The RESTORE Program of Restorative Justice for Sex Crimes

Abstract

The article reports empirical evaluation of RESTORE, a restorative justice (RJ) conferencing program adapted to prosecutor-referred adult misdemeanor and felony sexual assaults. RESTORE conferences included voluntary enrollment, preparation, and a face-to-face meeting where primary and secondary victims voice impacts, and responsible persons acknowledge their acts and together develop a re-dress plan that is supervised for 1 year. Process data included referral and consent rates, participant characteristics, observational ratings of conferences compared with program design, services delivered, and safety monitoring. Outcome evaluation used 22 cases to assess (a) pre–post reasons for choosing RESTORE, (b) preparation and conference experiences, (c) overall program and justice satisfaction, and (d) completion rates. This is the first peer-reviewed quantitative evaluation of RJ conferencing for adult sexual assault. Although the data have limitations, the results support cautious optimism regarding feasibility, safety, and satisfactory outcomes. They help envision how conferencing could expand and individualize justice options for sexual assault.

Keywords

restorative justice sexual assault rape sexual crime criminology victimology law and justice offender treatment victim–offender dialogue violence prevention

The term restorative justice (RJ) subsumes a variety of approaches to wrongdoing including crimes and student misconduct. RJ approaches share the viewpoint that violation of law and conduct codes causes negative impacts beyond those to the direct victim (for reviews, see McGlynn, Westmarland, & Godden, 2012) ; Naylor, 2010; Umbreit, Vos, Coates, & Lightfoot, 2006; for a short article aimed at general readers, see Koss & Achilles, 2008). From an RJ perspective, there are multiple victim constituencies including (a) direct victims, (b) family and friends of victims who suffer distress along with their loved ones, (c) family and friends of perpetrators who may experience shame, anger, and other emotions stemming from being part of an interpersonal relationship out of which the offense arose, and (d) community members who experience less safety and social connection when they perceive high levels of crime and low deterrence. RJ strives to balance fulfilling the justice expectations of victims with imposing perpetrator accountability.

Many RJ program formats exist such as sharing circles, victim–offender dialogue, victim impact panels, community reparation boards, circles of support, sentencing circles, and conferencing. The previously cited references provide more detail on these approaches as applied to a variety of juvenile and adult crimes including sexual assault. RJ programs are generally present and future oriented because they are intended for persons who acknowledge perpetration of wrong acts. Thus, RJ excludes processes that weigh evidence and deliberate fault. Instead, the emphasis is on opportunities for victims to make decisions about how their case proceeds, to express how the wrongdoing affected them, to experience acknowledgment of the wrongful act imposed on them, and to individualize the accountability that is imposed. RJ also aims to facilitate community affirmation of the norm violation and condemnation of the wrongdoers’ acts. Finally, RJ assumes that desistence from future offending is facilitated by maintaining wrongdoers’ connection with law-abiding citizens and supporting community re-integration if a period of exclusion has occurred.

RJ may be offered in various settings and at multiple time points. Within the criminal justice system, RJ approaches have been implemented at completion of police investigation, as pre-charging diversion, as components of post-charging plea agreements, post-conviction, during incarceration, immediately prior to or following release, and throughout the reintegration of the offender who has been returned to the community. The point where RJ options are offered is significant because progressively fewer victims are eligible as processing moves from crime occurrence through police report, law enforcement investigation, judicial review, and correctional supervision (prosecutor review, issuance of charges, plea negotiations, trials, sentencing, incarceration, and post-release). To date, RJ programs for adult sex crime have typically adopted victim–offender dialogue models. These interventions are offered when victims voluntarily request a meeting with their convicted or incarcerated offender (e.g., Miller, 2011; Patritti, 2010; Umbreit, Coates, Vos, & Brown, 2002). Victim–offender dialogue programs have not intentionally excluded sexual offenses, but inspection of the evaluation studies reveals that few such crimes were included and program designs were rarely adapted to the unique nature of sexual violation (reviewed in McGlynn, 2012; Naylor, 2010).

In contrast to methods for prison settings, RJ conferencing is typically conducted in law enforcement or community settings. It involves a face-to-face meeting where victims express harm, the perpetrator accepts responsibility, and participants develop an accountability plan. In the process of imposing accountability, conferencing strives to minimize negative social reactions and re-traumatization that may distance victims from their social network. Conferences have most often been used with juvenile crime (e.g., Daly, 2011; and for sexual abuse in therapeutic settings (e.g., http://www.brief-therapy.net/FinalRJreport.pdf). Search of scientific journals reveals few programs that focus on sex crimes involving adult victims and offenders. Those that do include the RESTORE Program in Pima County, Arizona, which is the focus of this article, Jülich and colleagues implementation of Project Restore-NZ in Auckland (2010), and Sten Madsen’s work in Copenhagen (2004, 2006).

Scholarly discourse on RJ for sexual assault has been hindered by lack of empirical data and is predominately conceptual and dialectic. Many commentators have raised concerns about the potential to reduce gender-based power dynamics, function safely, and exact sufficient accountability for wrongdoing (e.g., Cossins, 2008; Herman, 2005; Hudson, 2002; Matsui, 2011; Stubbs, 2007). A notable characteristic of this literature is the disproportionate focus on domestic violence or youth sex offending with inadequate attention to differences in crime characteristics from adult sexual assault (Hopkins, Koss, & Bachar, 2004). Other scholars have balanced risks with potential benefits (e.g., Curtis-Fawley & Daly, 2005; Daly, 2008a, 2008b, 2011; Daly & Curtis-Fawley, 2006; Daly & Stubbs, 2006; Jülich, Battle, Cummins, & Freeborn, 2010; Jülich, McGregor, et al., 2010; Koss, Bachar, & Hopkins, 2003; Koss, Bachar, Hopkins, & Carlson, 2004; McGlynn, 2011; McGlynn, Westmarland, & Godden, 2012; Nancarrow, 2010; Naylor, 2010; Stubbs, 2010; Vanseveren, 2010).

Just as there are few conferencing programs designed for sexual assault, published evaluations are scarce. The paucity of data has hindered the progression of dialogue from hypothetical to examination informed by program experience . The most extensive findings on RJ and gendered violence are based on archival analyses of the South Australia Juvenile Justice and Criminal Justice XXX on Conferencing and Sentencing and the and the South Australia Archival Study (Daly, 2006; Daly & Curtis-Fawly, 2006; Daly & Nancarrow, 2010; Daly & Wade, 2012; Daly, Bouhours, & Curtis-Fawley, 2007; Daly, Bouhours, Curtis-Fawley, Weber, & Scholl, 2007; Profetti, Scifoni, & Daly, 2011). BouhoursBouhoursThese projects involved youth conferencing programs that were not specific to sex crimes. The subset of sexual assault cases was often unique to youths such as sibling or peer abuse. Not reviewed here are unpublished evaluations of programs for familial sexual abuse or adult survivors of childhood victimization (e.g., http://www.brief-therapy.net/FinalRJreport.pdf). Published evaluations of adult conferencing programs to date have used qualitative data from case studies or file reviews to which quantitative methods have been applied (Bletzer & Koss, 2012, 2013; Jülich et al., 2010). A consistent limitation of this literature is small sample sizes ranging from approximately 5 to 10 cases.

This article contributes data from a quantitative process and outcome evaluation of RESTORE, a community-based RJ conferencing program for prosecutor-referred sex crimes involving adults. Process data include examination of recruitment flow and consent rates, conformance of conference components to the written guide book specifying how the program was designed to be delivered, and physical and psychological safety monitoring. Outcome data focus on participants’ self-reported reasons for choosing RJ, satisfaction with program components, procedural fairness, and completion rates. Henceforth in this article, RESTORE terminology will be used. The term survivor victim retains the empowerment conveyed by the word survivor and the outrage implied by the word victim. The term responsible person designates someone who committed an act that has been reported to police and viewed as a sex crime by prosecutors regardless of whether an arrest has been made or charges issued. Admitting responsibility is acknowledgment that the act occurred and is not synonymous with entering a guilty plea of guilty or self-identification as a rapist. By intention, designation of a survivor victim and a responsible person distributes power unequally to address concerns about deleterious influences of power dynamics. The term redress plan refers to the formal document of accountability that results from the conference and summarizes the activities that the responsible person will undertake to repair harm and rehabilitate.

RESTORE received referrals only from prosecutors. No self-referrals were permitted in this evaluation. Referrals included both misdemeanor and felony sex crimes. Felony crimes are sexual assaults, defined in Arizona statutes as oral, anal, or vaginal penetration against consent, forcibly or when incapacitated. Misdemeanor crimes are primarily indecent exposure with or without public masturbation. A primary rationale for implementing RESTORE was to provide an additional avenue that might reduce attrition in the criminal justice system. The term attrition refers to the large numbers of sexual assault cases that are closed at each stage of the justice system, cutting off survivor victims’ search for acknowledgment of their harm and a concrete response to it. It has been documented in every country that has been studied. Only 13% of reported rapes in the United States result in a finding of guilt (e.g., Daly & Bouhours, 2010). Even among this minority, many find the process re-traumatizing (Seidman & Pokorak, 2011; Seidman & Vickers, 2005). Survivor victims say that they desire a justice process that validates their status as legitimate victims, focuses on the offender’s behavior and not on theirs, provides a forum to voice the harm done to them, accords them influence over decisions about their case, and incorporates their input into the consequences imposed (Monroe, Kinney, Weist, Spriggs-Dafeamekpor, Dantzler, & Reynolds, 2005). Herman described the marginal role of sexual assault victims in the criminal justice system where their experiences constitute “humiliation reminiscent of the original crime” (p. 574). With reservations, she recommends that sexual assault advocates envision justice where survivor victims are “protagonists” rather than “peripheral actors” (Herman, 2005, p. 574).

