Submitted to University of Edinburgh for Msc Psychology of Mental Health qualification
The completion of this dissertation was facilitated through partnership with Police Scotland, and Edinburgh Police in particular, in addition to the VOW project peer mentors, I thank them for the opportunity and am grateful for their cooperation throughout the research process. In addition, my thanks to my colleagues Lily, Sylph, Roisin, Laura and Shanti for their mutual effort, support and guidance which has been an essential component in this project’s success.
Thanks also to our dissertation supervisor Dr. Suzanne O’Rourke for her effort in producing the research proposal to evaluate VOW and establishing a relationship with Edinburgh Police that has made this research project possible, and also for overseeing the process as a whole.
I would also like to thank my friends, my in-laws, my Twitter followers, and Considerit café for keeping me motivated and for always offering a more positive perspective when I’ve been neck-deep in statistics.
Finally, I would like to thank my husband Harry for whom I am eternally grateful. For inspiring and empowering me, helping me to get where I am today and for his patience, love and support - forever my rock and companion, no matter what challenges face me.
Contents
Acknowledgements i
Contents ii
List of Tables iv
Abstract 1
1.1 Life-Course Persistent versus Adolescent-Limited Theories of Offending 2
1.2 Social Control Theories 3
1.3 Desistance and Readiness for Change 5
1.4 Interventions 9
1.5 Present Study 11
1.6 Research Questions and Hypotheses 12
2.1 Design 14
2.2 Intervention 14
2.2.1 Ethical agreement document 14
2.3 Participants 15
2.3.1 Recruitment 15
2.3.2 Inclusion and Exclusion Criteria 15
2.3.3 Sample 16
2.4 Measures 16
2.4.1 Outcome Variables 16
2.5 Analysis 18
2.5.1 Descriptive Statistics of Sample 18
2.5.2 Survival Analysis 18
2.5.3 Chi Squares and Mann-Whitney U Tests 19
3.1 Sample 19
3.2 Testing for Normality 20
3.3 Intention to Treat Survival Analysis 21
3.4 Per Protocol Survival Analysis 22
3.5 Age Analysis 23
3.5.1 Reoffending 23
3.5.1.1 Chi-Square 23
3.5.1.2 Mann-Whitney U Test 24
3.5.2 Participation 24
3.5.2.1 Chi-Square 24
3.5.2.2 Mann-Whitney U Test 24
5. References 29
6. Appendix 40
6.1 Ethical Application 40
6.2 Data Sharing Agreement 47
8950 Words
List of Tables
Table Page1Frequencies & Descriptive Statistics192Normality Test (Kolmogorov-Smirnov)213Intention to Treat Survival Analysis (Kaplan-Meier)224Per Protocol Survival Analysis (Kaplan-Meier)235Chi-Square – Reoffending by Age Groups246Chi-Square – Participation by Age Groups24
Abstract
Introduction: The VOW project aims to tackle recidivism in Edinburgh. Current understanding of offending has been multi-faceted but developmental taxonomy and social control theories are particularly psychologically relevant, although there is little empirical research supporting offending reduction interventions.
Aims: To evaluate the effectiveness of VOW in preventing recidivism in participants, and assess the influence of age on intervention effectiveness and likelihood of signing up.
Method: A retrospective cohort study of an anonymised dataset featuring a sample of 251 participants and non-participants were deduced by focusing on cases with at least one year of data. An Intention to Treat (ITT) analysis regarded any case that signed with VOW as participants whereas the Per Protocol (PP) analysis required participant commitment for at least three months without disengaging or being rejected.
Results: ITT survival analysis produced a non-significant result (p = .141), but the PP survival analysis was significant (p = .006). Age was not significantly related to reoffending or participation in both the Chi-Squares (Age group analysis) and Mann-Whitney U tests (Mean age).
Conclusions: Once participants progress beyond the first three months, the VOW project intervention significantly prevents reoffending compared to non-participants indicative of both readiness for change and the cumulative effects of social support, and age is not necessarily indicative of offending behaviour.
Introduction
The Violent Offender Watch (VOW) scheme is an intervention designed to reduce violent and prolific crime in the city of Edinburgh. It was founded in 2013 by Edinburgh Police as a signposting and mentoring service designed to support repeat prolific offenders to reintegrate into society on an individual and bespoke basis, with two mentors with lived experience being introduced in 2016. This dissertation aims to evaluate the effectiveness of VOW in its ability to address criminogenic needs and encourage desistance from crime, and also to analyse the relationship between age and likelihood of participation and recidivism.
Prior to the evaluation of VOW itself, it is first important to gauge current understanding of the wider field. Therefore, the following literature review will entail contemporary research on offending and the underlying theories they draw upon, followed by an analysis of major formulas designed to encourage desistance and issue of ‘readiness for change’. Finally, a number of comparable interventions will be compared and evaluated, in addition to addressing the contribution that peer support and lived experience offer to rehabilitation interventions for prolific offenders.
1.1 Life-Course Persistent versus Adolescent-Limited Theories of Offending
Moffitt’s (1993) developmental taxonomy describes a pathological personality type, predisposing individuals to a life of crime as a result of neuropsychological deficiencies precipitated by unresponsive or inadequate parenting. Otherwise known as life-persistent anti-social behaviour, it is compared with adolescent-limited delinquency - an adaptive response to the ‘maturity gap’ between physical maturation and access to adult privileges (e.g. employment, sex and alcohol). Moffitt’s addressing of criminal causality has been highly influential in psychological and criminological research (e.g. Buffkin & Luttrell, 2005; Slobogin, 2017; Schofield et al., 2015), although pathologising antisocial behaviour is arguably deterministic - despite recognising varying criminal manifestations across the lifespan, life-course-persistent offenders are donned delinquent indefinitely, undermining their capacity to make rational, prosocial choices. Additionally, despite the certainty with which Moffitt presents these findings, retrospective data usage cannot with empirical certainty confirm neuro-psychological impairments or childhood adversity directly causal of antisocial behaviour (Fairchild et al., 2013). Nevertheless, recent treatise on life-course-persistent offending offers convincing support to Moffitt’s developmental taxonomy (Petkovsek et al., 2016; Barnes, Beaver & Boutwell, 2011; Jolliffe, 2017).
Raine (2018) reviews research into the relationship between antisocial personality disorder and areas of the brain including the pre-frontal cortex, amygdala, striatum believed to be associated with decision-making, emotion regulation, and impulsivity respectively (e.g. Faria, 2013; Sullivan-Toole et al., 2018; Verdejo-Román et al., 2018). The article proposes a link between pre/peri/post-natal-established brain irregularity, and antisocial behaviour symptoms be recognised, measurable and considered within criminal justice systems. However, more studies are required to reinforce this entreaty as cross-sectional data analysed in the study offers little insight into long-term brain functioning and behaviour (Schaie, 2005). Baglivio et al. (2016), drawn from temperament-based theory (DeLisi and Vaughn, 2014), assesses criminality and recidivism via the ‘big-five’ personality framework, particularly effortful control and negative emotionality – traits believed to be innate and driven by the prefrontal cortex and limbic systems of the brain. Thought to be in direct opposition to one another, offenders with low effortful control (high impulsivity) and high negative emotionality (emotion regulation) ‘exacerbate antisocial behaviour and negative interactions with and within…criminal justice systems’ (p. 380).
Like Raine (2018), Baglivio et al. (2016) proposes the criminal justice system should consider individual differences, specifically how personality types may differentially interact with sentencing measures. To be justifiably implementable in measuring risk of recidivism or determining initial penalisation it is necessary that predictive power of personality measures be empirically tested to further establish temperament-based theory in this field. A third perspective of life-course-persistent offending is attributed to psychopathology - Vachon et al. (2018) believed the condition to be established early in life, and stressed the importance of early detection and intervention. They identified the stability of psychopathic traits such as callousness, manipulation and impulsivity throughout adolescence into adulthood in this population compared to adolescent-limited individuals whose behaviours and traits varied to a much greater extent. Consequently, the fear of branding non-psychopathic teenagers with the condition is uncorroborated by evidence and thus supports the measurement of psychopathy in adolescents displaying chronic conduct behaviour issues. However, this study requires longitudinal model replication, ideally in a range of cultural contexts, to confirm age trend findings as established in this research which instead used cross-sectional cohort sampling of a specific White and middle-class demographic.
