To identity relevant literature for this report I searched three criminology databases (AGIS, CINCH and Criminal Justice Abstracts), two criminology journals (British Journal of Criminology and the Australian and New Zealand Journal of Criminology), two social work journals (Australian Social Work and The British Journal of Social Work) and one general database (ProQuest) using a combination of the following key terms: recidivism, meta-analysis, criminology, community, systematic review and social work. I also manually searched relevant articles and abstracts to identify appropriate material.
The search focused on studies that involved adults and young people, but excluded results that examined particular subgroups of offenders, such as sex offenders, racial groups or those suffering from severe mental illness. It also excluded studies that focused on specialist approaches such as drug treatment, restorative justice and mental health programmes. While it did include international studies, it excluded studies published prior to 1990. This ensured that a range of modern studies were included in the analysis.
Of the studies initially identified only those published in peer-reviewed journals were chosen for this paper. This step was taken to ensure that the quality of the studies met with the highest scientific standards (Hans-Dieter 1993). All of the studies employ recidivism as the primary measure of effectiveness and focus predominantly on community rehabilitation programs rather than those delivered in a residential setting. This is in keeping with the scope of the research question. The Literature Five studies met all of the criteria above and have been included in this paper. These studies are summarised below.
In their 1990 study, Andrews, Zinger and Hoge predicted that appropriate treatment – defined as ‘treatment that is delivered to higher risk cases, that targets criminogenic need, and that is matched with the learning styles of offenders’ – would reduce recidivism (Andrews et al. 1990, p. 377). To investigate this hypothesis they subjected 45 studies included in the Whitehead and Lab (1989) review to content and meta-analysis along with 35 studies from their own research files. Together, these studies ranged in date from the 1950s through to 1989 and included data on both juvenile and adult offenders.
The results supported their hypothesis and also revealed that community-based treatments led to higher reductions in recidivism than those delivered within a residential setting. In their 2006 study, Lowenkamp, Latessa and Holsinger examined the relationship between adherence to the risk principle and recidivism rates. To study the link between these two factors, the authors analysed data from two separate studies that provided information about 97 community-based rehabilitation programs that delivered services to 13,676 American offenders.
Instead of focusing on the different types of treatment delivered to lower-risk offenders compared to higher-risk offenders, Lowenkamp et al. studied the number and duration of services. They found a consistent pattern: if higher-risk offenders were targeted and provided with more services for a longer period of time than those in the lower-risk category their recidivism rates fell. They also found that cognitive behavioural based programs were more effective at reducing recidivism than all other treatment modalities.
Lipsey and Cullen (2007) studied over twenty meta-analyses, which between them had analysed hundreds of correctional intervention studies, to compare the effects of deterrence and rehabilitation programs on the recidivism rates of offenders throughout America. In their research they identified two consistent findings: programs based on a deterrence model resulted in either a small decrease in recidivism rates or, more commonly, an increase in recidivism rates while those programs founded on a rehabilitation model consistently resulted in reduced levels of recidivism.
The difference between these two models was so significant that the smallest mean reduction in the rehabilitation group was greater than the largest mean reduction in the deterrence group. They were, however, unable to draw any consistent conclusions about what type of rehabilitation program was the most effective. Instead, they hypothesised that factors other than the treatment type were influencing the effectiveness of these interventions and suggested that further research needs to be conducted to identify what these factors are.
In her 2011 Joan McCord Prize Lecture, (the transcript of which was published in the Journal of Experimental Criminology (2012)), MacKenzie reviewed current research findings about how correctional policies and programs impact on the later criminal activities of offenders. She studied a range of systematic reviews, meta-analyses and Maryland method scores to draw conclusions about the effectiveness of several types of programs, policies and management strategies. McCord’s work was published in her 2006 book, What Works in Corrections, and a summary of her findings were provided in this paper.
