The Problem of Reoffending

The Social Exclusion Unit (SEU) declared in 2002 that 'Prison sentences are not succeeding in turning the majority of offenders away from crime' (Social Exclusion Unit, 2002: 5). 58 per cent of prisoners released in 1997 were reconvicted of another offence within 2 years of being released from custody, and of those, 36 per cent also went on to receive an additional prison sentence (Social Exclusion Unit, 2002: 5). And although the Ministry of Justice (2009: 9) declare that the proportion of ex-prisoners reoffending is actually falling (noting a decrease in re-offences, from 43% in the year 2000 cohort to 39% in the 2007 equivalent), reoffending rates still remain persistently high and there is clearly much more to be done to help tackle the problem, and reduce the substantial costs associated with reoffending. The Social Exclusion Unit's report to the Office of Deputy Prime minister (2002: 5) details the many costs associated with reoffending and notes that they are mostly not quantifiable but identifies the groups it most obviously impacts on as; the victims of crime themselves (who can often be repeat victims), victim's families and predominantly communities in general. The financial costs of re-offending by ex-prisoners are also confounding and widely experienced. The Social Exclusion Unit (2002: 5) notes that in terms of costs to the criminal justice system in their dealings of the consequences of crime, recorded crime committed by ex-prisoners totals at least £11 billion per year. The Social Exclusion Unit (2002: 5) estimate that a reoffending ex-prisoner is expected to be responsible for crime totally an average of £65,000 and when their offending leads to a prison sentence being imposed at a crown court this can cost in the region of £30,500. The costs of actually keeping a prisoner in custody can vary significantly according to the type of institution they are held in but cost a staggering £37,500 per year on average (Social Exclusion Unit, 2002: 5). And it must be remembered that these figures only report on crimes that have been recorded, which account for only a small fraction of total crimes. The Social Exclusion Unit (2002: 5) estimates recorded crime to be between a tenth and a quarter of the real figure. Ex-prisoners are also likely to be prolific offenders who would be responsible for a large amount of unrecorded crime, as well as the costs it incurs (Social Exclusion Unit, 2002: 5). In conjunction with punishment, the reform of offenders is key to reducing reoffending and delivering justice to the public. But when these figures quoted from the Social Exclusion Unit (2002) and the Ministry of Justice (2009) are combined they make for shocking reading and imply that current rehabilitative interventions which aim to reform offenders and reduce reoffending, are thus 'failing'. This raises questions about how current rehabilitative interventions can be improved to achieve their aims, or if new interventions, such as mentoring schemes for example, should be further developed, researched and put into practise in place of present strategies.

Rehabilitation

Criminal Justice theories of rehabilitation broadly take the stance that crime is best prevented by working directly with offenders to address their criminogenic needs; the personal, social and economic factors most closely associated with their offending behaviour (Canton and Edie, 2008: 93). Millie and Erol (2006: 2) identify some examples of criminogenic needs as, 'substance misuse, poor educational and vocational skills, poor cognitive and interpersonal skills, and antisocial attitudes'. By focussing on these risk factors it is more likely that an offender will successfully rehabilitate (Millie and Erol, 2006: 2) and desist from committing further crime. Rehabilitation as a method for reducing re-offending has not always been held in high esteem, with Martinson in the 1970s claiming that 'education at its best, or that psychotherapy at its best, cannot overcome, or even appreciably reduce, the powerful tendency for offenders to continue in criminal behaviour' (1974: 49). Martinson's controversial article 'What works? Questions and answers about prison reform' (1974) assisted in the development of an era in UK government policy commonly known as 'Nothing Works' which took Martinson's article as proof of the failure of rehabilitative methods, with these being replaced by more centralized, administrative approaches to law and order (Barry, 2000: 4-5). A so-called 'What Works' era of literature (including for example, Lipsey, 1992; Palmer, 1992; Gendreau and Ross, 1980) then evolved in the 1980s as a direct response to Martinson's controversial proclamation, deriving its name from his article of the same title. The general message of the meta-analyses that took place during the 'What Works' movement was that when rehabilitative treatment was used with offenders it could have small but noteworthy effects in terms of reducing reoffending, as long as some key areas are taken into consideration in order to appropriately turn research into practice, with the aim of delivering programs to reduce recidivism. McGuire and Priestly (1995) outline their interpretation of these key areas through a set of guiding principles, concluding that if followed they could lead to greater effectiveness in program content and delivery. These guiding principles are;

