Skip to main content
NCBI home page
Psychiatry, Psychology, and Law logoLink to Psychiatry, Psychology, and Law
. 2020 Jun 23;28(2):310–323. doi: 10.1080/13218719.2020.1775154

Recidivism rates of sex offenders managed under the Dangerous Prisoners (Sexual Offenders) Act 2003: an evaluation of actuarial justice

MT Rowlands 1,✉, G Palk 1, R McD Young 1
PMCID: PMC8553259  PMID: 34720646

Abstract

According to the Dangerous Prisoners Sexual Offenders Act 2003 (DPSOA), an offender is considered ‘dangerous’ if there is an ‘unacceptable risk’ that he will commit ‘serious sexual harm’. Current legislation operates within an actuarial justice framework, whereby increasing resources are spent on those considered at greater risk. There is limited research on the efficacy of this approach. The current study examines sexual recidivism rates of a sample of DPSOA offenders. Court files of 104 community-supervised dangerous sex offenders (Mage = 50.7 SD = 10.8) were examined to determine date and type of re-offending. Recidivism was operationalised as time until arrest (for a sexual conviction/contravention). The overall level of sexual recidivism was low (7.69%). Kaplan–Meier analyses of survival curves identified no difference in rates between risk categories. While this likely suggests that they are not dangerous or an unacceptable risk, the strict conditions of supervision may be effective in preventing sexual re-offending. Further, limitations in empirically understanding the construct need to be considered.

Keywords: actuarial justice, dangerousness, Dangerous Prisoners (Sexual Offenders) Act 2003, dangerous sex offenders, DPSOA, recidivism, sex offender legislation, sex offending

The following paper is an analysis of sexual recidivism rates of offenders placed on Dangerous Prisoners (Sexual Offenders) Act 2003 (DPSOA) orders (Queensland, Australia). The physical and psychological, not to mention the financial and social, impact of sexual violence makes it one of the more egregious types of crimes. It could therefore be argued that sex offenders constitute a more dangerous type of offender. However, once a sex offender has completed their sentence is it certain that they remain dangerous to the community based on past sexual violence? In Australia, the Dangerous Prisoners (Sexual Offenders) Act 2003 (Queensland) (DPSOA) defines dangerousness as ‘[a] prisoner is a serious danger to the community as mentioned in subsection (1) if there is an unacceptable risk that the prisoner will commit a serious sexual offence – (a) if the prisoner is released from custody; or (b) if the prisoner is released from custody without a supervision. . . . On hearing the application, the [Supreme Court] may decide that it is satisfied as required under subsection (1) only if it is satisfied – (a) by acceptable, cogent evidence; and (b) to a high degree of probability; that the evidence is of sufficient weight to justify the decision’ (pp. 10–11). Like legislation in the UK and Germany, dangerousness is therefore equivalent to there being an unacceptable risk of further serious sexual offending.

A review of the Queensland Supreme Court DPSOA sentencing remarks by the authors noted a range of offender classifications from low to high risk. Here, ‘risk’ refers to the likelihood that the individual will engage in serious sexual violence at some point in the future, although it does not equate to a probabilistic rating (i.e. high is >50%; McSherry & Keyzer, 2009). The modern approach to determining individuals as dangerous is considered to utilise an actuarial justice model (McSherry, 2014; McSherry & Keyzer, 2009; Petrunik, Murphy, & Fedoroff, 2008; Scott & Resnick, 2006). This approach involves a judiciary decision that an offender is a low, moderate or high level of risk, with more systematic attention (i.e. intervention programmes, electronic monitoring, therapy) given to the higher risk categories (Hood, Shute, Feilzer, & Wilcox, 2002; Prentky, Janus, Barbaree, Schwartz, & Kafka, 2006). The level of risk, however, is not equivalent to that assessed by an actuarial risk scale (i.e. STATIC-99). The Supreme Court relies upon expert testimony from two psychiatrists, which will be a statement on an assessment of the level of risk (not necessarily the probability of), and recommendations for interventions (therapeutic and programmes). However, the level of risk will most be often be a judicial one, which determines whether there is ‘unacceptable risk’ of ‘serious sexual offending’. The court therefore considers the risk level as unacceptable if the probability of harm is beyond doubt in the absence of an order (McSherry, Keyzer, & Freiburg, 2006).

Importantly, ‘dangerousness’ is not an empirically tested construct (and remains a legal term) and, hence, should not be used in considering sexual recidivism (Nicholaichuk, Olver, Gu, & Takahashi, 2013). The classification is a judicial one, and hence normative (and moral) due to public and government needs (i.e. reflecting moral panics, Cipolla, 2011), which is not necessarily empirical (McSherry, 2014). In comparison, evaluations of risk (by clinicians) often rely on mathematical instruments (actuarial scales and structured clinical judgement) and clinical understanding of risk factors and interventions. While risk assessment measures are limited (sensitivity and specificity ratings, as well as issues with the instruments themselves – that is, early career offenders having lower static risk scores than those with more extensive histories; Vrieze & Grove, 2010), their outcomes (future offending) are tangible; risk (of a clinical understanding) can consequently be validated. Actuarial justice, or dangerousness, is more qualitative and simply denotes the level of concern (and, hence, the level of systematic resources utilised) held by the court once the individual is released into the community. Given this, is it justifiable to also include offenders classified as ‘low’ or ‘low–moderate’? If the actuarial justice model is valid, it is expected that offenders rated at a high risk are more likely to sexually re-offend sooner than low-risk groups.