RESTORE was intended as a justice process that expanded on justice options and responded in the ways survivor victims say they would like to be treated. The RESTORE Program is discussed in depth elsewhere (Koss, 2010). The following brief overview is provided to contextualize the present evaluation. RESTORE has four stages. They are presented as a flow model in Figure 1. Stage 1 is referral by prosecutors and informed consent to participate. Only on survivor victim consent is the program offered to the responsible person. Both parties are provided free access to legal counsel if desired to reach a decision. For the survivor victim, the choices include remaining in criminal justice, exploring civil justice options, or opting into RESTORE. For the responsible person, the decision is whether to participate in RESTORE or continue with standard criminal justice. Final enrollment is contingent on forensic assessment of the responsible person by an independent provider certified for this role in state and federal courts. Even while assessment is pending, stay away orders are implemented immediately. The purpose of forensic assessment is to exclude perpetrators whose undetected prior offenses or psychological characteristics make them unsuitable for a community-based program. Assessment consists of a guided clinical interview covering psychological symptoms and psychosexual life history. Widely used standard inventories are also administered including the Multiphasic Sex Inventory (Nichols & Molinder, 1996), the Millon Clinical Multitaxial Inventory–III (Millon, Millon, Davis, & Grossman, 2009), and the Sex Offenders Risk Appraisal Guide (Hanson & Thornton, 1999). Risk assessment after enrollment continued quarterly by the RESTORE staff using the Static-2002 (Hanson & Thornton, 1999; Helmus, Hanson, Thornton, Babchishin, & Harris, 2012).

Figure 1.

Operational process of RESTORE.

The second stage of Figure 1 is conference preparation. Here the goal is to ready all participants to meet face-to-face in a safe and constructive conference. Preparation consists of describing what will happen at a conference, answering questions, helping plan what to say, and guiding decisions about redress. The length of the second stage varies by each participant because survivor victims each have their own timetable to recover from initial trauma before they are able to speak and to contain their distress without humiliating loss of control. Responsible persons must achieve sufficient understanding of their acts to participate without traumatizing others through denial or blame. Finally, they must be familiar with the requirements that comprise their 12-month commitment to redress to avoid counter-productive resistance. The RESTORE redress plan consists of survivor victim-driven and program-imposed components. Required accountability and reparation includes sex offender therapy and any other intervention recommended by the forensic assessment (e.g., alcohol, anger management), monthly face-to-face meetings with a case manager, weekly check-up phone calls, quarterly meeting with the Community Accountability and Reintegration Board, community service, and compliance with stay away orders. Survivor victim–added activities are those with individual significance and constitute their input into accountability. Examples include selection of the type of community service, replacement of damaged property, contributions to charity in the survivor victim’s name, input into rehabilitative activities required of the responsible person, and payment of expenses for survivor victim therapeutic or reparative interventions.

The third stage in Figure 1 is the face-to-face conference. It is professionally facilitated by screened, trained, and compensated persons from various professions such as social work, law enforcement, counseling, and probation. Conferences are conducted according to a standard agenda but do not follow a script. Clear rules are stated and imposed to equalize communication opportunities, to prevent re-abuse of survivor victims, and to avoid excessive verbal shaming of responsible persons that might elicit dangerous or counter-productive anger and aggression (Massaro, 1997). Responsible persons begin the conference by describing their acts in sufficient detail to portray their offense. Survivor victims then speak about the distress and other impacts they experienced as a result of the responsible person’s acts. Next, their family and friends speak about their reactions followed by comments from those persons attending with the responsible person. Conferences conclude by finalizing the redress plan for the responsible person. Not all survivor victims desire a face-to-face meeting. When they prefer minimal participation and contingent on their consent, RESTORE proceeds using a surrogate victim. These are persons designated by the survivor victim or recruited by staff to attend conferences as a spokesperson for the direct victim including delivering an impact statement and participating in planning redress.

The final stage is monitoring of the responsible person’s redress plan fulfillment. Monitoring includes weekly phone contact and monthly face-to-face meetings with the RESTORE staff as well as and quarterly appearances before the Community Accountability and Re-Integration Board. This board consists of volunteers who represent the community in supporting the responsible person’s progress or in the case of non-compliance, terminating participation. RESTORE concludes with an exit meeting with the board where the responsible person presents a statement of accountability and reintegration that summarizes lessons learned and constitutes their formal apology.

Method

RESTORE was conducted in Pima County, Arizona (with a population of 989,569 people) by a collaboration of law enforcement, prosecution, sexual assault advocates, and public health professionals. Referrals were made by county and city prosecutors. Prosecutors’ referral criteria excluded repeat sexual offenders, persons with police reports for domestic violence, or individuals with arrests for any crimes involving non-sexual forms of physical assault. Enrollment criteria were subject to policies of the University of Arizona Institutional Review Board. The adult justice system often processes the cases of youths aged 15 to 17 years. Although they are adults from the justice perspective, under human subjects’ protection policy, they are viewed as children. The institutional review board deemed the safety record of restorative conferencing with juvenile offenders insufficient to outweigh the risks of including these youths. Therefore, all victims and offenders in the present study were 18 years or older.

Sample

Recruitment and consent

The flow of survivor victims and responsible persons through RESTORE is illustrated in Figure 2. The program operated from March, 2003, to August, 2007, and closed at the end of federal funding. Referrals were accepted during 2.5 years of this time due to two 6-month periods when all activities except supervision were suspended awaiting funding decisions. No new referrals were accepted in the last year to ensure completion of supervision. Figure 2 illustrates the progression of survivor victims and responsible persons from referral to program exit for both felonies and misdemeanors. This figure demonstrates the multiple steps involved in arriving at a consented case and provides raw numbers to estimate consent rates. Figure 2 illustrates that 64% of the 66 referrals were felonies and 36% were misdemeanors. Most referrals were pre-charging although five misdemeanors were post-plea. This article is based on the 22 cases where both survivor victim and responsible person consented to RESTORE. Each case involved multiple participants. A total of 109 individuals provided data at intake and 100 at post-conference (92% retention). Sample sizes for the subgroups of case participants at intake and post-conference were as follows: responsible persons, n = 20, 20; survivor victims, n = 11, 7; surrogate victims, n = 11, 11; minimal participation victims, n = 15, 13; responsible person family and friends, n = 23, 20; survivor victim family and friends, n = 19, 18, and volunteers, n = 10, 11, respectively.

Figure 2.

Participant flow from referral to program exit by crime type.

The first step after prosecutor referral was contacting survivor victims. Most could be reached (88% for felonies, 87% for misdemeanors). Among those contacted, the survivor victim consent rate was 63% for felonies and 70% for misdemeanors. The most common options selected by survivor victims who declined were standard criminal justice (38%) and civil justice (7%). Other reasons included lost desire for any form of criminal justice or the belief that too much time had passed. Responsible persons were contacted only after survivor victims’ consent was obtained. Lacking consent, responsible persons remained in criminal justice. The consent rate was 100% for misdemeanors and 90% for felonies among those responsible persons whose survivor victim consented, who could be reached, and met inclusion criteria. The felony consent rate would be 57% if calculations included the offenders who were not offered RESTORE because they denied responsibility for the incident. These persons maintained their lack of responsibility even though they did not have to use the word “guilty” or label the incident as “rape.” Deniers were excluded out of concern that they might pose safety risks to survivor victims including potential intimidation, verbal abuse, and/or retaliation.

Demographic characteristics

Table 1 displays the demographic characteristics of both referrals from prosecutors (n = 66) and consented cases (n = 22). The first section of Table 1 contains data on referrals. The data show that 4% of felony and 24% of misdemeanor survivor victims were men. One third of the male survivor victims were security guards who witnessed crimes on video surveillance. All but one referred offender was male (99%). Racial/ethnic data demonstrate that Caucasians comprised half or more of the referrals (54% of responsible persons and 64% of survivor victims). Responsible persons referred for felonies were younger (M = 31 years, range = 19-67 years) than those referred for misdemeanors (M = 39 years, range = 18-65 years). Regardless of crime type, survivor victims were younger than responsible persons (felony, M = 28 years, range = 18-49 years; misdemeanor, M = 31 years, range = 18-56 years). The percentage of youthful responsible persons aged 18 to 25 was 59% for felonies and 20% for misdemeanors. Survivor victim referrals in the 18 to 25 year age group were similar for felonies (62%) but approximately twice as high for misdemeanors (38%). Felonies as contrasted to misdemeanors were more likely to occur when the responsible person was drinking alcohol (50% vs. 21%). Felonies also more often involved friends or romantic partners than misdemeanors (57% vs. 21%). Among misdemeanors, however, acquaintances were limited to people such as a handyman, body worker, and school crossing guard. Few referrals involved intimate relationships between the survivor victim and responsible person (15% of felonies and 13% of misdemeanors).

Table 1.

Demographics of Referrals and Cases by Crime Type in Numbers and Percentages.