• Risk Classification- effective risk assessment is said to be required for the accurate matching of the clients with the level of delivery of certain rehabilitative programmes.
• Focus on criminogenic needs (those which directly influence ones propensity to offend).
• Responsivity- matching styles of learning between worker and service user.
• Community based interventions
• Treatment modality- a combination of skills-orientated, cognitive behavioural and other methods.
• Programme integrity- that ensures programme aims are reflected in the methods used (McGuire and Priestly, 1995).

And although there is increasing evidence of what does work with offenders in tackling re-offending for example, the Social Exclusion Unit identified that Offending Behaviour programmes which aim to 'change the way offenders think...and to teach positive techniques to avoid the situations that lead to offending', can reduce reconviction rates by up to 14% (2002: 7), little still has been learnt about more innovative forms of intervention with offenders, such as mentoring.

Resettlement

Connected to a model of rehabilitation is a similar concept known as 'resettlement' which although does hold out hope for the rehabilitation of offenders, it focuses most of its attention on the pressing practical problems faced by many ex-prisoners, with the government recently highlighting 'through the gate' interventions as a significant factor of its wider strategies to reduce reoffending (Maguire and Raynor, 2006: 2). The Social Exclusion Unit (2002: 6) identified 9 key areas that influence the likelihood of reducing reoffending; 'education, employment, drug and alcohol misuse, mental and physical health, attitudes and self control, institutionalisation and life-skills, housing, financial support and debt, and family networks'. For example 'being in employment reduces the risk of re-offending by between a third and a half' with having stable accommodation reducing the risk by a fifth (Social Exclusion Unit, 2002: 6). The Social Exclusion Unit's (2002) recommendations were translated into policy through the Government's Reducing Reoffending National Action Plan (Home Office, 2004), with its core focus on the resettlement of prisoner's after release. This National Action Plan required the production of 'Reducing Re-offending Strategies' and linking 'Action Plans' for the delivery of key services. Similarly to The Social Exclusion Unit's (2002) nine key factors, these action plans were divided into seven separate pathways including; 'accommodation, education, training and employment (ETE), mental and physical health, drugs and alcohol, finance, benefit and debt, children and families of offenders, and attitudes thinking and behaviour' (Maguire and Raynor, 2006: 4). The delivery of these services enacted by the Home Office (2004) now takes place in a dramatically different organisational framework after creation of NOMs (National Offender Management Service) which brings Probation and Prison under one management system. The use of a nationally standardized Offender Assessment System (OASys) has also been introduced as well as the availability to sentencers of 'custody plus' and a new system of 'end to end offender management', meaning an offender will be under the supervision of one 'manager' throughout the whole of their sentence. This new concept of 'end to end' offender management implies 'the close involvement of partner agencies in service planning and provision' (Maguire and Raynor, 2006: 5), and is an important move for third sector organisations such as mentoring projects, who are now increasingly seen to play an influential role in the resettlement of ex-prisoners and reducing reoffending. This optimism for third sector organisations is supported by the Ministry of Justice in reports such as 'Working with the third sector to reduce reoffending: securing effective partnerships 2008-2011' (2007) and NOMs in its consultation paper 'Volunteers Can: Towards a volunteering strategy to reduce re-offending (2007), which identifies the potential for volunteers to reduce reoffending in a number of ways;

• By helping to deliver the key 'change process' for offenders for example, gaining new accommodation.
• Improving the quality of life for service users by providing practical support.
• Providing a route to change and progression for those offenders who get involved in volunteering (NOMS, 2007: 15).

