At a crossroads? Offender rehabilitation in Australian prisons

Abstract

This article presents a commentary on the current status of offender rehabilitation in Australian prisons. Prompted by recent debates concerning the value of rehabilitation programs in reducing rates of re-imprisonment, it considers the current evidence to support the effectiveness of Australian correctional programs highlighting the importance of the prison social climate, the intensity and quality of treatment and the need for psychological treatment to be integrated within broader models of rehabilitation and reintegration.

One of the presentations at the most recent International Corrections and Prisons Association (ICPA; https://icpa.org) conference captured the attention of delegates more than most. In this, Sir Martin Narey, the former head of Her Majesty’s Prison Service in England and Wales, talked about how he had been responsible for what was at the time a significant investment in the mass delivery of prisoner rehabilitation programs. This was based on the expectation that long-term benefits to the community would accrue as a result of lower rates of re-offending and diminishing rates of return to prison (Narey, 2019). However, the anticipated reductions in re-offending largely failed to materialize, and this, in part, led Narey to form the view that prisons, by their very nature, served mainly to demoralise and institutionalise rather than to rehabilitate. He pointed specifically to the social climate of prisons at the time, arguing that rehabilitation programs were often delivered in environments where there was what he described as a ‘shameful’ culture of contempt for the incarcerated and a tolerance for brutality that doomed them to failure. His conclusion was that this investment in the mass delivery of rehabilitation yielded poor returns, and the money would have been better directed to simply making prisons better, more decent and more moral places. Whilst this conclusion was viewed by some of the audience at the conference as overly pessimistic, it nonetheless reflects an emerging view that the ‘what works’ movement in offender rehabilitation (Andrews & Bonta, 2010) has largely failed to deliver on some of the ‘promises’ that were made and that new approaches to correctional policy and practice are now needed. The aim of this paper is to consider at least some of these arguments from an Australian correctional perspective in the context in which there has been an apparent lack of progress in reducing rates of re-incarceration for people who identify from Aboriginal and Torres Strait Islander cultural backgrounds.

Context

Whilst it is important from the outset to state that there is no suggestion here that the apparently shameful culture that Narey described in England and Wales is characteristic of contemporary Australian prisons, the broader argument that the prison rehabilitation cannot succeed on an industrial scale is an important one in an era that has seen the emergence and embedding of rehabilitation programs in prisons across the country (see Heseltine et al., 2011). These days correctional services in all Australian jurisdictions espouse rehabilitation as part of their core purpose, contrasting markedly with some of the correctional policy of the past. Historically, the earliest Australian prisons were undoubtedly brutal places, with the establishment of the British Penal Colony in New South Wales in 1788 seeing the introduction of a highly punitive prison system that was underpinned by the practice of deportation. And yet, even in the early days of colonisation, there was evidence of efforts to rehabilitate, drawing largely on the work of social reformers of eighteenth-century Britain. Notable Quakers, such as John Howard and Elizabeth Fry,¹ were among the first to try to change prisons from what were termed institutions of ‘deep despair and cruel punishment’ to places that were more humane and had the potential to reform. Their thinking informed some of the earliest attempts to rehabilitate in Australian institutions, most notably in 1840, when Captain Alexander Maconochie, as the prison governor of Norfolk Island, introduced the idea of indeterminate rather than fixed sentences, implemented a system of rewards in which prisoners could earn marks that counted towards their early release and advocated for a new system of aftercare and community resettlement. These ideas informed the development of rehabilitation and parole systems around the western world (see Morris, 2002) and have provided the legal foundation for the development of a much wider range of programs that have been designed to improve the chances of prisoners successfully desisting from crime following their release. In addition to programs that provide psychological treatment to address the causes of offending behaviour, there are also educational and/or vocational training programs and an array of transitional support (end of sentence pre-release or community reintegration) programs which attempt to respond to the difficulties that many prisoners experience after release from custody. Heseltine and Day (2017) have observed that the delivery of offending behaviour programs in Australian prisons has come a long way in recent years, notably in relation to the delivery of programs of greater intensity, a clear commitment to staff training, supervision and ongoing professional development and improved information exchange and sharing of resources. Nonetheless, with programs for moderate and high-risk sexual and violent offenders requiring professionally qualified staff to deliver small-group programs that often involve over 180 hours of face-to-face contact, they do represent a reasonably significant investment for correctional services in the context of limited existing budgets.