1.2 Social Control Theories
Alternative to the perception of criminality causality as being primarily resultant of intrinsic deficiencies within individuals is a body of research that forms social control theories. The premise of these works instead focuses on the effect that environment and social context has in guiding people’s actions both prosocial and antisocial. There are a number of major contributions made to social control theory over the past century however for the purposes of this literature review a basic summary of consensus amongst theories will be described, followed by an examination of contemporary offers within the field. The Oxford Dictionary of Sociology (2009) recognises social control to be an umbrella term, but which fundamentally represents, and its researchers desire to investigate, the notion of conformity by way of formal and informal societal forces which reinforce norms and values regarded as fundamental to a particular community. Reckless (1961) claimed that without these controls crime would be rampant within society as he believed humans to be fundamentally criminal – driven by selfish, impulse-driven desire. Sykes and Matza (1957) likewise identifies the pervasive nature of social controls and hypothesised that crimes emerged as a result of individuals successfully justifying deviant acts through a number of ‘neutralising techniques’ used to quell the guilt and shame of violating these internalised codes of conduct. In contrast, Hirschi (1969) transposed existing thought on social control and postulated the benefits it offers individuals, binding them to conform through attachment, commitment, involvement and personal belief of what is right and wrong.
More recently, researchers have investigated facets of these social controls and their prohibitory influence on committing crime and recidivism. For example, Schnepel (2016) examined the protective effect employment had against offending and reoffending, a factor well established in research (e.g. Kling, 2006; Bhuller et al., 2016). Not only was this finding reiterated in this study but also addressed the wider social impact that labour markets have on its citizens, finding that as job opportunities for low-skilled workers increased, the rates of recidivism declined. More specifically, jobs that were well-paid and offered sufficient reward appear to offset the desire to offend by achieving a similar payoff but by legitimate means. Therefore, it is crucial that labour markets be incentivised to accommodate parolees with ample employment and training prospects to prevent a revolving-door effect within the criminal justice system where upon release offenders may negatively react to their reduced chances at attaining a decent quality of life due to stigma and lack of appropriate reintegration into society. Whilst this is a well-researched observational study into labour market effects on crime rates, empirical, multi-faceted evidence that could identify the component elements within this relationship in a group of newly-released offenders would further support these claims.
Another major aspect of society studied by researchers is the effect of social connections on offending. Atkin-Plunk & Armstrong (2018) analysed the association between prison visitation and recidivism but found them to be independent of one another, as the fundamental underlying aspect was in-fact pre-existing relationships offenders possessed prior to incarceration that appeared to determine offending, especially the quality of them, particularly those shared with their mothers. Although this study did not investigate why the maternal relationship was so important it did recognise the social and emotional benefits such connections are for support and self-concept. Similarly, the role peers play in encouraging or discouraging offending has been investigated based on the theory that delinquency is a contagion that spreads through affiliation (Sutherland, 1947; Bandura, 1977). However, Gallupe et al. (2018) found that selection of peers and their influence were equally weighted, or in other words offenders not only adjusted their delinquent behaviour to match that of their peers but also selected friendships based on pre-existing behavioural similarity and therefore not necessarily a causal factor of initial offending. It would have been advantageous to for future studies to further this research by conducting longitudinal investigations on how selection and influence affects alter over time.
A final major societal contribution hypothesised to contribute to offending is the immediate social surrounds, or neighbourhood, in which an individual resides. Houser et al. (2018) considered a large number of potential variables including proximity to offenders, racial homogeneity, and employment opportunities based on previous research (Mennis & Harris, 2011; Johnson, 2016; Makarios, Steiner, & Travis, 2010), and also considered potentially protective factors such as provision of community support and places of worship (neither of which had a major effect), however the most significant finding of the study established alcohol outlets to be the highest predictor of recidivism. Houser and colleagues argue that it is not just access to alcohol that is problematic but the differential standards of behaviour from everyday conduct, and as a result recommend that the findings be considered prior to releasing offenders into ‘high-risk’ neighbourhoods. As identified by the researchers, within-individual research on traits and criminogenic needs that determine personal vulnerability to neighbourhood factors would strengthen the claims made within this study.
1.3 Desistance and Readiness for Change
Having considered theories investigating the origins of offending and recidivism, it is important to then apply this knowledge to the understanding of how to reduce offending within individuals, as well as crime as a whole. Desistance has a large body of theories in its own right and a few select examples will be discussed here along with recent discussion on the subject. However, what also needs to be taken into consideration is that no amount of theorising about desistance can bring about any real change in offenders without accounting for their readiness to change – contemporary research of this topic will also be examined. Interestingly, desistance and readiness for change arguably express the aforementioned social context and individual difference concepts of offending respectively, and act as mediators from theoretical understanding to policy making.
A principal theory of desistance with enormous influence in contemporary understanding is the risk-needs-responsivity (RNR) model proposed by Andrews and Bonta in 2006. The publishing of this concept marked a major shift in focus from being able to distinguish between who are and are not more likely to offender, and towards active means of addressing the realistically complex matter of how to successfully encourage an offender to desist from crime altogether. RNR aims to address the extent of an individual’s criminogenic needs and along with its personal and social precipitants, and then consider the best intervention on this basis. The notion of customising rehabilitation to an individual’s specific requirements is a relatively new concept although this maiden approach in its application to policy has brought with it limitations that are now being addressed today.
For example, theorists such as Ward and Stewart (2003), co-creator of the Good Lives Model (GLM), believe that the primary focus on an individual’s ‘risk’ is unnecessarily negative and appears to serve community protection more than actual offender reform (Ward, Yates & Willis, 2012; Hampson, 2018). GLM is a strengths-based system, based on RNR, that incorporates all aspects of rehabilitation and operates via a series of ‘good lives’ targets that are important to an offender’s core values or primary goods – the goal is to offer alternative secondary goods, otherwise known as the instrumentation of those values. Similarly, Newsome and Cullen (2017) saw value in RNR as a means of understanding desistance but recognised that rapid biosocial research developments require such a system to be regularly subject to evaluation and incorporate recent findings concerning neurological functioning and heart rate into risk assessment procedures. Understandably though, research into the ‘biomarkers’ of human behaviour is in its infancy (e.g. Henshilwood & Marean, 2003; Lakhan, Vieira & Hamlat, 2010; Boksa, 2013), and so must be subject to rigorous testing and replication before such factored can be reasonably implemented into leading policy and practice.
A recent effort in the consolidation of desistance theories has been posed by Braithwaite (2018) which they refer to as ‘Minimally Sufficient Deterrence’ (MSD) (as opposed to deterrence minimalism). This theory attempts to take the most effective ingredients from classic concepts including recovery capital (Laudet & White, 2008), dynamic concentration (Kleinman, 2009), restorative justice (Ness et al., 1989), shame management (Ahmed et al., 2001), responsive regulation (Braithwaite, 1992), responsivity (Andrews, Bonta & Hoge, 1990), indirect reciprocity (Alexander, 1987), and traditional incarceration. Braithwaite argues that these theories of the past thirty years should not be considered in isolation but form part of a comprehensive framework with empirical support from numerous perspectives – from this he offers a list of seven suggestions constituting a foundation from which effective desistance can occur that incorporates society-wide involvement, as Ward and Stewart (2003) intended in their GLM concept. Due to the relative recency of this theory’s publication the only citation made of it so far is featured in an article sentencing principles in the US, positively reflecting on Braithwaite’s ‘benefit-of-the-doubt’ approach to offenders and the recognition of nuance in character of the repeat offender population when considering appropriate responsivity (Tonry, in press).
In addition, it is essential that this exploratory crime prevention strategy is tested to establish whether such a variety of different theories can work effectively together in practice, whilst continuing to replicate and review the findings of theoretical components. Having said this the practical limitations on piloting such a scheme are extensive due to its enormous scope in attempting to tackle an entire community’s engagement with crime and offenders. Therefore, the most appropriate immediate application of MSD could perhaps be implemented within computer simulation software (Gilbert & Doran, 2018), or tested in small or isolated communities with minimal interference from neighbouring areas, which can control for existing crime reduction strategies (Crawford & Novak, 2018).
Numerous studies over the past decade have confirmed that a primary contingent of desistance is the motivation or readiness for change an offender possesses. Mossiere and Serin (2014) identified that the terminology used to describe the place from which an offender makes the decision to desist from violent or antisocial behaviour, or participate in rehabilitation programmes is vastly interchangeable, identifying this as one of a few issues the area of research has in terms of standardisation that can translate into practice. Having said this, it is important to highlight the major contributors of the readiness for change concept, and most notably this has been the Stages of Change theory proposed by Prochaska and DiClemente (1992).
