Intervening to Prevent Repeat Offending Among Moderate- to High-Risk Domestic Violence Offenders

Abstract

Clear directions about best strategies to reduce recidivism among domestic violence offenders have remained elusive. The current study offers an initial evaluation of an RNR (Risk, Needs, and Responsivity)-focused second-responder program for men accused of assaulting their intimate partners and who were judged as being at moderate to high risk for re-offending. A quasi-experimental design was used to compare police outcomes for 40 men attending a second-responder intervention program to 40 men with equivalent levels of risk for re-offense who did not attend intervention (comparison group). Results showed that there were significant, substantial, and lasting differences across groups in all outcome domains. In terms of recidivism, rates of subsequent domestic-violence-related changes were more than double for men in the comparison group as compared with the intervention group in both 1-year (65.9% vs. 29.3%) and 2-year (41.5% vs. 12.2%) follow-up. Changes in the rates of arrest were consistent with reductions in men’s general involvement with police, with men in the intervention group receiving fewer charges for violent offenses, administrative offenses, and property offenses over the 2 years following intervention than men in the comparison group. Not surprisingly, these differences result in a much lower estimated amount of police time with intervention men than for comparison men. Results are discussed with reference to the possible impact of sharing information with men about their assessed risk for re-offending within a therapeutic justice context.

Keywords

domestic violence second responder therapeutic justice domestic violence risk assessment policing re-assault

Over the past 30 years, there have been many developments in ideas about how to best prompt non-recidivism among men who have assaulted their intimate partners. It was originally hoped that initial police response, specifically arrest, would be sufficient in and of itself to protect victims from recurrence of abuse (L. W. Sherman & Berk, 1984). As evidence accumulated that arrest alone was insufficient to curb men’s assaults against their intimate partners, a variety of specialized court procedures and intervention programs were developed. Many jurisdictions in Canada, the United States, the United Kingdom, and other developed nations now have dedicated domestic violence police and court services, specialized prosecution streams, targeted probation services, and court-linked batterer intervention programs in place for men who have been charged with assault against an intimate partner. Court-based advocacy and second-responder intervention programs for women victims of abuse are also common. There is ongoing controversy in the field about the efficacy of these court-linked interventions for reducing recidivism among offenders (Smedslund, Dalsbø, Steiro, Winsvold, & Clench-Aas, 2009) and of the implications of these developments for the penal justice system generally (e.g., Hannah-Moffat & Maurutto, 2012).

Although definitive conclusions about the efficacy of police or therapeutic interventions to curb domestic violence remain elusive, a great deal has been learned since the early 1980s about the nature and course of domestic violence offending. Studies using a range of informants (i.e., men, women, police) have found that, following identification, most men will curb their abusive behaviour. Specifically, about two thirds of households that report a domestic incident to the police do not report a subsequent incident of physical violence (to police or in regular follow-up interviews with victims) over the following 6 to 12 months (Babcock, Green, & Robie, 2004; Gondolf, 2012; Klein & Tobin, 2008). Emotional and psychological abuse also tends to decrease over time for the majority of couples (Gondolf, 2012). It is also clear that, among men who re-offend, there is a subgroup who are chronic and repeat offenders likely to cause their partners the highest level of physical injury and likely to pose the greatest threat to women’s lives. This population of more dangerous, repeatedly assaultive batterers also overlaps substantially (though not exclusively) with the general criminal population. These men are likely to be involved with police as a result of past and concurrent assault, theft, property, and administrative offenses that are not domestic in nature and they share many of the risk factors for re-assault characteristic of a general population of offenders (e.g., antisocial attitudes, employment problems) (Cattaneo & Goodman, 2005; Echeburúa, Fernández-Montalvo, de Corral, & López-Goñi, 2009).

Ironically, in many jurisdictions, higher risk batterers are least likely to receive the swift, sure, and balanced justice and therapeutic response touted by many as being most likely to lead to change in behaviour (Gondolf, 2012). Because of the severity of their offense and/or history of violence, offenders deemed to be at higher risk for re-offense are often ineligible for “fast-track” prosecution streams that link admission of responsibility and completion of treatment with reduced sentences or conditional discharge. Instead, following arrest, these men are most often released on bail or, much less frequently, remanded into custody to await trail. Although men’s bail conditions typically include conditions not to associate with the victim, there is widespread awareness within the domestic violence and policing community that, even with no-contact orders, the waiting period between arrest and trail (which is often many months) is a time of significant potential danger to victims from a potential re-assault by the offender (Office of the Chief Coroner, 2011).

Inquests and death reviews across jurisdictions have made a number of recommendations to attempt to address potential danger to victims while high-risk batterers await trial. These recommendations can be organized around the themes of containing men’s risk through court and police-based services (e.g., implementing risk assessment, ensuring efficient communication of risk information, priority-response policing of offenders) and of addressing the safety needs of women victims of men’s abuse (e.g., immediate availability of court-based victim advocates to ensure women’s safety concerns are communicated to the courts, better funding for shelters). More recently, another potential response to moderate- to high-risk domestic violence offenders has been considered—that of using secondary responders to address risk directly with offenders. Ontario’s recent Death Review Committee report (Office of the Chief Coroner, 2011), for example, included the recommendation that police-based risk assessment

. . . should be used not only to indicate the presence of risk-enhancing factors towards violence, but also to identify those areas where case management could mitigate the risk for future violence. When risk factors such as substance abuse, mental health concerns, employment issues etc. are identified, efforts should be made to provide appropriate references or involve appropriate services to alleviate those risk factors. (p. 17)

The current study was designed to explore this alternate, “second-responder” intervention response as another potential method for intervening with men at moderate to high risk of re-offending against their partners. Herein, we discuss the background and rationale for our intervention and outline the research questions and hypotheses that guided our analyses.

Second-Responder Programs: An Option for Offenders?

Within the domestic violence field, there is a strong history of attempting to combine and coordinate justice and community services to intervene with victims immediately following a police-reported incident. Second-responder programs for women victims of domestic violence began in the late 1980s and early 1990s and are now an established component of domestic violence response in many jurisdictions (Davis, Taylor, & Titus, 1997; Gamache, Edleson, & Schock, 1988; Lane, Greenspan, & Weisburd, 2004). Recognizing both the risk and potential opportunity for change during this critical period, second-responder programs aim to provide immediate short-term intervention that will help victims of abuse protect themselves from subsequent victimization. Programs generally involve home visits and/or follow-up phone contact by social workers or specialized teams of police. During these contacts, focus is placed on providing women with information on services and legal options, sharing information about their level of risk for subsequent victimization, improving women’s understanding of court orders, increasing access to therapeutic support, and engaging in immediate safety planning. The intent of these interventions has been to empower victims to access social service, mental health, and advocacy resources, so that they can take action to reduce their risk for future victimization (Davis & Smith, 1994).