	No.			%^a
Descriptor	Felony	Misdemeanor	Sample	Felony	Misdemeanor	Sample
Referrals, n = 66
Gender
SV male/71 SVs^b	2	6	8	4	24	11
RP male/70 RPs^c	42	27	69	98	100	99
RP age 18-25 years	17	5	22	59	20	43
SV age 18-25 years	26	9	35	62	38	53
RP race/ethnicity
White, non-Hispanic	17	20	37	40	74	54
Hispanic of any race	10	7	17	24	26	25
African American	6	0	6	14	0	9
American Indian	0	0	0	0	0	0
Unknown/did not say	9	0	9	21	0	13
SV race/ethnicity
White, non-Hispanic	31	15	46	63	65	64
Hispanic of any race	10	5	15	20	22	21
African American	1	0	1	1	0	1
American Indian	0	0	0	0	0	0
Unknown/did not say	7	3	10	14	13	14
RP and SV were acquaintances	24	5	29	57	21	44
RP was drinking before offense	21	5	26	50	21	39
Cases, n = 22
SV male	0	6	6	0	55	27
RP male	11	11	22	100	100	100
RP age 18-25 years	9	2	11	82	18	50
SV age 18-25 years	6	2	8	55	18	36
RP race/ethnicity
White, non-Hispanic	9	8	17	82	83	77
Hispanic of any race	0	3	3	0	17	14
African American	2	0	2	18	0	9
American Indian	0	0	0	0	0	0
SV race/ethnicity
White, non-Hispanic	13	8	21	100	72	88
Hispanic of any race	0	3	3	0	27	13
African American	0	0	0	0	0	0
American Indian	0	0	0	0	0	0
RP and SV were acquaintances	11	0	11	100	0	50
RP was drinking before offense	8	2	10	83	20	46

Note. SV = survivor victim; RP = responsible person.

Some percentages do not total to 100 because only one response option is included in the table.

Includes two multi-victim felonies and one multi-victim misdemeanor were included. Excepting gender, the data represent the primary survivor victim and responsible person for each referral or case.

Includes one multi-offender misdemeanor.

The second section of Table 1 describes the 22 consented cases. All felonies involved female survivor victims but men comprised half of those victimized by misdemeanors. All responsible persons were men. As in the referral sample, a wide range of ages were served by RESTORE (range = 18-66 years). Likewise, felony survivor victims were slightly younger (M = 28 years) than responsible persons (M = 31 years) and the age difference was more notable in misdemeanor cases where survivor victims were on average 10 years younger (M = 32 years) than responsible persons (M = 42 years).

Young men were more highly represented among felony cases than they were in the referral sample (82% of cases vs. 59% of referrals). This trend was not observed for misdemeanors and the reverse was seen for survivor victims. RESTORE cases involved somewhat fewer young survivor victims than the referrals. The difference was small for felonies but pronounced for misdemeanors (38% of referrals vs. 18% of cases).

Acquaintance, intimate partnership, and alcohol use by the perpetrator before the crime were all higher among felony cases compared with misdemeanors. All felony cases involved acquaintances or romantic partners whereas all misdemeanor cases involved strangers. The representation of acquaintance crime among referrals compared with cases differed both with and between crimes. RESTORE participation compared with cases reveals that the program was selectively more attractive to acquaintances for felony crimes (57% of referrals, 100% of cases) and progressively less appealing for misdemeanors (21% of referrals, 0% of cases). Alcohol use preceded the crime more often in felonies than misdemeanors (83% vs. 20%). Program participation appeared to be selected by more responsible persons who had been drinking before the offense (83%) compared with the percentage among referrals (50%). There was no difference among misdemeanors. Most responsible persons had completed high school and 14% were higher education students. Many (45%) were unemployed and 14% were disabled. Forty percent were married and 60% were separated, divorced, or never married. Racial/ethnic data identify show that most RESTORE cases involved Caucasians (77% of responsible persons and 88% of survivor victims).

Process Monitoring

Process monitoring is intended to ensure that services are accorded equably, the assets required for implementation in staff time and community capacity are estimated, and the interventions are delivered as designed. Without a formal manual stipulating the components of the intervention and a method for assessing compliance, outcomes might be internally invalid because individual staff could deliver services idiosyncratically.

Service delivery

Data sources for process evaluation included clinical and research files, and nonparticipant observation of conferences. Variables extracted from files included demographics and service characteristics such as time delay from police report to referral, duration of each RESTORE stage, length of conferences, and number of staff hours devoted to each case. Variables assessed through conference observation included (a) implementation of specified physical arrangements (8 items), (b) facilitator conformance with stipulated behaviors and procedures (20 items), and (c) facilitator enforcement of rules for participant behavior (15 items).Training for data extraction and observational ratings consisted of written manuals, didactic presentations, role-playing, and observed practice. Inter-rater reliability was not obtained as the checklists involved minimal subjectivity. Raters included six persons (staff members = 3; graduate students = 3). Each individual rated a mean of three conferences.

Observer ratings demonstrated that the stipulated conference design was followed closely. Physical arrangements were virtually 100% in conformance with minor exceptions. For example, 3 of 20 conferences lacked a pre-arranged seating plan, seat labels, or tissues. Conference procedures also revealed close to 100% compliance. Among the exceptions were 9 of 104 conference attendees who were rated as lacking input into the redress plan. Facilitator compliance was also close to 100% including items such as covering all points on the agenda, guiding discussion of reparation, refraining from intimidation of participants, and avoiding reprimands of responsible persons or survivor victims. In a few instances, facilitators exhibited annoying mannerisms or failed to redirect discussion. They were coached or not included again.

Safety

Case managers’ clinical notes and conference observations were used to identify incidents of verbal re-abuse, intimidation, or physical danger. Psychological safety for survivor victims was monitored at intake and post-conference with the 17-item Post-Traumatic Symptoms Scale (PSS; Foa, Riggs, Dancu, & Rothbaum, 1993). This scale was developed with assault survivor victims. It yields a total score and a dichotomous classification of whether formal diagnostic criteria for post-traumatic stress disorder (PTSD) are met. Respondents reported the frequency of various symptoms during the prior month on the 4-point scale used by the original authors (0 = not at all, 1 = a little bit, 2 = somewhat, and 3 = very much). Examples of item content are “having bad dreams or nightmares about the trauma,” “trying not to think about, talk about, or have feelings about the trauma,” and “having trouble falling asleep.” The full scale score that was used in the present study had a Cronbach’s alpha of .91 and test–retest reliability over 1 month of .74 (Foa et al., 1993). Diagnosis of PTSD based on PSS scores was compared with clinical interview and concurrent validity was established. PTSD diagnosis was assigned in this study using the authors’ criteria (Foa et al., 1993) . These include the requirement that the trauma occurred 3 months or more before assessment and endorsement at a level of 1 (“a little bit”) or higher with at least one re-experiencing symptom, three avoidance symptoms, and two arousal symptoms. Foa and colleagues (1993) reported that 94% of rape victims met diagnostic criteria for PTSD at 2 to 3 weeks after rape and 65% still did so 3 months post-assault (Rothbaum, Foa, Riggs, & Murdock, 1992). Psychological and physical health were also monitored at intake and post-conference among responsible persons and survivor victims using author-constructed scales for which internal consistency data are not available. The number of items and sample wording included mental distress (four items, “sudden tearfulness”), arousal (seven items, “problems concentrating”), physical symptoms (three items, “loss of appetite”), and avoidance (three items, “feeling like you wanted to hide from family and friends”).

Outcome Assessment

Data collection was done by self-report with measurement points at intake and immediately post-conference except for survivor victims and responsible persons. They provided post-conference data 1 week later at the RESTORE office. In addition to data collection, this meeting allowed clinical assessment of any negative effects. Respondents provided ratings that represented how they felt at the moment with the exception of psychological assessment where the recall period was 1 week. Measurement from the South Australia Juvenile Justice project is acknowledged for guidance in developing the item content that is described next (Daly et al., 2006).

Reasons for choosing RJ

These variables were assessed by five items at intake and nine post-conference. Each item was preceded by the question “How important were each of the following issues when deciding to participate in RESTORE?” Participants responded by selecting strongly disagree (1), disagree (2), agree (3), or strongly agree (4). Item content is provided in an abbreviated form in Table 3. There were a few differences in wording for responsible persons. For example, most groups were asked if they selected RESTORE to hear an apology. Responsible persons were asked if they participated because they felt they should apologize. The alternative wordings are separated by a slash in Table 3.

Table 2.

Pre to Post Monitoring of Psychological Status.

Scale	Intake M	SD	Post-conference M	SD	df	Dependent t test (t)	p Value
Survivor victims^a
Mental distress	8.40	4.50	9.20	3.56	4	−0.825	.456
Arousal	18.50	7.89	15.75	8.57	4	1.33	.340
Physical symptoms	7.00	3.67	6.40	3.13	4	0.418	.697
Avoidance	5.60	3.97	5.40	3.78	4	0.196	.854
Responsible persons
Mental distress	7.17	2.54	6.50	2.93	18	2.06	.055
Arousal	9.22	3.40	8.22	3.00	18	1.98	.064
Physical symptoms	5.33	2.42	4.83	1.91	18	0.94	.360
Avoidance	5.17	2.20	4.67	2.45	18	1.14	.269

Survivor victims who elected not to participate in a conference were not asked to complete the psychological status measures. In addition, six survivor victims chose not to provide psychological status information.