Sexual recidivism

There is public concern that sex offenders (those with child or stranger victims) are highly likely to sexually re-offend (Hood et al., 2002; Shackley, Weiner, Day, & Willis, 2014; Walters, Knight, & Thornton, 2009). Meta-analyses identify low rates of re-offending overall, suggesting that this fear may be misplaced (Doren, 1998; Hanson & Bussière, 1998). Hanson and Bussière (1998) analysed 61 studies (28,972 sex offenders) and reported an average rate of sexual recidivism of 13.4% over an average follow-up period of 5.8 years. More recent meta-analyses (Hanson & Morton-Bourgon, 2005, 2009) have identified similar rates of re-offending, even with nearly twice the number of participants (45,398) in the 2009 analysis. It should be noted, however, that the meta-analyses sampled research with a wide range of follow-up periods, so the results should be considered cautiously. Generally, low sexual recidivism rates (between 5% and 15%) are reported for studies with survival periods of less than 15 years (Freeman, 2007; Greenberg, Firestone, Bradford, & Greenberg, 2002; Heil, Harrison, English, & Ahlmeyer, 2009; Marshall, Eccles, & Barbaree, 1991; Olver, Nicholaichuk, Gu, & Wong, 2013; Rettenberger, Boer, & Eher, 2011). When considering different types of sex offenders, rates vary even within the shorter follow-up periods: for example, Proulx et al. (1997) reported that 21.2% of rapists sexually recidivated within 5.4 years, as opposed to the 13% of child sex offenders. The low rates would suggest that the term ‘dangerousness’ is incorrect by the legal definition. However, the model of actuarial risk used in the courts is concerned with sex offenders not as a whole but an individual level (McSherry, 2014).

Sex offenders may be dangerous over longer periods of time; however, this is expected given that life changes, antisocial life trajectories and stressors are more likely to occur the longer the period of time that is observed (Hood et al., 2002). When increasing the length of the follow-up period, rates of re-offending increase. Langevin et al. (2004) reported that 61% of their sample had sexually recidivated within 25 years. Generally, research with follow-up periods between 15 and 25 years has re-offending rates double that of research with shorter follow-up times (Hill, Habermann, Klusmann, Berner, & Briken, 2008; Kingston, Firestone, Wexler, & Bradford, 2008; Prentky, Lee, Knight, & Cerce, 1997; Soothill & Gibbens, 1978). These studies mostly relate to offenders released into the community on orders (or not) distinct from DPSOA conditions, and hence they are not classified as dangerous offenders.

Dangerous sex offenders and recidivism

There is a dearth of articles on dangerous sex offenders. A review of the literature did not find articles on re-offending rates of Australian offenders. There is also limited research worldwide on this unique population of offenders. This is problematic, as research could validate the use of the legislation and evaluate the belief that the community is at risk. Among the studies on dangerous offenders, Neller and Petris (2013) used an estimation model to consider risk level of offenders likely to be classified as sexually violent predators (United States). From a population of 20,000 sex offenders, 740 persons were assumed to be committed under the sexually violent predator (SVP) law. Their estimate suggested a high rate of recidivism (65%) within 10 years. Müller, Haase, and Stolpmann (2013) conducted a study of a cohort of German offenders released prior to changes in legislation. The offenders were not managed under the legislation at release, but they would later have met the criteria for a dangerous offender. Nearly a third had re-offended within two years with a severe offence (sexual and non-sexual violence). Both studies have serious limitations, with the former being an estimation (and inaccurate) and the latter having a small sample (N = 25) and being retrospective in nature. However, both studies suggest that the samples were dangerous offenders. In contrast, using the Brief Actuarial Risk Scale (BARS), Nicholaichuk et al. (2013) matched a sample of Canadian sex offenders categorised as dangerous with a non-dangerous cohort of sex offenders. The two groups had similar scores on the BARS, suggesting that the offenders were unlikely to be dangerous (more at risk). Further, the sexual recidivism rate was similar to the base rate in Canadian sex offenders.

With the contradictory re-offending rates here, and the lack of research, serious ethical and legal questions remain for offenders released on strict supervision orders.

Summary and aims

As noted above, there is limited research on re-offending rates of dangerous or serious sexual offenders. It therefore remains to be determined whether high-risk or serious sex offenders are more dangerous than other populations of sex offenders who commit lower level sex offences (or classified as a lower level of risk). The current paper therefore examines recidivism rates of sex offenders placed on DPSOA orders and released into the community under strict supervision conditions. Hence, it is expected that offenders categorised at a high level of risk should sexually recidivate at a higher rate than lower risk (low, low–moderate and moderate risk level) offenders. The high-risk group is predicted (a) to have a higher proportion of sexual recidivists, and (b) to sexually re-offend sooner than lower risk groups.

Method

Participants

The current study sampled 104 DPSOA offenders (males = 104). The mean age at the time of release was 43.5 years (SD = 11.1); it was 50.7 years (SD = 10.8) at the time of the current study. The participants were all classified as dangerous sex offenders under the Dangerous Prisoners (Sexual Offenders) Act 2003. They were recruited from a search of public databases – the Queensland Supreme Court online (http://www.sclqld.org.au/caselaw) and the Australasian Legal Information Institute (http://www.austlii.edu.au) databases, using the search terms ‘dangerous sex offenders’, ‘Dangerous Prisoners (Sexual Offenders) Act 2003’, and ‘DPSOA’. The searches were restricted to Queensland and only those DPSOA offenders with complete documentation (transcript) at the hearing for dangerousness were included. Documents related to judge’s statement, summaries, sentencing and contraventions were then downloaded and reviewed for relevant information (see Measures).