The rise of Mentoring in the UK

Joliffe and Farrington note that mentoring is used in the criminal justice context to increase 'the life successes of individuals who are at risk of reoffending' by providing direct practical assistance (for e.g. filling out housing applications, assisting in searching for employment) and indirect support (for e.g. by acting as a positive role model) (2007a: 2). By providing individuals with both of these forms of support, mentoring aims to assist in reducing reoffending and increasing positive life outcomes (Joliffe and Farrington, 2007b: 10). And for the purpose of their systematic review Tolan et al (2008: 6) identify the following 4 central characteristics of mentoring;

• Interaction of two individuals over an extended period of time
• Inequality of experience or power between the mentor and the mentee (recipient) with the mentee possessing a greater share.
• The mentee is in a position to imitate and benefit from the knowledge, skill, ability, or experience of the mentor
• The absence of the role inequality that typifies other helping relationships and is marked by professional training, certification, or predetermined status differences such as parent-child or teacher-student relationships.

Mentoring is a relatively new concept to the UK and especially to the criminal justice sector. It is identified by Newburn and Shiner (2006: 1) that mentoring originated as a 'formal response to social exclusion and social welfare problems in the US', with one of the earliest mentoring programmes being identified as the Big Brothers Big Sisters of America (BBBSA) which was established in 1904, influenced by a children's court judge who recruited delinquent boys who came before him, onto the programme (Big Brothers Big Sisters, n.d). The Big Brothers/Big Sisters programme has since expanded rapidly in the US (Big Brothers Big Sisters, n.d) and has been recreated by other mentoring organisations with similar visions elsewhere in the World including the UK. The Dalston Youth Project (DYP) and Big Brothers Big Sisters UK were established in the mid to late 1990s drawing heavily on the US expansion of mentoring (Newburn and Shiner, 2006: 2). The Dalston Youth Project remains one of the most widely known mentoring programmes in the UK focussing on 'at risk' young people aged between 11 and 18 years, beginning with a residential session followed by almost a year long programme which combines mentoring with sessions focussing on education and personal development (O'Sullivan, 2000: 1). Since New Labours arrival in 1997, they have continuously displayed concern with social exclusion leading to numerous calls for the use of mentoring in new areas of action and when the Social Exclusion Unit was established shortly after Labours election they also recommended the use of mentoring in a number of its reports (1998, 2000, 2001, 2002). Newburn and Shiner (2006: 3) note that mentoring took off most significantly in the area of Youth Justice in the late 1990s, particularly after the establishment of the Youth Justice Board who issued guidelines for mentoring with young offenders. By the year 2000 the Youth Justice Board had funded almost 100 mentoring schemes and the Home Office had also become a significant funder of local mentoring programmes via a capital grants scheme (Newburn and Shiner, 2006: 3). But despite its increasing popularity mentoring continues to be associated with a number of difficulties. Defining the term 'mentoring' becomes difficult due to the vast set of practises it can cover including one or all of the following; coaching, facilitating, counselling, befriending, tutoring, teaching, role-models, buddying or life-styling (Phillip, 1999; Clutterback, 2002). It is these definitional problems combined with mentoring's contemporary nature that creates further difficulty in assessing the value of mentoring which is why there is little empirical evidence evaluating its effectiveness, particularly concerning how/why it is/isn't of benefit in promoting desistance (i.e. what aspects of mentoring impact on desistance the most) and not simply does it reduce recidivism. This lack of research into mentoring effectiveness implies that most mentoring projects in the UK have begun due to a 'leap of faith'. Although a number of evaluations of the impact of mentoring on reductions in reoffending have been conducted with varying results in the US and some from the UK, they mostly focus on mentoring in youth justice or 'at risk' individuals. There also appears to be a significant focus on quantitative measures of success rather than qualitative approaches, which may provide a better measure of the benefits of mentoring and qualitative studies exploring mentoring schemes from the voice of the service user are also scarce, and as it is ultimately only the offenders themselves who can stop reoffending, it is important that their views on how this should be done are taken into account.

