Scolaris Content Display Scolaris Content Display

Restorative justice conferencing for reducing recidivism in young offenders

Esta versión no es la más reciente

Contraer todo Desplegar todo

Abstract

This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:

To evaluate the effects of restorative justice conferencing programmes for reducing recidivism in young offenders.

To explore process and implementation issues in relation to programme effects on recidivism and improving participants’ self‐concept.

Background

Description of the condition

A recent survey on the extent of youth crime in England and Wales indicated that 22% of young people aged 10 to 25 years admitted to committing an offence in the past 12 months, with close of half of these offences classified as severe (Roe 2008). According to the United States Department of Justice, while the overall percentage of juvenile arrests may have decreased slightly from 2007 to 2008, the overall rates remain problematic, with 16.2% of violent crime arrests and 26.1% of property crime arrests involving those under the age of 18 (U.S. Department of Justice 2009).

A range of factors have been identified that appear to increase the risk that a young person will engage in criminal behaviour, and that shape the frequency, duration and persistence of offending. These include genetic and biological (Garland 2002; Arsenault 2003), environmental (Rutter 2006; Rutter 2007) and psychosocial factors such as socio‐economic deprivation, family conflict, families with a history of criminal or antisocial behaviour and substance misuse (Farrington 1989; Farrington 1996). In general, most crime and most serious crime is committed by males more than females, although more attention has been paid to offending among females in recent years (Whyte 2009).

The age of criminal responsibility varies according to geographical location, with some countries setting it as low as seven years and others at 18 (Siegal 2008). In some jurisdictions young people can be viewed as 'minors' up until age 21, and in some countries the age of criminal responsibility may vary with gender, as in Iran where the age of criminal responsibility begins at nine for females, but 15 for males (Palme 1997). There is therefore no one definition of 'young offender'.

In New Zealand, 67% of offenders aged from 16 to 19 were reconvicted within two years of their previous offence (Triggs 2005). In England and Wales, the frequency of re‐offences for young offenders ranged from 40.2% in 2000 to 37.8% in 2007 (Ministry of Justice 2009). In Northern Ireland, re‐offending rates were reported as rising from 39.3% in 2005 to 41.8% in 2006 (Tate 2009). Reported recidivism rates are also likely to be an underestimate of the actual occurrence of the problem, as they only provide information on offences recorded by the judicial system, and do not take into account offences for which offenders have not been identified, or which victims have failed to report. Recidivism rates are also likely to be constrained further by the fact that even when a crime is brought to police attention, there is no guarantee it will be recorded. This decision will reportedly depend on three broader contexts: the current political context, the organisational context of policing priorities, and the situational context of how the crime was reported (Coleman 1996). A young person who has committed an offence can face a variety of outcomes, from informal warnings, through restorative cautions, to referral for prosecution. The evidence of effectiveness of interventions as measured by recidivism rates is varied.

Description of the intervention

Restorative justice is "a process whereby parties with a stake in a specific offence resolve collectively how to deal with the aftermath of the offence and its implications for the future" (Marshall 2003). Stakeholders include the victim, the offender (or perpetrator) and the broader community (Morris, 2002). Different names have been given to the process of restorative justice, including Victim‐Offender Mediation, Community Reparative Boards, Referral Orders, Family Group Conferencing and Circle Sentencing. Family Group Conferences are distinguished by the involvement of the young person’s family or support network in the overall process (Bazemore 2003).

Typically the process will begin with a conference facilitator contacting both the victim and the offender to explain the process and inviting them to attend, making it clear that their attendance is voluntary. At this time, both the victim and offender are asked to name an individual who will also be invited to attend as their supporter. All willing parties then meet with the facilitator, often in a neutral location. The conference proceedings begin with the offender being asked to describe the incident, followed by all other participants describing the impact of the incident on their lives. The victim is asked to identify the desired outcomes from the conference, followed by all participants contributing to the process of determining how the offender may best repair the harm caused by their actions. The conference ends with all participants signing an agreement which outlines the expectations and commitments decided upon as a result of this conference (Bazemore 2003).

