Face-to-Face Restorative Justice Conferences for Intimate Partner Abuse: an Exploratory Study of Victim and Offender Views

In what proportion of intimate partner abuse cases (in which offenders have made some level of formal admission of wrongdoing) would victims, offenders or both be willing to attend a face-to-face restorative justice conference, at least in principle? The study included offenders, victims and dyads in 75 low-harm cases of intimate partner abuse (IPA), defined as punishable by 10 or fewer days of Cambridge Crime Harm Index value for the offence. Eligible cases were limited to couples involving a male offender who had admitted responsibility for the harm and a female victim, both age 18 years or over, in cases finalised within 6 months of the incident and with no new incidents since the presenting offence. Of the 75 eligible cases of IPA for which contact was attempted with 150 people, 54 people agreed to grant interviews: 34 victims (45% of those successfully contacted) and 20 offenders (49% of successful contacts) completed interviews, including 13 cases in which both offender and victim in the dyad gave interviews. Structured phone interviews were conducted by police officers or staff within 10 months of case reporting. The key feasibility measure was a statement that the interviewee would be ‘likely’ to agree to participate in a face-to-face restorative justice conference (RJC) for crimes in an IPA relationship, based on a 5-point Likert scale. There were 14 (41%) victims and five (25%) offenders who indicated they would likely or definitely be willing to participate in a RJC if one were offered to them now. Seven of the 13 dyads matched on some degree of willingness to participate between both parties. Victims and offenders with some reason for ongoing contact with the other party tended to be more interested in RJCs than those who had severed contact. Victims reporting lower current levels of fear of the offender and who self-reported fewer prior offences were more willing to participate in RJCs. Cases in which victims indicated willingness had more often resulted in a caution than a charge. These findings demonstrated an appreciable level of interest in RJCs amongst victims and offenders in IPA cases with low harm. Based on the results of this study, a controlled trial could be considered offering this intervention to couples fitting the eligibility criteria of this study, with emphasis on couples found most likely to accept the offer.


Introduction
The use of face-to-face restorative justice conferences for domestic incidents has been widely seen as inappropriate by many stakeholders in intimate partner abuse (IPA). Yet outside of New Zealand (Hayden 2014;Kingi 2014), there is no evidence as to the views of victims and offenders on the prospect of such conferences. Given the rarity of effective responses to IPA (see Strang et al. 2013 for an exception), it could be in the best public interest to test a wider range of responses. The fact that restorative justice conferences (RJCs) have reduced repeat violence in other relationships (Sherman et al. 2015) should make it especially promising as a method to test for domestic abuse.
The RJC option might seem especially appropriate for couples who decide to continue to cohabitate, especially since most arrests are followed by immediate release of the abuser. It is clear from existing research that victims frequently call the police in domestic abuse cases for immediate protection, without intending or desiring that the perpetrator be prosecuted (Smith 2000;Curtis-Fawley and Daly 2005;Cornelius 2013). Problems with an unsupported arrest can be compounded by police pressing for thirdparty prosecutions, ignoring victims' wishes and potentially alienating them in the event of future offences (Smith 2000). Overly intrusive and punitive criminal justice system (CJS) interventions may deter victims from reporting abuse and therefore be more dangerous in the longer term (Smith 2000;Iyengar 2009).
It seems that more could be done to work with victims and offenders in IPA to determine an appropriate course of action in each case, rather than seeking a one-sizefits-all intervention. One such approach might be extending the use of face-to-face RJC to IPA cases.