Evidence for the success of these victim-focused second-responder programs is mixed. The most consistently positive results have been increased rates of calls to police. In numerous studies, it has been found that, compared with randomly assigned or matched control women, victims who receive second-responder intervention call the police sooner and more often during a follow-up 6- or 12-month study period (Davis, Weisburd, & Hamilton, 2010; Davis, Weisburd, & Taylor, 2008; Stover, Berkman, Desai, & Marans, 2010). Less evidence has been found in support of the efficacy of these programs in reducing subsequent victimization (Casey et al., 2007; Davis et al., 2010; Stover, Poole, & Marans, 2009).

Remarkably, second-responder programs have generally not extended services to perpetrators of abuse. In the few programs where contact with perpetrators is discussed as a possibility, focus is placed on providing perpetrators with reminders and warnings about the consequences of engaging in abusive behaviours (Davis et al., 2010). Theoretically, such reminders are based on the idea that future violence will be deterred by increased social and personal costs of abusive actions to the perpetrators; a perspective that is not consistent with current thinking on best practices that attempt to combine the criminal justice system’s strength in accountability with a social service focus on providing the intervention necessary to address risk factors for criminal involvement and promote positive behaviour change (Marlowe, 2002; Prendergast, Pearson, Podus, Hamilton, & Greenwell, 2013).

Risk, Needs, and Responsivity (RNR) as an Organizing Framework

An alternate, and potentially more useful, theoretical model for perpetrator-based second-responder program is the RNR model of intervention (Andrews, Bonta, & Wormith, 2011; Bonta, Andrews, & Wormith, 2006; Bourgon, Hodgson, Hanson, & Helmus, 2009; Polaschek, 2012; Wormith & Luong, 2011). Applied to second-responder programs for domestic violence, the RNR model suggests a number of features for program design (Connors, Mills, & Gray, 2012; Polaschek, 2012). First, this model suggests that more intense intervention be targeted toward domestic violence offenders assessed as being at higher risk for re-offending. Second, focus of intervention should be firmly placed on those needs most closely related to men’s offending, or their criminogenic needs. Consideration of the set of central criminogenic needs identified by Andrews and Bonta (2006) and the literature on dynamic risk factors for domestic violence offending (e.g., Capaldi, Knoble, Shortt, & Kim, 2012) identifies a number of core promising areas for intervention. In the relationship domain, these include men’s responses to recent separation or recent change in contact between accused and children. In the school/work domain, likely criminogenic needs are whether the accused believes he is at risk of losing employment as a result of charges as well as high rates of stress as a result of conditions such as unemployment, financial stress, or loss of job. In the antisocial cognitions domain, cognitions specifically recognized as relevant to domestic violence risk include jealousy, possessiveness, disrespect of women, disrespect of legal authority, and minimization or denial of abusive behaviour. Substance use as well, in terms of alcohol and drug dependence or abuse, noticeable increase in drug or alcohol use, or a tendency to become angry or violent when using drugs or alcohol, has been related to recidivism among perpetrators. In addition, the perpetration of lethal domestic violence has been associated with depression, desperation, and high levels of hopelessness. Finally, the RNR model specifies that interventions be evidenced-based and, to the extent possible, tailored to the learning style and motivational profile of the participants (Responsivity principle). Consistent with this principle, a second-responder program for men should be based on cognitive-behavioural principles and focused on immediate, practical actions that could be taken to reduce dynamic risk for re-offending.

Research Questions

Building on these theoretical perspectives, the current study was designed to report on the initial implementation and evaluation of an RNR-focused second-responder programs for men accused of assaulting their intimate partner and who were assessed as being at moderate to high risk for re-offending. We were interested in the feasibility of this project in terms of its acceptability to men and service providers. We were also interested in examining program efficacy. We used a quasi-experimental design, comparing police involvement in the first and second years following intervention for men who participated in the intervention as compared with men who did not participate in this program. Finally, we estimated the time saving to police services over 2 years as a result of this intervention.

Method

Setting

The current research was conducted in London, Ontario, Canada, a midsized city in the southwestern part of the province with a population of approximately 370,000 citizens. Involved partners included the London Police Service (LPS), Changing Ways, Women’s Community House, John Howard Society, the Centre for Research and Education on Violence against Women and Children, Western University, and the University of Toronto.

In Ontario, there are two streams of prosecution for individuals arrested and charged with domestic-violence-related offenses; one for accused who present a moderate to high risk of recidivism (called Enhanced Prosecution) and the other for lower risk offenders (called the Early Intervention Program). The LPS use a domestic violence risk assessment screening tool (Domestic Violence Supplementary Report [DVSR]) to assist in making decisions regarding bail eligibility and conditions and to place the accused in the appropriate process. The Early Intervention Program affords the accused the opportunity to enter an early guilty plea, gain immediate entry into a specialized domestic violence intervention program, and, upon successful completion of this program, have the charge disposed of by way of a conditional discharge. To be eligible for the Early Intervention Program, the accused must not have a record for violent offenses or criminal harassment, must not have caused injury to the victim, must not have perpetrated the assault in the presence of a child, and must plead guilty to the charge.

If an accused is ineligible for the Early Intervention Program, they enter the Enhanced Prosecution stream. Offenders in this stream are either released on bail, typically with conditions to not associate with the victim, or much less frequently, remanded in custody to await trial. During this waiting period, men are ineligible for domestic violence intervention and are screened out of most mental health and addiction services; thus, men typically receive no intervention between the time they are released on bail and the time their case is heard in court. Over 2010-2011, when this study took place, typical wait times for prosecution were 6 months to 1 year.

Sample

To be eligible for inclusion in this study, men must have been (a) 18 years or older, (b) charged by the LPS with committing a domestic-violence-related offense, (c) rated by the responding officers as moderate to high risk for domestic-violence-related re-offending on the DVSR (see description below) and thus not eligible for the Early Intervention program, and (d) not remanded into custody. As such, the samples of men studied were those offenders deemed at moderate to high risk for re-offending and who remained in the community. Importantly, this project was not designed to provide any feedback to the court process, and men were clearly informed at all recruitment contacts that participation in this project would not lead to any leniency in sentencing.

Over the 4-month period of the study, 229 men met study eligibility criteria. Once the case was determined to be eligible, a letter was sent to the accused, outlining the current project and inviting the man to participate and/or to contact the Family Consultant/Victim Services Unit (FC/VSU) of the London Police for support services. In no cases did men respond to this letter-based invitation. Letters were followed up one week later with a telephone call made by a police-based clinician (FC/VSU) to ensure that the letter had been received and to assess the man’s willingness to participate. Direct phone contact was successfully made with 63 men. Upon personal invitation, 76% of the 63 men contacted agreed to participate (n = 48). Those who refused participation did so for two main reasons, either because they were moving out of the area or because they indicated that they were already connected to services and support. Of the 48 men who consented to the project, 40 attended at least one appointment and were considered the intervention group.

