Restorative Justice is internationally recognised as one of the most powerful tools available to authorities in the justice sector who are seeking to re-empower the victims of crime, and discourage future offending by those who have committed crimes.
Previous research in New Zealand showed that 74% of victims of crime who had engaged in restorative justice conferences said they "felt better" after the process, and 80% said they would recommend restorative justice to others in similar situations.
In New Zealand, restorative justice is usually delivered through voluntary conferences, where the victim, the offender and community representatives meet and attempt to reach agreement on an appropriate response to the offending. Clearly addressing the harm which was caused allows victims to feel empowered, and challenges offenders to consider the consequences of their actions.
While recidivism (reoffending) by those who have committed crimes is a difficult behaviour to influence, international research has shown that restorative justice has a statistically significant effect in reducing future offending.
This new study accords with that body of research. It indicates that participation in restorative justice conferences reduces not only the rate, but also the frequency, of reoffending.
A previous study by the Ministry of Justice found that restorative justice conferences reduced reoffending rates and the frequency of reoffending (Reoffending Analysis for Restorative Justice Cases: 2008 and 2009, June 2011). This is a follow up to that study, examining reoffending rates and frequency of reoffending over the period from 2008 to 2011.
Based on the current research, restorative justice has increased the number of offenders who did not reoffend in the year following their conference (as opposed to a control group), and reduced the frequency of reoffending by nearly a quarter.
Methodology
Like the 2011 study, this follow-up study assessed the impact of restorative justice conferences on reoffending by comparing conferenced offenders to matched offenders who went through the Police diversion or court process.
Rigorous data verification, statistical modelling and matching processes were used. Conferenced offenders were compared with:
- offenders referred for restorative justice but who did not receive a restorative justice conference as the victim declined or the case was otherwise considered unsuitable (non-conferenced);
- other offenders meeting the restorative justice eligibility criteria who were not referred (other eligible); and
- a matched comparison group of offenders (a sub-set of the other eligible offenders, selected to match the demographic and offending characteristics of those who completed a restorative justice conference).
The primary point of comparison used for conferenced offenders was the matched comparison group. The study compared 2,323 conferenced offenders with 6,718 matched offenders.
Offenders in the study included those who:
- had been charged with an imprisonable offence involving a victim
- were aged 17 or over
- entered a guilty plea
- did not receive a custodial sentence, and
- were charged in a district court from which referrals were received by restorative justice providers included in the study.
The measures of reoffending used in this study were frequency of reoffending, rate of reoffending (the number of offenders within a group who reoffended), seriousness of reoffending and subsequent imprisonment rates.
This study used the same samples of conferenced cases from 2008 and 2009. The 2008 group was followed up over one to four years and the 2009 group was followed up over one to three years. In addition, cohorts of conferenced offenders from 2010 and 2011 were compared with matched comparison groups over the following 12 months and, for the 2010 cohort, two years.
The quality and completeness of restorative justice data varies from provider to provider. The recording of the numbers which uniquely identify a particular criminal charge (CRNs) was particularly variable. Data from restorative justice providers was only used where an accurate match was able to be made with the Courts case management system, which contains data on all criminal charges. This means the number of conferenced cases recorded in this study for a particular year is lower than the total number of conferences that occurred in that year.
As data quality has improved over time, however, a greater proportion and number of conferenced cases were able to be used in the analysis for 2010 and 2011. The net effect was that data from 2008 was provided by 7 restorative justice providers, with data from 12 providers being included for 2009 (in both years from a total of 26). Data from 2010 and 2011 was provided by 26 and 30 restorative justice providers respectively.The method employed in this study was the same as that used in the original study, except in 5 instances:
- The 2009 cohort increased in size as it included 67 additional cases that had a final court hearing date in 2010. This had only a minor impact on the final results.
- Reoffending was counted if the case was finalised 6 months after the end of the reoffending period, for all cohorts and follow-up time periods. This had a small impact on the final results.
- Diversions were included as proved reoffending. This had very little impact as very few of those who reoffend receive a diversion for a second or subsequent offence.
- Court cluster was included as a factor in the logistic regression models for predicting probabilities of reoffending over each time period. This change was made as there was substantial variation in overall reoffending rates by court cluster. This was the most important change from the method employed in the original report.
- Matching with the conferenced group was carried out on the basis of predicted probabilities of reoffending over the following 12 months and the case being in the same age group. This was used to improve the comparability between the conferenced and matched comparison groups, while not greatly reducing the number of cases which could be matched to a conferenced case.
The reoffending results for both the 2010 and 2011 cohorts of offenders were measured over a period when the overall number of prosecutions declined (as shown in Tables A1 to A4). This also affected the three and four year reoffending rates for the 2008 and 2009 cohorts. Much of this reduction was due to Policing excellence, which meant that a proportion of lower level offences that would previously have resulted in a prosecution were resolved in an alternative way. The extent to which this has affected the comparison between conferenced cases and comparable offenders, as well as cases that were referred but not conferenced, is unknown.
Whether a result between conferenced and matched offenders is statistically significant or not depends on the size and scale of difference between the two rates, and the sample size of the two groups². The standard level of confidence used throughout the report to determine significance is 90 percent, unless otherwise stated.
Key Results
The key findings of the report are that restorative justice had a statistically significant impact on the number of offenders who reoffended and, for those who did reoffend, the frequency of that offending.
On average, offenders who participated in a Police or court-referred restorative justice conference ("conferenced offenders"):
- committed 23 percent fewer offences than comparable offenders over the following 12 month period; and
- had a 12 percent lower rate of reoffending than comparable offenders over the following 12 month period.
Those results were statistically significant.
The percentage difference in the frequency of reoffending remained stable over the four-year period of the study. Although the two to four-year results did not meet the threshold for statistical significance, nevertheless the findings suggest that restorative justice may continue to have a positive impact on the number of offences committed over time.
The study also suggested that conferenced offenders were 28 percent less likely to be imprisoned for reoffending over the following 12 month period than comparable offenders. However, again that result was not statistically significant. It also needs to be viewed in light of the reoffending rates for high-level offending, which suggest restorative justice has no significant impact on the seriousness of reoffending.