Her research led to three main findings. Firstly, the most effective programs at reducing recidivism addressed the thought-processes of prisoners such as cognitive-behavioural therapy. Secondly, she found that interventions which addressed environmental opportunities such as housing or employment services, had no significant impact on recidivism rates. And, thirdly, she found that methods of deterrence, incapacitation and control increased recidivism rates. Unfortunately, the article did not elaborate on the number or types of studies examined, so it is hard to know how thorough her investigations were.
In their 2015 study, Chadwick, Dewolf and Serin studied the effect that officer training has on the recidivism rates of offenders in several different countries, including Australia, Canada, The United Kingdom and the United States. A meta-analysis of 10 studies that included a total of 8,335 male and female offenders revealed that offenders supervised by officers who received training aligned with the Core Correctional Practices (CCP) outlined by Dowden & Andrews (2004) had lower rates of recidivism than those supervised by untrained officers.
Considering the relatively small collection of studies chosen for this meta-analysis, these findings provide a preliminary overview of the effectiveness of officer training and further study is needed to gain more accurate results. What Works The above five studies identified several key factors that are consistently associated with lower recidivism rates. These factors are discussed below. The Risk-Need-Responsivity Model Several of the studies refer to the effectiveness of the risk-need-responsivity model in reducing recidivism. The risk-need-responsivity (RNR) model is based on three main principles (Patterson 2012).
The risk principle states that the most intensive treatment should be reserved for those offenders who are at the highest risk of reoffending (Lowenkamp et al. , p. 88); the need principle states that practitioners should focus on changing criminogenic needs (such as drug dependency, antisocial attitudes and lack of self-control) to reduce future criminal behaviours (Andrews et al. 1990); and the responsivity principle states that treatment should be matched with the learning styles and characteristics of the offender treated (Lipsey & Cullen 2007, p. 310). Andrews et al. 990) found that when all of these principles were adhered to recidivism rates fell by a mean of 53. 06%.
In contrast, Lowenkamp et al. (2006) found that if the risk principle was not adhered to, that is, if lower risk offenders were targeted with intensive treatment, their recidivism rates rose. While Lipsey and Cullen (2007) did not specifically study the effects that the RNR model has on recidivism rates, they did hypothesise that the variability found in their analysis of treatment types could be caused by inconsistencies in applying the RNR model to offender treatment.
These findings seem to indicate that when treatment is tailored to the individual needs, behaviour and personality of offenders, their recidivism rates will fall at a higher rate than those assigned to a generic treatment program. Other studies outside of this report (for example, Luong & Wormith (2011), Wilson & Zozula (2012) and Polaschek et al. (2005)) concur with this finding. Cognitive Based Treatment Within the rehabilitation model there are several different types of treatment available to offenders.
These include family therapy, life skills training, mediation, group counselling and employment programs (Lipsey & Cullen 2007). Yet despite the wide variety of programs available, only one was deemed to be the most effective in several of the studies chosen for this report. Cognitive based therapy is a type of psychotherapy that helps to change the way a person thinks, feels and behaves (Kinderman 2014). In the field of corrections, this therapy aims to address the thought patterns and attitudes that are specific to anti-social behaviour (Lowenkamp et al. 2006).
By undertaking this type of treatment, offenders are able to see the link between their thoughts and actions and are then taught the skills to make better choices in the future (Andres et al. 1990). The most effective cognitive based treatments involve several distinct stages including modelling, graduated practice, rehearsal, role playing and reinforcement (Andrews et al. 1990; Lipsey & Cullen 2007). In her numerous studies of the available literature, MacKenzie (2011) has consistently found cognitive based treatment to be the most effective rehabilitation program to reduce offender recidivism.
Likewise, Andrews et al. (1990) and Lowenkamp et al. (2006) also found that treatment based on cognitive change is typically more effective than all other treatment modalities. While their study focused on a different aspect of offender treatment, Chadwick et al. (2015) found that those offenders who worked with officers trained in the use of behavioural intervention techniques had lower recidivism rates than those who worked with officers who did not receive this training. Despite these promising results, many of the researchers warn that the type of treatment alone cannot lead to a reduction in recidivism.