The other 166 eligible men could not be contacted by police. In approximately two thirds of these cases (n = 115), there was no phone number listed in the police database and no further contact attempts were made.¹ In 19 cases, the accused had a phone number listed with the LPS but did not have voice mail and could not be reached in person. In a further 32 cases, police-based clinicians left unreturned voice mail messages. To create a comparison group, 40 men were randomly selected from these 166 men with whom police had no direct contact. In total then, analyses for this study involved a final sample of 40 men who received intervention and a comparison group of 40 men in an equivalent policing and prosecution stream who did not receive intervention.

Measures

DVSR

The DVSR is a supplemental reporting form created by the Ministry of the Solicitor General and the Ontario Provincial Police Behavioural Science Section. It is completed by Ontario police on the basis of file information on the accused and interviews with the accused and the victim as part of response to all domestic violence occurrences. This instrument involves collection of data on the marital status of the perpetrator, on whether children were present during the incident, and on 24 “yes/no/unknown” questions recognized as being potential risk factors for lethal violence, herein organized for ease of presentation into sections of risk factors related to current and past domestic violence; relationship; mental health, stress, and additions; and general criminal involvement (see Table 1 for a full list of items). For the purposes of the current study, only “yes/no” responses were analyzed (i.e., ratings of unknown were coded as missing). Although this measure is supposed to be used in response to all domestic violence occurrences, it is sometimes not completed by responding officers. For our sample, data on this measure were available for the majority of men in the comparison (n = 35) and intervention (n = 33) groups.

Table 1.

Risk Factors for Domestic-Violence-Related Re-Assault at the Time of the Index Offense for Men in the Intervention and Comparison Groups.

	Comparison (n = 35)	Intervention (n = 33)	Statistic	p
Risk factors related to past and current domestic violence (% yes)
Has there been a recent escalation in frequency or severity of violence	19 (57.6%)	18 (54.5%)	χ²(1) = 0.06	.804
Has the accused ever demonstrated jealousy or obsessive behaviour	22 (66.7%)	29 (87.9%)	χ²(1) = 4.23	.040*
Has the accused ever assaulted or sexually abused the victim or previous partners	34 (100%)	33 (100%)	χ²(1) = 0.0	1.00
As the accused threatened to harm/kill the victim or family member	10 (29.4%)	14 (42.4%)	χ²(1) = 1.23	.267
Has the accused threatened to destroy property of victims	9 (26.5%)	13 (39.4%)	χ²(1) = 1.27	.260
Has the accused ever assaulted the victim in public view	6 (19.4%)	4 (12.5%)	χ²(1) = 0.55	.457
Does the accused minimize or deny his violent behaviour	24 (77.4%)	22 (73.3%)	χ²(1) = 0.14	.711
Risk factors in the relationship (% yes)
Does victim fear that accused will continue the assaults, seriously injure, or kill her or him or the children?	17 (54.8%)	22 (66.7%)	χ²(1) = 0.94	.332
Has the accused demonstrated stalking behaviour (toward victim)	11 (32.4%)	15 (45.5%)	χ²(1) = 1.21	.271
Has there been a recent separation or change in the relationship	16 (45.7%)	17 (51.5%)	χ²(1) = 0.23	.632
Has there been a recent change in the contact between the children and accused	3 (11.5%)	4 (13.8%)	χ²(1) = 0.06	.802
Risk factors associated with stress, mental health, addictions (% yes)
Has the accused experience any unusually high stress—financial, loss of job, health, and so on	18 (60.0%)	18 (56.3%)	χ²(1) = 0.09	.765
Does the accused have any known mental health problems	14 (46.7%)	17 (54.8%)	χ²(1) = 0.41	.523
Does the accused’s personality feature anger, impulsiveness, poor behaviour control, or disrespect toward women or authority	30 (85.7%)	24 (75.0%)	χ²(1) = 1.23	.268
Does the accused believe he is a risk of losing employment as a result of the charges	2 (9.1%)	7 (28.0%)	χ²(1) = 2.70	.100
Does the accused abuse drugs/alcohol	23 (71.9%)	25 (78.1%)	χ²(1) = 0.33	.564
Has there been a noticeable increase in drug/alcohol use	19 (61.3%)	19 (61.3%)	χ²(1) = 0.00	1.00
Has the accused threatened suicide	12 (37.5%)	17 (60.7%)	χ²(1) = 3.22	.073
Has the accused been a victim or witness to domestic violence as a child	15 (65.2%)	16 (66.7%)	χ²(1) = 0.01	.917
Risks associated with general criminal involvement (% yes)
Has the accused injured, abused, or killed pets	4 (12.9%)	3 (10.7%)	χ²(1) = 0.07	.795
Does the accused have access to firearms/weapons	2 (7.7%)	2 (6.7%)	χ²(1) = 0.02	.882
Has the accused used or threatened the use of firearms or weapons against the victim or other	3 (9.7%)	4 (12.1%)	χ²(1) = 0.10	.754
Has the accused ever assaulted a non-family person or committed a violent offense?	11 (40.7%)	11 (42.3%)	χ²(1) = 0.01	.908
Has the accused ever breached a court imposed condition	13 (54.2%)	6 (25.0%)	χ²(1) = 4.27	.039*

p < .05. **p < .01.

Homelessness/transience

The LPS keeps records of the past and current addresses of individuals with whom they have had contact. A manual file audit of this database was conducted to explore the address history of the men in this study to determine how many times men were listed as homeless and how many times they changed addresses during 2010 and 2011.

Recidivism

Re-arrests were coded if men were charged by the LPS with an offense during the year following intervention, that is, January 2011 to December 2011 for Year 1 follow-up and from January 2012 to December 2012 for Year 2 follow-up. Type of offense was classified using the Canadian standardized coding system (Uniform Crime Reporting) as (a) violence, (b) property, and (c) administration of justice. Violence included all occurrences that were violent in nature and where violence was directed at a person, such as assault, threats, criminal harassment, or sexual assault. Property included all occurrences where damage to or theft of property was the primary complaint. Administration of justice included all occurrences where the primary complaint was a breach of process related to the administration of the criminal justice system, such as breach of a recognizance or failure to attend court.

A separate process was used to code for recidivism that was domestic in nature. Because there is no specific UCR code for domestic violence offenses, a manual review of police reports was necessary to determine if the violent, property, or administration of justice offense was against a current or former intimate partner. Any charge that met this criterion was considered domestic in nature. Thus, the domestic offenses reported herein make up a subcategory of all charges.

General police involvement

In addition to coding for re-arrest, we examined the amount of contact between police and men in the intervention and comparison groups during the 2011 and 2012 year In addition to arrests, charges, and dispositions, the police administrative database records all interactions with citizens categorized into four types: (a) complaints, (b) occurrences, (c) tickets or provincial offense notices (PON), and (d) street checks. Complaints consist of any call for service received by police concerning a person’s welfare, criminal activity, neighborhood problem, motor vehicle accident, and so on. An occurrence is generated as a result of an investigation into a complaint. All individuals involved in the complaint are indexed to an occurrence (i.e., accused, witnesses, victims, etc.). Traffic violations and city by-law infractions (e.g., noise, trespassing) are recorded under the third category. Finally, a street check is used when recording informal information on suspicious persons, unusual activity, or addresses connected to previous criminal activity, etc. Because men in our sample were seldom involved with police for just one incident (domestic or otherwise), amount of police contact was not coded for whether it was or was not related to a domestic-violence-related incident.