The focus of the conference is not to determine guilt or innocence, but to consider an appropriate plan of action or “conference plan”. These plans generally include one or more of: an apology to the victim, reparation or restitution to the victim or community, work or service to the community, restrictions on conduct, and treatment, for example to help overcome an addiction. The plan will be tailored to the age of the young person.

These restorative conferences may be ordered by the courts (court‐ordered) or recommended as an alternative to court proceedings by the police or public prosecution officer (diversionary). The minimum participants in this conference are the young person who committed the offence, the conference co‐ordinator, a police officer (or a representative from the criminal justice system) and an appropriate adult, for example the young person’s parent or guardian. The victim is encouraged to attend but is under no obligation, and in some instances the victim may be represented by another party.

These conferencing models are also used in other areas, such as child maltreatment (Dignan 2001). In these circumstances, families or those with significant relationships with the child in question come together to work collaboratively to resolve issues and develop plans that address the child's safety and well‐being (Knoke 2009). This review however will focus only on conferencing procedures aimed at young people who have committed a crime.

How the intervention might work

There are a number of theories as to how restorative conferences bring about change. Some have suggested that young people justify their behaviour and participation in a crime internally, using "techniques of neutralization" (Matza 1964), such as denying that any real injury was caused by their actions, for example, viewing theft as ‘borrowing’, or a gang fight as a ‘private dispute’. The process of a restorative conference is thought to make it more difficult to maintain such justifications; face to face contact with the victim and the subsequent discussion, together with increased awareness of the harm their actions caused, is likely to challenge such denials and justifications. Bandura conceptualised this process as the Reversal of Moral Disengagement (Bandura 1990). Others have suggested that the process of restorative conferencing provides a remedial opportunity for moral development in young offenders who may hitherto have had limited exposure to morally formative experiences (Barton 2003). The offender’s apology and the victim’s forgiveness are generally regarded as essential parts of the process of emotional healing and the key to the successful outcomes of satisfaction and reducing recidivism (Retzinger 1996).

Sherman's Defiance Theory suggests that a possible cause of recidivism is the individual's perception that their punishment was 'unfair', leading to defiance of the law. Conversely, if they had perceived their punishment to be fair, it is more likely to have resulted in compliance with the law (Sherman 1993). A related theory comes from Tyler's theory of procedural justice, in which a young person who feels they have been treated fairly by the justice proceedings may be less likely to break the law again (Tyler 2005). For most individuals, a fair procedure is not just related to how favourable the outcome may be perceived to be, but also to factors such as neutrality, lack of bias, honesty, efforts to be fair, politeness and respect. There is evidence from young people to that effect, which they attribute to being part of a collaborative endeavour, rather than a bystander in normal court proceedings.

The stigmatisation of offenders that occurs within conventional justice procedures is thought to be associated with the experience of shame. Braithwaite (Braithwaite 1989; Braithwaite 2001) suggests that this stigmatising shame can be counterproductive, potentially alienating the offender. Restorative justice emphasises social processes that involve a disapproval of offending, rather than the offender, as this is thought more likely to have a more positive impact on a young person’s cognitive and social development and self‐esteem than conventional justice processes. By focusing on the action, rather than the person, restorative justice prevents stigmatisation. Although restorative justice aims to shame offenders by confronting a young person with the impact of his or her actions, the requirement that they make reparation provides the means of forgiveness (Maruna 2007). Shame in restorative justice is therefore 'reintegrative' and thought to bring about better outcomes, but by definition it hinges on the presence of the victim.

All of the above suggests that the restorative justice process may have a differential impact on a young person’s self‐concept (that is, the multi‐dimensional construct of identity, self‐esteem, self‐efficacy and personal agency (McAdams 1996; Eccles 2002; Ward 2007)) than traditional criminal justice approaches.

In Family Group Conferencing all parties are encouraged to accept responsibility for their actions, including the young person’s family, who play a part in the creation of and adherence to the conference plan (Walgrave 2003) . Previous evaluations of the conferencing model have also indicated a 38% decrease in recidivism for young offenders, compared to a 6% increase in recidivism for adults convicted of drink‐driving (Sherman 2000).