Responses

So how should Australian prison-based rehabilitation providers react to Narey’s comments at a time in our history when the cost of running prisons has been calculated to be in the region of three billion dollars per year, when prison numbers continue to rise and when the current rate of re-imprisonment is approaching (or greater than) 50% in all jurisdictions (Productivity Commission, 2019, see www.pc.gov.au)? It might be argued, for example, that these statistics show that our current system of rehabilitation is having, at best, a negligible impact on these key indicators of performance. It is also surprisingly hard to locate evidence that current offending behaviour programs are resulting in lowered rates of re-offending. In fact, we still know relatively little about the impact of programs on re-offending, with recent reviews of both sexual offender rehabilitation treatment (Polaschek & Blackwood, 2018) and violent offender rehabilitation (Papalia, Ogloff, & Daffern, 2019) failing to identify any studies conducted in Australian correctional settings that meet widely accepted criteria for inclusion. In fact, a range of methodological problems have been evident in most internal department evaluations (including those related to sampling and design procedures, the psychometric properties of outcome measures, the use of small or unrepresentative samples, an inadequate description of demographic information and the absence of a comparison group or random allocation), and the majority of these remain outside of the public domain.² And so, whilst program evaluation remains firmly on the agenda for correctional agencies across Australia, it is probably fair to say that there is a continuing reliance on overseas research to inform rehabilitation policy and practice.

Contemporary debates in the international research and policy literature about the effectiveness of offender rehabilitation can be traced back to the original ‘nothing works’ position of Martinson and colleagues in the early 1970s (Martinson, 1974) through to the ‘what works’ era of contemporary correctional practice (Andrews & Bonta, 2010). Whilst the international evidence that rehabilitation programs can bring about positive change in rates of re-offending has accumulated over the last three decades, it may nonetheless also be the case that the effects of program participation are simply not strong enough to impact on overall rates of return to prison. In fact, the most recent reviews conclude that even when the best-designed, -implemented and -evaluated programs are considered, the evidence of rehabilitation success is only modest. Polaschek and Blackwood (2018), for example, concluded that the current evidence for sex offender treatment shows that only 1 in 15 participants will not re-offend as a direct result of treatment, whereas Papalia et al. (2019) report a success rate of 1 in 10 for those receiving violent offender treatment. Such figures clearly suggest that there is scope to improve treatment outcomes, but perhaps not to the extent that programs might be expected to impact on overall rates of re-imprisonment. This is, of course, not particularly surprising, as rates of return to custody will inevitably be influenced by a range of factors that are simply beyond the control of correctional services and program providers. These might include, for example, changes in sentencing or policing practice which potentially lead to more people entering the criminal justice system, or perhaps to the introduction of ‘tighter’ bail, parole and post-sentencing laws which have the potential to increase the use of custodial sentences (eg McMahon, 2019). And yet, at the same time, policy responses to high-profile offences committed by ex-prisoners (some of whom were thought to have been successfully rehabilitated; see Callinan, 2013), have given impetus to new calls to develop longer-term and more intensive treatment regimens (eg Applegate & Strong, 2019).