It is important to note here that whilst other theories on this subject exist (e.g. Bandura, 1977; Markus & Nurius, 1986; Fogg, 2009), the concept presented laid fundamental groundwork in the development of readiness for change understanding (Hollar, 2018). This theory suggests that offenders graduate through a series of five phases from completely disengaged ‘precontemplation’ through to fully reformed ‘maintenance’, in-between these stages are ‘contemplation’, ‘preparation’, and ‘action’. Despite the fact that this formulisation of progression to reform has received marked criticism for its lack of appreciation of nuance and non-linear relationships between the stages (Casey, Day & Howells, 2010), for the absence of key criteria of behaviours falling under each bracket (Whitelaw et al., 2000), and the apparent disregard for social and environmental context (Armitage, 2010), still today the theory has been enormously influential in current rehabilitation interventions and forms much of the foundation underlying current readiness for change discourse (Kwasnicka et al., 2016).
Another major focus in contemporary study of discerning readiness for change is the implementation of measures that enable appropriate matching of offender to intervention, both in terms of its intensity and its contents (e.g. Baglivio et al., 2018) – this aspect was another concern raised by Mossier and Serin (2014) who ascertained in their review that there was a distinct differences in defining criteria within the various measures they evaluated which not only has a limited effect in the evaluating of offender motivations and intervention effectiveness, but also in the furthering of the concept as a whole due to the lack of agreement of over what readiness for change looks like in forensic settings.
Prior to the publication of this article Day et al. (2009) felt there to be too little investment in the measurement of treatment readiness despite finding its looming influence over the rehabilitation process – they believed this function to be crucial as a means facilitating cost-effectiveness of interventions by first assessing offender mindset before appropriately pairing them with the programme catering best to their needs, usually funded by the public purse (Barry, 2013; Welsh & Farrington, 2000). In their evaluation of readiness for change measure they designed and implemented within four violent offender interventions noted found it to be a reliable, valid and consistent means of assessment prior to offender placement, although they recognise that their study in itself had limited generalisability having used a relatively small sample.
There are other theorists that believe that so much emphasis should not be placed solely on internal thought processes and motivations, but should also consider the effect that extrinsic factors can have on their confidence and willingness to desist from offending. Sturgess, Woodhams and Tonkins (2016) introduced the Multifactor Offender Readiness Model which accounted for lack of opportunity and social capital in addition to emotion regulation, self-efficacy and perception of control. In their systematic review of intervention effectiveness, they recognised an apparent cyclical interaction between intrinsic and extrinsic factors. For example, lack of educational opportunities may negatively affect an offender’s intellectual ability, which practitioners then need to acknowledge when assigning intervention to them which may or may not require considerable metacognitive ability (Bosker, Witteman & Hermanns, 2013). This is reiterated in Yong’s (2017) thesis whose thorough evaluation of current understanding of readiness of change made it clear that both researchers and practitioners have to determine the potential adverse effects that interventions, which in themselves act as extrinsic factors, can have on an offender’s readiness for change – if not carefully and accurately handled an individual’s genuine motivation can be overestimated and they may simply comply as a bargaining chip to earn parole or use as a ruse to avert suspicion of continued illicit activity (Berenji, Chou & D’Orsogna, 2014).
From causation theory, to the mediatory instruments of desistance and readiness for change frameworks it is next important to appraise current interventions designed to apply empirically-supported understanding into practice. The importance of these interventions in the present day has never been more vital as crime, and more particularly recidivism remain at a high level. However, the figures for Scotland are significantly lower than those of England and Wales over a similar time period (Office for National Statistics, 2018): 238,651 total recorded criminal offences committed in Scotland in the year 2016/7 (The Scottish Government, 2017), and 28% of offenders go on to reoffend within 12 months of release, based on my recent cohort figures from 2014/15 (Scottish Offenders Index & Scottish Government Justice Analytical Services, 2017). Having said this, it is essential to make the distinction between offences that are reported and result in conviction, and actual crime levels (Youngs & Canter, 2014). Therefore, when evaluating the effectiveness of interventions, it is necessary to account for the integrally flawed methodology of measuring crime reduction using statistics, and the extent of influence any given rehabilitation programme can claim for it (Skogan, 1975).
1.4 Interventions
A recent analysis of descriptive statistics undertaken by Mears and Cochran (2018) investigated US policy shift towards tougher punitive sanctioning for repeat offenders under the assumption that this would reduce recidivism and protect the community (Hardy, 2018; Bates, 2018; Buller, et al, 2018). However, their findings could not find sufficient support for this hypothesis and this is supported by desistance research literature, some even stating it increases the likelihood of recidivism (Cullen, Jonson & Nagin, 2011; Bales & Piquero, 2012). Although this study is based on US justice system policy which differs from the UK who emphasises the importance of rehabilitation strategies, that is not to say that the findings do not contribute to the conversation regarding the negative effects incarceration as a means on intervention can have on offenders (Lacey, 2008; Crewe, 2011).
Meanwhile, the mere threat of imprisonment based been used, particularly in the US, as a means to deter delinquent behaviour in vulnerable adolescents. A systematic review conducted by Wilson, Brennan and Olaghere (2018) evaluated such interventions including the ‘Scared Straight’ programme that involves visits to prisons to observe daily life there as well as interacting with prisoners (Royster, 2012). They found that such crime prevention strategies to be harmful to participants’ wellbeing and likely increase their risk of offending but also identify a considerable lack of useable data for the purposes of the review due to a lack of standardisation in what was available – in order to improve future appraisals of fear-inducing incarceration programmes, introducing a measure that records various crime rates featuring precise, validated and collaborated set of criteria amongst law enforcement organisations which then record data at various intervals of the intervention (Cinar, Trott & Simms, 2018; Urwin, 2018; Aagard, 2012).
Another example of intervention strategy are skill-based training designed to provide offenders with the necessary tools to effectively navigate everyday life without having to resort to crime. Jolley (2018) conducted a multi-method analysis of such programmes using questionnaires, focus groups, semi-structured interviews with staff, and direct observation of the intervention in action. They identified in their preliminary research that educational skill-based training to be notably lacking in the UK system which they argue neglects to consider the wider context surrounding an offender’s antisocial behaviour, such as their thinking and reasoning ability in addition to self-esteem and communication capability – such interventions are believed to facilitate long-term sustainable change in behaviour and motivation (Pike & Hopkins, 2018; Szifris, Fox & Bradbury, 2018). Their study involved the appraisal of one-such programme, named Chrysalis Lite, in a low-security English prison and identified clear benefit of such an intervention on subsequent engagement by offenders towards rehabilitation, although they recognised that its rigid structure may need greater flexibility in order to account for individual differences in ability and experience (Criminal Justice Alliance, 2014). Most importantly they recognised the importance of offenders taking control of their ownership of their own behavioural reform in a way that plays to their strengths and preferences (Prison Reform Trust, 2017).
In a recent doctoral thesis by Perrin (2017), their comprehensive evaluation of peer-support programmes highlight another valuable and empirically-supported intervention designed to reduce crime and recidivism. They define such a strategy as ‘a system of social and emotional support that rests on the core tenets of mutual reciprocity, shared problem-solving, and empathy’. The research focus for the thesis is based on prison-based peer-support intervention and emphasise the benefits both in terms of receiving direct support from fellow inmates, and the cost-effective nature of having ‘helper’ offenders volunteer to support peers, which in turn develops their own rehabilitative skills, including altruism (Fletcher & Batty, 2012). Despite the focus on the helper and the prison environment in this research it certainly evidences the benefit of peer support programmes, and indeed such an intervention has been applied to many different social support fields (e.g. Narain & Adcock, 2017; Jain & Jain, 2018).
Prior to this study, back in 2012 Fletcher and Batty published a document evaluating current understanding of peer-support programmes implemented to reduce recidivism. They recognised that despite cost-effectiveness there has not been a substantial roll-out of standardised peer-support frameworks for offenders in and out of prison due to a relative lack of empirically validated research confirming the effectiveness of current interventions in practice, in addition to the newness of the field altogether resultant from the paradigm shift in criminological enquiry from ‘Nothing Works’ to ‘What Works?’ (Cullen & Gendreau, 2001). They cite a few existing peer support schemes including the St Giles Peer Advice Project, and the Routes Out of Prison project, finding that narrow inclusion criteria for would-be offender recruits and high attrition rates of peers proved to be major issues that render such strategies less reliable than more traditional and structured means of rehabilitation (Smith et al., 2011).
These weaknesses of peer-support were not the only identified within Fletcher and Batty’s review, alluding to the inevitability of boundary and confidentiality-related problems associated with the use of former offenders to guide and support participants, rather than highly trained professionals (Barker et al., 2018; Fuhr et al., 2014). Having said this, the value of peer support as a means to rehabilitate cannot be understated in its ability to truly resonate with offenders: with greater contribution from psychological and criminological research and through the development of a comprehensive understanding of what does and does not work within this approach it is likely that implementation of peer-support interventions will increase as public-spending budgets continue to tighten in the UK and importance of health and social care integration increases (The Scottish Government, 2016).
