Recidivism in Intimate Partner Violence Among Antisocial and Family-Only Perpetrators

Abstract

The aim of this study was to compare antisocial and family-only intimate partner violence (IPV) perpetrators in terms of IPV recidivism rates, type of recidivism crime, and time to recidivism. A total sample of 628 perpetrators was categorized as antisocial (n = 327) or family-only (n = 301) based on general criminality. Results demonstrated that antisocial perpetrators recidivated to a larger extent than family-only perpetrators (27.2% vs. 12.9%). Antisocial perpetrators were more prone to recidivate in both physical and nonphysical violence. Of the perpetrators who recidivated, the majority did so within the first year after their index crime. However, antisocial perpetrators recidivated more than family-only perpetrators after the first year. Key findings consisted of the subtypes’ differing propensity to reoffend and their different critical time periods for recidivism. These findings suggest the need for different risk management strategies depending on perpetrator subtype to prevent future violence.

Keywords

intimate partner violence recidivism perpetrators antisocial family-only

Introduction

It has previously been stated that intimate partner violence (IPV) perpetrated by males toward their female partner constitutes a stable and systematic behavioral pattern (e.g., Garcia-Moreno et al., 2013). This implies that recidivism rates in IPV are similar for all such perpetrators. However, due to the demonstrated heterogeneity among IPV perpetrators, it is likely that there also may be differences in recidivism rates between perpetrator subtypes (Cantos & O’Leary, 2014; Cavanaugh & Gelles, 2005). To date, there is a scarcity of research examining whether such recidivism differs between subtypes of IPV perpetrators. Furthermore, the scant literature examining this topic has been inconclusive, as well as demonstrated a lack of detailed analyses of recidivism patterns. As such, the present study sought to address these gaps in the previous literature by outlining the IPV recidivism patterns of antisocial and family-only perpetrator subtypes with greater precision.

In general, recidivism rates among IPV perpetrators have been demonstrated to range between 15% and 60% across studies (e.g., Hilton, Harris, Popham, & Lang, 2010; Klein & Tobin, 2008; Lin et al., 2009; Loinaz, 2014). Considering the World Health Organization’s global estimate that nearly 30% of ever-partnered women have been victimized by IPV (Garcia-Moreno et al., 2013), the extent to which male perpetrators of such violence recidivate has profound negative impacts on women’s health and well-being. In a recent national survey conducted in Sweden, where the present study was carried out, 25% of women aged 16 to 79 years reported ever being victimized by IPV (National Council for Crime Prevention [NCCP], 2014). Furthermore, the IPV recidivism rates among perpetrators of such violence reported to the Swedish police have been found to range between 36% and 42% (Belfrage & Strand, 2012; Svalin, Mellgren, Torstensson Levander, & Levander, 2014). The most commonly reported IPV recidivism crimes in Sweden consist of physical violence (38%) such as assault and sexual violence (Svalin et al., 2014). Nonphysical violence in the forms of illegal threats, harassment, and stalking were found to be the second most commonly reported type of IPV recidivism crimes (33%). From a victim safety perspective, the high extent of recidivism highlights the need to reliably identify those perpetrators constituting a high risk for repeatedly exerting such violence.

Identification of IPV perpetrators most at risk for such recidivism is especially important considering that studies suggest the existence of a critical time period where the risk for IPV recidivism is elevated (e.g., Loinaz, 2014). Based on previous research, it seems that this risk is heightened within the first year after the index crime (Bowen, Gilchrist, & Beech, 2005; Goldstein, Cantos, Brenner, Verborg, & Kosson, 2015; Klein & Crowe, 2008; Lin et al., 2009; Loinaz, 2014; Petrucci, 2010; Richards, Jennings, Tomsich, & Gover, 2014; Stansfield & Williams, 2014). More specifically, several studies have concluded that the highest risk for IPV recidivism is within the first 100 days postintervention (Klein & Crowe, 2008; Lin et al., 2009; Petrucci, 2010; Richards et al., 2014; Stansfield & Williams, 2014; Svalin et al., 2014).