The Prison Social Climate

The key argument that Narey (2019) put forward was, in essence, that even the best-designed and -delivered programs should not be expected to effect significant change on participants when they are provided in prisons where rehabilitative learnings are not reinforced outside of the programs room. As Schalast, Redies, Collins, Stacey, and Howells (2008) have argued, the extent to which the prison environment is perceived as supportive of therapy and therapeutic change is often in sharp contrast with the level of tension and perceived threat of aggression and violence that exists. Others, such as Wong and Gordon (2003), have further observed that institutional environments that do not support pro-social attitudes and behaviour and fail to substitute positive peer-group pressure for negative peer-group pressure are unlikely to be successful in rehabilitating offenders. Ware and Galouzis (2019), in writing about the rehabilitation of sex offenders in Australian prisons, have similarly suggested that developing a positive prison climate is needed to create the conditions necessary to counteract what they refer to as a hypermasculine and hostile environment that creates an atmosphere of fear and isolation and presents considerable challenges for therapeutic efforts that aim to address the antecedents and risk factors associated with sexual offending.³ Day, Casey, Vess, and Huisy (2012) have summarised this general view in the following way:

It is possible that features of institutional life, particularly in prisons, work against engagement in therapeutic programs in some cases and that the services are thus ‘unready’ [to deliver effective treatment]. The provision of therapy is, typically, not a primary goal for prison systems. Even where therapeutic goals are acknowledged as important, for example, in specialist therapeutic prisons, they are secondary to the custodial and deterrence functions of imprisonment (p.147).

Schalast et al. (2008) draw attention to three facets of correctional settings that define the rehabilitative potential: (a) perceived inmate safety; (b) support provided by correctional staff and inmates; and (c) the inmate’s perception of the climate as conducive to therapeutic change. In their view, then, a positive prison climate is one that is supportive, safe and maintains opportunities for prisoner personal growth and development. For Day and Vess (2017), it is the extent to which prisoners feel safe that has the potential to have the most profound impact on rehabilitation program outcomes. They suggest that, at the most basic level, some prisoners will decline to participate in treatment if they believe that this will place them at risk, whilst others will pursue accommodation changes (to different units or different prisons) when they are afraid, thereby disrupting treatment efforts. They go on to observe that the impact of threats to personal safety on the behaviour of those who actually attend treatment may be even more powerful, especially if this inhibits the ability to engage in a therapeutic process.

The general idea that the prison environment can, by its very nature, have a major impact on the behaviour of both prisoners and staff is best illustrated by the findings of a landmark study in the early 1970s that came to be widely known as the Stanford Prison Experiment (Zimbardo, 1972). A group of psychologically healthy male university students were randomly assigned to roles as either ‘guards’ or ‘prisoners’ in a mock prison environment set up in a university building. The experiment was designed to run for two weeks but had to be terminated after only six days because of the escalating levels of harassment and abuse from the ‘guards’ to the ‘prisoners’. This study was important, despite its methodological limitations, in so far as drew attention to the importance of situational factors rather than personality factors in creating a negative and damaging institutional climate; an argument also put forward to the investigation into abuses of detainees by US armed forces in Iraq, where some shocking abuses of prisoners occurred. Factors such as poor training, high levels of environmental stress, insufficient staffing, inadequate oversight, confused lines of authority, evolving and unclear policy and a generally poor quality of prison life were all identified as key features of the Abu Ghraib prison environment (see Schlesinger, Brown, Fowler, & Horner, 2004). In an important sense the broad observation that can be made here is that any attempt to rehabilitate in an institutional environment that is primarily harmful would appear misplaced.

There have been few attempts to formally assess the social climate of Australian prisons, although this is a topic that has interested the prison inspectorates that operate in some jurisdictions (eg Harding, 2014). The only published empirical research in this area used a self-report measure of social climate (the EssenCES survey; Schalast et al., 2008) which was administered to 253 people (109 staff members and 144 prisoners) at two correctional facilities in one jurisdiction (Day et al., 2012). A small, but statistically significant, positive association was reported between prisoner scores of the social climate and a measure of readiness to engage with offender programs. The authors of this study argued that the main value of this survey was to allow the repeated assessment to measure changes in climate over time, and they did not report data which allowed social climate ratings to be related to any metrics of rehabilitative success (beyond a self-reported readiness to participate in programs). Other assessment tools, including the Measure of Quality of Prison Life (MQPL; and the Staff Quality of Life, SQL; see Liebling & Arnold, 2004), are now also available to assess a broad range of factors that influence prison life and determine a prison’s climate, but the association between ratings on these measures and re-imprisonment rates have yet to be reported. The MQPL, for example, measures four key dimensions of prison climate: (i) Harmony (respect, staff-prisoner relationships, humanity); (ii) Professionalism (staff professionalism, bureaucratic legitimacy, fairness); (iii) Security (policing and security); and (iv) Well-Being and Development (addressing offending behaviour, preparation for release) (see Liebling, Hulley, & Crewe, 2012).