As the databases were public, each offender was also searched by name, so that the authors could track subsequent court appearances/sentencing. Offender details, such as age, index offence, prior offending, psychiatric diagnoses, dates of sentencing/court appearances and substance history, were coded into a database. Offender names were then replaced with unique codes. While there were some missing entries (approximately 10%) relating to court hearings and sentencing, the authors were able to track most offender trajectories. Importantly, the authors did not include a matched sample of non-dangerous sex offenders. Given that there are significant differences between DPSOA and non-DPSOA offenders, in terms of post-release restrictions, we felt that it would have introduced a confound into the study.

The sample was then refined, as only those who had been released into the community were included (n = 128). As the average time from arrest until court hearing is 9.4 months, it is expected that some offenders may have re-offended, yet the event will be unlikely to be recorded in court by the time data collection was completed. Hence, offenders released within 9.4 months before December 2016 (31 March 2016) were excluded (n = 24) from analysis given that the average amount of time before an arrest was recorded in the courts was greater than this period. The study therefore only included 104 participants. The average time that offenders were on supervision was 9.25 years. The authors only considered recidivism to be any conviction or arrest post-release, and any offence prior to or during index offending was not considered.

Measures

Recidivism and survival period

There was no set length of time during which each offender was observed for re-offending. As offenders were released on different dates, the length of follow-up time varied; for instance, some offenders were released soon after the start of the DPSOA legislation (2004) and others closer to the observation cut-off date (March 2016). Therefore, the survival period varied between offenders, although it was not greater than 12 years (144 months). Time until arrest was determined by the difference, in months, between the date of release (or DPSOA hearing for release if the date was not specified, as these were most often the same) and the date of arrest. For survival analysis data, only the time (arrest date) until the first recidivism event was recorded, regardless of whether it was a contravention or conviction. The current study operationalises recidivism as re-arrests for sexual convictions and contraventions.

Sexual recidivism was defined as any conviction (determination of guilt) classified by the Penalties & Sentencing Act, 1992 (i.e. rape, sexual assault and maintaining a sexual relationship with a child). Contraventions (contravening a community-based order, as per the Act 1992) are less clear in the classifications of offending behaviour (there is no equivalent act). Hence, the researchers classified the contravention if it was sexual in nature (i.e. possessing sexually explicit material), related to the nature of the offender’s index offence (i.e. child sex-offender contacting, or attempting to contact, a child), or a technicality specific to the offender’s order (i.e. not declaring a relationship, or soliciting sexual activity with a sex worker).

Risk level

The offender’s level of risk was determined by the judge’s final statement on risk level (low, low–moderate, moderate, moderate–high and high) during the hearing for the determination of dangerousness. If a statement was not available, the most common level (or range) noted by the clinicians was utilised. To account for lowered risk level resulting in a decision to release from custody, only the risk level at time of release was coded, and prior rating (if the offender had been detained in custody) was ignored if this had changed. Actuarial scores were not used to determine overall risk level at release, as (a) clinicians use a range of measures, and it is not advisable to combine them into a single score (Hanson & Morton-Bourgon, 2009; Seto, 2005; Vrieze & Grove, 2010), and (b) clinicians and judges consider other factors (i.e. structured clinical judgements and intervention programmes) in assessing overall risk level.

Design

This study utilises a mixed-methods within- and between-subjects longitudinal design. It was not possible to create a control group due to the possibility of confounding variables, as offenders must meet certain criteria for placement on a DPSOA order; it was therefore not possible to have a matched, non-dangerous, sample of sex offenders. The data were analysed using a Kaplan–Meier test, examining a Weibull distribution of recidivism events (Maltz, 2001).

Procedure

Court hearing documents were obtained from the Queensland Supreme Court online database and the Australasian Legal Information Institute (see earlier). This involved reviewing dangerous sex offender hearings, coding each offender separately. Only hearings (and appeals) related to an offender being classified as a dangerous sex offender (DPSOA), trials for further offences and breaches/contravention hearings were used. The dates of release from custody on a supervision order and subsequent arrest were recorded, along with the nature of the re-offence (sexual conviction or breach/contravention).

Results

Demographics

Table 1 displays the demographics of the 104 offenders included in the study. Of the sample of offenders followed up for recidivism, all were males, 31.7% of whom were Indigenous Australians. For index offences, most were against children and young adults (50.0%), with the second largest being adult victims (42.3%), and then mixed adult, prepubescent and pubescent victims (5.77%). Extrafamilial child offending (49.0%) was the most common index offence, with rape (adult victim; 42.3%) second most common. The lowest index offending counts were intrafamilial child offending (5.77%) and possessing child pornography (0.96%; Table 1).

Table 1.

Demographics of DPSOA offender sample.

Characteristic DPSOA offenders
(n = 104)
n %
Male 104 100
Indigenous 33 31.7
Index offence    
  Rape (of adult) 44 42.3
  Sexual assault 3 2.88
  Intrafamilial child sex offender 6 5.77
  Extrafamilial child sex offender 51 49.0
  Child pornography 1 0.96
Risk level (non-actuarial)    
  Low 2 1.92
  Low–moderate 11 10.6
  Moderate 14 13.5
  Moderate–high 46 44.2
  High 31 29.8
Open in a new tab

Note: DPSOA = the Dangerous Prisoners Sexual Offenders Act 2003.