Instead, other factors such as treatment integrity and adherence to the RNR model can significantly impact on the effectiveness of this treatment type (Lipsey & Cullen 2007; Andrews et al. 1990; Chadwick et al. 2015). Program Integrity Several of the studies also focused on how program integrity impacts on the effectiveness of treatment. Program integrity refers to the quality of the treatment being offered to offenders and can encompass such things as how well the program was developed and also how well it was delivered. In their 2015 study, Chadwick et al. cused on the effects that worker training had on recidivism rates.
They found that those workers who were trained in five dimensions of correctional practice – use of authority, prosocial modelling, problem-solving strategies, use of community resources and relationship factors (p. 979) – were able to deliver more effective programs than those who received no training. The results were significant, with those offenders supervised by trained officers achieving a recidivism rate of 36. 22% compared with a recidivism rate of 49. 91% for those supervised by untrained workers.
Lipsey and Cullen (2007) also identified the skills of the worker as one of the main determinants in program effectiveness. Another aspect of program integrity referred to in several of these studies was the importance of the client-worker relationship. Andrews et al. (1990) noted that a warm, flexible and enthusiastic approach to service delivery was central to reducing recidivism rates, while Chadwick et al. (2015) referred to a study by Kennealy et al. (2012) which found that the higher an offender rated the quality of their relationship with their supervising officer, the greater the reduction in their recidivism rate.
These results show that in order to deliver effective programs, workers can’t simply be trained in a set of objective skills: they must also be enthusiastic about the corrections field, authentic in their interactions with offenders and have a strong commitment to, and belief in, the ability of rehabilitation to reduce recidivism rates. Moving beyond the role of the worker, another significant aspect of program integrity identified by Lipsey and Cullen (2007) was the frequency and duration of treatment.
They found that the effectiveness of treatment was dependent upon how long an offender engaged with the service. They hypothesised that if an offender was given the correct treatment (for example cognitive behavioural therapy) but did not receive it for the correct amount of time, recidivism rates would be affected. Lowenkamp et al. (2006) also studied the effects of program duration on offender recidivism. They found that the length of treatment needs to be tied to the level of risk, with higher risk offenders receiving more frequent and long-term support than those deemed to be a lower risk.
What Doesn’t Work In the same way that these five studies were able to identify several key factors that contribute to reduced recidivism, they were also able to highlight other factors that consistently increase recidivism rates. These factors will be discussed now. The Deterrence Model The deterrence model is designed to deter, incapacitate and control offenders through the use of several methods including increased prison sentences, boot camps, intensive supervision, urine testing and electronic monitoring (MacKenzie 2011; Andrews et al. 1990).
In contrast to the community programs based on the rehabilitation model, deterrence programs do nothing to address the criminogenic needs of offenders nor their anti-social cognitions (MacKenzie 2011). Instead, proponents of this approach believe that the severity of the treatment will, in itself, deter people from committing future crimes (Patterson 2012). However, all of the studies chosen for this report concluded that treatment based on a deterrence model did not lead to a significant reduction in recidivism rates and, in some cases, even led to an increase in offender recidivism.
In their 1990 study, Andrews et al. concluded that, to their knowledge, not a single study had linked criminal sanctions to lower rates of recidivism and, over twenty years later, MacKenzie (2011) proclaimed that there was still “no evidence” (p. 8) linking increased punishment and control to reduced recidivism. Lipsey and Cullen went one step further and analysed several studies that have linked deterrence methods to an increase in recidivism rates.
Spohn and Holleran (2002) found that people sent to prison had higher recidivism rates than those placed on probation; Petersilia and Turner (1993) found that offenders placed on intensive supervision programs (ISPs) had a 4% higher chance of reoffending within one year of release than those placed on probation; and Aos et al. (1999) found that initiatives such as the Scared Straight program increased rates of recidivism by a substantial amount. So, while punitive treatment is often the method of choice for those working in the corrections field, the evidence suggests that rehabilitation is a far superior way to reduce recidivism rates.