1.5 Present Study
The present study aims to evaluate VOW, an example of an intervention that incorporates peer support that aims to eliminate recidivism of prolific offenders in Edinburgh, in coordination between Police Scotland and peer mentors. This partnership between law enforcement and community support is highly beneficial as it represents a consolidation of crime and desistance eradicating the need to involve third-party organisations for rehabilitation. Having said this, the primary role of VOW is to support offenders, facilitated by lived experience of peer mentors and signpost to services that extend beyond their skill base – this provides a service tailored to their specific criminogenic needs via a network of partnerships with voluntary and community organisations across the city. Additionally, the peer mentors support the VOW police officers in their outreach work in recruiting individuals by enhancing initial trust between staff an offenders, in a way that solely-police operated approaches may struggle to achieve (Hohner, 2017).
The VOW project incorporation of peer mentors is unique, and based on its own findings, estimate to have achieved a 79.5% reduction in reoffending in participants, measured by comparing equivalent pre-VOW and current or post-VOW activity, and also approximate cost-savings of £6,155,974.79 since the intervention began in 2013. These figures generated via cost-benefit analyses are crude sums that amass estimated costs of an individual’s crimes recorded per offence. However, offending is defined and addressed by VOW as any offending incident, irrespective of it resulting in arrest, but even this does not necessarily reflect the actual frequency of offences committed (Gomes, Maia & Farrington, 2018). Additionally, by only referencing offending rate in like-for-like temporal intervals, factors unique to an individual in terms of their personal and social circumstances are not taken into consideration which may have a direct effect on recidivism and consequently the effectiveness of the intervention (King et al., 2014). It is the objective of this study to empirically evaluate the effectiveness of VOW as a peer-support intervention in its ability to prevent reoffending.
1.6 Research Questions and Hypotheses
The present study will aim to address the following four research questions:
Were participants in the VOW programme, on average, less likely to reoffend than non-participants?
Intention to Treat:
Null Hypothesis 1
H0 - There is no significant difference between participants upon signing up to VOW and non-participants in their desistance from offending.
Alternative Hypothesis 1
H1 - There is a significant difference between participants upon signing up to VOW and non-participants in their desistance from offending.
Per Protocol
Null Hypothesis 2
H0 - There is no significant difference between participants having committed to three months of involvement with VOW, and non-participants in their desistance from offending.
Alternative Hypothesis 2
H1 - There is a significant difference between participants having committed to three months of involvement with VOW, and non-participants in their desistance from offending.
Did those who chose to engage with VOW differ significantly with regards to their age?
Reoffending:
Null Hypothesis 1
H0 – Age does not significantly contribute to the effectiveness of VOW in preventing reoffending within participants.
Alternative Hypothesis 1
H1 - Age significantly contributes to the effectiveness of VOW in preventing reoffending within participants.
Participation:
Null Hypothesis 2
H0 – Age does not have a significant effect on whether an offender with sign up with VOW once approached.
Alternative Hypothesis 2
H1 - Age has a significant effect on whether an offender signs up with VOW once approached.
In addressing these research questions in turn, the researchers hypothesise that VOW participants reoffend less than non-participants, and also less than themselves had they not participated. In addition, the study aims to investigate the relationship between intervention effectiveness, and likelihood to participant, and offender age.
Methodology
2.1 Design
This study adopted a retrospective cohort analysis utilising routinely collected data in a spreadsheet maintained by VOW police officers. The data features an anonymised record of participants, and offenders that declined to sign up in their offending behaviour prior to, during and after having engaged with the intervention, in addition to basic demographic information. All cases were required to have at minimum 12 months of recorded data to be considered as part of the sample to adequately capture follow-up within a measurable 12-month time-frame.
2.2 Intervention 2.2.1 Ethical agreement document
VOW’s focus is supporting participants facilitated by peer support, in addition to monitoring their progress, and routine data collection of both participants, and of those that were approached but declined to partake. Due to the bespoke nature of the VOW intervention itself, and the restricted access the researchers had to the highly sensitive qualitative details of individual cases (Quick & Choo, 2014), the programme can only be summarised here (the full details are described in the Level 1 Ethics Application Form submitted for this dissertation, and approved by the University of Edinburgh; see appendix).
Monitoring Participants
This process is enabled through the daily review of ‘GRIP’ sheets that summarise the daily criminal activities that occur throughout Edinburgh – this allows VOW staff members to target offenders for recruitment onto the programme, as well as surveying the activities of individuals currently subscribed to the intervention.
Routine Data Collection
Through consultation of various Police databases such as the Scottish Criminal Database, UNIFI, and Police National Computer, VOW staff collate demographic and forensic information about participants and potential recruits for their own records, from which they record subsequent details of offending and rehabilitation activity.
2.3 Participants 2.3.1 Recruitment
The participants for this study were drawn from a population at the time of writing of 305 prolific offenders within Edinburgh, whom VOW themselves recruit through a number of different locations and services. The data provided for the purposes of this research identify five locations (in addition to a minority ‘other’ category) that VOW are in partnership and regularly engage with offenders that are processed through these channels, and they are as follows:
Edinburgh Sheriff Courts (ESC)
Prison
Polmont Young Offender’s Institute (PYOI)
St Leonard’s Police Station (St. Leo’s)
At the participant’s home address
In addition to their own outreach work, VOW frequently receives referrals from third-party organisations - the data details information indicating these sources as follows:
Social Work
3rd Sector
Police
Prison
Self
Other (not specified in the document)
2.3.2 Inclusion and Exclusion Criteria
VOW primarily target young adults with a violent or prolific offence history, however VOW also offers places to older offenders as long as they are prepared to cease offending entirely. Having said this, VOW have highlighted a number of exclusionary criteria due to pre-existing specialist resources that are better equipped to appropriately support certain categories of offenders. These include:
Sexual offenders
Perpetrators of persistent domestic abuse
Serious mental health diagnosis (e.g. schizophrenia)
‘white collar crime’ (e.g. fraud, embezzlement)
2.3. 3 Sample
Of the total population (n=305), the following were excluded from analysis resulting in a remaining total sample of 251:
11 participants with ‘pending’ status
1 Case (#106) with data input error
42 with less than 365 days of data on VOW records
2.4 Measures 2.4.1 Outcome Variables
To be able to establish the effectiveness of VOW the researchers needed to isolate variables that could be measured and could be compared between participants and non-participants. In addressing the primary research question as to whether VOW resulted in less offending, the following variables were identified:
Days until Offending – time calculated between from date of approach of signing up (depending on whether cases represent participants or non-participants, respectively), and date of first crime committed since having come into contact with VOW.
Level of Participation - All cases in the dataset are categorised by status indicating the extent of success within individuals, these statuses are as follows:
Released – Successfully participated and completed their time with VOW having ‘turned a corner’ in their ability to desist.
Engaged – currently involved in the programme.
Rejected – VOW removes participants from the programme based on poor behaviour or no indication of readiness for change.
Disengaged – chose to leave intervention of own volition (reason not provided in the dataset).
Declined: chose not to participate in VOW, this group from the primary non-participants in this study.
NB: In the ITT analysis, released, engaged, rejected and disengaged status cases were all considered participants on the condition that they had signed up with VOW. In the PP analysis, rejected and disengaged cases with less than 3 months participation were reclassified as non-participants.
Length of Participation – Calculated for participants only by measuring the difference in days between date of signing up and date of disengagement - this variable provides a good indication of attrition rate.
The second research question specifically focused on age - a within-participant variable that may potentially interact with intervention effectiveness. Initially, the researchers assessed which of the available variables would be the most viable for analysis as little personal demographic data had been provided. These variables provided were as follows:
Age – often cited in criminological research as a crucial factor in relation to offending, this variable provided the strongest potential for secondary analysis.
Gender – this variable provides opportunity to compare the factor of gender between participants and non-participants although the sample is approximately 90% male but only 10% female.
Number of offences – as VOW’s mission is to prevent reoffending altogether in their intervention although welcome any positive change. The present study did not focus on the actual number of offences committed before, during or after it took place although there may be an opportunity to correlate this with other secondary factors such as age or gender.
Date of Birth – this variable would offer interesting insight as to how socio-political or cultural events coincide with an individual’s development and criminogenic risk, although there is no scope to explore this path in the present study.
Area – as there is much research indicating a contagion effect of crime based on geography and socioeconomic factors this variable offers a promising research avenue however as the composition of Edinburgh was not specifically investigated this variable was not considered for analysis.
2.5 Analyses
2.5.1 Descriptive Statistics of Sample
In preparation of the main analysis in addressing the primary research question of gauging VOW’s effectiveness in reducing offending, the researcher’s needed to define and refine the sample and then run a series of descriptive statistics regarding between-participant variables including age, gender, status and length of participation.