Restorative justice conferences
A commonly used definition of RJCs is 'a process whereby all the parties with a stake in a particular offence come together to resolve collectively how to deal with the aftermath of the offence and its implications for the future' (Braithwaite 1999:1). These meetings involve direct communication between those harmed and those admitting responsibility for the harm, in the presence of a trained facilitator. Victims have the opportunity to participate in the resolution of the offence they have experienced whilst offenders are held accountable, leading to an agreement focused on how to repair the harm.
Although well-researched and implemented in other offence types (Strang et al. 2013), face-to-face RJCs have largely been banned for responses to IPA, at least in the UK, the USA and Australia. Opponents have raised concerns regarding victim safety, both physical and emotional, and fear that RJCs might be an opportunity for offenders to continue to exert control over their victims (Coker 1999;Curtis-Fawley and Daly 2005). Proponents of research in this area believe that RJCs could present an opportunity to reduce re-offending in an extremely harmful area (Sherman et al. 2015) and to heal victims emotionally and psychologically (Angel et al. 2014).
There is presently little empirical evidence on the effectiveness of RJCs in IPA (Curtis-Fawley and Daly 2005). Even less is known about the threshold question of whether there is any appetite for it amongst those who might be involved. Hayden (2014) and Kingi (2014) sought views in New Zealand, both regarding use of RJCs in principle and actually following participation, but there has been little research elsewhere. This is a major knowledge gap. If there is no interest in RJCs from IPA victims or offenders, then the arguments about its appropriateness are irrelevant. If there is an interest, however, then the people who would actually participate in such conferences are a vital source of information regarding format, safety concerns and practicalities.

Research questions
The key research question in this study concerns the likelihood of IPA victims and offenders participating in a restorative justice conference, once the process is explained to them. Interviewees responded to questions about their willingness to take part on a 5-point Likert scale. Those indicating some level of willingness were asked follow-up questions about how they thought RJCs should be run in IPA cases. All participants were asked background questions that assisted in establishing whether there were particular categories of people more inclined towards RJCs than others.

Data
This research aimed to establish the views about RJCs of both (female) victims and (male) offenders in recent low-harm IPA offences. Their views were gathered by the use of telephone interviews, taking place between 1 and 10 months after the offence that made them eligible.
Potential participants were identified through searches of Surrey Police's Niche Records Management System, which is an IT system for storing intelligence and crime reports. Whilst it is recognised that police data is not the best method of assessing total crime, and particularly in IPA where major underreporting exists (Gelsthorpe 2014), using police data is appropriate for this research because RJCs are almost always instigated by police or other justice agencies. A report to police is therefore the necessary starting point to get to a RJC, so it would not be feasible to offer one to anyone who had not reported an offence to police.
Efforts were made to keep the sample as homogeneous as possible, meeting eligibility criteria likely to be required if RJCs were offered. The criteria were set as follows: & Both parties age 18 or older at the time of the presenting offence & Male offender, female victim (the study was not large enough to attempt to gain meaningful results from other types of relationships) & Offender formal admission to the offence, resulting in either a police caution or a guilty plea at court (a standard pre-requisite for RJCs in other offences) & Both parties fluent in English (necessary for delivery of the interview schedule) & 'Low-harm' offences only, as indicated by the Cambridge Crime Harm Index (Sherman et al. 2016) threshold of 10 days or less of potential imprisonment & Recent but closed cases & No new incidents having occurred since the presenting offence The total number of crimes recorded by Surrey Police between 1 December 2016 and 31 July 2017 flagged as IPA was 2082. After applying the eligibility criteria, a total of 75 cases were eventually identified as eligible for inclusion in the study. This represented 3.6% of IPA crimes occurring in the time frame used. Of the 75 cases, 45 couples had no prior incidents recorded, 19 had one, eight had two to four and three had five to seven. Presenting offence types are shown in Table 1.
Of the total sample, 23% of victims and 39% of offenders could not be contacted. Of those with whom interviewers managed to establish some level of dialogue and explanation of what the call was about, 34 victims (59% of those contacted) were willing to complete the interview (45% of the overall victim sample) and 20 offenders (49% of those contacted) were willing (27% of the total offender sample). These totals included 13 complete dyads.