A Swedish study of police-reported IPV perpetrators subjected to a risk assessment for IPV recidivism concluded that such perpetrators were most likely to recidivate within the first month after being subjected to the risk assessment, where 40% of the perpetrators who recidivated did so during this time period (Svalin et al., 2014). In a longitudinal study of perpetrators arrested for IPV-related crimes, Richards et al. (2014) demonstrated that nearly half of the perpetrators who were rearrested for such violence recidivated within 66 days of their prior arrest. These results are consistent with the findings reported by Klein and Crowe (2008) who examined IPV recidivism among perpetrators on probation. They reported that approximately one third of all IPV perpetrators who recidivated in such violence did so within 2 months. In a similar vein, examining perpetrators of IPV who were court-ordered to treatment, Petrucci (2010) concluded that the highest risk for IPV recidivism was within 60 to 90 days of postsentencing of the initial IPV crime. Furthermore, IPV recidivists were rearrested much sooner than IPV perpetrators who recidivated in non-IPV-related crimes. Moreover, Lin et al. (2009) reported that perpetrators of IPV who received court-ordered treatment were more likely to recidivate in such violence within 3 months posttreatment compared with a control group who did not receive such treatment. In addition, although examining family violence recidivism in general among arrested perpetrators, Stansfield and Williams (2014) noted that recidivism in such violence predominantly concerned IPV. More specifically, recidivism in IPV was most likely to occur within the first 100 days after the perpetrators had been subjected to a family violence risk assessment for their index crime of family violence.

It should be noted, however, that some studies report a short-term cessation of violence among IPV perpetrators (e.g., Klein & Tobin, 2008). This short-term cessation of violence, however, may not be indicative of a stable change of violent behavior. As such, Klein and Tobin (2008) reported in their longitudinal study that 32% of the perpetrators recidivated by perpetrating another act of IPV within the first year, whereas this figure rose to 60% within the following decade. Thus, it was concluded that these findings might reflect the chronic abusive nature of such perpetrators. However, in terms of the well-studied domain of IPV perpetrator typologies (e.g., Cavanaugh & Gelles, 2005; Dixon & Browne, 2003; Holtzworth-Munroe & Stuart, 1994), it is reasonable to believe that this conclusion may not be applicable to such perpetrators in general. Due to the heterogeneity of IPV perpetrators, it has been suggested that the subtypes likely will demonstrate different propensities to recidivate in such violence (Cantos & O’Leary, 2014; Cavanaugh & Gelles, 2005). Thus, the conclusion regarding the suggested chronic abusive nature of IPV perpetrators in general (Garcia-Moreno et al., 2013; Klein & Tobin, 2008) warrants further scrutiny.

Originally, Holtzworth-Munroe and Stuart (1994) proposed that IPV perpetrators could be categorized on three descriptive dimensions: (a) severity and frequency of violence, (b) generality of violence, and (c) psychopathology and personality disorder. It was further suggested that categorizing perpetrators according to these dimensions would result in the identification of three different subtypes (Holtzworth-Munroe & Stuart, 1994). Family-only perpetrators were described as using low levels of less severe violence, being violent only within the family, and demonstrating little or no psychopathology and either no personality disorder or a passive-dependent personality disorder. Dysphoric/borderline perpetrators were described as using moderate to high levels of severe violence (including psychological and sexual abuse), being violent predominantly within the family, as well as demonstrating traits of severe depression and borderline personality disorder. Finally, generally violent/antisocial perpetrators were described as frequently using high levels of severe physical violence (including psychological and sexual abuse) and being violent also toward others outside of the family context (e.g., friends and strangers). Moreover, they were the most likely subtype to demonstrate traits of antisocial personality disorder and psychopathy.