Critics argue that what began as an organic, community process has become an increasingly streamlined, police‐led process, with the risk that the police will improvise and act as judge and jury, thus eroding the legal rights of the young offender (Young 2001). In fact, a young offender remains entitled to consult a lawyer during this process, but more efforts are being made to have conferences facilitated by non‐police bodies and held in more neutral venues (Campbell 2006). Other criticisms of restorative conferencing include: net‐widening, in which those who commit very minor offences and who might previously have received a warning are caught up in a more substantial process (O'Mahony 2004); the potential trivialisation of crime, with the result that acts such as male violence towards women becomes a private matter to be dealt with away from the courts; and uncertainty about what the term ‘restorative justice’ means or what this process is able to ‘restore’: a sense of offender responsibility, offender sense of control or a belief that the overall process and outcomes were fair (Morris, 2002). In response others have pointed out that restorative procedures take crime more seriously than court proceedings by focusing on the harm caused by the offender’s actions and finding ways they can specifically make amends for their actions (Morris, 2002).

There is uncertainty about the effectiveness of a brief intervention to counter the years of social disadvantage and exclusion that many of these offenders are likely to have faced. A range of literature has explored the factors linked with youth offending and desistance from crime. In particular, the interrelationships between lack of educational, training and employment opportunity and involvement in youth crime have been well documented (Burnett 2004). Interventions that aim to address these areas have demonstrated effectiveness (Hayward 2004). Similarly, the links between substance misuse and youth offending have been well documented (Britton 2008). It is argued that addressing such needs where identified is essential to reduce future offending (Burnett 2004). One of the aims of restorative justice interventions is to address the implications of offending for the future. It therefore stands that addressing such issues may be relevant for the ‘successful’ outcome of an intervention.

There is also uncertainty about the impact of restorative conferences on the victim, when he or she attends. Some victims report feelings of empowerment and resolution of their grief and distress, leading to emotional healing, and state that the 'humanising' of the offender has, for some, minimised the fear they had as a 'victim'. Others, however, report more negative experiences, for example, when faced with a conference facilitator who was perceived to have an inappropriate style or be insufficiently well prepared; or when the offender had an non co‐operative attitude (was aggressive, argumentative or uninformative); or when the victim felt coerced into taking part (Umbreit 1994). It is possible that the victim’s experience and attitude towards the conference could have a contributing effect on the overall experience of the young offender, and a negative victim outcome could be considered an important adverse effect to take into account in the overall evaluation of the intervention.

There is also uncertainly about some aspects of process. The introduction of court‐ordered conferences mean that some offenders agree to participate because they feel that they have no choice, which may affect their motivation and compliance (Campbell 2006). The effectiveness of conferences may be a function of the severity of the young person's initial offence. Previous studies suggest that while restorative conferences can lead to a large drop in post‐intervention offending rates in severe offences of youth violence, this was not the case for those convicted for less severe offences such as shoplifting or property offending (Sherman 2000).

Why it is important to do this review

Despite the increasing use of restorative justice programmes, there is a relative lack of up to date, high quality evaluation of the effectiveness of these programmes on young offenders in particular. Most previous reviews have also taken a broad focus, for example including results concerning both adult and juvenile offenders (Miers 2001).

Although reduction of recidivism is not a primary aim of restorative justice, it is a key policy concern. Furthermore, while recidivism is one measure of ‘effectiveness’, it is by no means the only focus and the emergent evidence suggests that there are other ancillary benefits that are also relevant to policy‐makers.  As restorative conferencing in particular is becoming the more common course of action for young offenders across many countries, it is clearly of importance to determine primarily whether this approach does hold any demonstrable benefits specifically for this population in terms of reducing re‐offending over the standard judicial proceedings, and to explore whether there are any particular issues regarding implementation of the programmes that play a vital role in the outcome of the procedure.

Objectives

To evaluate the effects of restorative justice conferencing programmes for reducing recidivism in young offenders.

To explore process and implementation issues in relation to programme effects on recidivism and improving participants’ self‐concept.