Table 3.

Reasons for Choosing Restorative Justice in Percentages by Participant Type.

	Item Responses (%)^a
	SV Present^b (n = 11)			SV Absent^b (n = 15)			SV Surrogate^b (n = 11)			SV Family/Friends^b (n = 19)			RP^b (n = 20)			RP Family/Friends^b (n = 23)
Item: Please rate your agreement	1-2	3	4	1-2	3	4	1-2	3	4	1-2	3	4	1-2	3	4	1-2	3	4
Baseline reasons for participating in RESTORE
Say how I was affected/explain my side	9	45	45	—^c	—	—	46	50	14	30	45	25	10	80	10	40	50	10
Making the RP accountable/taking responsibility to make things right^d	0	20	80	6	27	67	7	39	54	5	5	90	0	50	50	—	—	—
Have input into punishment/getting punished	18	46	36	—	—	—	36	50	17	20	60	20	25	35	40	55	40	5
Hear an apology/apologize	27	9	64	40	40	20	0	29	71	20	35	45	5	53	42	50	40	10
Participate in an alternative to court	20	30	50	0	33	67	0	21	79	25	40	35	10	0	90	0	30	70
Post-conference reasons for participating in RESTORE
Say how I was affected/explain my side	0	0	100	8	23	69	7	31	62	0	25	75	0	45	55	11	42	53
Making the RP accountable/taking responsibility to make things right	15	15	70	—	—	—	0	46	54	6	26	68	0	28	72	10	55	35
Have input into punishment/getting punished	33	17	50	38	23	38	58	16	25	—	—	—	22	50	28	—	—	—
Hear an apology/apologize	14	14	71	—	—	—	17	33	50	—	—	—	5	26	69	—	—	—
Participate in an alternative to court	0	33	67	23	31	46	0	45	55	—	—	—	5	21	74	—	—	—
Make sure the RP gets help/to understand what happened	0	14	84	15	23	62	8	25	67	0	47	53	15	50	35	5	58	37
Making sure that the RP doesn’t do this to anyone else	0	0	100	0	23	79	8	8	83	—	—	—	5	0	95	—	—	—
Put this behind me	36	18	46	13	20	67	—	—	—	5	67	29	0	65	30	17	61	22
Take back my power	0	0	100	39	15	46	—	—	—	—	—	—	16	47	37	—	—	—

Note. SV = survivor victim; RP = responsible person.

Percentages reflect the proportions within each participant type who selected 1 = strongly disagree, 2 = disagree, 3 = agree, and 4 = strongly agree.

SV present are those who attended their conference. SV Absent are those selecting minimal participation. SV Surrogates are volunteers who represented survivor victims with their consent at conference. Family and friends are support persons. The sample size at intake was n = 109. Data for 10 community volunteers are omitted but available.

This item was not asked among this group of participants.

Alternative RP wordings are separated by a slash.

Satisfaction with preparation and conference experiences

Preparation experiences were measured with eight items covering the extent to which participants perceived that staff had readied them to meet face-to-face. The response scale was strongly disagree (1), disagree (2), agree (3), or strongly agree (4). Item content is provided in an abbreviated form in Table 4. Table 5 contains the same information for the 13 items that assessed conference experiences. Question format and response scale were the same as for preparation ratings. Four items in Table 5 are expressed in the negative so that higher numbers throughout the table uniformly represent desirable outcomes. These items are indicated by the presence of the word “NOT” in several Table 5 items. That word was not present in actual administration. Satisfaction data were obtained by six items that focused on the individual components of RESTORE as well as assessments of the overall justice experience. The response format was very dissatisfied (1), dissatisfied (2), satisfied (3), and very satisfied (4). The abbreviated text is found in Table 6.

Table 4.

Preparation Experiences in Percentages by Participant Type.

	Item Responses in Percentages^a
	SV Present (n = 7)^b			SV Absent (n = 13)			SV Surrogate (n = 11)			SV Family/Friends (n = 18)			RP (n = 20)			RP Family/Friends (n = 20)			Volunteer Supporters (n = 11)
Please Rate Your Agreement	1-2	3	4	1-2	3	4	1-2	3	4	1-2	3	4	1-2	3	4	1-2	3	4	1-2	3	4
Listened to me express feelings	0	0	100	8	23	69	7	31	62	0	40	60	0	55	45	10	45	45	0	20	80
Answered all my questions about participation	0	33	67	8	46	46	0	29	71	0	30	70	0	53	47	10	35	55	—^c	—	—
Answered my questions about RESTORE	0	0	100	8	23	69	0	15	85	0	25	75	5	30	65	5	42	53	0	40	60
Explained the purpose of the redress plan	0	17	83	8	38	54	0	31	69	0	20	80	0	40	60	10	35	55	0	30	70
Explained how the redress plan would be developed	0	17	83	8	46	46	7	23	69	0	30	70	0	45	55	10	45	45	0	30	70
Helped in figuring out what to say	0	0	100	—	—	—	9	46	46	5	40	55	0	35	65	17	50	33	—	—	—
Received good preparation	17	0	83	—	—	—	0	15	85	0	10	90	0	30	70	0	30	70	0	30	70
Treated me with respect	0	0	100	8	8	85	0	0	100	0	5	95	10	30	60	0	20	80	0	20	80

Note. SV = survivor victim; RP = responsible person.

Percentages reflect the proportions within each participant type who selected 1 = strongly disagree, 2 = disagree, 3 = agree, and 4 = strongly agree.

The sample sizes provided are from the post-conference assessment point that reflects 92% retention of the intake sample.

This item was not asked among this group of participants.

Table 5.

Conference Experiences in Percentages by Participant Type.

	Item Responses (%)^a,b
	SV Present			SV Surrogate			SV Supporters			RP			RP Supporters			Volunteer Supporters
Item: Please Rate Your Agreement	1-2	3	4	1-2	3	4	1-2	3	4	1-2	3	4	1-2	3	4	1-2	3	4
Felt safe	0	33	67	8	0	92	0	20	80	5	50	45	5	33	61	0	20	80
Felt listened to	0	16	83	—^c	—	—	0	15	85	0	60	40	5	37	58	10	20	70
Support during conference	0	17	83	—	—	—	0	40	60	0	42	58	6	50	44	—	—	—
Everyone was treated fairly	0	50	50	0	27	73	0	17	83	5	53	42	0	33	67	0	10	90
RP seemed to accept responsibility	33	0	66	25	50	25	15	45	40	—	—	—	12	35	53	0	30	70
RP seemed sincerely sorry/felt sincerely sorry	50	16	33	14	46	36	26	47	26	0	74	26	6	41	53	10	40	50
RP (DID NOT)^d apologize due to group pressure/felt group pressure	16	16	67	0	0	100	11	47	42	5	10	85	11	17	72	0	0	100
RP will (NOT) hurt again/I will not hurt someone again	17	33	50	0	18	82	5	5	90	5	10	85	0	5	95	0	5	95
It was (NOT) hard to talk openly	16	16	67	25	25	50	35	30	35	30	40	30	0	33	67	0	10	90
Felt like I was (NOT) blamed	0	0	100	0	0	100	15	20	65	21	26	53	17	11	72	0	0	100
Was (NOT) expected to do more than I should have	0	0	100	0	0	100	10	10	80	5	30	65	11	6	83	0	0	100
Was treated with respect	0	33	67	0	8	92	5	20	75	0	55	45	0	28	72	0	20	80
Redress plan was fair	0	40	60	33	25	42	5	40	55	32	37	32	16	37	47	0	20	80
The conference was a success	0	66	33	8	50	42	0	76	24	0	38	62	5	53	42	20	20	60

Note. SV = survivor victim; RP = responsible person.

Percentages reflect the proportions within each participant type who selected 1 = strongly disagree, 2 = disagree, 3 = agree, and 4 = strongly agree.

The group sizes are the same as presented in Table 4.

This item was not asked among this group of participants

Parentheses (NOT) indicate items that were reverse scored so that higher scores on all items represented positive experiences.

Table 6.

Overall Satisfaction in Percentages by Participant Group.

	Item Responses (%)^a,b
	SV Present			SV Absent			SV Surrogate			SV Supporters			RP			RP Supporters			Volunteer Supporters
	1-2	3	4	1-2	3	4	1-2	3	4	1-2	3	4	1-2	3	4	1-2	3	4	1-2	3	4
Preparation	0	17	83	0	62	38	0	25	75	0	38	62	0	55	45	5	50	45	0	40	60
Conference	0	17	83	—	—	—	0	17	83	0	33	67	0	55	45	5	40	55	0	40	60
Redress plan	0	33	67	0	56	46	0	33	67	0	24	76	10	40	50	10	45	45	0	40	60
How RESTORE handled the case	0	0	100	0	39	62	0	33	67	—^c	—	—	0	26	74	—	—	—	0	30	70
I feel justice was done^d	17	33	50	—	—	—	30	20	50	16	53	32	5	68	26	10	40	50	20	10	70
I would recommend RESTORE^d	0	0	100	8	31	62	16	50	33	0	33	67	0	50	50	5	47	47	0	30	70

Note. SV = survivor victim; RP = responsible person.

Percentages reflect the proportions within each participant type who selected 1 = very dissatisfied, 2 = somewhat dissatisfied, 3 = satisfied, and 4 = very satisfied.