Current Research into Mentoring Effectiveness

As mentioned most evaluations of mentoring schemes have originated from the US and report generally favourable findings. Becker (1994) for example studied delinquent youth who were involved in the Partners Inc. Mentoring programme and reported a 65-75% reduction in recidivism. Lewis et al (2007) lend an explanation for this by stating that mentoring can benefit ex-prisoners through contact with people who have time to pay attention to individual needs and whose unique input is often the provision of emotional and personal support. This outcome backs up Maruna's (2000) argument that by paying attention to offenders' mental processes is a fundamental factor in affecting change in their offending behaviour. Here in the UK, Newburn and Shiner (2006) conducted a study of the 'Mentoring Plus' schemes set up by Crime Concern. Similarly to some other studies, Newburn and Shiner took a cautious approach in their conclusion of their evaluation. They state that 'mentoring faces an uncertain future' (2006: 17) but that it has 'real potential' (2006: 16). They found the biggest impact of mentoring was strongest in relation to engagement with education, training and work and similarly Clancy et al (2004) reported that offenders found assistance completing forms and handling finances the most useful feature of mentoring. Clancy et al (2005) have also associated 'through the gate' work with lower reconviction rates as the volunteer mentor had already established a positive relationship with the prisoner in custody which was continued after release (Hudson et al, 2007; Lew is et al, 2003). This is supported by McGuire et al (2007) who highlight mentoring schemes as a promising strategy to assist ex-prisoners in settling in at their initial accommodation and assisting their progress afterwards. St. James-Roberts et al (2005) reported gains in many of the programmes included in their study, such as improving literacy and numeracy, and particularly with BME targeted schemes, accommodation and family relationships. They also identified that programmes lasting over ten months, including 15 meetings on average, and had a steering group, were considered to be most successful. The characteristics of the mentee also had effects on outcomes, reporting that mentees who were younger, had a lack of offending history and were female (particularly when engaged with female mentors) were most successful. When Clancy et al (2004) completed a review of the Transitional Support Scheme (TSS) which provided mentoring for offenders in Wales. The scheme produced statistically significant changes in offenders' attitudes to crime, through consistent, high quality training focusing upon practical needs and cognitive motivational deficits of offenders, according to Clancy et al (2004) and this guaranteed that contact was maintained post release, the integrity of the scheme was maintained and offenders' commitment to change was preserved. In Joliffe and Farrington's rapid evidence assessment of the impact of mentoring on re-offending (2007), a range of studies on mentoring were analysed in order to assess how successful mentoring is in reducing reoffending. 18 studies were included in their analysis with 7 of these showing a 'statistically significant' positive impact on reoffending, demonstrating a 4 to 11 per cent reduction of subsequent offending for those involved in these mentoring schemes (although it was noted that this result was driven primarily by those studies with lower methodological value, with the best studies that were designed to provide the most accurate assessment of the impact of mentoring actually suggesting that mentoring did not cause a statistically significant reduction in reoffending). Those mentoring programmes that were successful in reducing reoffending were those where the mentee and mentor met at least once a week and spent longer periods of time together when they met. When mentoring occurred as part of a multi-modal programme, with a larger number of interventions involved, these programmes tended to be more successful in reducing reoffending. And another interesting observation made by Joliffe and Farrington was that it was only studies in which mentoring was still being given during the follow up period which proved a statistically significant reduction in reoffending, suggesting the benefits of mentoring did not persist after the relationship had ended. Joliffe and Farrington do note that due to the restricted time-period that their study had to be completed in, it was unrealistic to be able to include unpublished studies, difficult to obtain materials and foreign language studies which can reduce confidence and also result in 'publication bias' due to there being a greater tendency for statistically significant findings to be published (those showing the interventions had an impact) over studies that show non-statistically significant findings. The most recent systematic review of mentoring schemes has come from the Campbell Collaboration in the US (Tolan et al, 2008). Tolan et al evaluated the effects of the mentoring interventions included in their study 'on delinquency outcomes for youth...and key associated outcomes' (2008: 5), with 39 studies meeting the criteria for inclusion in their quantitative analyses. The Review concluded that when the 39 included studies were analysed for four outcomes measuring delinquency or closely related outcomes it suggests 'mentoring for high-risk youth has a modest positive effect for delinquency, aggression, drug use, and achievement' (2008: 8). They noted that effects were strongest when 'emotional support was a key process in mentoring interventions, and when professional development was an explicit motive for participation of the mentors' (2008: 8). But Tolan et al are wary not to make any sweeping conclusions about what elements of mentoring make it an effective intervention by stating that 'the valuable features and most promising approaches cannot be stated with any certainty', putting this down to the' remarkable lack of description of key features or basic program organization' in the studies included in their review (2008:8). They finally call for 'more careful design and testing of mentoring effects to provide the needed specificity to guide effective practice of this popular approach' (2008: 5) On the whole mentoring seems to be a promising intervention and the elements which have been identified as making it effective, have been highlighted. But most studies into mentoring's effectiveness so far have been overly concerned with quantative analyses of statistically positive outcomes whilst only being able to provide some general and tentative statements about what might affect the impact of mentoring. With most of the studies generally using reconviction rates as a measure of success they are ignoring the fact that desistance is a process rather than a single event, and that mentoring can play a key role in encouraging this process of desistance, meaning just because an offender does re-offend whilst being mentored or after having been mentored, the intervention has thus not failed or been ineffective.