Despite the absence of local data to demonstrate the impact of Australian prison social climates on longer-term rehabilitation outcomes, interest in providing institutional environments that are more supportive of rehabilitative goals has been evident in the development of a number of specialist prisons and units that adopt some of the principles of the therapeutic community model of treatment. In different ways, each of these seeks to explicitly address institutional or environmental impediments to the delivery of rehabilitation, as well as aiming to directly address some of the interpersonal factors that potentially contribute to risk of re-offending. In brief, the therapeutic community model uses the prison ‘community’ to provide a range of life situations in which members (prisoners) can re-enact and re-experience their relationships in the outside world, with opportunities provided through group and individual therapy to examine and learn from any difficulties that arise. This is what De Leon (1997) has referred to as ‘community as method’ or the ‘purposive use of the peer community to facilitate social and psychological change in individuals’ (p. 5). There are currently four prisons/prison units in Australia that have attempted to implement some form of a therapeutic community model of practice, with the commissioning of a 300-bed therapeutic prison in Victoria the most ambitious of these (Morison & Craig, 2002). Specialist treatment units have since been developed to treat substance use (the Compulsory Drug Treatment Correctional Centre [CDTCC] in NSW, Birgden & Grant, 2010; and the Solaris unit in the ACT, Cox & Rosenberg, n.d.) and, more recently, for the post-sentence treatment of dangerous offenders in Victoria (Applegate & Strong, 2019). However, to date, there have not been any published reports of the outcomes of any these programs (although data from CDTCC appears promising), and the therapeutic integrity of the models that have been implemented is largely unknown. This is important, given Wexler’s (1997) identification of a number of challenges to the successful delivery of therapeutic communities in prison settings.

In summary then, it can be argued that there is a lack of evidence to demonstrate that the rehabilitation programs offered by Australian correctional services are having a positive impact on re-offending outcomes. Whilst an ‘absence of evidence’ should not be confused with ‘evidence of absence’, Narey’s proposition that prisons are simply environments where therapeutic work is doomed to fail should be considered at least plausible. There are, of course, other explanations, beyond the impact of the institutional environment, that are potentially important. Two of these, alluded to above, are considered next.

Rehabilitation programs are simply not powerful enough to produce systemic benefits

It may be that prison rehabilitation programs in Australia are simply not being delivered to sufficient numbers of prisoners to produce a measurable reduction in returns to custody. Although data on the numbers who successfully complete rehabilitation programs are, somewhat surprisingly, not readily available, most current rehabilitation programs are only available to those assessed as at ‘moderate’ or ‘high’ risk of re-offending, or to specific cohorts, such as convicted sexual offenders. A number of stand-alone briefer programs that potentially provide some form of intervention to greater numbers of offenders (or, at least, prepare them for more intensive rehabilitation) are available in some jurisdictions (Day & Collie, 2013), but it is difficult to assess how many people actually complete these. As a result, it is difficult to judge if this represents the ‘mass delivery’ of offender rehabilitation of the type that Narey spoke about, especially when one considers that the provision of rehabilitation is largely a matter of State and Territory policy, with no underpinning legislative mandate to provide access to programs.⁴