The group sizes are the same as Tables 4 and 5.

This item was not asked in this group.

These two items were scored 1 = strongly disagree, 2 = disagree, 3 = agree, 4 = strongly agree.

Completion rates

Two types of completion rates were obtained from clinical files. The first was the percentage of cases that resulted in a conference being held. The second was the percentage of responsible persons who successfully completed all components of their redress plan and supervision and did not re-offend during the 12 months they were monitored by RESTORE.

Data Analysis

The items analyzed in the present study are those that were asked in a virtually identical form across participants. No data replacement was done. Daly (2006) reported disaggregated findings on satisfaction for misdemeanors and felonies but found no differences. Therefore, except for demographics, the present data were not disaggregated by crime type. Dependent t tests were conducted on the four psychological distress scales. Tests were two-tailed to allow for either positive or negative change. The probability level of p < .05 was adopted. All other results are descriptive. The data presented in Tables 3 through 5 collapse the percentage of respondents selecting ratings of 1 and 2 (any degree of dissatisfaction or disagreement) and separately present ratings of 3 (agree or satisfied) and 4 (strongly agree or very satisfied). The intent is to present the results conservatively by allowing the reader to use one percentage to note any degree of negative reaction and to highlight positive ratings that were not maximal.

Results

Process Monitoring

Service delivery

Figure 1 illustrates that the yield of RESTORE cases that resulted from prosecutor referrals, including all persons regardless of whether they could be found and qualified, was 46% for misdemeanors and 26% for felonies. More than 3 months elapsed from police report to RESTORE referral (M = 107 days, range = 21-220 days). Overall, 55% of cases were referred in the same year as the police report, 35% were referred within 12 to 24 months, and 10% of cases were referred after more than 2 years. On receipt of referral, the mean number of days to secure survivor victims’ consent was 11 (range = 1-37). Locating, screening, consenting, and forensic examination of responsible persons required a mean of 24 days (range = 5-73). The preparation of conference attendees lasted approximately 2 months (M = 67 days, range = 25-156). The length of the RESTORE Program from referral to conference for survivor victims was close to 3 months (95 days, range = 25-253). It was slightly shorter for felonies (95 days, range = 25-170) than misdemeanors (M = 117 days, range = 31-253). The mandatory 12 months of supervision for responsible persons began on the day of the conference. Conferences were approximately 45-min long (misdemeanors, M = 40 min, range = 20-68; felonies, M = 47 min, range = 27-64). The case manager effort involved in each case averaged 48 hr or the equivalent of 6 workdays (range = 3-7).

Safety

The PSS administered to survivor victims revealed a decrease in PTSD symptoms from intake to post-conference. At intake, 82% of survivor victims met diagnostic criteria for PTSD compared with 66% post-conference. The results from dependent t tests of pre–post psychological and physical symptoms for survivor victims and responsible persons are summarized in Table 2. No comparisons among survivor victims achieved or approached significance. Pre–post comparisons among responsible persons revealed two symptom scores that approached but did not meet the pre-specified significance criterion (p < .05). These included mental distress (M = 7.17 at intake, and M = 6.50, post-conference; t = 2.06, p = .55) and arousal (M = 9.22, at intake, and M = 8.22, post-conference; t = 1.98, p = .064). Case notes and conference observations revealed no physical safety issues before, during, or after the conferences. Across all conferences, a punitive statement was made by a responsible person for a survivor victim only once. However, punitive or blaming comments toward responsible persons occurred in half the conferences and were made by their own supporters as well as by survivor victims and their family and friends.

Outcome Evaluation

Reasons for choosing RESTORE

Table 3 shows that the highest consensus reason at intake was “to make the responsible person accountable” (>93% agreement). All responsible persons agreed with the comparable wording they received (“taking direct responsibility for making things right”). Most groups agreed that they selected RESTORE to have an alternative to court (>75%). The most likely groups to disagree were survivor victims and their family and friends. Across groups, 20% to 50% of respondents disagreed that apology was important. The exception was surrogate victims who all agreed that apology was significant to them. Virtually all (95%) responsible persons agreed or strongly agreed that “apologizing to the person I harmed” was a major reason they chose RESTORE.

At the post-conference measurement, two items achieved consensus. They were “making sure the responsible person doesn’t do what he did to anyone else” (>92% agreement) and “making sure the responsible person gets help” (>85% agreement). Many survivor victims agreed that having input into punishment was important (66% of survivor victims attending conferences, 61% of those with minimal participation). Surrogate victims more often disagreed (58%). Comparing intake with post-conference ratings revealed one reason for choosing RESTORE that became more important than what respondents had initially perceived. The item was “having the opportunity to express how the incident affected me.” Disagreement with this item fell from 9% to 0% in survivor victims, 46% to 7% in surrogate victims, 30% to 0% in their supporters, and 40% to 11% in responsible person supporters.

Preparation and conference experiences

Table 4 contains the responses in percentages to the items that solicited opinions on the preparation activities that preceded the conference. Most participants (>90%) agreed or strongly agreed that the preparation achieved its intended goals. An exception was survivor victims who participated in their conference. They either strongly agreed that their preparation was good (83%) or they strongly disagreed (17%). The weak area in preparation identified by these items was that 1 in 6 (17%) of responsible person family and friends disagreed that they received help in figuring out what to say.

Table 5 presents the data on participants’ experiences during the conference. Across all groups, more than 90% of the participants agreed or strongly agreed that they felt safe, listened to, supported, treated fairly, treated with respect, and not expected to do more than they anticipated. A clear consensus emerged that the conference was a success (>90% agreement across participants) with the exception of community volunteers. No survivor victims felt blamed but some responsible persons did (21%) as did some family and friends of both survivor victims and responsible persons (15% and 17%, respectively). Among all groups, survivor victims who attended their conference most often disagreed or strongly disagreed with items based on the envisioned aims of the RESTORE conference. However, required disagreement was on only 4 of 14 items for this distinction. Three of the four items on which survivor victims expressed negative opinions focused on the sincerity, genuineness, and likelihood of re-offending by the responsible person. Survivor victim reactions were mirrored to a somewhat lesser degree by their family and friends. Those viewing the conference from the survivor victim perspective clearly differed from the opinions of responsible persons. All responsible persons indicated that “I felt sincerely sorry for what happened.” However, 50% of survivor victims and 26% of their supporters disagreed. Likewise, several groups disagreed that responsible persons accepted responsibility (33% of survivor victims, 25% of surrogate victims, and 15% of their family and friends, and 12% of responsible person supporters). There were also variable opinions about the fairness of the redress plan. Those who most agreed that it was fair to everyone included survivor victims attending their conference (100%), their family and friends (95%), and volunteers (100%). The largest percentage of disagreement over redress fairness occurred among surrogate victims (33%), and responsible persons (32%) and 16% of their supporters. Some conference attendees found it difficult to speak openly (16% of survivor victims attending their conference, 25% of victim surrogates, 35% of survivor victim supporters, and 20% of responsible persons). Although 17% of responsible person supporters disagreed that their preparation equipped them to speak, none reported actual difficulty in speaking openly once they were in the conference.

Satisfaction

Table 6 contains satisfaction ratings for each of the RESTORE stages. More than 90% of participants were satisfied with their preparation, the conference, and the redress plan. The most satisfied group was survivor victims who attended their conference (100% satisfied or very satisfied on five of six items) and their supporters (100% satisfied or very satisfied on four of five items). All participants except 21% of responsible persons were satisfied with how RESTORE handled their case. Some disagreed that justice was done including 17% of survivor victims, 30% of surrogate victims, 16% of survivor victim supporters, and 20% of community volunteers. Nevertheless, more than 90% of all participants and 84% of surrogate victims would recommend RESTORE to others.

Completion Rates

Once consented, virtually all cases led to a conference (n = 20 of 22, 91%). From the 20 conferences that were held, 10 of 11 misdemeanor and 6 of 9 felony responsible persons completed RESTORE (80%). The two terminations and one withdrawal were all felonies. The terminations were for non-compliance related to alcoholism, financial distress, or homelessness. The withdrawal occurred when the offender recanted responsibility. The sole re-arrest was an elderly person arrested for exposure who was in declining health throughout his enrollment. At the time of the re-offense, he was beginning to show symptoms suggestive of dementia.

Discussion

Consideration of RJ for crimes against women has focused primarily its use in domestic violence cases. This article contributes empirical data on sexual assault to augment the dialogue. The following discussion begins with an examination of the feasibility, fairness, and safety of RESTORE. Following these comments, participants’ experiences are compared with the program vision. Next, service delivery is reviewed to aid planners to better anticipate the demands of program implementation. Subsequently, outcome evidence is scrutinized to reach conclusions about the extent to which the program achieved its aims. Finally, lessons learned are reviewed.

Feasibility

RJ conferencing for sex crimes would be infeasible unless certain pre-conditions were present. RESTORE-type programs depend on assumptions that (a) survivor victims desire face-to-face contact with their perpetrator and (b) responsible persons will accept an alternative to standard justice given the low likelihood of prosecution. The existing literature including studies that did and did not include sexual assault suggests that three quarters of victims agree to meet (Sherman et al., 2005; Umbreit et al., 2002). Herman’s (2005) study is an exception where many adult survivors of child abuse did not want face-to-face time with their perpetrator. In RESTORE among felony survivor victims, the desire to meet was consistent with earlier findings across crime types. Three quarters of the felony sexual assault survivors wished to meet face-to-face and this figure dropped only slightly when attention was limited to those cases where the survivor victim and responsible person were romantic partners (67%). The opposite was true for misdemeanors where four of five survivor victims did not wish to meet the responsible persons who were all strangers. Thus, the appeal of the RESTORE option increased as relationships became more intimate.