Inferential statistics

Sexual recidivism

The sexual recidivism count is shown in Table 2. Figure 1 displays the survival curve for first sexual re-offence for the sample (not by risk categories). The recidivism rate is much less than 50%, so quartile and median data were not listed. The sample continue to sexually re-offend over the 6-year follow-up period, although the total rate is low (7.69%).

Table 2.

Sexual recidivism count.

Category DPSOA offenders
(n = 104)
n %
Sexual recidivism: total arrests 8 7.69
First new sexual offence 4 3.85
First new sexual contravention 4 3.85
Sexual recidivism by risk category    
 Low–moderate 0 0a
 Moderate 1 6.67a
 Moderate–high 2 3.51a
 High 5 9.52a
Open in a new tab

Note: DPSOA = the Dangerous Prisoners Sexual Offenders Act 2003.

aThe percentage reflects the proportion of each risk category-group that have re-offended, not the proportion out of the total sample.

Figure 1.

Figure 1.

Open in a new tab

Count of offenders on the Dangerous Prisoners Sexual Offenders Act 2003 (DPSOA) not committing further sexual re-offences over 72 months.

Figure 2 displays the survival curves (sexual recidivism) for the risk categories – high, moderate–high, moderate and low–moderate – over 6 years. The low–moderate DPSOA offenders were the only group not to sexually re-offend during this time. Only the moderate, moderate–high and high-risk groups committed further offending. (For purely visual reasons, the vertical axis has been set to start at 90, as it would be otherwise difficult to see the differences.) A log rank (Mantel–Cox) test did not find a significant difference of survival times for the risk categories, χ2(3, N = 104) = 4.94, p = .18. The power of the test was too low because of the small number of events, to detect differences between survival times for the high, moderate–high and moderate risk categories. A visual inspection of Figure 2 notes a trend for the high-risk group to have fewer offenders surviving until 6 years, but this is not a significant result. Last, as the sexual re-offending rate is too low, it was not possible to generate median numbers for the quartiles; hence, no table displaying estimates and standard errors is displayed here.

Figure 2.

Figure 2.

Open in a new tab

Count of offenders on Dangerous Prisoners Sexual Offenders Act 2003 (DPSOA) order, by risk category, who have not committed a further sexual offence over 72 months.

Most re-offending (just over 50%) occurs within the nine-month mark for the high-risk category. This group also had the longest period of offending, as it lasted until the 69-month point (almost 6 years). Similarly, the moderate–high risk group had over 50% of sexual re-offending occur within a year; however, there were no further offences after this point. There was only one re-offence for the moderate group, and this occurred at the 54-month point (4.5 years).

Discussion

It is important to note that the rates of re-offending likely under-represents the true rate of offending. Given the limitations of the study, and that most offences are not reported by the victims or the offender, the numbers are possibly higher (Doren, 1998; Håkansson & Berglund, 2012; Heil et al., 2009; Hood et al., 2002; Maltz, 2001; Tollenaar & van der Heijden, 2013). In addition, how recidivism is operationalised (and the length of follow-up) will affect rates (Maltz, 2001); the current study also included contraventions. The numbers may therefore be inflated in comparison to those studies measuring only re-convictions, where the criteria are stricter. The authors therefore caution against interpreting the rate (7.69%) as evidence that the cohort are engaging in serious crime.

There were very few instances of sexual re-offending: over the 6-year follow-up period, only eight offenders had committed further sexual offences: four contraventions and four re-convictions. The lower count of re-convictions suggests that the sexual recidivism is less severe here, given that a more serious offence will likely result in a conviction, and less likely for a contravention (Zara & Farrington, 2016), and that the sample is therefore not as dangerous overall. Further, as the study also included contraventions that are, using the above reasoning, reflective of less serious offending, the reported rate of re-offending should be considered, in any case, indicative of the dangerousness of the cohort.

There are reasons to argue that the sample was not actually dangerous. First, sex offenders tend to have a low rate of sexual re-offending in general (Doren, 1998; Hanson & Morton-Bourgon, 2009; Hood et al., 2002; Nicholaichuk et al., 2014). Second, age has been reported to be a factor in desistance of re-offending, as offenders older than 30–40 years of age are less likely to engage in violent or sexual recidivism (Hanson & Bussière, 1998; Nicholaichuk et al., 2013; Olver et al., 2013; Zara & Farrington, 2016). The average age of the current study’s sample at release was 44.9 years, suggesting that due to cultural or maturational factors, the DPSOA offenders were highly unlikely to sexually recidivate.

Half of the sexual recidivism had occurred within 9 months, and none of the risk groups offended at a significantly higher rate than others. Considering Figure 2, the high-risk group appears to recidivate sooner and continues to sexually re-offend for a longer period than the other groups. Most of the offending occurs within this group (and none within the low–moderate group). However, the result is likely spurious due to the low count of sexual re-offences. Perhaps if the follow-up period was longer (14 years or more) or by increasing the geographical scope of the study (i.e. nation-wide), and hence the sample size, more sexual re-offences may be detected. However, differences in legislation and arrest protocols between Australian states may create analysis issues.