Methods

Criteria for considering studies for this review

Types of studies

Randomised controlled trials or quasi‐randomised trials, where allocation to intervention was determined, for example, by day of week, are eligible. We will include studies comparing restorative justice conferencing programmes to management as usual.

Management as usual

Depending on the area in which the study has taken place, the definition of 'management as usual' may differ. For example, in Bethlehem, PA, management as usual would include formal adjudication through the magistrate court system (McCold 1998). In Indianapolis, IN, the usual course of action may include alternative diversionary practices, such as victim‐offender mediation, volunteer services, teen court or garden projects (McGarrell 2000).

Types of participants

Young people aged between seven and 21 who have admitted to or been found guilty of any offence and are eligible for a restorative justice programme.

While this wide age bracket is a necessary criterion to ensure this review remains internationally inclusive, it is noted that most trials in this area will concern young offenders in their teenage years.

Types of interventions

Restorative justice conferences that follow the previously outlined definition of a restorative justice programme are eligible, specifically programmes whereby "parties with a stake in a specific offence resolve collectively how to deal with the aftermath of the offence and its implications for the future" (Marshall 2003).

Programmes that include all of the following features are eligible.

  1. Minimum attendees must include the young person, a conference co‐ordinator, an appropriate adult (for example, the young person's parent or guardian) and a representative from a key stakeholder group (for example, the victim of the crime, a representative of the victim of the crime, a member of the community affected or a representative of the justice community).

  2. Young person is attending of their own free will.

  3. All parties are involved in the construction of a conference plan with a primary aim of reparation.

  4. Where appropriate, plans are approved by the appropriate authorities.

Types of outcome measures

Primary outcomes

A. Recidivism rate as measured by administrative data *

B. Post‐intervention rates of offending as measured by self‐report *

C. Young person’s sense of self following conference *, specifically measures reporting:

  • remorse;

  • recognition of wrong‐doing;

  • self‐perception.

Secondary outcomes

D. Characteristics of the post‐intervention offence *, specifically:

  • post‐intervention offending with the same offence as they were originally brought to attention for;

  • severity of the post‐intervention offence.

E. Young person’s satisfaction with overall process *

F. Victim/key stakeholder satisfaction with overall process *

G. Change in social circumstances, including:

  • employment;

  • education;

  • training;

  • substance misuse.

H. Adherence to conference plan

Where feasible, comparisons will be made at the following specific follow‐up periods:

  • one year after conference takes place;

  • two years after conference takes place;

  • more than two years after conference takes place.

We will present outcomes indicated by an asterisk (*) in a 'Summary of findings' table. Where data are insufficient, we may provide a narrative account of the outcomes. It is also acknowledged that attention must be paid to recidivism rates that reflect crimes committed prior to the intervention, but for which judicial proceedings have only begun after the intervention. Care will be taken to ensure that these rates are not falsely reflected as post‐intervention offending.

Search methods for identification of studies

Electronic searches

We will search the following databases:

  1. Cochrane Central Registrar of Controlled Trials;

  2. Applied Social Sciences Index and Abstract (ASSIA);

  3. Bibliography of Nordic Criminology;

  4. National Criminal Justice Reference Service (NCJRS);

  5. Index to Theses;

  6. PsycINFO;

  7. Social Sciences Citation Index;

  8. Sociological Abstracts;

  9. Social Services Abstracts;

  10. Social Care Online;

  11. CareData;

  12. Restorative Justice Online;

  13. MEDLINE;

  14. Scopus;

  15. Science Direct;

  16. LILACS;

  17. Education Resources Information Centre (ERIC);

  18. WorldCat.

We will use the strategy shown in Appendix 1 for CENTRAL and modify it appropriately for all other databases.

We will also use randomised controlled trial filters where appropriate. We will apply no language or date restrictions.

Searching other resources

We will scan bibliographies of included and excluded studies for possible additional references of interest.

We will also review the American Society of Criminology Conference Programme in order to identify relevant trials.