Problematically, subsequent efforts to validate this trifold typology have shown inconsistent results. For example, using Holtzworth-Munroe and Stuart’s (1994) approach, researchers have identified a four subtype-solution (e.g., Holtzworth-Munroe, Meehan, Herron, Rehman, & Stuart, 2000, 2003; Serie, van Tilburg, van Dam, & de Ruiter, 2015; Thijssen & de Ruiter, 2011). More problematically, studies have reported difficulties distinguishing dysphoric/borderline perpetrators from generally violent/antisocial perpetrators on key features such as traits of borderline personality disorder (Eckhardt, Holtzworth-Munroe, Norlander, Sibley, & Cahill, 2008; Holtzworth-Munroe et al., 2003; Huss & Ralston, 2008). On a similar note, the majority of proposed typologies have been created using complex statistical methods and instruments requiring clinical expertise. This has called into question the practical utility of such typologies for criminal justice professionals (Cantos, Goldstein, Brenner, O’Leary, & Verborg, 2015; Stoops, Bennett, & Vincent, 2010). Holtzworth-Munroe and Stuart (1994) stated that “ultimately, typologies of batterers will prove relatively useless if they remain descriptive; their potential fruitfulness comes from the ability to use them to better understand the causes and functions of various types of marital violence” (p. 493). In light of the above-mentioned difficulties, combined with the conclusion that the distinction between, as well as existence of, family-only and antisocial perpetrators is particularly strong across studies (Cantos & O’Leary, 2014), this study opted to focus on the family-only and the generally violent/antisocial perpetrators. For the sake of brevity, generally violent/antisocial perpetrators will hereafter be referred to as “antisocial” perpetrators.

Previous studies have demonstrated that the family-only and the antisocial perpetrators can be identified using only the generality of violence dimension suggested by Holtzworth-Munroe and Stuart (1994; Boyle, O’Leary, Rosenbaum, & Hassett-Walker, 2008; Cantos et al., 2015; Goldstein et al., 2015; Petersson, Strand, & Selenius, 2016). As such, the resulting subtypes have shown strong resemblance to the originally proposed family-only and generally violent/antisocial subtypes described by Holtzworth-Munroe and Stuart (1994). More importantly, using only generality of violence to classify IPV perpetrator subtypes provides a practically useful approach for nonclinically trained criminal justice professionals (e.g., Cantos et al., 2015). As suggested by Holtzworth-Munroe and Stuart (1994) and described in subsequent studies (e.g., Boyle et al., 2008; Goldstein et al., 2015; Holtzworth-Munroe et al., 2003; Petersson et al., 2016; Serie et al., 2015), family-only perpetrators are defined as perpetrators who are violent only toward their family, specifically their partner (i.e., nongenerally violent), whereas antisocial perpetrators are defined as those perpetrators who are violent also toward others outside of the family context (i.e., generally violent).

Although antisocial and family-only perpetrators have been described in the literature as a high-risk and a low-risk subtype for IPV, respectively (e.g., Cavanaugh & Gelles, 2005; Holtzworth-Munroe & Stuart, 1994), research on IPV recidivism among these subtypes is inconclusive. In general, studies on this topic have failed to report significant results between the subtypes’ IPV recidivism rates (Eckhardt et al., 2008; Holtzworth-Munroe et al., 2003; Loinaz, 2014; Stoops et al., 2010; Thijssen & de Ruiter, 2011). Nevertheless, there is some evidence that the antisocial perpetrators have a significantly higher propensity to recidivate in such violence compared with family-only perpetrators (Goldstein et al., 2015; Huss & Ralston, 2008). Huss and Ralston (2008) reported that 39.1% of antisocial and 10.6% family-only perpetrators recidivated by perpetrating another act of IPV. Moreover, it was also concluded that antisocial perpetrators recidivated more frequently and more rapidly in such violence compared with family-only perpetrators. However, the authors did not specify time to recidivism among the subtypes (Huss & Ralston, 2008). In addition, Goldstein et al. (2015) demonstrated that antisocial perpetrators were more likely to recidivate compared with family-only perpetrators. However, they did not report on the proportion of perpetrators in each subtype that recidivated.