Methods
This study involved structured phone interviews with victims and offenders designed to last 15-20 min and conducted by one of three police officers or one member of police staff from a joint team of Surrey and Sussex Police. The three interviewers were chosen either for their experience of working in IPA or in criminological research and included male and female officers. Two separate interview schedules were designed, one for victims and one for offenders, with substantial overlap between the two. The questions were mostly forced-choice, but there were some open-ended questions which participants could answer at length. For example, after interviewees were asked how serious they felt the presenting offence had been in terms of its impact on the victim on a scale from 1 to 10, they were asked to explain in narrative form why they had given that score. In developing the schedules, ideas were drawn from the work of Coker (1999), Stubbs (2002), Busch (2002) and Curtis-Fawley and Daly (2005). All questions were reviewed by Surrey Police Domestic Abuse leads and by Victim Support Services.
The interview schedules started with background questions to gather information that was not consistently held by police, such as parties' ethnicities, whether they were still in some form of contact with the other party and whether they had engaged in any kind of relationship counselling or specialist IPA support previously.
The key research question concerning willingness to participate in a RJC was measured by use of a 5-point Likert scale from 'definitely not' to 'definitely would'. There were then follow-up questions for anyone other than the 'definitely nots' about what practical arrangements might help them to participate and what services or agreements they would like to see made available in the offender post-conference contract. Both sides were asked who they would want to support them at a RJC. The victims were asked what safety measures they would like to see in place during a conference and whether there was anyone they would not want the perpetrator to bring with them, in case of concerns around intimidation or other negative impacts on the conference.
Unless they said they were definitely not interested, interviewees were asked what they personally would hope to get out of attending a RJC. In the event that participants had indicated they would be unlikely to want to participate, or were unsure, they were asked to explain why, including whether they felt that a conference might make their situation worse.
At all times, the study was conducted with a keen awareness of safety concerns. It was subject to ethical oversight by the joint Surrey and Sussex Police Research Board and review by force Domestic Abuse leads and a key representative from a local domestic abuse Outreach service. The scripted sections read to participants by the interviewers included reassurances of confidentiality: in particular, assurances were given that their responses would not be discussed with the other party, or even whether the other had taken part. Checks were made at the start of each call about whether the call could be overheard by anyone. All victims were offered details of their local Outreach and Victim Support services at the end of the interview. Offenders were offered a referral to a Surrey Police voluntary perpetrator programme called Stepping Up. All victims were given the opportunity to make official reports regarding any new offences they disclosed and to discuss how they wanted any reports dealt with (in fact, only one new report was created; as this was a current and ongoing risk to the victim, she was informed that it needed to be recorded and investigated). None of the victims who disclosed previously unreported offences wanted them recorded or investigated.

Findings
Of the 34 victims and 20 offenders ultimately interviewed (including 13 complete dyads), 14 victims (41%) and five offenders (25%) said they were likely or definitely willing to participate in an RJC. One victim was unsure about whether she might participate, as were three offenders. There were several interviewees, however, who indicated that, although they were not interested in a RJC at the time of the interview, they would have been if it had been offered closer to the time of the offence. Those whose presenting offence involved some sort of threat or unwanted contact tended not to be interested in RJCs.

Current contact and RJCs
At the time of their interviews, 70% (38) of participants still had some form of contact with the other party: 58% (22) of those were still in the same relationship and 34% (13) had contact for children. Victims tended to be more interested in RJCs where there was ongoing contact but offenders were more evenly divided.
There were 22 participants who said they were definitely not interested in a RJC: 14 victims (41%) and eight offenders (40%). Seven of these victims had no ongoing contact at all with the offender; only one offender was no longer in contact with the victim.
Of the 20 participants who were probably not willing to participate, unsure about participation, or probably willing to participate, only one interviewee no longer had any contact at all with the other party. The others were mostly either continuing the relationship or in contact for children.
There were 12 participants who said they would definitely take part in a RJC if it were offered to them: nine victims and three offenders. Three victims were still in a relationship with the offenders, and three more had contact for children, whilst one offender had contact just in relation to his children.
Overall, victims were less interested in RJCs when they had stopped all contact with the offender. However, two victims and two offenders with no further contact were very interested in RJCs.