2.5.2 Survival Analysis
The researchers chose to run a survival analysis as this would provide a clear demonstration of the effectiveness of VOW by measuring ‘survival time’ or duration before committing an offence over the course a fixed time period. Having gauged that the majority of cases had a minimum of 12 months follow-up period (cases with less than 12 months were removed) and still captured 251 participant and non-participant cases, and so an artificial census point was created for each case within which to frame the analysis. The type of survival analysis that was to be run depended on whether the survival time amongst participants and non-participants was equally distributed i.e. parametric – a Kolmogorov-Smirnov test will be run as there is over 100 cases within the sample. If the data were parametric, the researchers intended on running a Cox Proportional Hazards regression analysis which is suitable as it is designed to measure more quantitative and or categorical variables (Fox & Weisberg, 2010). If non-parametric, Kaplan-Meier would instead be used which generally measures ‘lifetime’ data in medical research but can also measure other events over a period of time – a log-rank test would be run alongside this to measure significance in a way that weights each time point equally (as opposed to weighting time points with increasing or decreasing significance) (Goel, Khanna & Kishore, 2010).
2.5.3 Chi-Squares and Mann-Whitney U Tests
A series of subgroup analyses was also run to address the second research question of measuring within-group variables against VOW effectiveness. These included:
Chi-square tests: As per categorical assumptions of this statistic, case ages were grouped into two categories – under and over 25 years old as per research defining the former group as ‘young offenders’. These age groups have been measured against their likelihood to participate, and whether they offended or not having been approached and or signed by VOW. Subsequent analyses focusing just on participants is also included.
Mann-Whitney U test: Age as a continuous variable (as opposed to categorical in chi-square tests) measured against whether participants and non-participant did or did not offend.
Results
3.1 Sample
A range of frequency and descriptive statistics were run to examine the variance in a number of variables within the sample (See Table 1).
Frequencies ParticipationIntention to Treat AnalysisPer Protocol Analysis
Participants
(n = 96; 38%)
Non-Participants (n = 155, 62%)
Total: 251
Participants
(n=80; 32%)
Non-Participants
(n=171; 68%) Total: 251Level of participation
Released (7%)
Engaged (7%)
Disengaged (23%)
Rejected (1%)
Declined (62%).
Age
Under 25 years old (72%)
Over 25 years old (28%)
Gender
Male (93%)
Female (7%)
Types of first offence following engagement with VOW as per Police Scotland crime categories
Robbery & Assault (4%)
Sexual Offences: (0% - not applicable)
Theft (26%)
Vandalism & Arson (4%)
Drugs, Weapons & Bail Offences (23%)
Threatening and Abusive Behaviour (21%)
Accidental Offences (failure to act) (10%)
None (no offences committed) (12%)
Descriptive statisticsAge (M = 25, SD = 8.15, range = 37)Days until offending (M = 160.22, SD = 140.02)Length of participation [ITT participants only, n=96] (M = 259, SD = 128.58)
Table 1
3.2 Testing for Normality
Determining which survival analysis to use required ascertaining whether the data was parametric or not – even though the sample was large it was not randomised, and so assumptions of normal distribution do not apply. Therefore, entire sample (n=251) was measured against days until first offence, otherwise referred to as survival time in days, and produced a significant result (D (251) = 0.307, p = .00), establishing the data as non-parametric (see Table 2).
Table 2
3.3 Intention to Treat Survival Analysis
A Kaplan-Meier survival analysis was conducted to measure days of survival, or days until first offence from each case start date (date of signing for participants, and date of approach for non-participants). The results were insignificant (χ 2 (1, N = 251) = 2.17, p = .141) thus the null hypothesis is retained (See Table 3).
Table 3
3.4 Per Protocol Survival Analysis
A Kaplan-Meier survival analysis was performed incorporating the conditions of VOW protocol involving the reclassification of ‘disengaged’ and ‘rejected’ participants with less than 3 months participation as non-participants, thus measuring survival time or days until first offence accounting for a probationary period. The results of this analysis were significant (χ 2 (1, N = 251) = 8.86, p = .003) and so the null hypothesis is rejected for the per protocol analysis (See Table 4).
Table 4
3.5 Age Analysis
A series of tests were run to measure the relationship existing between offender age and reoffending, and its effects on participation.
3.5.1 Reoffending
3.5.1.1. Chi-Square
A Chi-Square test was conducted to establish whether a significant relationship existed between reoffending and age groups (under or over 25 years old) in participants only. The result was non-significant (χ 2 (1, N = 96) = .016, p = .544) and the null hypothesis is retained as there is no significant relationship (See Table 5).
Offences vs 0 Offences TotalDidn’t offendDid offendAge GroupsUnder 25234972Over 2561624Total316596
Table 5
3.5.1.2 Mann-Whitney U Test
A Mann-Whitney U test was done to ascertain whether mean age of participants was significantly related to reoffending. The results were found to be insignificant (Didn’t offend (Mdn = 22), Did offend (Mdn = 21), U = 5447, p = .16, r = -.09), therefore the null hypothesis is retained.
3.5.2 Participation
3.5.2.1 Chi-Square
A Chi-Square test was run to gauge whether offender age significantly affected an offender’s decision to sign up with VOW, measured by age groups (under/over 25 years-old) in the total sample. The results were insignificant (χ 2 (1, N = 93) = .645, p = .256) meaning the null hypothesis is retained (see Table 6).
Signed or not Total Didn’t signSignedAge GroupsUnder 2510972181Over 25462470Total15596251
Table 6
3.5.2.2 Mann-Whitney U Test
A Mann-Whitney U test was performed to investigate if mean offender age in the total sample significantly related to their decision to sign up with VOW. The findings were insignificant (Didn’t sign (Mdn = 21), Did sign (Mdn = 21), U = 7152.5, p = .61, r = .03), and so the null hypothesis is retained.
Discussion
The present retrospective cohort study evaluated the effectiveness of the VOW project delivered by Edinburgh Police through two primary research questions: Was recidivism lower in participants of the intervention than non-participants, and whether age in any way contributed to this result, and also whether a relationship existed between age and signing up.
The results from the first survival analysis investigating the first research question from an Intention to Treat (ITT) perspective were found to be insignificant, or in other words there was no significant difference in recidivism reduction between participants and non-participants, when measured from the first day of involvement with VOW. Having said this, the second survival analysis which adopted an as Per Protocol (PP) approach did prove to be significant. This latter test involved reclassifying cases with a ‘rejected’ or ‘disengaged’ status with less than three months of participation as non-participants as this more accurately reflected VOW’s own policies – a cost-benefit analysis is drawn up by them for each participant that has consistently committed to the intervention for three months. In-line with readiness for change research, the first three months of participation acts as a probationary period that gauges dedication to the intervention and reform more generally, hence those that disengaged or were rejected within this time period were not ready (Baglivio et al. 2018). However, having compared the findings from each of these tests it would appear to suggest that a ‘minimum dosage’ is required before the VOW intervention becomes effective, as opposed to being indicative of change readiness due to the high attrition rate in survival even with the per-protocol adjustment – it appears that the VOW intervention becomes effective in preventing recidivism after three months of participation and becomes increasingly so with time.
The implications of this in accordance with social control research is that there may be a cumulative protective effect as participants are gradually introduced to more and more opportunities to walk away from crime such as rehabilitation from drugs, education, employment opportunities, and perhaps most importantly the therapeutic relationship maintained between the individual, and the peer mentors and police officers supporting them (Houser et al., 2018). Conversely, the results imply that ‘readiness for change’ did not appear to factor in the act of signing up with VOW as evident from the steep decline in survival in the ITT test – there were comparatively much fewer participants that did not offend in the first three than those that did over the 12-month period measured. Therefore, if the offenders of this study were ready to desist from crime upon agreeing to participate in VOW this arguably would have been reflected in a much more graduated decline in the survival curve.
Age was measured to ascertain relationships between participation and desistance from crime using Mann-Whitney U and Chi-Square tests. The Chi-Squares were conducted by grouping cases into two age categories (under or over 25 years of age), formed on the basis of developmental criminological theories that consistently establish a boundary between younger and older offenders in this way (e.g. Sawyer et al., 2018), particularly adolescent-limited versus life-persistent offending – that is not to say however that there is not a level of arbitrariness in making this distinction as age does not necessarily reflect individual experience or circumstances (Gillespie, 2012). The results of the Chi-Square test conducted to measure desistance between these two age groups within participants was insignificant, counter to this age-related understanding of offending: within the young offender group, 72% of people reoffended but 76% of over 25-year-olds reoffended in the 12-month period measured.