Policing time

Total number of police hours devoted to men in the intervention and comparison groups was estimated on the basis of time spent by uniformed officers and by officers in the Criminal Investigation Division (CID). Uniformed officer time was calculated through the computer aided dispatch system that allows for an actual allocation of police hours/minutes spent responding to the initial occurrence with a particular offender. Cases referred to the CID were assigned an average time per event based on an experienced estimate by officers of how many minutes/hours are spent per occurrence. Similar to the category of general involvement with police, policing time was not coded for whether it was or was not related to a domestic-violence-related incident.

Intervention Protocol

Men in the intervention condition attended individual appointments with therapists at Changing Ways, a local social service provider specializing in work with men who have perpetrated domestic violence. The overarching aim of all therapeutic contact was to reduce identified dynamic risks for domestic violence perpetration. Men’s initial appointment was devoted to completion of the Brief Spousal Assault Form for the Evaluation of Risk (B-SAFER), an assessment of dynamic risk for domestic violence offending (Kropp & Hart, 2004). Men were provided with clear and direct feedback on their assessed level of risk (which in all cases was moderately to highly urgent) and on the areas of dynamic risk contributing most to their overall rating. Men were subsequently invited to attend individual sessions with a therapist to address dynamic risk factors. Men could attend as many sessions as required until their charges were heard in front of the court. Sixty-five percent of men elected to attend additional sessions. Among these men, the average number of additional sessions attended was 5.8 with a range of 1 to 20. More specifically, 8 men attended 1 to 4 sessions, 16 men attended 5 to 10 sessions, and 2 men attended more than 10 sessions. The types of services provided to participants included linking to community resources (such as housing and legal advice) for practical assistance, referral to addictions and mental health services, practical support such as accompanying men to a food bank or shelter, and short-term cognitive-behavioural therapy to address abuse-supporting cognitions.

In addition to these services, partners of men in both the intervention and comparison groups received support through an established second-responder program for victims of domestic violence. Specifically, the Family Consultants/Victim Services Unit of the London Police provided on-scene intervention and follow-up support to all victims of domestic violence. This support included feedback on risk for victimization, crisis counselling, safety planning, and referrals to ongoing supports where indicated.

Data Analyses

Analyses were completed using IBM SPSS Statistics Version 21. Comparisons between groups were done using Fisher’s Exact test when expected cell sizes were small, with chi-square for analyses with more balanced categorical data, and with unpaired t tests for continuous data. Analyses examined both occurrence (e.g., arrested at least once) and frequency of offending. In no case were results substantively different; thus, for simplicity and ease of interpretation, only the occurrence results are presented.

Results

Equivalency of the Intervention and Comparison Groups

To ensure equivalence of the intervention and comparison groups, men were compared on demographics, current risk factors for domestic violence, and history of general police involvement. On average, men in both groups were in their low 30s, comparison M = 32.8, intervention M = 34.8, t(78) = .79, p = .432, with more than half employed (comparison 57.6%, intervention 60.6%) and the remainder fairly evenly divided as students (18.2% in both groups) or unemployed (24.2% comparison, 18.2% intervention), and with fewer than five retired men across groups, χ²(3) = 1.31, p = .726. More than half of the men were in committed (i.e., married or common law as opposed to dating or separated) relationships at the time of the index offense—comparison 78.0%, intervention 61.0%, χ²(3) = 3.17, p = .367. Because phone contact with a police-based clinician was the most important determinant of men’s participation in intervention, we were concerned about the potential for meaningful group differences based on men’s homelessness (which may relate to both men’s likelihood of being recruited and their risk for re-offending). Results showed that neither group was likely to be homeless/marginally housed (three men in the intervention group and two in the comparison group recorded as having no fixed address at some point during 2010) and that rates of transience were equivalent across groups, with an average of 2.56 addresses for the intervention group and 2.12 addresses for the comparison during the 2010 year, t(78) = −1.22, p = .226.

Men in the intervention and comparison groups were next compared on their profile of current risk for domestic violence offending, as judged by the responding officer to the target offense on the DVSR. Results, as shown in Table 1, confirm that the intervention and comparison groups showed similarly high levels of risk for re-offense. The victims of close to half of the comparison group and two thirds of the intervention group responded positively to the item “fear that the accused will continue the assaults, seriously injure or kill her or him or the children.” Approximately half reported recent escalation in frequency or severity of violence. One third to one half of the victims reported that the accused had threatened to commit suicide and similar numbers reported that the accused had threatened to harm or kill the victim or her family members. Mental health difficulties were noted for approximately half of the accused and drug or alcohol problems in approximately three quarters of men. There were only two statistically significant differences between the comparison and intervention groups: victims of men in the intervention group were more likely to report that the accused had ever demonstrated jealousy or obsessive behaviour as compared with victims of men in the comparison group, 87.9% versus 66.7%, χ²(1) = 4.23, p = .040, and men in the intervention group were less likely than men in the comparison group to have ever breached a court imposed condition, 25.0% versus 54.2%, χ²(1) = 4.27, p = .039. Given that differences were few (i.e., 2 out of 24) and in opposite directions, no DVSR variables were controlled for in subsequent analyses.

Finally, comparison of past police involvement found that men in the intervention and comparison groups had virtually equivalent involvement with police in the 8 months prior to the index offense (i.e., January 1-August 31, 2010). As expected, given the selection of moderate- to high-risk offenders, more than 90% of men in both the comparison and intervention groups had been charged at least once in 8 months preceding the index offense (see Table 2). The majority of these offenses (85.0%-87.5%) were violent in nature, though a third of men were charged with property offenses and approximately 40% for administration of justice offenses. In consideration of the more global variables of general police involvement and police time, there were minor differences in the comparison and intervention groups, with the comparison group significantly more likely to have been subject to a street check than men in the intervention group. This difference was controlled for in subsequent analyses.

Table 2.

Police Involvement Among Men in the Comparison and Intervention Groups in the Year Prior to the Project Entry (January 1-August 31, 2010).