The overall rate of sexually violent re-offending is similar to those reported in other sex offender research. For studies with similar follow-up periods, such as Greenberg et al. (2002), Hanson and Morton-Bourgon (2009), Heil et al. (2009), Hood et al. (2002), and Rettenberger et al. (2011), the rate of sexual re-offending was between 5% and 15%. Although these studies use convictions, which will therefore have lower rates, the range of sexual re-offending falls between 4% and 11%. The current study is therefore reporting an inflated rate of sexual recidivism. Freeman (2007) study used re-arrests but reported a lower rate (5.5%), which was not surprising, as the survival period was 3 years.

Higher rates have been reported in other studies. Looman, Morphett, and Abracen (2013) reported a rate over twice as much as that of the current study (15.4%). In Hanson and Morton-Bourgon’s (2005) meta-analysis, with an average survival period of 5.8 years, they reported a rate almost twice that of the current study (13.7%; although the criteria for recidivism were broader). Note that both studies also used convictions as the criterion for re-offending.

The results of the current study provide an interesting comparison with results of other research on sex offenders classified as dangerous. First, the rate is much lower than what was reported by Neller and Petris (2013) and Müller et al. (2013). Both studies reported high rates of sexual re-offending (although with the caveat of their methodological limitations), indicating potentially dangerous offenders in their sample. Nicholaichuk et al. (2013) noted a rate of sexual re-offending similar to that in the current study. Of interest, their study had a different spread of risk profiles among the dangerous and pseudo-dangerous samples, with there being a greater proportion of lower level risk levels (according to the BARS) than in the current study’s sample. Given that Nicholaichuk et al. matched (via an actuarial scale) the sample to sex offenders not classified as dangerous, the study is better placed to make an evaluation of the suitability of dangerousness. Assessing persons by level of risk, at least purely using actuarial scales, indicates that the actuarial justice method is not really indicative of dangerous persons – that is, gauging an offender as having a high chance of future offending does not translate to being an ‘unacceptable risk’, at least in terms of an (judicial) arbitrary cut-off.

Perhaps it is difficult to argue that the sample represent a more dangerous cohort of sex offenders. They simply do not re-offend at a higher rate than has been noted in other studies. However, this should be interpreted with caution; it is difficult to compare samples as (a) the current study’s sample is much smaller, and (b) the supervision conditions on which the offenders are released are different from those that other populations are placed on. Regardless, it was not feasible to analyse this, given the difficulties in creating a control group that could be matched to remove confounding variables.

There are ethical issues with DPSOA orders. The low–moderate risk groups had the lowest level of re-offending compared to the other risk categories. While there were no significant differences between the groups for sexual violence, there was no sexual re-offending for the low–moderate risk group (and low levels for all groups). Consequently, the sample were simply not re-offending in a serious manner. It is therefore necessary to consider why these offenders were placed on the order (and strict conditions) when they are possibly not a risk. The decision-making process likely results in more Type I errors (Maltz, 2001; Petrunik et al., 2008). It is important to note that, within a consideration of the actuarial justice model, appropriate resources may have been spent to reduce risk; those classified as dangerous likely received sufficient treatment and restriction to reduce re-offending.

An important consideration is that the order may have weakened the predictive element of the actuarial justice model. By placing restrictions on the ‘dangerous offender’ due to there being an ‘unacceptable risk’ of future sexual offending, the expected reality of offending has not eventuated (Norko & Baranoski, 2008). This implies that the decision may be unwarranted, but given that the prediction of harm is difficult there may have been no circumstance where the court could have been certain (Murray & Thomson, 2010); in this regard, erring on the side of caution, in protecting the community by restricting the offender’s liberties, is a safe decision (and a utilitarian one; Norko & Baranoski, 2008). Importantly, decisions to enact the order also undermine the accuracy of risk assessments and the clinical assessment process (see earlier), given the low rates of re-offending of the high-risk group. However, this is an issue for all risk assessment tools regardless of actuarial justice or clinical models, as some level of response is required. Clinicians are also not expected to be prognosticians of risk, as the process of assessing offenders utilises a case formulation (idiopathic) model that identifies intervention targets (Rowlands, Palk, & Young, 2017). In this sense, identifying and preventing risk is within the actuarial justice model, and clinically and ethically appropriate.

As dangerousness is a legal and normative term, not an empirically validated construct, it is difficult to argue that a clinician’s decision is anything other than an understanding of risk and treatment. This is an important understanding that needs to be emphasised in the findings of the current study. Further, it is not possible to evaluate how much the judicial decision of dangerousness is based on forensic expert evidence or whether the level of ‘unacceptable risk’ is a probabilistic one. Given the limitations with long-term forecasts of risk, and that risk is idiopathic and dynamic, it is likely that the court is unable to determine dangerousness with any accuracy. Hence, the current study is a critique not so much of the construct of dangerousness as of the suitability of the legislation in framing ‘unacceptable risk’. Again, there may be ethical problems with placing offenders on an order given the limitation in determining the level of unacceptable risk (not to ignore the vagueness and lack of clarity of the term).

Another issue, the extra ‘time’ (or parole condition outside the length of sentence) that the offender has served on the order in response to that reality has violated the principle of proportionality in that the sentence does not reflect the offence (or reality of an offence; McSherry et al., 2006). In response, it could be argued that as the DPSOA order utilises a civil commitment model, whereby the conditions should be therapeutic in nature, the offender is not technically being punished. The current study only considers offenders released on supervision conditions, so that issues of severe restrictions being in custody are not addressed here.

