-depth look at the few older individuals who were known to reoffend sexually would be helpful to see whether Mattek and Hanson’s postulated “vim and vigor” influence might account for these recidivism events.

Finally, this study intended to provide context to recidivism rates as they relate to relevant aspects of the release environment. The constraints of different release environments likely played a role in suppressing risk for many individuals in the current study. Limiting the focus to those who did not reoffend sexually, many spent substantial portions of their first 3 years (and total follow-up time more generally) under varying levels of administrative constraint, such as parole supervision, or residing in assisted living facilities. This was especially true for dismissed individuals; 10 of the 14 (71%) who were sexual offense free had more than half of their first 36 release months marked by constraint. This was the case for only nine of 35 (26%) discharged individuals who did not sexually recidivate. Three of the four sexual reoffenses were committed while the individual was free of any administrative or behavioral constraints. Recidivism rates observed in the current study should be interpreted in the light of the release environments. The impact of administrative and behavioral constraints warrants further investigation. Seto et al. (2017) treated supervision status as a time-varying covariate to capture the impact of the release environment throughout the follow-up time and found that being under supervision was associated with an over 50% reduction in the hazard of sexual recidivism even after controlling for risk factors (including the Static-99R). This highlights the marked effect of supervision on sexual recidivism.

Observations of release environments in the current study also have implications for interpreting earlier findings. Some studies acknowledge the potential impact of postrelease supervision (e.g., Boccaccini et al., 2009). Detailed information about levels of risk constraint covering entire follow-up periods is typically not available. This seems particularly important for SVP recidivism studies of individuals released from commitment consideration during the screening processes. Such individuals may often be released into environments where substantial portions of their early release years fall under supervision due to sanctions still in effect from their index offense. In the current study, 80% of the dismissed subsample was released to parole supervision. As such, recidivism studies of “near miss SVPs” may be less generalizable to SVPs discharged following lengthy periods of commitment. One SVP program reported revising their screening procedures to account for the protective effective of parole supervision (Freeman & Sandler, 2012). This change in practice contributed to a decline in referrals for civil management while allowing individuals to be referred once again for civil management when their parole supervision expired.

It is important to recognize that 39.6% of participants in this study died within the follow-up period. The average age at the time of death was 76 for the overall sample. The total follow-up time of about 9 years was reduced to 6.76 years when adjusted to subtract follow-up time accumulated after date of death. A history of offending is associated with increased mortality risks (e.g., Piquero et al., 2014). Forensic evaluators are often responsible for assessing lifetime risk and may extrapolate beyond 5- or 10-year recidivism estimates provided by current actuarial measures. While beyond the scope of this study, evaluators may want to consider incorporating findings comparing life expectancies with observed mortality rates of individuals within their jurisdictions.

Individuals released from prisons in general, and older individuals in particular, face substantial postrelease challenges, including housing, transportation, employment, and managing chronic health conditions (Williams et al., 2012). Individuals with a history of sexual offending face additional barriers associated with the social stigmatization (Booth, 2016). While not a primary focus of this current study, it is noteworthy that one quarter of the released sample in this study had been homeless at some point during their follow-up. Bolstering protective factors could be especially important to offset effects of lengthy periods of institutionalization. A clear argument has been made for the value of intermediate options, such as supervised release (Duwe, 2014; Montaldi, 2015). Supporting community reintegration through such programs would appeal to patients concerned about finding housing or navigating rapidly changing health care services in the community. Such programs should also appeal to policy makers and community stakeholders considering public safety implications and efficient management of public funds.

This study focused on individuals 60 years and older released from one state’s SVP program. Results may not generalize to older individuals released from other programs. The current study likely underestimates recidivism due to exclusion of offenses that were not formally charged or sanctioned, or not identified because they occurred outside Wisconsin. However, development of detailed postrelease timelines using chronological address histories available through the Sex Offender Registry should have minimized the extent to which serious offenses or custody events were missed. Neither registration requirements nor GPS tracking in and of themselves were coded as a type of constraint. The current study involved a relatively small sample and identified few sexual recidivists. This is a welcome finding in terms of community safety, but it also presents statistical challenges and limits conclusions that may be drawn, especially with regard to Static-99R calibration.

Larger samples and replication of results are needed. Methodological trade-offs are often made in recidivism studies seeking larger sample sizes. Larger samples allow the use of more sophisticated analyses and increased power, but may be achieved at the cost of less reliable or precise data based on easily accessible historical records and offense databases. Such records may not account for actual street time or varying levels of supervision and behavioral constraints, which are relevant for interpreting results. Collecting precise postrelease timelines is tedious. Extracting such data for large samples, if even possible, may not be feasible.