We will also search the European Forum for Victim‐Offender Mediation and Restorative Justice and the International Centre for Criminal Law Reform and Justice Policy, in order to identify additional relevant grey literature.

We will contact authors and key scholars to identify any additional ongoing or missed studies.

Data collection and analysis

Selection of studies

Three authors (GM, NC and NL) will independently select and assess studies to determine whether they meet the inclusion criteria for this review.  Any disagreements between the authors will be resolved through discussion with the review team.

Data extraction and management

Three authors (GM, NC and NL) will extract data independently and enter data into a piloted data extraction form (see Appendix 2). Any disagreements between the authors will be resolved through discussion with the review team. Data to be extracted will include the following.

  • Study characteristics: study author(s), year of publication, citation and contact details, study design, study duration, details of attrition, and risk of bias concerns.

  • Participant characteristics: total number randomised in each study, age of participants, gender distribution, geographical location of study, and type and severity of offence.

  • Intervention characteristics: aim of the intervention, who was present at the intervention, time between crime and intervention, duration of the intervention, number of interventions, source of the intervention (i.e. diversionary or court‐ordered), construction of a conference plan, and completion of a conference plan.

  • Comparison characteristics: form of 'management as usual', frequency of 'management as usual', duration of 'management as usual', time between crime and 'management as usual'.

  • Outcome characteristics: details on all primary and secondary outcomes, measures used, length of follow up and summary data, including means, standard deviations, confidence intervals and significance levels for continuous data and proportions for dichotomous data.

Assessment of risk of bias in included studies

Using the data extraction form, authors will independently assess each study for risk of bias and assign each selected study to one of the following categories as outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2009):

(A)   High risk of bias

(B)   Low risk of bias

(C)   Unclear or unknown risk of bias

We will base assessment of each study for risk of bias on the following criteria as outlined by the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2009):

  1. Sequence generation (was the allocation sequence adequately generated?)

  2. Allocation concealment (was allocation adequately concealed?)

  3. Blinding of participants, personnel and outcome assessors (was knowledge of the allocated intervention adequately prevented during the study?)

  4. Incomplete outcome data (were incomplete outcome data adequately addressed?)

  5. Selective outcome reporting (are reports of the study free of suggestion of selective outcome reporting?)

  6. Other sources of bias (was the study apparently free of other problems that could put it at a high risk of bias?)

One example of an 'other source of bias' may be studies including participants who have received more than one restorative conference for separate offences. This could lead to bias, as having received a higher dosage of the intervention than others, they may be more likely to demonstrate an effect.

Measures of treatment effect

For dichotomous outcome data (recidivism, re‐offending and adherence to conference plan), we will calculate effect sizes as odds ratios (OR) with 95% confidence intervals. We will convert continuous outcome data (self‐concept/satisfaction scores) into standardised mean differences (SMDs) and present with 95% confidence intervals.

Unit of analysis issues

Cluster‐randomised trials

Cluster‐randomised trials are possible in this research, as allocation to the intervention group may occur by jurisdiction or by community as opposed to by individual offender. We anticipate that investigators will have controlled for a clustering effect when presenting their results. We will contact authors for further information if this is unclear. If the clustering effect was not controlled for, we will request individual participant data to calculate an estimate of the intra‐cluster correlation coefficient (ICC). If individual participant data are not available, we will obtain external estimates of the ICC from similar studies or available resources. Once established, we will use the ICC to re‐analyse the trial data to obtain approximate correct analyses. We will then enter these data into RevMan (RevMan 2008) to analyse effect sizes and confidence intervals using the generic inverse variance method (Higgins 2009).

If insufficient information is available to control for clustering in this way, we will enter data into RevMan using individuals as the unit of analyses. We will then perform sensitivity analyses to assess the potential bias that may have occurred as a result of the inadequately controlled clustered trials. We will also perform sensitivity analyses if the ICCs were obtained from external sources.

Cross‐over trials

Cross‐over trials are unlikely in this area as participants will only be allocated to one option. Those who are part of both a restorative intervention and full standard court proceedings are likely to do so because the conference was terminated and court proceedings were implemented as an alternative. We will exclude those who are part of this group from this review.