Another possibility is that current programs are simply not intensive enough or perhaps not delivered with sufficient therapeutic integrity to effect change. One recent meta-analytic review of the outcomes of sex offender treatment, for example, has reported evidence that program are most effective when they are delivered by professionally qualified staff (Gannon, Olver, Mallion, & James, 2019). Indeed, the ‘treat them longer, treat them better’ argument has been evident in a number of state governmental policy reviews, especially in relation to the management of high-risk violent offenders (the Victorian Government’s Review of the Serious Sex Offenders [Detention and Supervision] Act, 2009). The notion that longer treatment programs will result in better outcomes has, however, been questioned in two recent reviews; both of which expressed caution about prescribing minimal treatment dosage (eg 300 hours for high-risk offenders) and concluded that the focus of rehabilitation efforts should really be on measuring change rather than counting the time spent in programs (Day et al., 2019; Simourd & Olver, 2019). Thus, although there may be a prima facie argument that longer treatment will have a better chance of being effective, the current evidence to support such claims is, once again, surprisingly weak. In relation to the suggestion that the quality of current treatment delivery could improve, there is a small professional literature documenting some of the problems that practitioners face in forming effective therapeutic relationships with prisoners (eg Kozar & Day, 2009), as well as delivering high-quality cognitive-behavioural therapy (eg Marshall, Marshall, Serran, & O’Brien, 2011). There have, however, been few empirical examinations of differential outcomes achieved by ‘low’ and ‘high’ integrity programs and, as such, the extent to which investment in the further professionalisation of rehabilitation staff will yield benefit remains unclear. More generally, however, there is some acceptance of the need to incorporate a better understanding of both gender and culture into therapeutic practice, as well as genuine uncertainty about the quality of prison rehabilitation programs for participants from Aboriginal and/or Torres Strait Islander cultural backgrounds (Mals, Howells, Day, & Hall, 2000). Some of the constraints here are systemic, whilst others are therapeutic, with Heseltine, Sarre, and Day (2011) arguing that:

There are inherent difficulties in custodial environments in the development and implementation of programmes for Indigenous offenders. These include, but are not limited to, short custodial sentences, serving sentences away from family and community support networks, mental health and substance use co-morbidity, language barriers and low level of English literacy, educational difficulty, grief and loss issues, trans-generational trauma, kinship difficulty with group composition, and the lack of Indigenous facilitators (p. 48).

Other challenges to the delivery of high-integrity rehabilitation services in Australian correctional services come from the (at times) inconsistent and fragile political and public support that exists for program delivery, as well as the very real practical challenges inherent in efforts to deliver programs to those serving short sentences or who are incarcerated in rural and/or geographically remote areas.

The focus of current programs is too narrow

A second explanation of the apparent low impact of rehabilitation programs relates to the underlying logic of the rehabilitation model which is predicated on the notion of reducing levels of dynamic risk resulting in reduced rates of re-offending (Andrews & Bonta, 2010).⁵ There is emerging, but consistent, evidence that change in dynamic risk does occur in programs, but that this change does not translate directly into behavioural change following release, with one recent study conducted in NSW by Howard and van Doorn (2018) reporting that measures of within-treatment change were not significantly associated with subsequent re-offending. In other words, although a reduction in dynamic risk was evident following completion of the program, this was not directly associated with any change in offending. A similar conclusion was drawn from another Australian study, conducted in Victoria, by Klepfisz, O’Brien, and Daffern (2014) who reported that raw change, reliable change and clinically significant change for scores on measures of three dynamic risk factors did not predict either re-offending or the time taken to re-offend (violently) in a sample (n = 42) of violent offenders. They reported that only one participant achieved reliable change on all three of the risk factors assessed, and only four achieved reliable change on at least two of these. Along with similar findings from a growing body of international research (eg Barnett, Wakeling, Mandeville-Norden, & Rakestrow, 2013; Beech & Ford, 2006; Olver, Kingston, Nicholaichuk, & Wong, 2014), the conclusion that has to be drawn here is that change scores on measures of dynamic risk are, at best, only weakly and inconsistently related to re-offending (see Heffernan, Wegerhoff, & Ward, 2019 for a review). Although there are a number of possible reasons for this (eg treatment is simply not targeting the right variables or current measures of change lack validity), this challenges the idea that we can successfully identify a set of relevant dynamic risk factors, deliver interventions in custody that address these and then expect that a decrease in offending will happen. Thus, even when rehabilitation programs are successful in reducing assessed levels of dynamic risk, this is still insufficient to have a substantial impact on re-offending.