The results of the second Chi-Square test measuring the two age groups against recruitment also produced an insignificant result, evidencing little difference in whether individuals signed up or not but still indicative of older individuals being less ready to desist offending: 60% of under-25’s declined to participate, 66% of over-25’s declined. The findings of this analysis when compared with the desistance Chi-Square contradict developmental taxonomy theories that would have supported a hypothesis of under 25-year-olds offending more and also less likely to comply with an intervention such as VOW. Mann-Whitney U analyses were conducted to measure mean age as opposed to age groups and so offered a more objective perspective on the relationship between age and participation or recruitment. However, the results of these tests reiterated the Chi-Squares delivering insignificant figures in both instances. In addition, the first Mann-Whitney conducted measuring reoffending generated a negative effect size (-.09) emphasising older participants reoffending fractionally more than younger participants. The participation Mann-Whitney U effect size was similarly very small (.03) but positive, indicating older offenders signed up less than younger offenders.
Findings from the age analyses lack validity due to the disproportionate frequency between the two groups within the sample as there are 43 more younger offenders (177, compared to 74 over-25’s) and therefore an unequal comparison of younger and older offenders. This is likely due to VOW’s primary focus on the younger offender demographic and as a result target them more in their outreach work within prisons, social services and schools. On the other hand, the results may simply reflect and support statistics suggesting there are generally higher numbers of prolific and violent offenders under the age of 25 (Office for National Statistics, 2018; The Scottish Government, 2017).
A number of limitations have arisen regarding the research project as well as the intervention itself, in addition to strengths. The primary difficulty in evaluating the VOW intervention was the extensive restrictions the researchers had in accessing sample data. As information related to forensic populations is highly sensitive in nature due to data protection law, Edinburgh Police require any individuals wanting to access such data be extensively vetted (Police Scotland, 2018). As the researchers comprised a total team of eight this would have taken a substantial amount of time, much longer than the duration to complete the research project could afford. The consequence of this being that the data received for analysis was very basic and retrospective in nature, therefore although this allowed the researchers to address the primary research questions, does not offer any insight as to why various patterns emerged the way that they did (Sedgwick, 2014). For example, as the team did not have access to the qualitative details of individual case files there is a total reliance on age at time of engagement with VOW as being a valid representation of reoffending or participation as part of age-crime curve research, without accounting for any individual differences (Shulman, Steinberg & Piquero, 2013).
Another issue that prevented a more thorough examination of the data available for analysis was related to time constraints – the majority of time assigned for the research project was spent conducting preliminary secondary research on offending and interventions, as well as drawing up the data sharing agreement (see appendix) and subsequent ethical approval process. This resulted in a timeframe of three weeks from receiving the data until the deadline to become familiar with it, conduct various analyses, interpret the results and produce the final write-up. Consequently, the present study was very focused on addressing the research questions using survival analysis, Chi-Squares, Mann-Whitney U tests, and descriptive statistics – future research could invest more time performing more exploratory analyses and provide a more in-depth investigation of the data.
The data received by the researchers represents a record of the information that VOW Police officers maintain to track participants, in addition to offenders that declined to participate, in their offending however does not offer explanation for offending or lack thereof. If VOW were to feature more individual context to their records such as their ethnicity, education and employment history, social support circumstances (e.g. friends and family), and mental health and wellbeing situation this would allow future analysis to not only establish a level of causation for offending trends both within and between participant and non-participant groups, but also would offer a more valuable contribution to psychological understanding of recidivism (Ttofi et al., 2016).
In addition, implementing greater standardisation in protocol would strengthen the construct validity of the data collected (Rycroft-Malone et al., 2009) - for example, a deadline of 30 days to sign up with VOW once approached would reduce the contamination effects of an individual’s environment affecting the decision of whether or not to sign up. Also, although the sample size reaped from the dataset was large, the disproportionate quantities of individuals in terms of their age and gender prevented a close examination of offending variance between subgroups (see Kish, 1965). Therefore, to greater improve the empirical basis of the intervention, VOW could increase their efforts to recruit women and older participants, or otherwise focus on specialising in rehabilitating young male offenders which comprises the majority of their current population.
Having said this, as the intervention proved to be significantly effective in the per protocol analysis for all participants, VOW could instead focus on improving the peer support element of the intervention by adding richness to the data they record, detailing elements within protocol of peer mentors and offender interactions, even if generally informal in nature (Gillard et al., 2017). Moreover, analysis of the peer support element of VOW would be made possible with a considerably larger sample that could sufficiently power a sub-analysis of the intervention prior and subsequent to the introduction of peer mentors in January 2016 (Columb & Atkinson, 2016).
Aside from strengths and limitations of the present study and current VOW intervention protocol, there are several opportunities for future research directions to be considered. First of all, accounting for the time required to pass the vetting process required to gain full access to case files, a qualitative micro-study of the VOW project could offer rich insight into not only participant progress throughout their involvement in the intervention but also a greater understanding as to what does and does not work within and between individuals (Gaudet & Robert, 2018) – offering the recipient perspective in the appraisal of VOW would place the intervention in a unique position by being informed both by empirical research and service user feedback.
Another potential research avenue in the evaluation of the VOW project could be to conduct a longitudinal study that measures offender progress throughout their participation in the intervention and for an extensive period of time once they have disengaged that could measure its effectiveness in maintaining a state of total desistance from crime at routine time intervals (e.g. every 6 months for 5 years) (Farrington, 2013). This approach could offer a major contribution to offending intervention research in being able to much more accurately ascertain whether VOW is only effective whilst participating, due to the therapeutic relationship maintained throughout and the support they offer in providing opportunities, or if there is a lasting effect on participants having learned the skills needed to support themselves (Joseph et al., 2017).
In conclusion, the present retrospective cohort study has established that the VOW project is significantly effective in preventing recidivism in participants once they have received a minimum three-month ‘dose’ of the intervention, indicating that readiness for change is not necessarily established upon signing up but is evident subsequent to this probationary period. In addition, the present study does not substantiate age as being a relevant contributing factor in the effectiveness of VOW within participants or their likelihood of signing up. This research offers insight as to how the VOW project can develop and greater support their client group, in addition to contributing to empirical research and current understanding of readiness for change, developmental theories of offending as well as social control models. In time, it is the hope of the researchers that with gradual development and expansion of the VOW project, future research can add to the wider body of peer-support-oriented interventions.
References
Aagard, P. (2012). Drivers and Barriers of Public Innovation in Crime Prevention. The Innovation Journal. 17(1). 1-17.
Armitage, C. J. (2010). Is there utility in the transtheoretical model? British Journal of Health Psychology. 14(2). 195-210.
Atkin-Plunk, C. A., & Armstrong, G. S. (2018). Disentangling the Relationship Between Social Ties, Prison Visitation, and Recidivism. Criminal Justice and Behavior. 1-2-. DOI: 10.1177/0093854818772320.
Baglivio, M. T., Wolff, K. T., DeLisi, M., Vaughn, M. G., & Piquero, A. R. (2016). Effortful control, negative emotionality, and juvenile recidivism: an empirical test of DeLisi and Vaughn’s temperament-based theory of antisocial behavior. The Journal of Forensic Psychology and Psychiatry. 27(3). 376-403. Baglivio, M. T., Wolff, K. T., Howell, J. C., Jackowski, K., & Greenwald, M. A. (2018). The search for the holy grail: Criminogenic needs matching, intervention dosage, and subsequent recidivism among serious juvenile offenders in residential placement. Journal of Criminal Justice. 55. 46-57. Bales, W. D., & Piquero, A. R. (2011). Assessing the impact of imprisonment on recidivism. Journal of Experimental Criminology. 8(1). 71-101.
Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review. 84 (2): 191–215. Barker, S. L., Maguire, N., Bishop, F. L., & Stopa, L. (2018). Peer support critical elements and experiences in supporting the homeless: A qualitative study. Journal of Community & Applied Social Psychology. 28(4). 213-229.
Barnes, J. C., Beaver, K. M., & Boutwell, B. B. (2011). Examining the Genetic Underpinnings to Moffitt’s Developmental Taxonomy: A Behavioural Genetic Analysis. Criminology. 49(4). 923-954.
Barry, M. (2013). Desistance by Design: Offenders' Reflections on Criminal Justice Theory, Policy and Practice. European Journal of Probation. 5(2). 47-65. Bates (2018). A Phenomenological Examination of Prisonization and the Psychological Effects of Incarceration. Doctoral thesis, Walden University. Available at: https://scholarworks.waldenu.edu/dissertations/5053/. Berenji, B, Chou, T., & D’Orsogna, M. R. (2014). Recidivism and Rehabilitation of Criminal Offenders: A Carrot and Stick Evolutionary Game. PLoS ONE. 9(1). 1-13.