Responsible persons were willing to enter RESTORE at very high rates (85% of felonies and 100% of misdemeanors). The consent rate for felonies dropped to 57% if calculations included felony offenders who denied responsibility and were excluded from eligibility. This decision was justified on safety grounds, but it is an empirical question whether preparation time could have moved them to a point where they could have participated constructively in a conference. The inclusion of youthful offenders aged 15 to 17 years who were in the adult system would have boosted the number of referrals and potentially left out people who could have benefited from RESTORE. Hopefully, this exclusion on the grounds of human subject policy will prove specific to the site of the present study. Taken overall, the feasibility data were encouraging.

Fairness of Allocation of RJ

Equity has been a significant concern in the RJ literature generally. The findings present sobering data on potential bias in how RESTORE was allocated. Tasca, Rodriquez, Spohn, and Koss (2012) collected ethnicity/race data from all sexual assault reports processed by the largest police department in RESTORE’s jurisdiction during the calendar year prior to opening. These data allow for the estimation of racial/ethnic composition as justice progressed from police report to RESTORE consent. Disturbing racial/ethnic trends occurred. Caucasian responsible persons comprised 33% of police reports, 54% of prosecutor referrals, and 77% of RESTORE cases. In contrast, African Americans offenders constituted 25% of police reports, 9% of referrals, and 9% of cases. The comparable figures for persons of

Hispanic ethnicity were 42% of police reports, 25% of referrals, and 14% of cases. The patterns among survivor victims were similar. For example, Caucasian survivor victims comprised 64% of police reports, 64% of prosecutor referrals, and 88% of RESTORE cases. Readers should note that RESTORE staff could not directly control the race/ethnic makeup of consented cases because the referred persons themselves selected or declined the program. Care was taken to make RESTORE attractive to diverse groups by soliciting input from focus groups, nurturing partnerships with community agencies representing minority groups, arranging physical accessibility of facilities, and staffing the program ethnically and linguistically to reflect the community (Sniffen, Sisco, Bachar, & Koss, 2004). In addition, the initial presentation of the RESTORE option was guided by standard procedures and formal documents. Nevertheless, minority persons could have been discouraged by verbal or nonverbal staff behavior. However, it is more likely that the program was less attractive for minority groups for the cultural reasons that key informants had anticipated. Explanations for non-participation include norms about disclosure of negative information and reluctance to discuss what some may perceive as intimate matters. Structural factors could also have impacted consent. For example, immigration documentation practices in the state of Arizona might encourage Hispanic persons to avoid enlarging their exposure to institutional systems, especially if there are persons without papers in their extended families. The findings suggest that future programs should enhance cultural competence training among those who investigate and prosecute sexual assaults and implement ongoing monitoring of referral and consent demographics to correct imbalances as soon as they are documented.

There is also a concern that disproportionate access might result from bias against low-income persons. The demographics of RESTORE participants suggest the opposite. Prosecutors disclosed to an independent evaluator (Stubbs, 2009) that they selectively referred responsible persons who lacked resources. Half of them were unemployed non-students, and one in seven was disabled. If economically privileged offenders were less likely to be offered RESTORE, further research should document whether the criminal justice system was imposing any accountability on them at all.

Safety

Case notes and observational data from RESTORE conferences documented one isolated incident of survivor victim re-abuse that was stopped in midstream by the facilitator. The psychometric assessment of survivor victims revealed that many had symptoms of PTSD on entry. Symptoms did not exacerbate during participation and fewer met PTSD criteria at post-conference 3 months later. This reduction in symptom severity is consistent with data on survivor victims seen at sexual assault service centers. The decrease has been attributed to the natural recovery course of PTSD (Rothbaum et al., 1992). Statistical pre–post comparisons revealed no significant negative or positive impact on survivor victims’ emotional or physical health. Responsible persons had trend level evidence of symptom reductionThe mean scores of responsible persons were in the direction of improvement on half of the health outcomes. These analyses must be interpreted cautiously because of low power. Differences might have been detected with larger samples.

Many modifications of standard RJ conferencing models were incorporated into RESTORE in the interest of safety. These included focusing on prosecutor referral of offenders deemed safe for community-based resolution, using clinical risk assessment prior to enrollment, during preparation, and throughout supervision, determining that parties were ready before placing them face-to-face, holding conferences in police stations, establishing conduct rules for conferences, employing trained facilitators, supervising responsible persons for 12 months with either help or prodding as needed to complete the redress plan, and enforcing stay away orders. With these stipulations, RESTORE was conducted safely. RESTORE planners proceeded cautiously to avoid outcomes that would damage the prospects of future programs. Replication of identical procedures may be unnecessary. However, without greater attention to safety than is typical in conferencing programs, their use for sex crimes would be ill-advised.

Justice Experience Compared With Program Vision

Survivor victims say that they want justice that validates the legitimacy of their victimization, gives voice to their harm, empowers them to influence how their case is conducted, focuses on offender behavior and not on theirs, and involves them in determining the consequences imposed on the responsible person. RESTORE was envisioned as a justice process that responded to their expectations. The findings suggest that victim survivors did select RESTORE for many of these reasons. However, there were also unanticipated results. For example, RESTORE was described in brochures as “Justice That Heals.” Yet, the findings reveal that one third of survivor victims disagreed that they selected the program to put the experience behind them. It may be that some survivor victims feel that closure after sexual assault is not something one seeks from RJ, is not likely with any justice model, or is not possible under any circumstances because the impact of rape is life changing and lifelong. The data cannot weigh these alternatives.

The vision to empower survivor victims was better validated. All survivor victims strongly agreed that taking back their power was a major reason to select RESTORE over other justice options. Most also agreed that it was particularly important to have input into the consequences for the responsible person. Scholars have expressed fears that the accountability imposed by RJ is insufficient in comparison with criminal penalties. Participants knew that some forms of punishment could not be achieved through RESTORE, yet they still elected to enroll. Thus, it is unlikely that participants themselves devalued the forms of accountability offered.

Apology is thought to be an integral component of both accountability and healing (reviewed in Blecher, 2011). A meta-analysis of juvenile conferencing across crime types concluded that apologies are “extremely important” to victims and that RJ conferences produce sincere apologies (Sherman et al., 2005, p. 388). The RESTORE findings differed. Almost one third of survivor victims disagreed that they selected RESTORE to hear an apology. RESTORE actively discouraged apology until program exit where it would constitute an earned opportunity for responsible persons. Nevertheless, many responsible persons apologized at conferences. Survivor victims and supporters often viewed their apologies as insincere. Only one person was observed to forgive. Program designs that mandate apology at conferences are common in the literature. Forcing apology may be misguided with sexual assault survivors.

Contrary to expectations that a public apology is validating, no survivor victims chose to attend the exit meeting where the responsible person presented a letter he had written expressing his reflections over his acts, the harm he caused, and the changes he had made to avoid hurting others in the future. RESTORE case managers were in contact with survivor victims throughout the year following the conference to maintain compliance with Arizona victims’ rights statutes and check for the delayed negative impact. Survivor victims received updates on their responsible persons’ progress and notification of all public meetings of the community board involving him. No survivor victims asked to suspend contact, and so alienation from RESTORE is unlikely to explain their absence from these events that were intended to contribute to their recovery. Limits on the capacity to forgive have been hypothesized to constrain what any justice response may achieve (Blecher, 2011). Bletzer and Koss (2012) suggested that the survivor victims’ may prefer private closure. In any case, their choice not to attend precludes critics’ concern that survivor victims may be ill-used in RJ as absorbents of apologies who primarily serve the needs of responsible persons to feel better about themselves.

Service Delivery

The service delivery data may be helpful to future program planners including anticipating what the caseload might be. In RESTORE, prosecutorial case disposition was the single most important determinant of referral flow, although the high rate of police case closure was also problematic. Prosecutors disclosed to an independent evaluator (Stubbs, 2009) that they had used a “provable at trial” standard to select referrals. Given the average conviction rate for rape in the United States is approximately 13% of reported cases (Daly & Bouhours, 2010), offering restorative options only in those cases deemed likely to prevail at trial could not achieve the intent of RESTORE to enlarge the pool of offenders who are held accountable. In retrospect, a conversation should have occurred about the standard of evidence that would be applied. Possibly, other standards are legally defensible and would have produced a genuine expansion of justice options.

The present data also highlight the time that survivor victims are kept waiting for a justice response and the likelihood of achieving a justice outcome that they perceive as fair and responsive to their needs. The existing literature establishes that the juvenile sex offense court process is longer than RJ conferencing. For example, the South Australia Archival Study found that court process was 6 months from police report to finalization compared with 2.5 to 3 months for conferences (Daly, 2006). Likewise, RESTORE conferences also occurred approximately 3 months after receiving referrals. Prior to referral however, nearly half of the cases had languished more than a year after the crime before prosecutors referred them. From the survivor victim perspective, this delay is dismaying but it is actually an improvement over the data for victim–offender dialogue in prisons. Marshall found that the average time lapse between crime commission and victim–offender dialogue was 9.5 years (as quoted in Naylor, 2010). Miller (2011) asserted that post-conviction victim–offender dialogue is the only acceptable RJ approach for rape. However, premising RJ on conviction restricts it to a small number of cases where guilt is established. Many of these cases involve child abuse or stranger rape and therefore are not reflective of the greater prevalence of acquaintance rape. In addition, Miller’s (2011) proposal fails to consider the trauma of adversarial process and would not shorten law enforcement delay.