Criticisms of the current study

The re-offending rates shown by the current study are likely to be conservative. It is believed that some of the files (hearings) were missing from the online database, and the researchers were therefore not able to record all re-offending. There were missing entries in the database, which meant that it is possible that not all reconvictions and re-arrests were available. The authors did not have access to the Queensland Corrective Services database, which tracks every offender. Hence, while it is unlikely that the authors would have uncovered further recidivism, it does mean that the current study is limited in its capacity to fully explore the trajectories of the cohort. This is not unique to the study, as missing records are a common limitation of recidivism research (Heil et al., 2009; Maltz, 2001; Tollenaar & van der Heijden, 2013). Further, the re-offending rates of sexual and non-sexual violent offending were very low, and this affected the power of the log rank analyses. This likely impacts the validity of the critique of dangerousness, as the study may have too many Type II errors; hence, the authors’ conclusions here have incorrectly assumed the null hypothesis as true. Perhaps measuring unreported criminal offending would increase this count, but this has serious ethical and legal issues. Further, there is no guarantee that the participants would disclose further offending, as more serious or dangerous offenders have been reported to be less likely to provide information to researchers (Delisi et al., 2016; Harsch, Bergk, Steinert, Keller, & Jockusch 2006). The study could be extended beyond 6 years, which may capture more re-offending, but most new offences occurred within 18 months for sexual re-offending, with the curves tending to plateau after. The follow-up time is too short, given that other studies, with longer follow-up periods, reported greater rates of sexual and non-sexual violent re-offending (Hill et al., 2008; Langevin et al., 2004; Nicholaichuk, Olver, Gu, & Wong, 2014; Soothill & Gibbens, 1978). However, the current study could only provide a short follow-up as the legislation was only enacted in 2003.

Another limitation is the smaller count of offenders categorised as low–moderate risk. Most offenders were at least classed a moderate–high risk. This is not surprising, given the actuarial justice model, as higher risk offenders are more likely to be placed on an order. However, the consequence of this is that the variance of follow-up time and recidivism is likely to be reduced, impacting on the study’s validity. Last, the authors did not consider non-dangerous sex offenders or general offenders. This may have provided information on whether, after comparing re-offending rates, the cohort are a more dangerous group. However, given that the supervision conditions would not be comparable to what non-DPSOA offenders experience when released, it would be problematic to compare the findings. It may have been that, lacking strict conditions, monitoring, and therapeutic and intervention supports, non-DPSOA offenders may re-offend at higher rates and sooner. But this would provide a useful analysis of the utility of DPSOA conditions, something that the current research is not able to provide.

Conclusion

The rates of sexual re-offending are within the same range of re-offending rates as that reported in other studies. The offenders in the current study therefore cannot be said to be engaging in a greater (or serious) rate of harm to the community. On rates of re-offending alone, the sample does not appear to be a more dangerous cohort of sex offenders as they simply do not commit further sexual re-offences at a higher rate than has been noted in research on dangerous offenders. In this sense, the legislation appears to be a valid means of containing harm, although it is difficult to assess the effects of supervision conditions from the population without using a matched control sample. However, it is difficult to argue that dangerousness is a valid construct, or relevant for clinicians in assessing and treating sex offenders; it remains a judicial (normative decision), and not easily translated into psychological practice.

Ethical standards

Declaration of conflicts of interest

Michael T. Rowlands has declared no conflicts of interest

Gavan Palk has declared no conflicts of interest

Ross McD. Young has declared no conflicts of interest

Ethical approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of Queensland University of Technology and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Informed consent

Informed consent was obtained from all individual participants included in the study