	Comparison (n = 40)	Intervention (n = 40)	Statistic	p
Prior charges (one or more)
Any charge and arrest	38 (95.0%)	37 (92.5%)	Fisher’s Exact	1.00
Violent offense	34 (85.0%)	35 (87.5%)	χ²(1) = 0.10	.745
Property offense	13 (32.5%)	13 (32.5%)	χ²(1) = 0.0	1.00
Administrative offense	16 (40.0%)	17 (42.5%)	χ²(1) = 0.05	.820
General police involvement
General occurrence	38 (95.0%)	36 (90.0%)	Fisher’s Exact	.675
Street checks	12 (30.0%)	4 (10.0%)	χ²(1) = 5.00	.025*
Complaints	28 (70.0%)	21 (52.5%)	χ²(1) = 2.58	.108
Tickets	11 (27.5%)	8 (20.0%)	χ²(1) = 0.62	.431
Police time
Uniformed Officer Time	19.2 (22.2)	16.1 (16.4)	t(78) = 0.86	.392
Criminal Investigation Division Time	24.6 (20.2)	23.0 (28.1)	t(78) = 0.98	.333

p < .05. **p < .01.

Recidivism in the Intervention and Comparison Groups

To examine the efficacy of intervention, we compared the 1- and 2-year recidivism and police involvement rates for men in the intervention and comparison groups. As shown in Tables 3 and 4, there were significant, substantial, and lasting differences noted across groups across outcome domains. In terms of the subset of re-offending that was domestic in nature, analyses showed that the frequency of subsequent charges was more than double for men in the comparison group as compared with the intervention group in 1 year (65.9% vs. 29.3%) follow-up. The majority of next domestic-violence-related offenses were against the same victim (97.4%) and administrative in nature (66.7% for comparison, 91.7% for intervention), though substantial proportions were violent (29.6% comparison, 8.3% for intervention). Differences were also noted in the more general profile of men’s recidivism. Over a 1-year period, police charged 85.0% of comparison men with at least one additional offense as compared with 45.5% of men in the intervention group. When the nature of these offenses is compared, we found particularly large differences in rates of violent offending, where recidivism was 52.5% for men in the comparison group and only 10% for men in the intervention group. Significant differences favouring the intervention group in 1-year follow-up were also noted for administration of justice offenses (50.0% vs. 20.0%). In contrast, no significant differences were found for property offenses with about one third of men in both the intervention and comparisons groups charged with property offenses in 2011. In terms of general contact with police, compared with the control group, men who received intervention were involved in significantly fewer general occurrences (95.0% vs. 60.0%), received fewer tickets (52.5% vs. 25.0%), and were less likely to be identified by police as part of a street check (55.5% vs. 20.0%). Complaints to police followed the same general pattern. Together, these differences in the amount of contact with police translated into large and meaningful differences in the amount of time police spent interacting with men in the intervention and comparison groups. The mean police time for Uniformed Officers was 18.6 hr for intervention men and 29.6 hr for comparison men; for Criminal Investigation Detectives, mean times were 7.8 and 11.0 hr for the intervention and comparison groups, respectively.

Table 3.

Police Involvement Data for Men in the Comparison and Intervention Groups in Year 1 Follow-Up (January 1-December 31, 2011).

	Comparison (n = 40)	Intervention (n = 40)	Statistic	p
Charges domestic in nature (one or more)
Any charge	27 (65.9%)	12 (29.3%)	χ²(1) = 11.00	.001**
General charges (one or more)
Any charge	34 (85.0%)	18 (45.5%)	χ²(1) = 14.07	.000**
Violent offense	21 (52.5%)	4 (10.0%)	χ²(1) = 16.82	<.001**
Property offense	10 (25.0%)	10 (25.0%)	χ²(1) = 0.0	1.00
Administrative offense	20 (50.0%)	8 (20.0%)	χ²(1) = 7.91	.005**
General police involvement
General occurrence	38 (95.0%)	24 (60.0%)	χ²(1) = 14.05	<.001**
Street checks	22 (55.0%)	8 (20.0%)	χ²(1) = 10.45	.001**
Complaints	30 (75.0%)	22 (55.0%)	χ²(1) = 3.52	.061
Tickets	21 (52.5%)	10 (25.0%)	χ²(1) = 6.37	.012*
Police time
Uniformed Officer Time	18.6 (15.9)	7.8 (10.6)	t(78) = 4.49	<.001**
Criminal Investigation Division Time	29.6 (29.0)	11.0 (14.9)	t(78) = 3.26	.002**

p < .05. **p < .01.

Table 4.

Police Involvement Data for Men in the Comparison and Intervention Groups in Year 2 Follow-Up (January 1-December 31, 2012).

	Comparison (n = 40)	Intervention (n = 40)	Statistic	p
Charges domestic in nature (one or more)
Any charge	17 (41.5%)	5 (12.2%)	χ²(1) = 8.94	.003**
Prior charges (one or more)
Any charge	21 (52.5%)	10 (25.0%)	χ²(1) = 6.37	.012*
Violent offense	10 (25.0%)	6 (15.0%)	χ²(1) = 1.25	.264
Property offense	13 (32.5%)	2 (5.0%)	χ²(1) = 9.93	.002**
Administrative offense	10 (25.0%)	2 (5.0%)	χ²(1) = 6.28	.012*
General police involvement
General occurrence	34 (85.0%)	21 (52.5%)	χ²(1) = 9.83	.002**
Street checks	12 (30.0%)	3 (7.5%)	χ²(1) = 6.65	.009**
Complaints	19 (47.5%)	7 (17.5%)	χ²(1) = 8.20	.004**
Tickets	10 (25.0%)	6 (15.0%)	χ²(1) = 1.25	.264
Police time
Uniformed Officer Time	11.8 (15.6)	4.3 (7.8)	t(78) = 3.24	.002**
Criminal Investigation Division Time	15.4 (21.6)	4.7 (9.3)	t(78) = 2.94	.004**

p < .05. **p < .01.

Followed over 2 years, the same significant and substantial differences were found in re-offending for the intervention versus comparison groups (Table 4). Just over 40% (41.5%) of men in the comparison group had at least one re-offense in Year 2 that was judged as domestic in nature as compared with just over 10% (12. 2%) of men in the intervention groups. Once again, the majority of these offenses were against the same victim and administrative in nature, with a substantial proportion violent. Differences in re-offending were also generally preserved when more general categories of offending are considered. General recidivism rates for men in the intervention program were approximately half of those of comparison men (52.5% vs. 25.0%). Moreover, only 5% of men in the intervention group had offenses in the property and administration of justice categories, as compared with 32.5% and 25.0% for the comparison group. The exception to this pattern was violent offending, with 15% of intervention men and 25% of comparison men having a violent offense change, a difference that was not statistically significant. In terms of general contact with police, effects of intervention were completely retained, with men who received intervention involved in significantly fewer general occurrences (85.0% vs. 52.5%), less likely to be identified by police as part of a street check (47.5% vs. 17.5%) and less likely to be the subject of a complaint (47.5% vs. 17.5%) than men in the comparison group. Once again, differences in recidivism and police involvement translated into large and meaningful differences in the amount of time police spent interacting with men in the intervention and comparison groups, with mean year two time for comparison men for uniformed officers and the Criminal Investigation detectives at 11.8 and 15.4 hr and for the intervention group at 4.3 and 4.7 hr, respectively.