The implication here might be that prison rehabilitation should be considered as something broader than the treatment provided through offending behaviour change programs. In the spirit of the Mandela Rules (ie that a period of imprisonment should be used to ensure the reintegration of prisoners into society so that they can lead a law-abiding and self-supporting life), it seems useful to view psychological treatment as a necessary, but often not sufficient, condition for behavioural change to occur; particularly in light of evidence that other forms of intervention are rehabilitative. For example, Cale et al. (2019) have reported data from Australian correctional services across four states relating to a total of 10,834 Australian prisoners released from custody from 2010 to 2011. Their analysis showed that participating in prison vocational education and training made a significant contribution to the likelihood of remaining custody free at two- and five-years post-release for both male and female prisoners, with the relationship moderated by risk level for male prisoners. This study highlights the importance of thinking about rehabilitation as involving the combined effects of a range of other programs, including the throughcare programs that are routinely available across Australian correctional settings. The underlying logic here would draw on a broader psychosocial approach of the type described in theories of desistance, rather than solely relying on reducing dynamic risk at the individual level. This would conceptualise the process of change to an offending-free lifestyle in terms of the development of human capital, social capital and psychological capital. Whereas human capital is largely concerned with problem awareness, education and skills about offending – of the type that is discussed in current offending behaviour programs – social capital largely concerns the post-release space (although the importance of maintaining relationships whilst in prison should not be under-estimated). Psychological capital invokes the idea of identity and how important this is to an offence-free lifestyle (Maruna, 2001). McNeill (2006) has also described the importance of providing opportunities for offenders to see themselves in ways that are less compatible with a criminal lifestyle and increase a sense of belonging to a community that also offers acceptance.

The suggestion here then is that rehabilitation programs, as currently designed and structured, should not be expected to have the impact on overall rates of re-offending that some expect, and that more integrated approaches that include socially focussed interventions should also play a key role in the successful delivery of rehabilitation. In this way it might be possible to articulate both a rehabilitation theory and program logic that could then be applied to larger numbers of people in prison and result in some of the more systemic changes that rehabilitation providers aim to achieve.

Conclusion

This article, prompted by a recent call for prison administrations to largely abandon rehabilitation efforts in favour of improving basic prison conditions and culture, identifies a number of challenges facing the model of rehabilitation currently favoured across Australian jurisdictions. It is suggested that rehabilitation programs – even the best-developed and highest-integrity ones – are probably not quite as powerful as was once hoped and, even in the absence of Australian evidence about their outcomes, there is probably only limited reason to expect them to produce substantive reductions in the rate of re-imprisonment. Australian rehabilitation programs can, it is suggested, be reasonably expected to make a significant difference to perhaps 1 participant in a group of 12 prisoners who have committed serious violence offences. That is not to suggest that current programs should be abandoned or that they are economically unviable (see program cost-benefit analyses by the Washington State Institute for Public Policy, https://www.wsipp.wa.gov), but rather that the ‘mass delivery’ of rehabilitation through offending behaviour programming is unlikely to provide a public policy solution to the problems created by a growing criminal justice system.

In conclusion, it seems reasonable to argue, as Narey (2019) did, that it is the social climate and culture of prisons that can render attempts to rehabilitate futile, but that there are other aspects to rehabilitation service provision that also need to be considered. These include the scope and breadth of those programs that are considered to be rehabilitative and attending to the quality and integrity of program delivery. Thus, whilst improving the material and physical environments in which prisoners are housed is clearly important, we can aspire to do more than simply provide safe and humane care. The crossroads that we currently face relates to the decision to either invest more heavily in the existing programs model or think about how best to integrate more socially focused interventions that strengthen connection to family, community and culture. In his talk, Narey compared the success of prison rehabilitation to the futility of watering a few flowers in a desert, whereas what is really needed to reduce prison populations is an intensive and sustainable approach to farming.

Ethical standards

Declaration of conflicts of interest

Andrew Day has declared no conflicts of interest.

Ethical approval

This article does not contain any studies with human participants or animals performed by any of the authors.