Conclusion

The current study examined recidivism and other outcomes among individuals 60 years of age or older released after consideration for civil commitment as SVPs. The results highlight the importance of distinguishing between individuals dismissed from consideration during SVP screening from those discharged following civil commitment. These groups differed on relevant age-related variables and release environments that may impact recidivism rates. Four of the 53 participants (7.5%) sexually reoffended over about 9 years of follow-up. Notably, only two of the four committed a statutorily defined sexually violent offense. Thus, the sexual recidivism rate relevant to the statutory focus concerning sexually violent offenses was actually 3.8%. Additional studies are needed to confirm whether protective effects of aging also apply to specialized SVP samples, and whether commonly used actuarial measures are calibrated for use in these settings. Future research should also explore whether the relatively low sexual recidivism rate observed in the current study reflects a successful selection process, whereby older SVPs who continue to be at high risk for sexual reoffending continue to remain civilly committed.

Footnotes

Authors’ Note:

The views expressed in this article are those of the authors and not necessarily those of the Wisconsin Department of Health Services.

ORCID iD

Gina Ambroziak

Gina Ambroziak, BS, is the quality improvement and research supervisor for the Wisconsin Department of Health Services at Sand Ridge Secure Treatment Center. Her research focuses on risk assessment, treatment, and recidivism of sexually violent persons.

Rachel E. Kahn, PhD, is a psychological associate for the Wisconsin Department of Health Services at Sand Ridge Secure Treatment Center and a licensed psychologist. Her most recent work has focused on adverse childhood experiences among those who sexually offend, biases in risk determinations for civil commitment evaluations, and using electroencephalogram (EEG) to help accurately identify treatment needs in sexual offending populations.

James C. Mundt, PhD, is a senior research analyst for the Wisconsin Department of Health Services at Sand Ridge Secure Treatment Center. His research focuses on the assessment, treatment, and prevention of sexual violence and recidivism among civilly committed sexually violent persons.

Kerry L. Keiser, BA, is a graduate of the University of Wisconsin–Madison with a degree in psychology. She was an intern research assistant with Sand Ridge Secure Treatment Center before becoming a project assistant at Innovenn, Inc.

David Thornton, PhD, is a psychologist in private practice in Wisconsin, USA, and holds a part-time position as a professor in the Department of Clinical Psychology at the University of Bergen in Norway. He has carried out research into the assessment of risk for sexual and violent recidivism.

References

American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). https://doi.org/10.1176/appi.books.9780890425596

Barbaree

H. E.

Blanchard

(2008). Sexual deviancy over the lifespan. In Laws

D. R.

O’Donohue

W. T.

(Eds.), Sexual deviance: Theory, assessment, and treatment (pp. 37–60). Guilford Press. https://www.guilford.com/books/Sexual-Deviance/Laws-O’Donohue/9781593856052

Boccaccini

M. T.

Murrie

D. C.

Caperton

J. D.

Hawes

S. W.

(2009). Field validity of the Static-99 and Mn-SOST-R among sex offenders evaluated for civil commitment as sexually violent predators. Psychology, Public Policy, and Law, 15(4), 278–314. https://doi.org/10.1037/a0017232

Booth

B. D.

(2016). Elderly sexual offenders. Current Psychiatry Reports, 18(4), 34–40. https://doi.org/10.1007/s11920-016-0678-1

DeClue

Rice

(2016). Florida’s released “sexually violent predators” are not “high risk.” Open Access Journal of Forensic Psychology, 8, 22–51. https://www.oajfp.com/blank-3

Doren

D. M.

(2006). What do we know about the effect of aging on recidivism risk for sexual offenders. Sexual Abuse, 18(2), 137–157. https://doi.org/10.1177/107906320601800203

Duwe

(2014). To what extent does civil commitment reduce sexual recidivism? Estimating the selective incapacitation effects in Minnesota. Journal of Criminal Justice, 42(2), 193–202. https://doi.org/10.1016/j.jcrimjus.2013.06.009

Elwood

R. W.

Doren

D. M.

Thornton

(2010). Diagnostic and risk profile of men detained under Wisconsin’s sexually violent person law. International Journal of Offender Therapy and Comparative Criminology, 54(2), 187–196. https://doi.org/10.1177/0306624X08327305

Farrington

D. P.

(1986). Age and crime. Crime and Justice, 7, 189–250. https://doi.org/10.1086/449114

10.

Fazio

R. L.

Dyshniku

Lykins

A. D.

Cantor

J. M.

(2017). Leg length versus torso length in pedophilia: Further evidence of atypical physical development early in life. Sexual Abuse, 29(5), 500–514. https://doi.org/10.1177/1079063215609936

11.