Discussion

In this study, we examined the impact of providing immediate RNR-informed individual intervention focused on dynamic risk factors to men accused of assaulting their intimate partners and who were assessed as being at moderate to high risk for subsequent offending. Results were very promising in terms of the potential of this service to reduce police involvement with moderate- and high-risk domestic violence offenders. Despite being clearly informed that participation in this project would not lead to any leniency in sentencing, 76% of the men who police-based clinicians were able to contact and invite to participate in this intervention chose to do so. Level of follow-up by men offered intervention was also reasonable: 40 of 48 men invited came to a meeting where they were assessed and given feedback on their dynamic risk and two thirds of these men elected to attend additional sessions for support in addressing these risks.

Men who received intervention were significantly and substantially less likely to be arrested for another offense in the 2 years following intervention than a comparison group of men with an equivalent risk profile. In fact, both overall rates of re-arrest and rates of arrest for domestic-violence-related offenses for men in the comparison group were consistently twice as high as those in the intervention group. Group differences in arrest were consistent with reductions in men’s general involvement with police. Specifically, relative to the comparison group, men who attended intervention were involved in significantly fewer general occurrences, were less likely to be identified by police as part of a street check, and were less likely to be the subject of a complaint in the 2 years following intervention. Not surprisingly, these differences resulted in a much lower estimated amount of police time with intervention men over a 2-year perspective than for comparison men.

Although this study is preliminary, current results were surprisingly positive. Results are even more notable in light of historic difficulties in finding large intervention-related reductions in violent offending. The obvious question then is why such a brief intervention appeared to have such a large effect. There are at least three hypotheses to consider. First, results may be due to the success of tailoring intervention to the RNR of the clients. Research on the RNR model in general has been very positive, and at times, consistent with the magnitude of reductions in re-offending noted in the current study (Andrews & Bonta, 2006, 2010). Although attractive, there are problems with this explanation. Most importantly, it is difficult to argue that an intervention lasting an average of just under six sessions could be of sufficient intensity for higher risk clients. Other attempts in Canada and the United States to use the RNR model to organize intervention for domestic violence offenders involve many more hours of service. For example, high-risk domestic violence offenders incarcerated in the Canadian penitentiary system receive 29 three-hour intervention sessions (Connors et al., 2012). In Colorado, men judged as being at high risk to re-offend are mandated to complete twice weekly intervention over a period of months until the offender has developed a set of prescribed competencies (Colorado Domestic Violence Offender Management Board, 2010). Theoretically, this magnitude and intensity of intervention could have been provided to men in the intervention group in the current study in cases where there were frequent meetings between men and counsellors along with sessions with referred services such as housing, employment, and addictions. However, anecdotal reports from counsellors and community members suggest that this is unlikely. Counsellors spoke of an acute lack of available services for men and explained that, even when services existed, they were almost always “closed” to men with pending changes. Thus, lack of access to service and failed referrals were the norm.

A second hypothesis for the large treatment effects relates to the immediacy of intervention. In the study of mental health interventions more generally, longer wait times are associated with worse outcomes and, conversely, more immediate access to intervention is associated with better outcomes (Barrett, Chua, Crits-Christoph, Gibbons, & Thompson, 2008). Long wait times have been associated with higher rates of pre- and post-intake attrition, reduced effectiveness of therapy, and lower client satisfaction (DiMino & Blau, 2012; May, 1991; M. L. Sherman, Barnum, Buhman-Wiggs, & Nyberg, 2009). In the criminal justice field as well, early diversion options are associated with lower rates of re-offending (Schwalbe, Gearing, MacKenzie, Brewer, & Ibrahim, 2012). Given this foundation of research, it is likely that at least some of the positive impacts noted in the current study were due to the immediacy of intervention provided. Still, this explanation is unsatisfying as it fails to address the question of the nature and content of intervention most likely to be helpful.

These considerations lead to a final hypothesis about the potential nature and impact of intervention. Although there was considerable variation in the focus and goals of treatment sessions based on men’s unique dynamic risks, intervention for all men began with completion of the B-SAFER followed by clear feedback on their assessed level of risk for repeat offending. Is it possible that providing this feedback to men was a major contributor to reduced offending? Certainly, risk assessment has been an area of substantial growth in the field of domestic violence. Many jurisdictions now use risk assessment information to guide men’s progress through the criminal justice system. Moreover, best practice guidelines for work with victims of domestic violence generally include a recommendation that women be assessed for, and given clear feedback on, her level of risk for lethal victimization as part of safety planning activities (Campbell, 2004; Campbell & Glass, 2009). There is a fairly large body of literature on the validity of risk classifications (e.g., Campbell, Webster, & Glass, 2009) and a growing literature on the efficacy of matching intervention to the level of risk (e.g., Connors et al., 2012); however, there are virtually no studies of the advantages or potential disadvantages of sharing risk information with clients. An exception is a very small body of literature on the safety planning process for women that includes a few qualitative studies of women’s experiences working with advocates toward safety planning (e.g., Goodkind, Sullivan, & Bybee, 2004) and a single quantitative study of the effect of providing information on assessed risk for lethal violence on problem solving and safety planning of at-risk women (Glass, Eden, Bloom, & Perrin, 2010). There are no studies of the potential impact of providing risk information to offenders.

Outside the domestic violence and criminal justice literature, however, there is a great deal of study of the impact of risk information on subsequent behaviour. In the field of preventive medicine, virtually all theories propose a link between perceptions of personal risk and the likelihood of preventive action (e.g., Becker, 1974; Maddux & Rogers, 1983; Weinstein & Nicolich, 1993).² Moreover, there is a large body of literature supporting the supposition that provision of risk information changes behaviour, with documented associations between perceived risk for negative health outcomes and a range of associated prevention actions (see reviews by Brewer et al., 2007; Milne, Sheeran, & Orbell, 2000). For example, patients’ knowledge about their risk for coronary heart disease predicts lifestyle changes in risk factors such as diet and exercise and adherence to prescribed medication (Alm-Roijer, Fridlund, Stagmo, & Erhardt, 2006) and knowledge of risk for cancer predicts use of preventive screening (Codori, Petersen, Miglioretti, & Boyd, 2001; Murabito et al., 2001). In fact, on the basis of the strength of these theories and findings, prevention campaigns for chronic conditions often include a “knowing your risk” component of public education.

Given this research in the health field, it is possible to imagine that clear communication of risk information to offenders may contribute to reductions in later offending. To further explore this hypothesis, qualitative and process-oriented studies are needed. Information is required on how risk information is shared with offenders and how such information is perceived. It is also important to examine whether men’s knowledge of their risk predicts preventive actions such as recruiting a friend to help monitor and supervise contact with a former partner, avoiding drug or alcohol use when with a potential victim, or seeking supervision for exchanges. More pessimistic possibilities should also be explored, such as the potential for risk information to contribute to men’s justification for offending. Finally, in subsequent studies with larger samples, it would be useful to compare outcomes for a group who received risk assessment feedback only as compared with a group where more intensive intervention was made available.