References

  1. Cipolla, C. (2011). “Preventative corrections”: Psychiatric representation and the classification of sexually violent predators. The Journal of Medical Humanities, 32(2), 103–113. doi: 10.1007/s10912-010-9134-0 [DOI] [PubMed] [Google Scholar]
  2. Dangerous Prisoners (Sexual Offending) Act. 2003. (QLD, Australia) . Retrieved from https://www.legislation.qld.gov.au/LEGISLTN/ACTS/2003/03AC040.pdf
  3. Delisi, M., Caropreso, D.E., Drury, A.J., Elbert, M.J., Evans, J.L., Heinrichs, T., & Tahja, K.M . (2016). The dark figure of sexual offending: New evidence from federal sex offenders. Journal of Criminal Psychology, 6(1), 3–15. [Google Scholar]
  4. Doren, D.M. (1998). Recidivism base rates, predictions of sex offender recidivism, and the ‘sexual predator’ commitment laws. Behavioral Sciences & the Law, 16(1), 97–114 [Google Scholar]
  5. Freeman, N.J. (2007). Predictors of rearrest for rapists and child molesters on probation. Criminal Justice and Behavior, 34(6), 752–768. doi: 10.1177/0093854806298280 [DOI] [Google Scholar]
  6. Greenberg, S.R.R., Firestone, P., Bradford, J.M., & Greenberg, D.M. (2002). Prediction of recidivism in exhibitionists: Psychological, phallometric, and offense factors. Sexual Abuse: A Journal of Research and Treatment, 14(4), 329–347. doi: 10.1023/A:1019921720366 [DOI] [PubMed] [Google Scholar]
  7. Håkansson, A., & Berglund, M. (2012). Risk factors for criminal recidivism – a prospective follow-up study in prisoners with substance abuse. BMC Psychiatry, 12(1), 111–111. doi: 10.1186/1471-244X-12-111 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Hanson, R.K., & Morton-Bourgon, K.E. (2005). The characteristics of persistent sexual offenders: A meta-analysis of recidivism studies. Journal of Consulting and Clinical Psychology, 73(6), 1154–1163. doi: 10.1037/0022-006X.73.6.1154 [DOI] [PubMed] [Google Scholar]
  9. Hanson, R.K., & Morton-Bourgon, K.E. (2009). The accuracy of recidivism risk assessments for sexual offenders: A meta-analysis of 118 prediction studies. Psychological Assessment, 21(1), 1–21. doi: 10.1037/a0014421 [DOI] [PubMed] [Google Scholar]
  10. Hanson, R., & Bussière, M. (1998). Predicting relapse: A meta-analysis of sexual offender recidivism studies. Journal of Consulting and Clinical Psychology, 66(2), 348–362. doi: 10.1037//0022-006X.66.2.348 [DOI] [PubMed] [Google Scholar]
  11. Harsch, S., Bergk, J.E., Steinert, T., Keller, F., & Jockusch, U. (2006). Prevalence of mental disorders among sexual offenders in forensic psychiatry and prison. International Journal of Law and Psychiatry, 29(5), 443–449. doi: 10.1016/j.ijlp.2005.11.001 [DOI] [PubMed] [Google Scholar]
  12. Heil, P., Harrison, L., English, K., & Ahlmeyer, S. (2009). Is prison sexual offending indicative of community risk? Criminal Justice and Behavior, 36(9), 892–908. doi: 10.1177/0093854809338989 [DOI] [Google Scholar]
  13. Hill, A., Habermann, N., Klusmann, D., Berner, W., & Briken, P. (2008). Criminal recidivism in sexual homicide perpetrators. International Journal of Offender Therapy and Comparative Criminology, 52(1), 5–20. doi: 10.1177/0306624X07307450 [DOI] [PubMed] [Google Scholar]
  14. Hood, R., Shute, S., Feilzer, M., & Wilcox, A. (2002). Sex offenders emerging from long-term imprisonment: A study of their long-term reconviction rates and of parole board members’ judgements of their risk. British Journal of Criminology, 42(2), 371–394. doi: 10.1093/bjc/42.2.371 [DOI] [Google Scholar]
  15. Kingston, D.A., Firestone, P., Wexler, A., & Bradford, J.M. (2008). Factors associated with recidivism among intrafamilial child molesters. Journal of Sexual Aggression, 14(1), 3–18. doi: 10.1080/13552600802074924 [DOI] [Google Scholar]
  16. Langevin, R., Curnoe, S., Fedoroff, P., Bennett, R., Langevin, M., Peever, C., … Sandhu, S. (2004). Lifetime sex offender recidivism: A 25-year follow-up study. Canadian Journal of Criminology and Criminal Justice, 46(5), 531–552. doi: 10.3138/cjccj.46.5.531 [DOI] [Google Scholar]
  17. Looman, J., Morphett, N.A.C., & Abracen, J. (2013). Does consideration of psychopathy and sexual deviance add to the predictive validity of the static-99R? International Journal of Offender Therapy and Comparative Criminology, 57(8), 939–965. doi: 10.1177/0306624X12444839 [DOI] [PubMed] [Google Scholar]
  18. Maltz, M.D. (2001). Recidivism. New York: Pergamon Press Inc. [Google Scholar]
  19. Marshall, W.L., Eccles, A., & Barbaree, H.E. (1991). The treatment of exhibitionists: A focus on sexual deviance versus cognitive and relationship features. Behaviour Research and Therapy, 29(2), 129–135. doi: 10.1016/0005-7967(91)90041-Z [DOI] [PubMed] [Google Scholar]
  20. McSherry, B. (2014). Managing fear: The law and ethics of preventive detention and risk assessment. New York, N.Y: Routledge. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. McSherry, B., & Keyzer, P. (2009). Sex offenders and preventative detention. Sydney, Nsw: The Federation Press. [Google Scholar]
  22. McSherry, B., Keyzer, P., & Freiburg, A. (2006). Preventative detention for ‘Dangerous’ offenders in Australia: A critical analysis and proposals for policy development. Report to the Criminology Research Council 2005. Retrieved from http://www.criminologyresearchcouncil.gov.au/reports/200405-03.pdf