Prior offences and RJCs
Some measures of previous under-reporting were included in the victim interviews. Each victim was asked how many times in the 12 months before the presenting offence they had been physically hurt by, and how many times other offences had been committed against them by, the offender. These two figures were summed to give an estimated number of prior offences. Five victims said that offences had happened with such regularity they were unable to quantify the number of priors, but only one of them had more than one previous incident recorded on the police data base.
Of the 75 eligible couples, 45 had no prior incidents recorded in the preceding 12 months. Of those 45, 21 victims were interviewed: 12 confirmed that the presenting offence had been a one-off, with nothing else occurring before or since, and nine said that there had been incidents prior to the presenting offence, ranging from once before to an unquantifiable number of occasions. This resonates with prior findings (Bland and Ariel 2015;Barnham et al. 2017) that whilst most domestic abuse dyads have no subsequent reports to police, a small 'power few' percentage of all dyads represent a large proportion of incidents.
There was greater interest in RJCs amongst victims where there had been fewer prior offences between the couple, either officially recorded or disclosed by the victim at interview (see Tables 2 and 3). Table 2 shows the numbers of interested participants in the overall sample by number of prior police-recorded offences, and Table 3 shows the numbers interested by number of prior incidents reported by victims at interview. Table 4 shows that there was more support for RJCs from victims whose offenders had received a caution for the presenting offence than amongst victims whose offenders had been charged. The five offenders who were interested in RJCs were spread across several categories.

Impact and fear
Victims tended to be more interested in RJCs where they reported a current level of fear of 5 or less on a 10-point scale (Table 5). Importantly, when offenders were asked to estimate the impact of an incident on their victims, they tended to underestimate it substantially compared with their victims' ratings.

Demographics and RJCs
There was no apparent relationship between interest in RJCs and demographic factors of ethnicity, age or length of relationship. In this data set, the vast majority of participants  The relationship lengths ranged from a few months to 39 years, with a median of about 5 years, both for those who did not wish to participate in a RJC and for those who thought they would be willing to do so.

Preferred format of RJCs
Those respondents expressing any willingness to participate in an RJC (20 victims and 12 offenders) were asked what format they thought conferences should take. The majority requested a choice of location as their first priority. Other requests included child care and a choice of times and the possibility of assistance with transport. The issue of supporters for each party goes to the heart of many debates on the use of RJCs in domestic disputes. It has been suggested (see, for example, Stubbs 2002) that power imbalances in an intimate partner relationship may be replicated in the RJC and that supporters may contribute to this problem. This possibility was recognised by some of the respondents. Whereas most wanted a family member or friend present, seven of these 20 victims were concerned that the   Finally, a small number indicated that they would not want people personally involved with them to be present, owing to embarrassment or wanting to keep such affairs private. Some victims would consider including outreach workers, or Independent Domestic Violence Advisors (IDVAs). Some offenders suggested a mental health worker or someone from Citizens' Advice Bureau. Very few of the victims interviewed raised any practical safety concerns, perhaps because those who felt less safe around the offender had said they were unwilling to participate in a RJC. None raised concerns about being faceto-face with the offender during the conference, or needing measures such as pre-recorded messages or separate rooms, as suggested by Curtis-Fawley and Daly (2005).
Half of all respondents willing to participate in an RJC (10 victims and six offenders) said they would have preferred a RJC instead of the offender being dealt with formally, and about half of these volunteered that it would have worked well in tandem with the CJS measures that were imposed, such as a condition attached to a caution or a community order.
In terms of outcome agreements from the RJC, many respondents found it difficult to come up with ideas, yet some preferences emerged. The most popular services for both victims and offenders were some sort of behavioural programme for the offender, such as a domestic abuse perpetrator programme or anger management (15 victims and three offenders), some form of therapy, whether for the offender on their own or for the couple together (12 victims and five offenders), and substance rehabilitation (eight victims and three offenders). One victim and two offenders thought it was a good idea to agree a period of separation from their partner after the RJC to allow what had been discussed to be processed without the other party present. Other suggestions included additional support around childcare and housing, so that the couple did not feel forced to stay together if they decided it was best to separate.