Although thought provoking, there are a number of reasons to be cautious about the potential promise of this intervention. First, the current manuscript reports on a small preliminary study conducted in one police department in one community. It is unclear if results would generalize to other jurisdictions with different policing policies and community-based services. Second, from a process perspective, the experience of this group suggests that implementing a program such as this is going to face significant challenges in terms of recruitment. Surprisingly, such problems may not stem from men refusing service, but from difficulties in contacting offenders in the first place. More than one half of the moderate- to high-risk men eligible for this study had no phone number in the police file. Alternate methods of contacting clients, or alternate timing of recruitment (perhaps at the bail hearing), should be considered. Program implementation will also face challenges in terms of getting men to attend sessions. Herein, men attended an average of 5.8 sessions, but also missed an average of 2.2 scheduled sessions. Clearly, in organizing an intervention for these moderate to higher risk clients, models of service that anticipate missed sessions are necessary.

Finally, there are some limitations associated with research design that lessen confidence in current findings. The comparison group was not randomly assigned, but rather chosen at random from the group of men who could not be contacted by police. Although we attempted to establish equivalency of groups, there may have been differences in the intervention and comparison groups that we missed. The only way to entirely circumvent this problem is to randomly assign men at the time of recruitment, which would need to be done in a subsequent replication of this study. Second, because of disaggregation of data from the community and police services (required given our ethics protocol and consent process), we were not able to examine whether there was a relationship between the number of intervention sessions men attended and the magnitude of intervention effect. It will be important to find alternate methods of integrating data in a replication, so that such dosage effects can be examined in subsequent studies. Third, it is important to note that recidivism data were collected on the basis of men’s involvement with the local (not national) jurisdiction; thus, it is possible that some re-offending was missed. However, there is no reason a priori to expect higher mobility in the intervention group versus the comparison group (which is the main threat to misinterpretation of our findings). Finally, we note that no data were collected on other services that may have been received by men in the intervention and comparison groups over the 2-year study period. As discussed earlier, anecdotal reports from counsellors and community members suggest that very few alternate services were available to the accused; however, to be sure that differential effects for the intervention and comparison men were not due to other service provision, such data need to be collected in subsequent work.

Despite these limitations, the current study is certainly thought provoking in terms of a potentially hopeful direction for intervention with men accused of domestic violence and judged to be at moderate to high risk for re-offending. It is also relevant philosophically, as a contrasting form of “second-responder” intervention consistent with shifts occurring in the criminal justice system toward therapeutic justice practices (Hannah-Moffat & Maurutto, 2012). If replicated, this study has potentially large implications for further development of integrated justice and therapeutic services to address domestic violence.

Footnotes

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: Research was supported, in part, by a grant from the Ministry of Community Safety and Correctional Services.

Notes

References

Alm-Roijer

Fridlund

Stagmo

Erhardt

L. R.

(2006). Knowing your risk factors for coronary heart disease improves adherence to advice on lifestyle changes and medication. The Journal of Cardiovascular Nursing, 21, E24-E31. doi:http://urn.kb.se/resolve?urn=urn:nbn:se:hj:diva-9799

Andrews

D. A.

Bonta

(2006). The psychology of criminal conduct (4th ed.). Newark, NJ: LexisNexis/Matthew Bender.

Andrews

D. A.

Bonta

(2010). Rehabilitating criminal justice policy and practice. Psychology, Public Policy, and Law, 16, 39-55. doi:10.1037/a0018362

Andrews

D. A.

Bonta

Wormith

J. S.

(2011). The risk-need-responsivity (RNR) model: Does adding the good lives model contribute to effective crime prevention? Criminal Justice and Behavior, 38, 735-755. doi:10.1177/0093854811406356

Babcock

J. C.

Green

C. E.

Robie

(2004). Does batterers’ treatment work? A meta-analytic review of domestic violence treatment. Clinical Psychology Review, 23, 1023-1053. doi:10.1016/j.cpr.2002.07.001

Barrett

M. S.

Chua

Crits-Christoph

Gibbons

M. B.

Thompson

(2008). Early withdrawal from mental health treatment: Implications for psychotherapy practice. Psychotherapy: Theory, Research, Practice, Training, 45, 247-267.

Becker

M. H.

(1974). The health belief model and personal health behavior. Thorofare, NJ: Slack.

Bonta

Andrews

D. A.

Wormith

J. S.

(2006). The recent past and near future of risk and/or need assessment. Crime & Delinquency, 52, 7-27. doi:10.1177/0011128705281756

Bourgon

Hodgson

Hanson

R. K.

Helmus

(2009). The principles of effective correctional treatment also apply to sexual offenders: A meta-analysis. Criminal Justice and Behavior, 36, 865-891. doi:10.1177/0093854809338545

10.

Brewer

N. T.

Chapman

G. B.

Gibbons

F. X.

Gerrard

McCaul

K. D.

Weinstein

N. D.

(2007). Meta-analysis of the relationship between risk perception and health behavior: The example of vaccination. Health Psychology, 26, 136-145.

11.

Campbell

J. C.

(2004). Helping women understand their risk in situations of intimate partner violence. Journal of Interpersonal Violence, 19, 1464-1477.

12.

Campbell

J. C.

Glass

(2009). Safety planning, danger, and lethality assessment. In Mitchell

(Ed.), Intimate partner violence: A health-based perspective (pp. 319-334). New York, NY: Oxford University Press.

13.

Campbell

J. C.

Webster

D. W.

Glass

(2009). The danger assessment: Validation of a lethality risk assessment instrument for intimate partner femicide. Journal of Interpersonal Violence, 24, 653-674.

14.

Capaldi

D. M.

Knoble

N. B.

Shortt

J. W.

Kim

H. K.

(2012). A systematic review of risk factors for intimate partner violence. Partner Abuse, 3, 231-280. doi:10.1891/1946-6560.3.2.231

15.

Casey

R. L.

Berkman

Stover

C. S.

Gill

Durso

Marans

(2007). Preliminary results of a police-advocate home-visit intervention project for victims of domestic violence. Journal of Psychological Trauma, 6, 39-49.

16.

Cattaneo

Goodman

(2005). Risk factors for reabuse in intimate partner violence. Trauma, Violence, & Abuse, 6, 141-175. doi:10.1177/1524838005275088

17.

Codori

Petersen

G. M.

Miglioretti

D. L.

Boyd

(2001). Health beliefs and endoscopic screening for colorectal cancer: Potential for cancer prevention. Preventive Medicine: An International Journal Devoted to Practice and Theory, 33(2, Pt. 1 of 2), 128-136.

18.

Colorado Domestic Violence Offender Management Board. (2010). Standards for treatment with court ordered domestic violence offenders. Lakewood: Colorado Department of Public Safety.

19.

Connors

Mills

Gray

(2012). An evaluation of intimate partner violence intervention with incarcerated offenders. Journal of Interpersonal Violence, 27, 1176-1196. doi:10.1177/0886260511424499

20.