Defining 'Effectiveness'

The study of desistance had been somewhat neglected by modern criminology up until the last 10 to 20 years, prompted by the criminal career literature of the 1980's and 1990's (See Farrington, 1986; Paternoster, 1989; Moffitt, 1993), when more meticulous and sustained attempts at charting the process and factors associated with desistance have occurred, and been reviewed by writers such as Laub and Sampson (2001; 2003), Farrall (2000; 2002) and Maruna (2001). Farrall and Calverley (2006) describe desistance as 'the process of ending a period of involvement in offending behaviour'. This definition recognises that the study of desistance is not primarily concerned with whether an individual has committed further crime or not but rather the journey towards becoming a non-offender. Maruna (1998) who has also published some influential work on desistance backs this up by claiming desistance is not an 'event' but a 'process' (See Farrall, 2002: 65) and Sampson and Laub (1998) define desistance as 'a gradual movement away from criminal offending'. All this suggests that recidivism rates are not the only measure of effectiveness of criminal justice interventions such as mentoring or otherwise, the recidivism of an offender does not necessarily mean that the intervention has not been effective. Therefore Maruna and Farrall (2004: 174-175) note that they are not interested in short periods of crime free lulls (which they term 'primary desistance') but they are more concerned with charting 'secondary desistance' which refers to a process whereby offenders gradually assume the role of a non-offender or a 'reformed person'. Desistance forms part of the larger lengthy process of re-integration (otherwise known as 're-entry' or 'resettlement'). Maruna, Immarigeon and Lebel state it 'is both an event and a process' (2004: 5), accounting for the initial changes experienced on the day of release from prison and more broadly as a longer term process beginning prior to release from custody and continuing well afterwards also. Ideally prisoner assistance to their reintegration should begin as soon as they are taken into custody and continue until they are released and beyond. The UK Association of Chief Officers of Probation have defined 'resettlement' as: A systematic and evidence-based process by which actions are taken to work with the offender in custody and on release, so that communities are better protected from harm and re-offending is significantly reduced. It encompasses the totality of work with prisoners, their families and significant others in participation with statutory and voluntary organisations (cited in Morgan and Owers, 2001: 12). Essentially reintegration involves everything 'from literacy training to electronic monitoring' (Maruna, Immarigeon and Lebel, 2004:6) which are all intended to reduce recidivism, and since Martinson's (1974) 'What Works' literature, policy makers and practitioners have began to ask they how do offenders desist and what prompts them to do so.

Factors and Processes associated with desistance

Research undertaken so far into desistance has shed light on the role of social and personal factors associated with it, commonly being based on longitudinal data sets or on one-off retrospective research and in a few cases by following the careers of persons after they have been made subject to criminal justice interventions (Farrall and Calverley, 2006: 4). Employment has been identified as a factor in encouraging desistance by some researchers (See Uggen and Kruttschnitt, 1998; Mischowitz, 1994; Farrall, 2002) with Shover reporting how having a job generated '... a pattern of routine activities... which conflicted with and left little time for the daily activities associated with crime'(1983: 214). Although Ditton (1977) and Henry (1978) have demonstrated that full time employment does not inhibit opportunities to offend or actual offending and Graham and Bowling (1995: 56, Table 5.2) also found that for young males employment was not related to desistance. The initiation of a significant life partnership have also been found to occur at about the same time individuals cease to offend (See Chylicki, 1992). With a number of studies suggesting the experience of becoming a parent is also associated with desistance (See Trasler, 1979; Leibrich, 1993: 53; Hughes, 1997; 1998: 146).