Bhuller, M., Dahl, G. B., Løken, K. V., & Mogstad, M. (2016). Incarceration, recidivism, and employment. Bergen: University of Bergen. Available at: http://www.uib.no/sites/w3.uib.no/files/attachments/working_paper_07-16.pdf.
Bhuller, M., Dahl, G. B., Løken, K. V., & Mogstad, M. (2018). Intergenerational Effects of Incarceration. AEA Papers and Proceedings. 108 .234-240/
Boksa, P. (2013). A way forward for research on biomarkers for psychiatric disorders. Journal of Psychiatry & Neuroscience. 38(2), 75–77. Bosker, J., Witteman, C., & Hermanns, J. (2013). Do intervention plans meet criteria for effective practice to reduce recidivism? How probation officers forget about social capital and basic needs. European Journal of Probation. 5(1). 65-85.
Buffin, J. L., & Luttrell, V. R. (2005). Neuroimaging studies of aggressive and violent behavior: current findings and implications for criminology and criminal justice. Trauma Violence Abuse. 6(2). 176-191. Casey, S., Day, A., & Howells, K. (2010). The application of the transtheoretical model to offender populations: Some critical issues. Legal and Criminological Psychology. 10(2). 157-171. Cinar, E., Trott, P., & Simms, C. (2018). A systematic review of barriers to public sector
innovation process. Public Management Review. DOI: 10.1080/14719037.2018.1473477. Columb, M. O., & Atkinson, M. S. (2016). Statistical analysis: sample size and power estimations. BJA Education. 16(5). 159-161. Crawford, L. A., & Novak, K. B. (2018). Individual and Society: Sociological Social Psychology. New York: Routledge. Crewe, B. (2011). Depth, weight, tightness: Revisiting the pains of imprisonment. Punishment & Society. 13(5). 509-529.
Cox, S. M., Kochol, P., & Hedlund, J. (2018). The Exploration of Risk and Protective Score Differences Across Juvenile Offending Career Types and Their Effects on Recidivism. Youth Violence and Juvenile Justice. 16(1). 77-96. Cullen, F. T., & Gendreau, P. (2001). From Nothing Works to What Works: Changing Professional Ideology in the 21st Century. The Prison Journal. 81(3). 313-338. Cullen, F. T., Jonson, C. L., & Nagin, D. S. (2011). Prisons Do Not Reduce Recidivism: The High Cost of Ignoring Science. The Prison Journal. 91(3). 48-65. DeLisi, M., & Vaughn, M. G. (2014). Foundation for a temperament-based theory of antisocial behavior and criminal justice system involvement. Journal of Criminal Justice 42(1). 10-25. Faria, M. A. (2013). Violence, mental illness, and the brain – A brief history of psychosurgery: Part 3 – From deep brain stimulation to amygdalotomy for violent behavior, seizures, and pathological aggression in humans. Surgical Neurology International. 4. 91. Farrington, D. P. (1986). Age and Crime. Crime and Justice. 7. 189-250. Farrington, D. P. (2013). Longitudinal and Experimental Research in Criminology. Crime and Justice. 42(1). 453-527. Fletcher, D. R., & Batty, E. (2012). Offender Peer Intervention: What do we know? Sheffield: Sheffield Hallam University. Fogg, B. J. (2009). A Behavior Model for Persuasive Design. Proceedings of the 4th International Conference on Persuasive Technology. Persuasive '09. New York: ACM: 40:1–40:7. Fox, J., & Weisberg, S. (2010). An R Companion to Applied Regression (2nd edn). California: SAGE Publications. Fox, C., Fox, A., & Marsh, C. (2014). Personalisation in the criminal justice system: what is the potential? Policy briefing. London: Criminal Justice Alliance. Fuhr, D. C., Salisbury, T. T., De Silva, M. J., Atif, N., van Ginnekan, N., Rahman, A., & Patel, V. (2014). Effectiveness of peer-delivered interventions for severe mental illness and depression on clinical and psychosocial outcomes: a systematic review and meta-analysis. Social Psychiatry and Psychiatric Epidemiology. 49. 1691-1702. Gallupe, O., McLevey, J., & Brown, S. (2018). Selection and Influence: A Meta-Analysis of the Association Between Peer and Personal Offending. Journal of Quantative Criminology. 1-23. https://doi.org/10.1007/s10940-018-9384-y. Gato, W. E., Posick, C., Williams, A., & Mays C. (2018). Examining the Link Between the Human Microbiome and Antisocial Behavior: Why Criminologists Should Care About Biochemistry, Too. Deviant Behavior. 39(9). 1191-1201. Gaudet, S., & Robert, D., R. (2018). A Journey Through Qualitative Research: From Design to Reporting. London: SAGE Publications Ltd. Gilbert, N., & Doran, J. (2018). Simulating Societies: The Computer Simulation of Social Phenomena. London: Routledge. Gillard, S., Foster, R., Gibson, S., Goldsmith, L., Marks, J., & White, S. (2017). Describing a principles-based approach to developing and evaluating peer worker roles as peer support moves into mainstream mental health service. Mental Health and Social Inclusion. 21(3). 133-143. Gillespie, A., Howarth, C., & Cornish, F. (2012). Four problems for researchers using social categories. Culture and Psychology. 18(3). 391-402. Goel, M. K., Khanna, P., & Kishore, J. (2010). Understanding survival analysis: Kaplan-Meier estimate. International Journal of Ayurveda Research. 1(4). 274-278. Gomes, H. S., Maia, A., & Farrington, D. P. (2018). Measuring offending: self-reports, official records, systematic observation and experimentation. Crime Psychology Review. 4(1). 26-44. Hardy, J. (2018). Parental Incarceration’s Effect on Family: Effects on Mothers, Fathers, Marriage, Children, and Socioeconomic Status. Canadian Journal of Family and Youth. 10(1). 119-140. Henshilwood, C., & Marean, C. (2003). The Origin of Modern Human Behavior: Critique of the Models and Their Test Implications. Current Anthropology, 44(5), 627-651. Hirschi, T. (1969). Causes of Delinquency. California: University of California Press. Hohner, C. (2017). 'The environment says it's okay': The tension between peer support and police culture. Doctoral Thesis, The University of Western Ontario. Available at https://ir.lib.uwo.ca/etd/5034/. Hollar, D.W. (2018). Trajectory Analysis in Health Care. Basel: Springer International Publishing. Houser, K., McCord, E. S., & Nicholson, J. (2018). The Influence of Neighborhood Risk Factors on Parolee Recidivism in Philadelphia, Pennsylvania. The Prison Journal. 98(3). 255-276. Jacobson, J. & Fair, H. (2017). Sense of self and responsibility: a review of learning from the Winston Churchill Memorial Trust Prison Reform Fellowships – Part V. London: Institute for Criminal Policy Research at Birkbeck, University of London. Jain, Y. & Jain, P. (2018). Communitisation of healthcare: peer support groups for chronic disease care in rural India. The BMJ. 360(85). 1-5.
Jolley, M. (2017). Rehabilitating prisoners: the place of basic life skills programmes. Safer Communities. 17(1). 1-10.
Jolliffe, D., Farrington, D. P., Piquero, A. R., Loeber, R., & Hill, K. G. (2017). Systematic review of early risk factors for life-course-persistent, adolescence-limited, and late-onset offenders in prospective longitudinal studies. Aggression and Violent Behavior. 33. 15-23.
Johnson, L. T. (2016). Drug markets, travel distance, and violence: Testing a typology. Crime & Delinquency, 62, 1465-1487. Joseph, R. P., Daniel, C. L., Thind, H., Benitez, T. J., & Pekmezi, D. (2017). Applying Psychological Theories to Promote Long-Term Maintenance of Health Behaviors. American Journal of Lifestyle Medicine. 10(6). 356-368. King, M., Freiberg, A., Batasol, B., & Hyams, R. (2014). Non-Adversarial Justice (2nd edn). Sydney: The Federation Press. Kish, L. (1965). Sampling Organizations and Groups of Unequal Sizes. American Sociological Review. 30(4). 564-572.
Kling, J. R. (2006). Incarceration length, employment, and earnings. American Economic Review. 96. 863–76.
Kwasnicka, D., Dombrowski, S. U., White, M., & Sniehotta, F. (2016). Theoretical explanations for maintenance of behaviour change: a systematic review of behaviour theories. Health Psychology Review. 10(3). 277-296. Lacey, N. (2008). The Prisoners’ dilemma in England and Wales. In M. Hough, R. Allen & E. Solomon (Eds.) Tackling Prison Overcrowding: Build More Prisons? Sentence Fewer Offenders? (pp. 9-23). Bristol: Policy Press.