Davis

R. C.

Smith

(1994). Teaching victims crime prevention skills: Can individuals lower their risk of crime? Criminal Justice Review, 19, 56-68.

21.

Davis

R. C.

Taylor

B. G.

Titus

R. M.

(1997). Victims as agents: Implications for victim services and crime prevention. In Davis

R. C.

Lurigio

A. J.

Skogan

W. G.

(Eds.), Victims of crime (2nd ed., pp. 167-179). Thousand Oaks, CA: Sage.

22.

Davis

R. C.

Weisburd

Hamilton

E. E.

(2010). Preventing repeat incidents of family violence: A randomized field test of a second responder program. Journal of Experimental Criminology, 6, 397-418.

23.

Davis

R. C.

Weisburd

Taylor

(2008). Effects of second responder programs on repeat incidents of family abuse. Campbell Systematic Reviews. Advance online publication. doi:10.4073/csr.2008.15

24.

DiMino

Blau

(2012). The relationship between wait time after triage and show rate for intake in a nonurgent student population. Journal of College Student Psychotherapy, 26, 241-247.

25.

Echeburúa

Fernández-Montalvo

de Corral

López-Goñi

(2009). Assessing risk markers in intimate partner femicide and severe violence. Journal of Interpersonal Violence, 24, 925-939. doi:10.1177/0886260508319370

26.

Gamache

Edleson

Schock

(1988). Coordinated police, judicial, and social service response to women battering: A multiple-baseline evaluation across three communities. In Hotaling

G. T.

Finkelhor

Straus

M. A.

Kilpatrick

J. T.

(Eds.), Coping with family violence: Research and policy perspectives (pp. 193-209). Newbury Park, CA: Sage.

27.

Glass

Eden

K. B.

Bloom

Perrin

(2010). Computerized aid improves safety decision process for survivors of intimate partner violence. Journal of Interpersonal Violence, 25, 1947-1964.

28.

Gondolf

E. W.

(2012). The future of batterer programs: Reassessing evidence-based practice. Lebanon, NH: Northeastern University Press.

29.

Goodkind

J. R.

Sullivan

C. M.

Bybee

D. I.

(2004). A contextual analysis of battered women’s safety planning. Violence Against Women, 10, 514-533.

30.

Hannah-Moffat

Maurutto

(2012). Shifting and targeted forms of penal governance: Bail, punishment and specialized courts. Theoretical Criminology, 16, 201-219.

31.

Klein

Tobin

(2008). A longitudinal study of arrested batterers, 1995-2005. Violence Against Women, 14, 136-157. doi:10.1177/1077801207312396

32.

Kropp

P. R.

Hart

S. D.

(2004). The development of the Brief Spousal Assault Form for the Evaluation of Risk (B-SAFER): A tool for criminal justice professionals. Department of Justice Canada. Retrieved from http://publications.gc.ca/collections/Collection/J3-2-2004-1-1E.pdf

33.

Lane

Greenspan

Weisburd

(2004). The second responder program: A coordinated police and social service response to domestic violence. In Fisher

B. S.

(Ed.), Violence against women and family violence: Developments in research, practice, and policy (NCJ 199701, pp. III-2-1). Retrieved from www.ncjrs.gov/pdffiles1/nij/199701.pdf

34.

Maddux

J. E.

Rogers

R. W.

(1983). Protection motivation and self-efficacy: A revised theory of fear appeals and attitude change. Journal of Experimental Social Psychology, 19, 469-479.

35.

Marlowe

D. B.

(2002). Effective strategies for intervening with drug abusing offenders. Villanova Law Review, 47, 989-1025.

36.

May

R. J.

(1991). Effects of waiting for clinical services on attrition, problem resolution, satisfaction, attitudes toward psychotherapy, and treatment outcome: A review of the literature. Professional Psychology: Research and Practice, 22, 209-214.

37.

Milne

Sheeran

Orbell

(2000). Prediction and intervention in health-related behavior: A meta-analytic review of protection motivation theory. Journal of Applied Social Psychology, 30, 106-143.

38.

Murabito

J. M.

Evans

J. C.

Larson

M. G.

Kreger

B. E.

Splansky

G. L.

Freund

K. M.

. . . Wilson

P. W. F.

(2001). Family breast cancer history and mammography: Framingham offspring study. American Journal of Epidemiology, 154, 916-923. doi:10.1093/aje/154.10.916

39.

Office of the Chief Coroner. (2011). Eighth annual report: Domestic violence death review committee—2010. Toronto, Ontario, Canada: Author.

40.

Polaschek

D. L.

(2012). An appraisal of the risk-need-responsivity (RNR) model of offender rehabilitation and its application in correctional treatment. Legal and Criminological Psychology, 17, 1-17. doi:10.1111/j.2044-8333.2011.02038.x

41.

Prendergast

M. L.

Pearson

F. S.

Podus

Hamilton

Z. K.

Greenwell

(2013). The Andrews’ principles of risk, needs, and responsivity as applied in drug treatment programs: Meta-analysis of crime and drug use outcomes. Journal of Experimental Criminology. Advance online publication. doi:10.1007/s11292-013-9178-z

42.

Schwalbe

C. S.

Gearing

R. E.

MacKenzie

M. J.

Brewer

K. B.

Ibrahim

(2012). A meta-analysis of experimental studies of diversion programs for juvenile offenders. Clinical Psychology Review, 32, 26-33. doi:10.1016/j.cpr.2011.10.002

43.

Sherman

L. W.

Berk

R. A.

(1984). The specific deterrent effects of arrest for domestic assault. American Sociological Review, 49, 261-272.

44.

Sherman

M. L.

Barnum

D. D.

Buhman-Wiggs

Nyberg

(2009). Clinical intake of child and adolescent consumers in a rural community mental health center: Does wait-time predict attendance? Community Mental Health Journal, 45, 78-84.

45.

Smedslund

Dalsbø

T. K.

Steiro

Winsvold

Clench-Aas

(2009). Cognitive behavioural therapy for men who physically abuse their female partner. Cochrane Database of Systematic Reviews. Advance online publication. doi:10.1002/14651858.CD006048.pub2

46.

Stover

C. S.

Berkman

Desai

Marans

(2010). The efficacy of a police-advocacy intervention for victims of domestic violence: 12 month follow-up data. Violence Against Women, 16, 410-425.

47.

Stover

C. S.

Poole

Marans

(2009). The domestic violence home-visit intervention: Impact on police-reported incidents of repeat violence over 12 months. Violence and Victims, 24, 591-606.

48.

Weinstein

N. D.

Nicolich

(1993). Correct and incorrect interpretations of correlations between risk perceptions and risk behaviors. Health Psychology, 12, 235-245. doi:10.1037/0278-6133.12.3.235

49.

Wormith

J. S.

Luong

(2011). Applying risk/need assessment to probation practice and its impact on the recidivism of young offenders. Criminal Justice and Behavior, 38, 1177-1199. doi:10.1177/0093854811421596