Outcomes

Among the consented cases, 91% resulted in a completed conference. Jülich et al. (2010) found that in Restore-NZ, only 25% of self-referred adult survivors of child sexual abuse proceeded to conferences. One must be cautious in comparing these figures because it is unclear if their use of the term referral is equivalent to what was labeled a consented case in the present study. What appear to be stark differences in progression to conference may illustrate the extent to which elapsed time since offense, survivor victim expectations, crime type, and criminal justice system involvement may motivate or deter offenders’ participation.

In addition to completed conferences, successful program exit is another important justice outcome. Two thirds of felony and 91% of misdemeanor responsible persons fulfilled all re-dress plans and supervision requirements and exited RESTORE successfully. Each success offered a survivor victim the opportunity to experience validation and achieve accountability. In comparison with RESTORE outcomes, three quarters of the cases retained in the prosecution pipeline were closed without any consequences. Satisfaction with conferencing alternatives is typically high (Sherman et al., 2005). RESTORE was also well received. The most satisfied group was survivor victims who attended their conference. Responsible person supporters were the most dissatisfied group across all indicators. Even so, 90% were satisfied that justice was done and 95% would recommend RESTORE. Surrogate victims were least satisfied with the justice outcome. Most surrogate victims were advocates at local sexual assault centers. In Nancarrow’s study (2006), victim advocates including persons of both indigenous and non-indigenous heritage were ambivalent about RJ conferencing for crimes against women in general and especially when the crime was sexual. Although lower than those of other participants, surrogates’ satisfaction with RESTORE was actually encouraging. All of them felt that the conference was a success, 66% said the redress plan was fair, 70% believed that justice was done, and 84% would recommend RESTORE to others. Daly and colleagues (2007) concluded that RJ conferences would be seen as more advantageous for victims than court. In particular, conferences were more likely than court to result admission of responsibility and raised the likelihood that offenders would receive counseling. Although the results of the present study do not permit a direct comparison with court outcomes, these two advantages were clearly demonstrated because RESTORE conferences did not proceed without the acceptance of responsibility, and psychological evaluation and treatment were mandatory.

Limitations and Lessons Learned

RESTORE was a demonstration project and represents what was possible as a pioneering effort in its setting and with available resources. The most obvious limitation of the evaluation is the small sample size. However, the number of cases is actually large in the context of the available literature. The usefulness of the findings to the literature that is primarily archival or anecdotal is enhanced by including multiple participant viewpoints and supplementing self-report with observational and objective data. The data would be difficult to replicate in the United States today without substantial local funding. That is because RJ conferencing is now specifically forbidden by many U.S. government entities. For example, the document known as the “Dear Colleague Letter,” which is the U.S. Office of Civil Rights guidelines for higher education institutional response to sexual discrimination, harassment, and sexual assault, may be read to preclude RJ (U.S. Department of Education, 2011). This document as well as the funding guidelines of other federal agencies that respond to sexual assault confuses RJ with mediation. In practice, it is traditional mediation, especially the type used in divorce courts where participation is ordered without consent that is the basis for animus and prohibitions against face-to-face justice outside the courtroom. Heroic efforts were made to sustain funding for RESTORE through local, state, federal, and private philanthropy but they ultimately proved futile in the face of institutionalized opposition. More encouraging is that Restore-NZ in Auckland has received government funding and appears sustainable (S. Jülich, private communication, July 25, 2013).

With the limitations and cautions previously expressed, the findings of the present study demonstrate that a conferencing program like RESTORE is feasible, can be conducted safely, achieves acceptable levels of satisfaction, and attains many of the envisioned outcomes. However, the broader conversation about RJ for sexual assault is all too often about whether it should be done at all and not about how to do it (Matsui, 2011). Readers of this article are all undoubtedly deeply committed to the welfare of survivor victims and to ending sexual assault. Most of us are quite aware of the performance of criminal justice and the treatment accorded to sexual assault survivors. For these reasons, it is not productive to continue “oppositional contrasting” of programs like RESTORE and adversarial justice (Stubbs, 2010, p. 104). The sexual assault response agenda might be energized by planning a listening project to renew our understanding of the justice desires and interests of survivor victims so that we can align our priorities with theirs. RJ can be approached slowly and thoughtfully, recognizing that there are many forms and points in time where its principles are applicable. Insights may be incorporated into existing process or developed as freestanding alternatives. We can innovate within the comfort zone of individual settings. Taking more ownership of the justice response could be empowering for the anti-sexual-violence movement.

Footnotes

Acknowledgements

A large core of people invested their energy in the program. They include Quince Hopkins, Florida Atlantic School of Law; Karen Bachar, National Institute of Justice; Chad Sniffen, California Coalition Against Sexual Assault; Steve Sadler, Forensic Associates of Tucson; and Ralph Renger at the University of Arizona. Jolene Unruth oversaw administration and Misty Harvey headed the data team. Also acknowledged are CDC staff, members of the National Advisory Board, community volunteers, participants, individual police officers and prosecutors, meeting facilitators, students, administrative staff, community and university-based social workers and therapists, American Sign Language translators and psychotherapists. They are deeply thanked for coming together to realize a vision.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Initial planning was funded by the Tucson Police Department, and implementation and evaluation were funded by the National Center for Injury Prevention and Control, an entity within the U.S. Centers for Disease Control and Prevention (CDC; R49/CCR921709-03-3).

Author Biography

Mary P. Koss, PhD, is a Regents’ Professor in the Mel and Enid Zuckerman College of Public Health at the University of Arizona. She is co-editor of a two-book series for the American Psychological Association, Violence Against Women and Children (2011; American Psychological Association). She was the principal investigator of the RESTORE Program, a restorative justice option for selected sex crimes among adults. In 2010, she was the eighth recipient of the Visionary Award from the law enforcement training and technical assistance organization, Ending Violence Against Women International. Recently she has contributed to surveys of sexual assault prevalence and response conducted by the U.S. Air Force and participated in the documentary film, The Invisible War: Military Sexual Assault.

References

Blecher

N. J.

(2011). Sorry justice: Apology in Australian family group conferencing. Psychiatry, Psychology and Law, 18, 95-116.

Bletzer

K. V.

Koss

M. P.

(2012). From parallel to intersecting narratives in cases of sexual assault. Qualitative Health Research, 22, 291-303.

Bletzer

K. V.

Koss

M. P.

(2013). Restorative justice and sexual assault: Outcome appraisal through textual analysis. Open Area Studies Journal, 5, 1-11.

Cossins

(2008). Restorative justice and child sexual offences: The theory and the practice. British Journal of Criminology, 48, 359-378.

Curtis-Fawley

Daly

(2005). Gendered violence and restorative justice: The views of victim advocates. Violence Against Women, 11, 603-638.

Daly

(2006). Restorative justice and sexual assault: An archival study of court and conference cases. British Journal of Criminology, 46, 334-356.

Daly

(2008a). Seeking justice in the 21st century: Towards an intersectional politics of justice. Sociology of Crime, Law and Deviance, 11, 3-30.

Daly

(2008b). Setting the record straight and a call for radical change: A reply to Annie Cossins on “Restorative Justice and child sex offences.” British Journal of Criminology, 48, 547-566.

Daly

(2011). Conventional and innovative justice responses to sexual violence (ACSSA Issues No. 12). Melbourne, Victoria, Australia: Australian Centre for the Study of Sexual Assault, Australian Institute of Family Studies. Retrieved from http://www.griffith.edu.au/school/ccj/kdaly.html

10.

Daly

Bouhours

(2010). Rape and attrition in the legal process: A comparative analysis of five countries. Crime and Justice, 39, 485-565.

11.

Daly

Curtis-Fawley Bouhours (2003). Sexual offence cases finalized in court, by conference, and by formal caution in South Australia for young offenders, 1995-2001. Brisbane, Australia: Griffith University. http://www.griffith.edu.au/professional-page/professor-kathleen-daly/publications,

accessed December 5, 2013

12.

Daly

Bouhours

Curtis-Fawley

(2006). South Australia Juvenile Justice and Criminal Justice (SAJJ-CJ): In-depth study of sexual assault and family violence cases (Technical Report No. 4). Brisbane, Queensland, Australia: School of Criminology and Criminal Justice, Griffith University. Retrieved from http://www.griffith.edu.au/professional-page/professor-kathleen-daly/publications,

accessed December 5, 2013

13.

Daly

Bouhours

Curtis-Fawley

Weber

Scholl

(2007). South Australia Juvenile Justice and Criminal Justice (SAJJ-CJ): Sexual Assault Archival Study (SAAS), an archival study of sexual offence cases disposed of in youth court and by conference and formal caution (Technical Report No. 3). Brisbane, Queensland, Australia: School of Criminology and Criminal Justice, Griffith University. http://www.griffith.edu.au/professional-page/professor-kathleen-daly/publications,

accessed December 5, 2013

14.