Motivations and reservations
Amongst respondents who indicated an interest in participating in RJC, there was a large degree of overlap between victims' and offenders' motivations. Themes of understanding, closure and offender accountability and self-improvement emerged. Surprisingly, given Coker's (1999) and Busch's (2002) concerns over the use of apologies and forgiveness as a control tactic in RJCs, these elements were not a common hope amongst either offenders or victims. Some offenders, however, saw a conference as an opportunity to explain their behaviour to their victim and to try to reassure the victims.
There was evidence of dissatisfaction with criminal justice system processes. Interviewees saw conferences as a way of getting help for the offender, couple or family without resorting to prosecution. Some interviewees on both sides indicated that they lacked understanding of their interpersonal situation partly because formal criminal justice processes had prevented dialogue. The use of bail conditions or restraining orders did not allow discussion of problems. In cases where parties eventually reconciled, the CJS process had delayed the healing process. Interviewees hoped that RJCs would offer an alternative to such measures. Offenders showed some indications of stigmatising shame (Braithwaite 1989) and resulting dissatisfaction with court, as indicated by these quotations from responses to the interviewers: Court can mean part of the story remains untold. A conference would give an opportunity to go through everything properly so both sides know how the other is feeling.
It would be nice to get someone's objective view on how things got to the situation we hadat the moment, I did something bad and I've been punished. We haven't really examined what else there was in our relationship that led up to it. Court…doesn't solve the underlying issues.
I think it would have been much better for us to have this opportunity to talk, rather than me getting arrested and then prevented from contacting each other. It might have allowed a better resolution.
These views are at odds with assertions that all domestic abuse cases need to go to court in order to send a strong message to offenders (Coker 1999;Curtis-Fawley and Daly 2005;Hudson 2002).
The motivations for many respondents who were willing to accept a RJC were centred on couples getting help for their situation via an impartial adjudicator in a safe environment for discussion of matters like child contact, a topic of particular interest to several offenders. Several victims saw this as a way to engage the offender with support services without the need for them to go through court. This echoed Cornelius' (2013) finding that victims often call police to get help for the offender, rather than to prosecute them.
In terms of reservations about the RJC process, several victims and offenders said they simply wanted no further contact with the other party: they felt the relationship was over and there was no prospect of reconciliation so any further contact was unnecessary. Others who still had contact said that they had already sorted out their issues and did not want to dwell on the presenting offence. These included some who said that they would have engaged in the process if it had been offered to them at the time, but were not interested by the time of the interview. In addition, some victims felt their offenders were unsuitable for the process because they were manipulative or controlling and they believed they were unlikely to change so a RJC would be pointless.
There were some expressions of concern from victims that a RJC might make things worse between the parties: these included the anticipated trauma of discussing the offence again and anticipated negative reactions to contact from their offenders. All of those with these views indicated unwillingness to participate.

Conclusion
Despite the limitations of the small sample in this study, the results clearly reveal a substantial level of interest amongst both victims and offenders in IPA cases for a RJC. Amongst those who were interviewed, a quarter of offenders and more than two fifths of victims said it was a process in which they would be interested. If we include those who said they would have been interested in a RJC if it had been offered sooner after their presenting offence, this increased to more than half of each group.
The findings suggest that offers to participate in RJCs could be targeted to couples who would be most likely to accept the offer. Whilst there were too few willing offenders to target offers on their characteristics, the results suggest some useful evidence for targeting offers of RJCs to victims.
Formal criminal justice disposal outcome One of the most important factors in whether victims wanted to engage with a RJC was the offender's formal criminal justice disposal outcome-there was more interest when the offender received a simple caution rather than being charged to court. Whilst this may be a spurious correlation in a causal sense, it may still be a useful predictor. Victims may be more interested in RJC given the absence of opportunity with a caution for the victim to tell her side of events. Cautioned offenders are not required to hear how the incident affected victims, and after a caution, offenders may minimise what happened. In cases disposed by a caution, RJCs may seem to offer a real alternative: victims might actually be listened to not only by the offender, but by police as well.
Number of prior offences A second important factor for victims is the number of prior offences against them committed by the offender. Those who had experienced fewer offences were more likely to want an RJC, perhaps seeing it as an opportunity to draw a line under abusive behaviour early on, and to move on more positively. In these cases, RJCs appeared to take on a more mediation-like role for many of the people interviewed, replacing cautions as a first intervention. By contrast, of the five victims who were abused so often they had been unable to quantify it, four were definitely not willing to engage in a RJC.
Current contact with offender At the time of the interviews, 70% of participants were still in contact with the other party, with more than half of those continuing the relationship. Those victims were more likely to consider an RJC than those not in contact.
Other factors Victims were also more interested in RJC if the nature of the presenting offence was an actual assault, with less uptake after threats and unwanted communication. The victim's current level of fear of the offender also seem to influence victims' preference (less interest in RJC when fear was higher). Demographic characteristics, DASH risk assessment grading by police and perceived seriousness of impact of the presenting offence did not appear to influence participants' willingness to participate in RJCs.