  23. Müller, J.L., Haase, K.A., & Stolpmann, G. (2013). Recidivism and characteristics of highly dangerous offenders being released from retrospectively imposed preventive detention: An empirical study. Behavioral Sciences & the Law, 31(3), 359–380. doi: 10.1002/bsl.2069 [DOI] [PubMed] [Google Scholar]
  24. Murray, J., & Thomson, M.E. (2010). Clinical judgement in violence risk assessment. Europe’s Journal of Psychology, 6(1), 128. doi: 10.5964/ejop.v6i1.175 [DOI] [Google Scholar]
  25. Neller, D.J., & Petris, G. (2013). Sexually violent predators: Toward reasonable estimates of recidivism base rates. Behavioral Sciences & the Law, 31(4), 429–443. doi: 10.1002/bsl.2072 [DOI] [PubMed] [Google Scholar]
  26. Nicholaichuk, T.P., Olver, M.E., Gu, D., & Wong, S.C.P. (2014). Age, actuarial risk, and long-term recidivism in a national sample of sex offenders. Sexual Abuse: A Journal of Research and Treatment, 26(5), 406–428. doi: 10.1177/1079063213492340 [DOI] [PubMed] [Google Scholar]
  27. Nicholaichuk, T., Olver, M.E., Gu, D., & Takahashi, Y. (2013). Correctional careers of dangerous offenders. Criminal Law Quarterly, 59(4), 497–497. [Google Scholar]
  28. Norko, M.A., & Baranoski, M.V. (2008). The prediction of violence: Detection of dangerousness. Brief Treatment and Crisis Intervention, 8(1), 73–81. doi: 10.1093/brief-treatment/mhm025 [DOI] [Google Scholar]
  29. Olver, M.E., Nicholaichuk, T.P., Gu, D., & Wong, S.C.P. (2013). Sex offender treatment outcome, actuarial risk, and the aging sex offender in Canadian corrections: A long-term follow-up. Sexual Abuse: a Journal of Research and Treatment, 25(4), 396–422. doi: 10.1177/1079063212464399 [DOI] [PubMed] [Google Scholar]
  30. Penalties and Sentencing Act. 1992. (QLD, Australia). Retrieved from https://www.legislation.qld.gov.au/view/pdf/inforce/current/act-1992-048
  31. Petrunik, M., Murphy, L., & Fedoroff, J.P. (2008). American and Canadian approaches to sex offenders: A study of the politics of dangerousness. Federal Sentencing Reporter, 21(2), 111–123. doi: 10.1525/fsr.2008.21.2.111 [DOI] [Google Scholar]
  32. Prentky, R.A., Lee, A.F.S., Knight, R.A., & Cerce, D. (1997). Recidivism rates among child molesters and rapists: A methodological analysis. Law and Human Behavior, 21(6), 635–659. doi: 10.1023/A:1024860714738 [DOI] [PubMed] [Google Scholar]
  33. Prentky, R.A., Janus, E., Barbaree, H., Schwartz, B.K., & Kafka, M.P. (2006). Sexually violent predators in the courtroom: Science on trial. Psychology, Public Policy, & Law, 12(4), 357–393. doi: 10.1037/1076-8971.12.4.357 [DOI] [Google Scholar]
  34. Proulx, J., Pellerin, B., Paradis, Y., McKibben, A., Aubut, J., & Ouimet, M. (1997). Static and dynamic predictors of recidivism in sexual aggressors. Sexual Abuse: A Journal of Research and Treatment, 9(1), 7–27. doi: 10.1177/107906329700900102 [DOI] [Google Scholar]
  35. Rettenberger, M., Boer, D.P., & Eher, R. (2011). The predictive accuracy of risk factors in the Sexual Violence Risk–20 (SVR-20). Criminal Justice and Behavior, 38(10), 1009–1027. doi: 10.1177/0093854811416908 [DOI] [Google Scholar]
  36. Rowlands, M.T., Palk, G., & Young, R.M. (2017). Psychological and legal aspects of dangerous sex offenders: A review of the literature. Psychiatry, Psychology, and Law: An Interdisciplinary Journal of the Australian and New Zealand Association of Psychiatry, Psychology and Law, 24(6), 812–824. doi: 10.1080/13218719.2017.1315763 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Scott, C.L., & Resnick, P.J. (2006). Violence risk assessment in persons with mental illness. Aggression and Violent Behavior, 11(6), 598–611. doi: 10.1016/j.avb.2005.12.003 [DOI] [Google Scholar]
  38. Seto, M.C. (2005). Is more better? Combining actuarial risk scales to predict recidivism among adult sex offenders. Psychological Assessment, 17(2), 156–167. doi: 10.1037/1040-3590.17.2.156 [DOI] [PubMed] [Google Scholar]
  39. Shackley, M., Weiner, C., Day, A., & Willis, G. (2014). Assessment of public attitudes towards sex offenders in an Australian population. Psychology Crime & Law, 20(6), 553–572. doi: 10.1080/1068316X.2013.793772 [DOI] [Google Scholar]
  40. Soothill, K.L., & Gibbens, T.C.N. (1978). Recidivism of sexual offenders: A re-appraisal. The British Journal of Criminology, 18(3), 267–276. doi: 10.1093/oxfordjournals.bjc.a046912 [DOI] [Google Scholar]
  41. Tollenaar, N., & van der Heijden, P.G.M. (2013). Which method predicts recidivism best?: A comparison of statistical, machine learning and data mining predictive models: Which method predicts recidivism best? Journal of the Royal Statistical Society: Series A (Statistics in Society), 176(2), 565–584. doi: 10.1111/j.1467-985X.2012.01056.x [DOI] [Google Scholar]
  42. Vrieze, S.I., & Grove, W.M. (2010). Multidimensional assessment of criminal recidivism: Problems, pitfalls, and proposed solutions. Psychological Assessment, 22(2), 382–395. doi: 10.1037/a0019228 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Walters, G.D., Knight, R.A., & Thornton, D. (2009). The latent structure of sexual violence risk: A taxometric analysis of widely used sex offender actuarial risk measures. Criminal Justice and Behavior, 36(3), 290–306. doi: 10.1177/0093854808330341 [DOI] [Google Scholar]
  44. Zara, G., & Farrington, D.P. (2016). Criminal recidivism: Explanation, prediction and prevention. London: Routledge. [Google Scholar]

Articles from Psychiatry, Psychology, and Law are provided here courtesy of Taylor & Francis

RESOURCES