Studies with multiple intervention groups

Multiple observations are a possibility in this area, hence the decision to define specific follow‐up intervals.

Studies with multiple intervention groups are also a possibility in this area. If two or more interventions groups are compared to an eligible control group, the intervention group that most closely follows the previously outlined definition of a restorative justice conference will be included in the meta‐analysis. The decisions made during this process will be clearly outlined in the review. Some studies may also include more than one control group, who are undergoing different yet equally eligible forms of "management as usual". In this situation, the control groups will be combined to create a single pair‐wise comparison. If this strategy poses a problem for investigation of heterogeneity, each group will be compared separately as part of the subgroup analyses. The sample size for the shared comparator group will be split accordingly for subgroup analyses to prevent the same comparator participants being included twice (Higgins 2009, Section 16.5.5).

Dealing with missing data

In the case of missing data, where possible, we will contact the original investigators to request the missing data. Where this is not possible, we will make assumptions regarding whether the data are ‘missing at random’ or ‘not missing at random’ and we will follow the recommendations of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2009).

Data that are missing at random (due, for example, to postal disruptions) can be ignored as the reason behind these missing data is unlikely to be related to the outcomes of the missing data. In this case, we will analyse data using an available case analysis.

Data that are not missing at random cannot be ignored as they are likely to be missing for reasons related to the outcomes of the missing data. For example, if a participant agrees to take part in the trial, but was unhappy with the outcome, or goes on to commit a post‐intervention offence, it may be unlikely that they will be willing to complete any follow‐up interviews or questionnaires on their experience. In this situation, where possible, we will impute missing data with replacement values.

Where dichotomous data are missing, we will impute data with the assumption that the patients experienced the less favourable outcome (for example, recidivism did occur).

Where continuous data are missing, we will impute data using a 'last observation carried forward' approach. Where cases are missing from the first outcome measure, sample mean values will be imputed

Where studies have missing summary data, such as missing standard deviations, we will derive these data where possible using calculations provided in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2009).

We will clearly specify the methods used to address any missing data in the results tables. If imputation is not possible, we will not include these data in the pooled analysis and we will clearly outline reasons for this decision in the text.

We will consider the implications on results of the strategies used to account for the missing data as part of the sensitivity analyses to address how sensitive the results are to changes in the assumptions made about them. For example, where continuous data cases are missing from the first outcome measure and sample mean values have been imputed, the imputation of alternative mean values will be explored.

We will address the potential impact of the missing data on the findings of this review in the discussion section of the review.

Assessment of heterogeneity

We will assess statistical heterogeneity using the Q statistic and its P value, the I2 statistic, and by visual inspection of the forest plots.

Assessment of reporting biases

We will assess publication and other reporting biases through the use of visual inspection of funnel plots along with trim and fill analyses.

Data synthesis

Meta‐analysis will be performed on the results assuming that at least 2 studies suitable for inclusion are found. Due to the expected heterogeneity among included studies, we will use a random‐effects meta‐analysis. When meta‐analysis is inappropriate, we will provide a narrative description of the study results alone. We will perform both fixed and random‐effects analyses as part of a sensitivity analysis.

Subgroup analysis and investigation of heterogeneity

If the number of included studies is sufficient, subgroup analyses will examine:

  • the differential effects of interventions by the severity of the offence, specifically whether those with more serious offences respond differently to the conferences than those with more minor offences;

  • the differential effects of interventions by the gender of the young person who has committed the offence;

  • the differential effects of interventions by the presence of the personal victim versus the presence of a victim representative or no victim at the conference; and

  • the differential effects of interventions by diversionary versus court ordered conferences.

Sensitivity analysis

We will perform sensitivity analysis to assess whether the findings of this review are robust to the decisions made in the process of obtaining them. We will perform sensitivity analysis by reanalysis, excluding studies according to study quality issues, including those with low sample size, high risk of bias, or high attrition and drop out rate, and whether randomisation occurred pre‐charge, post‐charge but pre‐sentencing, or post‐sentencing, and before or after agreement to participate in the study.