Study limitations
This study has many limitations that must be acknowledged. There are issues around sampling from police data, especially in IPA where under-reporting is such a serious problem (Gelsthorpe 2014): the entire eligible sample of reports included in this study was 3.6% of IPA crimes reported to Surrey Police within that time frame, and only 36% of contacts attempted led to successful interviews. Thus, the views reported in this study come from just over 1% of the overall IPA crimes in Surrey for the period. Furthermore, the limited agreement to interview rate amongst successfully contacted victims and offenders (45% and 49% percent, respectively) means great caution must be applied in drawing conclusions from the available data.
Despite these limitations, this study did achieve its principal goal of interviewing more low-harm IPA victims and offenders about this issue than had been achieved in previous studies, such as Kingi's (2014) and Hayden's (2014). It also appears to be the first study of willingness of couples to consider restorative conferences after an IPA incident. The study would also be straightforward to replicate using the tools designed for the interviews.
Furthermore, a benefit of the research that cannot be measured was that of reaching out to this population to show them that the police are interested in their views and attempting to do things differently and better. Many participants expressed their gratitude to the researchers for taking an interest and looking for new ways to help those involved in IPA.
There is certainly some appetite for RJCs amongst offenders and more particularly amongst victims, in low-harm IPA incidents to which police have been called that result in a sanctioned detection. Indeed, a small number of respondents thought RJCs could be of great benefit to them and their relationship. Where discussions have previously focused on whether or not RJCs should even be permitted for use in IPA, perhaps the time has come to explore it empirically (rather than philosophically) as a possibility. Actual field trials could also try to address opponents' concerns, by, for example, implementation in such trials of the safety measures suggested by the victims interviewed in this study.
The clearest outcome of the present research is a recommendation to conduct replications. This sample largely represented a White British population living in relatively affluent areas, with no major cities: ideally, replications should be conducted in multiple force areas with a more ethnically diverse, higher density population, or subject to higher overall levels of deprivation. Replications might find interactions with demographic data that this study did not. Also, this research looked solely at male-on-female abuse: it would be helpful to see whether opinions vary in other relationship types.
Even without replications, it should be possible to begin considering the parameters of a pilot to test RJCs in IPA, based on the findings from this research about which cases might be most suitable. The first couples offered this intervention might be those whose presenting offence was 6 months ago or less, with no more than one previous 'low-harm' offence in the preceding 12 months, no indicators of coercive control within the relationship and with the RJC being offered in addition to whatever disposal the police might have sought anyway. Although many of the interviewees in this study felt a RJC would be best in their situation instead of formal action, that might not be advisable as a first step. Testing it in addition to formal processes in IPA seems more cautious and, if shown to be effective, could then be tested as a diversion.
The response rate for victims in any future study of views on RJCs could be improved by arranging an initial introduction to the researcher through officers investigating presenting offences, or through follow-up visits with support agencies. For offenders, the possibility of RJCs could be broached with them upon leaving custody or after a voluntary interview, with a request for a viable phone number on which to contact them.
The study found that RJCs were of more interest to those people continuing some form of relationship with the other party and those where victims reported less current fear of the offender. Whether to exclude cases not fitting those criteria from any subsequent pilot remains unclear, however. This study's sample size was not large enough to allow extrapolation of these findings to wider populations, and the findings around these indicators were not consistent across categories of willingness to participate in RJCs. Since consent is required from both parties in order for a RJC to happen, an ethical principle of restorative justice can govern any field test, just as it did in 12 prior tests of RJCs (Sherman et al. 2015); it should always be left to each person involved to decide whether they want to participate or not, after being fully informed about the process. Now that research has demonstrated that there is an interest in the use of RJCs for low-harm IPA in the UK, perhaps it is time to consider designing and testing a safe model of such a process. The victims and offenders interviewed in this study expressed hopes of better emotional and practical resolutions. Previous research on other offences has suggested that such hopes can be realised through the use of RJCs. It may be a continued disservice to keep denying willing participants an intervention that could be of huge benefit, in terms of both healing them in the aftermath of an incident and preventing further offences in the future.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.