Finkelhor

Jones

(2012). Have sexual abuse and physical abuse declined since the 1990s? (Research Report CV267). http://www.unh.edu/ccrc/pdf/CV267_Have%20SA%20%20PA%20Decline_FACT%20SHEET_11-7-12.pdf

12.

Freeman

N. J.

Sandler

J. C.

(2012). Risk assessment and sex offender screening in New York. Sex Offender Law Report, 13(2), 17–32.

13.

Hanson

R. K.

(2002). Recidivism and age: Follow-up data from 4,673 sexual offenders. Journal of Interpersonal Violence, 17(10), 1046–1062. https://doi.org/10.1177/08862605-0201710-02

14.

Hanson

R. K.

(2017). Assessing the calibration of actuarial risk scales: A primer on the E/O index. Criminal Justice and Behavior, 44(1), 26–39. https://doi.org/10.1177/0093854816683956

15.

Hanson

R. K.

Babchishin

K. M.

Helmus

L. M.

Thornton

Phenix

(2017). Communicating the results of criterion referenced prediction measures: Risk categories for the Static-99R and Static-2002R sexual offender risk assessment tools. Psychological Assessment, 29(5), 582–597. http://doi.org/10.1037/pas0000371

16.

Hanson

R. K.

Bussiere

M. T.

(1998). Predicting relapse: A meta-analysis of sexual offender recidivism studies. Journal of Consulting and Clinical Psychology, 66(2), 348–362. https://doi.org/10.1037/0022-006X.66.2.348

17.

Hanson

R. K.

Harris

A. J.

Helmus

Thornton

(2014). High-risk sex offenders may not be high risk forever. Journal of Interpersonal Violence, 29(15), 2792–2813. https://doi.org/10.1177/0886260514526062

18.

Hanson

R. K.

Thornton

(1999). Static 99: Improving actuarial risk assessments for sex offenders. Public Safety Canada. https://www.publicsafety.gc.ca/cnt/rsrcs/pblctns/sttc-mprvng-actrl/sttc-mprvng-actrl-eng.pdf

19.

Hanson

R. K.

Thornton

(2012, October 17–20). Preselection effects can explain group differences in sexual recidivism base rates in Static-99R and Static-2002R validation studies [Paper presentation]. Association for the Treatment of Sexual Abusers 31st Annual Conference, Denver, CO, USA.

20.

Hare

R. D.

(1991). Manual for the Hare Psychopathy Checklist–Revised. Multi-Health Systems.

21.

Hare

R. D.

(2003). Manual for the Hare Psychopathy Checklist–Revised (2nd ed.). Multi-Health Systems.

22.

Harris

D. A.

(2014). Desistance from sexual offending: Findings from 21 life history narratives. Journal of Interpersonal Violence, 29(9), 1554–1578. https://doi.org/10.1177/0886260513511532

23.

Helmus

Thornton

Hanson

R. K.

Babchishin

K. M.

(2012). Improving the predictive accuracy of Static-99 and Static-2002 with older sex offenders: Revised age weights. Sexual Abuse, 24(1), 64–101. https://doi.org/10.1177/1079063211409951

24.

Jackson

R. L.

Travia

Schneider

(2008, October). Annual survey of sex offender civil commitment programs: 2008 [Paper presentation]. Sex Offender Civil Commitment Programs Network Annual Conference, Atlanta, GA, USA.

25.

Kelley

S. M.

(2018, October 17–20). Protective factors and treatment change in high stakes evaluations. In A. J. R.

Harris

(Chair), Dynamic Smack down. Dynamic Assessment—Promise or Pretext? [Symposium]. Association for the Treatment of Sexual Abusers 37th Annual Conference, Vancouver, British Columbia, Canada.

26.

Levenson

J. S.

(2004). Sexual predator civil commitment: A comparison of selected and released offenders. International Journal of Offender Therapy and Comparative Criminology, 48(6), 638–648. https://doi.org/10.1177/0306624X04265089

27.

Freeman

N. J.

Sandler

J. C.

(2015). Predictors of the sex offender civil commitment trial outcomes in New York State. Law and Human Behavior, 39(5), 514–524. https://doi.org/10.1037/lhb0000143

28.

Mattek

Hanson

R. K.

(2018). Committed as a violent sexual predator in his 10th decade: A case study. Archives of Sexual Behavior, 47(2), 543–550. https://doi.org/10.1007/s10508-017-1041-2

29.