Daly

Curtis-Fawley

(2006). Restorative justice for victims of sexual assault. In Heimer

Kruttschnitt

(Eds.), Gender and crime: Patterns of victimization and offending (pp. 231-265). New York: New York University Press.

15.

Daly

Nancarrow

(2010). Restorative justice and youth violence toward parents. In Ptacek

(Ed.), Restorative justice and violence against women (pp. 150-174). New York, NY: Oxford University Press.

16.

Daly

Stubbs

(2006). Feminist engagement with restorative justice [Special issue on gender, race, and restorative justice]. Theoretical Criminology, 10, 9-28.

17.

Daly

Wade

(2012). South Australia Juvenile Justice and Criminal Justice (SAJJ-CJ): In-depth study of sexual assault and family (Technical Report No. 5). Retrieved from http://www.griffith.edu.au/__data/assets/pdf_file/0009/497484/Technical-Report-5-updated-27-Feb-2013.pdf,

accessed December 5, 2013

18.

Foa

Riggs

D. S. B.

Dancu

C. V.

Rothbaum

B. O.

(1993). Reliability and validity of a brief instrument for assessing post-traumatic stress disorder. Journal of Traumatic Stress, 6, 459-473.

19.

Hanson

R. K.

Thornton

(1999). Static 99: Improving actuarial risk assessments for sex offenders (Vol. 2). Ottawa, Ontario, Canada: Department of the Solicitor General of Canada.

20.

Helmus

Hanson

R. K.

Thornton

Babchishin

K. M.

Harris

A. J. R.

(2012). Absolute recidivism rates predicted by Static-99R and Static-2002R sex offender risk assessment tools vary across samples: A meta-analysis. Criminal Justice and Behavior, 39, 1143-1171.

21.

Herman

J. L.

(2005). Justice from the victim’s perspective. Violence Against Women, 11, 571-602.

22.

Hopkins

C. Q.

Koss

M. P.

Bachar

K. J.

(2004). Applying restorative justice to ongoing intimate violence: Problems and possibilities. St. Louis University Public Law Review, 20, 289-312.

23.

Hudson

(2002). Restorative justice and gendered violence: Diversion or effective justice? British Journal of Criminology, 42, 107-129.

24.

Jülich

(2010). Restorative justice and gendered violence in New Zealand: A glimmer of hope. In Ptacek

(Ed.), Restorative justice and violence against women (pp. 239-254). New York, NY: Oxford University Press.

25.

Jülich

McGregor

Annan

Landon

McCarrison

McPhillips

(2010). Yes, there is another way! Canterbury Law Review, 17, 222-228.

26.

Julich

Buttle

Cummins

Freeborn

(2010). Project Restore: An exploratory study of restorative justice and sexual violence. http://www.academia.edu/274691/Project_Restore_An_Exploratory_Study_of_Restorative_Justice_and_Sexual_Violence.

27.

Jülich

McGregor

Annan

Landon

McCarrison

McPhillips

(2010). Yes, there is another way! Canterbury Law Review, 17, 222-228.

28.

Koss

M. P.

(2010). Restorative justice for acquaintance rape and misdemeanor sex crimes. In Ptacek

(Ed.), Feminism, restorative justice, and violence against women (pp. 218-238). Thousand Oaks, CA: Sage.

29.

Koss

M. P.

Achilles , (2008). Restorative justice for sexual assault. Retrieved from http://www.vawnet.org/summary.php?doc_id=1231&find_type=web_desc_AR

30.

Koss

M. P.

Bachar

K. J.

Hopkins

C. Q.

(2003). Restorative justice for sexual violence: Repairing victims, building community, and holding offenders accountable. In Prentky

Burgess

A.W.

(Eds.) Understanding and managing sexual coercion (Vol. 989, pp. 384-396). Annals of the New York Academy of Sciences.

31.

Koss

M. P.

Bachar

Hopkins

C. Q.

Carlson

(2004). Expanding a community’s justice response to sex crimes through advocacy, prosecutorial, and public health collaboration: Introducing the RESTORE Program. Journal of Interpersonal Violence, 19, 1435-1463.

32.

Madsen

K. S.

(2004). Mediation as a way of empowering women exposed to sexual coercion. Nordic Journal of Feminist and Gender Research, 12, 58-61.

33.

Madsen

K. S.

(2006). A walk on the wild side (Victims & Mediation Project). Retrieved from http://www.apav.pt/pdf/Victims_Mediation_EN.pdf#page=108

34.

Massaro

T. M.

(1997). The meanings of shame: Implications for legal reform. Psychology, Public Policy, and Law, 3, 645-704.

35.

Matsui

(2011). Justice for the accused or justice for victims: The protection of victims’ rights in Japan. Asian-Pacific Law & Policy Journal, 54, 97-126.

36.

McGlynn

(2011). Feminism, rape and the search for justice. Oxford Journal of Legal Studies, 1, 825-842.

37.

McGlynn

Westmarland

Godden

(2012). “I just wanted him to hear me”: Sexual violence and the possibilities of restorative justice. Journal of Law and Society, 39, 213-240.

38.

Miller

(2011). After the crime: The power of restorative justice dialogues between victims and violent offenders. New York: New York University Press.

39.

Millon

Davis

Grossman

(2009). MCMI-III Manual (4th ed.). Minneapolis, MN: Pearson Education, Inc.

40.

Monroe

L. M.

Kinney

L. M.

Weist

M. D.

Spriggs-Dafeamekpor

Dantzler

Reynolds

M. W.

(2005) The experien e of sexual assault: Findings from a statewide victim needs assessment. Journal of Interpersonal Violence, 20, 767-776.

41.

Nancarrow

(2010). Restorative justice for domestic and family violence: Hopes and fears of indigenous and non-indigenous Australian women. In Ptacek

(Ed.), Feminism, restorative justice, and violence against women (pp. 123-149). Thousand Oaks, CA: Sage.

42.

Naylor

(2010). Effective justice for victims of sexual assault: Taking up the debate on alternative pathways. University of New South Wales Law Journal, 33, 662-683.

43.

Nichols

H. R.

Molinder

(1996). Multiphasic Sex Inventory II. Tacoma, WA: Nichols and Molinder Assessments.

44.

Patritti

(2010). Restoring human trafficking victims through victim-offender dialogue. Cardozo Journal of Conflict Resolution, 12, 217-244.

45.

Rothbaum

B. O.

Foa

E. B.

Riggs

D. S.

Murdock

Walsh

(1992). A prospective examination of post-traumatic stress disorder in rape victims. Journal of Traumatic Stress, 5, 455-475.

46.

Seidman

Pokorak

(2011). Justice responses to sexual violence. In Koss

M. P.

White

J. W.

Kazdin

(Eds.), Violence against women and girls, Vol. 2. Navigating the solutions (pp. 137-158). Washington, DC: American Psychological Association.

47.

Seidman

Vickers

(2005). The second wave: An agenda for the next thirty years of rape law reform. Suffolk University Law Review, 38, 457-490.

48.

Sherman

L. W.

Strang

Angel

Woods

Barnes

G. C.

Bennett

Inkpen

(2005). Effects of face-to-face restorative justice on victims of crime in four randomized trials. Journal of Experimental Criminology, 1, 367-395.

49.

Sniffen

Sisco

Koss

M. P.

Bachar

K. J.

(2004, May 26). Focusing on Community: Diversity and RESTORE. Paper presented at the 3rd National Sexual Violence Prevention Conference, Los Angeles, CA.

50.

Stubbs

(2007). “Beyond apology?” Domestic violence and critical questions for restorative justice. Criminology & Criminal Justice, 7, 169-187.

51.

Stubbs

(2009, November). Meanings of justice: Sexual assault and the appeal to restorative justice. Paper presented at the Australian & New Zealand Society of Criminology Conference, Perth, Western Australia, Australia.

52.

Stubbs

(2010). Relations of domination and subordination: Challenges for restorative justice in responding to domestic violence. University of New South Wales Law Journal, 16, 102-111.

53.

Tasca

Rodriquez

Spohn

Koss

M. P.

(2012). Suspect Identification and arrest. Journal of Interpersonal Violence. doi: 10.1177/0886260512468233.

54.

Umbreit

M. S.

Coates

R. B.

Vos

Brown

(2002). Executive summary: Victim offender dialogue in crimes of severe violence. A multi-site study of programs in Texas and Ohio. Minneapolis, MN: Center for Restorative Justice and Peacemaking. Retrieved from http://www.cehd.umn.edu/ssw/RJP/Resources/Research/Exec_Sum_TX_OH_VOD_CSV.pdf

55.

Umbreit

M. S.

Vos

Coates

R. B.

Lightfoot

(2006). Restorative justice in the twenty first century: A social movement full of opportunities and pitfalls. Marquette Law Review, 253-304. Retrieved from http://www.cehd.umn.edu/ssw/RJP/Resources/RJ_Dialogue_Resources/RJ_Principles/Marquette%20RJ%2021st%20Century%20Social%20Movement%20Full%20of%20Pitfalls%20and%20%20Opportunities.pdf

56.

U.S. Department of Education, Office for Civil Rights. (2011, April 4). Dear Colleague Letter: Sexual violence background, summary, and fast Facts. Washington, DC: Author.

57.

Vanseveren

(2010, June 17-19).Developments in Flanders and HERGO in sexual assault cases. Paper presented at the A Way Forward for Restorative Justice Conference in Europe, Leuven, Belgium.