Lakhan, S. E., Vieira, K., & Hamlat, E. (2010). Biomarkers in psychiatry: drawbacks and potential for misuse. International Archives of Medicine. 3(1). 1-6.
Makarios, M., Steiner, B., Travis, L. F. (2010). Examining the predictors of recidivism among men and women released from prison in Ohio. Criminal Justice and Behavior, 37, 1377-1391.
Markus, H., & Nurius, P. (1986). Possible selves. American Psychologist, 41(9), 954-969.
McCuish, E. C., Corrado, R. R., Hart, S. D., & DeLisi, M. (2015). The role of symptoms of psychopathy in persistent violence over the criminal career into full adulthood. Journal of Criminal Justice. 43(4). 345-356.
Mears, D. P., & Cochran, J. C. (2018). Progressively Tougher Sanctioning and Recidivism: Assessing the Effects of Different Types of Sanctions. Journal of Research in Crime and Delinquency. 55(2). 194-241.
Mennis, J., Harris, P. (2011). Contagion and repeat offending among urban juvenile delinquents. Journal of Adolescence, 34, 951-963.
Moffitt, T. E. (1993). Adolescent-limited and Life-Course-Persistent Antisocial Behaviour: A Developmental Taxonomy. Psychological Review. 100(4). 674-701.
Mossière, A., & Serin, R. (2014). A critique of models and measures of treatment readiness in offenders. Aggression and Violent Behavior. 19(4). 383-389. Narain, T., & Adcock, L. (2017). Peer Support Programs for Adults Who Have Experienced Sexual Assault, Abuse, Harassment, or Misconduct: A Review of Clinical Effectiveness and Guidelines. Ottowa: Canadian Agency for Drugs and Technologies in Health. Newsome, J. & Cullen, F. T. (2017). The Risk-Need-Responsivity Model Revisited: Using Biosocial Criminology to Enhance Offender Rehabilitation. Criminal Justice and Behavior. 44(8). 1030-1049. Office for National Statistics (2018). Crime in England and Wales: year ending December 2017. Newport: ONS. Perrin, C. (2017). The Untapped Utility of Peer-Support Programs in Prisons and Implications for Theory, Policy, and Practice. Doctoral Thesis, Nottingham Trent University. Available at http://irep.ntu.ac.uk/id/eprint/32821/. Petkovsek, M. A, Boutwell, B. B., Barnes, J. C., & Beaver, K. M. (2016). Moffitt’s Developmental
Taxonomy and Gang Membership: An Alternative Test of the Snares Hypothesis. Youth Violence and Juvenile Justice. 14(4). 335-349. Pike, A. & Hopkins, S. (2018). 'Education is Transformational': Positive identity through prison-based Higher Education in England and Wales. International Journal of Bias, Identity and Diversity in Education. (In Press). Police Scotland (2018). Vetting: Standard Operating Procedure. Fife: Professional Standards. Prochaska, J. O., & DiClemente, C. C. (1992). Stages of change in the modification of problem behaviors. Progress in Behavior Modification. 28. 183-218. Quick, D., & Choo, K. R. (2014). Impacts of increasing volume of digital forensic data: A survey and future research challenges. Digital Investigation. 11(4). 273-294. Raine, A. (2018). Antisocial personality as a Neurodevelopmental Disorder. Annual Review of Clinical Psychology. 14. 259-289. Reckless, W. C. (1961). A new theory of delinquency and crime. Federal Probation. 25. 42–46. Royster, M. (2012). The Success and Failure of Scared Straight: A Reassessment of Juvenile Delinquency Deterrent Methods and their Measurements. International Journal of Interdisciplinary Social Sciences. 6(8). 145-151. Rycroft-Malone, J., Fontenla, M., Seers, K., & Bick, D. (2009). Protocol-based care: the standardisation of decision-making? Journal of Clinical Nursing. 18(10). 1490-1500. Sawyer, S. M., Azzopardi, P. S., Wickrenmarathne, D., & Patton, G. C. (2018). The age of adolescence. The Lancet Child & Adolescent Health. 2. 223-228. Scottish Offenders Index & Scottish Government Justice Analytical Services. (2017). Reconviction Rates in Scotland: 2014-2015 Offender Cohort. Edinburgh: SOI. Schaie, K. W. (2005). What Can We Learn From Longitudinal Studies of Adult Development. Research in Human Development. 2(3). 133-158. Schnepel (2016). Good Jobs and Recidivism. The Economic Journal. 128(2). 447-469. Schofield, P.W., Malacova, E., Preen, D. B., D’Este, C., Tate, R., Reekie, J., Wandan, H., & Butler, T. (2015). Does Traumatic Brain Injury Lead to Criminality? A Whole Population Retrospective Cohort Study Using Linked Data. PLoS ONE. 10(7). 1-12. Sedgwick, P. (2014). Retrospective cohort studies: advantages and disadvantages. BMJ. 348. 1-2. Shulman, E. P., Steinberg, L., & Piquero, A. R. (2013). A Mistaken Account of the Age–Crime Curve. Journal of Adolescent Research. 29(1). 25-34. Slobogin, C. (2017). Neuroscience nuance: dissecting the relevance of neuroscience in adjudicating criminal culpability. Journal of the Law and the Biosciences. 4(3). 577-593. Smith, V., Devane, D., Begley, C. M., & Clarke, M. (2011). Methodology in conducting a systematic review of systematic reviews of healthcare interventions. BMC Medical Research Methodology. 11(15). 1-6. social control. Oxford Reference. Retrieved 19 Jun. 2018, from http://www.oxfordreference.com/view/10.1093/oi/authority.20110803100515340. Sullivan-Toole, H., Dobryakova, E., DePasque, S., & Tricomi, E. (2018). Reward circuitry activation reflects social preferences in the face of cognitive effort. Neuropsychologia. 1-47. https://doi.org/10.1016/j.neuropsychologia.2018.06.011. Sykes, G. M., & Matza, D. (1957). Techniques of Neutralization: A Theory of Delinquency. American Sociological Review. 22(6). 664-670. Szifris, K., Fox, C., & Bradbury, A. (2018). A Realist Model of Prison Education, Growth, and Desistance: A New Theory. Journal of Prison Education and Reentry. 5(1). 41-62. The Scottish Government (2016). Health and Social Care Delivery Plan. Edinburgh: The Scottish Government. The Scottish Government (2017). Recorded crime in Scotland: 2016-2017. Edinburgh: The Scottish Government. Tonry, M. (in press) Punishment and Human Dignity: Sentencing Principles for Twenty-First Century America. Crime and Justice. 47. 119-157.
Ttofi, M. M., Farrington, D. P., & Piquero, A. R., DeLisi, M. (2016). Protective Factors against Youth Offending and Violence: Results from Prospective Longitudinal Studies. Journal of Criminal Justice. 45. 1-100. Urwin, J. (2018). The Current State of Youth Justice. In J. Urwin (Ed) A Return to Social Justice: Youth Justice, Ideology and Philosophy. London: Palgrave Macmillan. Vachon, D. D., Lynam, D. R., Schell, S. E., Dryburgh, N.S., & Costa, P. T. (2018). Teenagers as Temporary Psychopaths? Stability in normal adolescent personality suggests otherwise. Personality and Individual Differences. 131. 117-120. Verdejo-Román, J., Bueso-Izquierdo, N., Daugherty, J. C., Pérez-García, M., & Hidalgo-Ruzzante, N. (2018). Structural brain differences in emotional processing and regulation areas between male batterers and other criminals: A preliminary study. Social Neuroscience.1-8. DOI: 10.1080/17470919.2018.1481882. Ward, T., Yates, P. M., & Willis, G. M. (2012). The Good Lives Model and the Risk Need Responsivity Model - A Critical Response to Andrews, Bonta, and Wormith (2011). Criminal Justice and Behavior. 39(1). 94-110. Skogan, W. G. (1975). Measurement problems in official and survey crime rates. Journal of Criminal Justice. 3(1). 17-31. Welsh, B. C., & Farrington, D. P. (2000). Monetary Costs and Benefits of Crime Prevention Programs. Crime and Justice. 27. 305-361. Whitelaw, S., Baldwin, S., Bunton, R., & Flynn, D. (2000). The status of evidence and outcomes in Stages of Change research. Health Education Research. 15(6). 707-718. Wilson, D B, Brennan, I, & Olaghere, A. (2018). Police-initiated diversion for youth to prevent future delinquent behavior: A systematic review. Campbell Systematic Reviews. Available at: https://doi.org/10.4073/csr.2018.5. Youngs, D., & Canter, D. (2014). When is an offender not a criminal? Instrumentality distinguishes self-reported offending of criminals. Journal of Criminal Psychology. 4(2). 116-128.