Meloy

M. L.

Saleh

Wolff

(2007). Sex offender laws in America: Can panic-driven legislation ever create safer societies? Criminal Justice Studies, 20(4), 423–443. https://doi.org/10.1080/14786010701758211

30.

Mercado

C. C.

Jeglic

Markus

Hanson

R. K.

Levenson

(2013). Sex offender management, treatment, and civil commitment: An evidence based analysis aimed at reducing sexual violence (DOJ Document No. 243551). U.S. Department of Justice. https://www.ncjrs.gov/pdffiles1/nij/grants/243551.pdf

31.

Milloy

(2007). Six-year follow-up of 135 released sex offenders recommended for commitment under Washington’s sexually violent predator law, where no petition was filed (Document No. 07-06-1101). Washington State Institute for Public Policy. http://www.wsipp.wa.gov/ReportFile/985

32.

Minnesota Department of Corrections. (2007). Sex offender recidivism in Minnesota. https://mn.gov/doc/assets/04-07_Sex_Offender_Report-Recidivism_tcm1089-272768.pdf

33.

Montaldi

(2015). A study of the efficacy of the sexually violent predator act in Florida. William Mitchell Law Review, 41(3), 780–868. https://open.mitchellhamline.edu/wmlr/vol41/iss3/4/

34.

Olver

M. E.

Mundt

J. C.

Thornton

Beggs Christofferson

S. M.

Kingston

D. A.

Sowden

J. N.

Nicholaichuk

T. P.

Gordon

Wong

S. C. P.

(2018). Using the Violence Risk Scale-Sexual Offense version in sexual violence risk assessments: Updated risk categories and recidivism estimates from a multisite sample of treated sexual offenders. Psychological Assessment, 30(7), 941–955. http://doi.org/10.1037/pas0000538

35.

Piquero

A. R.

Farrington

D. P.

Shepherd

J. P.

Auty

(2014). Offending and early death in the Cambridge study in delinquent development. Justice Quarterly, 31(3), 445–472. https://doi.org/10.1080/07418825.2011.641027

36.

Prentky

R. A.

Lee

A. F. S.

(2007). Effect of age-at-release on long term sexual re-offense rates in civilly committed sexual offenders. Sexual Abuse, 19(1), 43–59. https://doi.org/10.1177/107906320701900105

37.

Rice

M. E.

Harris

G. T.

(2014). What does it mean when age is related to recidivism among sex offenders? Law and Human Behavior, 38(2), 151–161. https://doi.org/10.1037/lhb0000052

38.

Sandler

J. C.

Freeman

N. J.

(2017). Evaluation of New York state’s sex offender civil management assessment process recidivism outcomes. Criminology & Public Policy, 16(3), 1–24. https://doi.org/10.1111/1745-9133.12310

39.

Sandler

J. C.

Freeman

N. J.

Socia

K. M.

(2008). Does a watched pot boil? A time-series analysis of New York State’s sex offender registration and notification law. Psychology, Public Policy, and Law, 14(4), 284–302. http://doi.org/10.1037/a0013881

40.

Schneider

J. E.

Jackson

R. L.

Ambroziak

D’Orazio

Freeman

Hebert

(2019, November 4–5). Annual survey of sex offender civil commitment programs [Paper presentation]. Sex Offender Civil Commitment Programs Network Annual Conference, Atlanta, GA, USA.

41.

Seto

M. C.

(2012). Is pedophilia a sexual orientation? Archives of Sexual Behavior, 41(1), 231–236. https://doi.org/10.1007/s10508-011-9882-6

42.

Seto

M. C.

Sandler

J. C.

Freeman

N. J.

(2017). The Revised Screening Scale for Pedophilic Interests: Predictive and concurrent validity. Sexual Abuse, 29(7), 636–657. https://doi.org/10.1177/1079063215618375

43.

Singer

J. D.

Willett

J. B.

(2003). Applied longitudinal data analysis: Modeling change and event occurrence. Oxford University Press.

44.

Thornton

(2006). Age and sexual recidivism: A variable connection. Sexual Abuse, 18(2), 123–135. https://doi.org/10.1177/107906320601800202

45.

Williams

B. A.

Stern

M. F.

Mellow

Safer

Greifinger

R. B.

(2012). Aging in correctional custody: Setting a policy agenda for older prisoner health care. American Journal of Public Health, 102(8), 1475–1481. https://doi.org/10.2105/AJPH.2012.300704

46.

Wilson

R. J.

Looman

Abracen

Pake

D. R.

(2013). Comparing sexual offenders at the regional treatment center (Ontario) and the Florida civil commitment center. International Journal of Offender Therapy and Comparative Criminology, 57(3), 377–395. https://doi.org/10.1177/0306624X11434918