The antisocial perpetrators’ higher propensity to recidivate in IPV may likely be attributed to their characterization as generally criminal (Petersson et al., 2016; Serie et al., 2015; Thijssen & de Ruiter, 2011). General criminality can be described as the engagement in persistent, frequent, or diverse antisocial behavior, which also includes the use of general violence (Kropp, Hart, & Belfrage, 2010). Importantly, perpetrators’ use of general violence is one of the strongest predictors for violent recidivism (e.g., Stalans, Yarnold, Seng, Olson, & Repp, 2004).

Based on previous findings demonstrating antisocial perpetrators as being more likely to recidivate in IPV (Goldstein et al., 2015; Huss & Ralston, 2008), as well as findings of general violence as a predictor for violent recidivism (Stalans et al., 2004), we hypothesized that antisocial perpetrators would be more likely to recidivate in IPV compared with family-only perpetrators. Moreover, as previously demonstrated by Huss and Ralston (2008), we also hypothesized that antisocial perpetrators would recidivate more quickly than family-only perpetrators. As previously mentioned, generally, criminal behavior consists of persistent and frequent antisocial behavior (Kropp et al., 2010), which would imply faster rates of recidivism among the antisocial perpetrators.

Being able to outline the IPV recidivism patterns of antisocial and family-only perpetrator subtypes with greater precision is important for several reasons. First, previous research has generally been unable to demonstrate significant differences between the two subtypes’ recidivism rates. Second, studies demonstrating differences between the subtypes have not described such differences in sufficient detail. Such a lack in detail mainly concerns analyses of time to IPV recidivism (Huss & Ralston, 2008), as well as the proportion of perpetrators within each subtype that recidivates (Goldstein et al., 2015). Taken together, there is a need to clarify, as well as to provide a more detailed analysis of, the relationship between family-only and antisocial perpetrators in terms of IPV recidivism. Furthermore, such clarification holds several potential practical implications. For example, this could facilitate the difficult task of matching perpetrators with adequate risk management strategies aiming to prevent such violence. As such, this could inform criminal justice professionals about the need for different protective measures depending on perpetrator subtype. For example, some perpetrators and victims might require higher levels of interventions and protective measures during a critical time period where the risk for IPV recidivism is elevated.

The aim of the present study was twofold. First, we sought to compare antisocial and family-only male IPV perpetrators in terms of IPV recidivism rates and type of recidivism crime. Second, we aimed to explore and compare temporal aspects of IPV recidivism between antisocial and family-only perpetrators.

Method

Study Design

The present study was a prospective cross-sectional study using a quantitative design. Furthermore, this study was conducted within the frame of a 7-year prospective research project. The aim of this larger research project was to implement and evaluate the use of structured violence risk assessments within one rural and one remote police district in Sweden. More specifically, the overall aim of the project was to implement and validate the use of structured professional violence risk assessments to assess risk for various forms of family violence. Moreover, the project also examined and evaluated the risk management strategies initiated by the police in such cases. Data from this project have also been used in Petersson et al. (2016).

Procedure

We collected data for all perpetrators who were reported for an IPV-related crime and subjected to a structured violence risk assessment in two Swedish police districts between 2011 and 2014. The Swedish police performed the violence risk assessments using the Brief Spousal Assault Form for the Evaluation of Risk (B-SAFER; Kropp et al., 2010). As such, the reported IPV included in this study was defined according to the Swedish version of this instrument as “the actual, attempted, or threatened physical harm, inflicted by a man or a woman with whom the victim has or has had an intimate, sexual, relationship” (Kropp, Hart, & Belfrage, 2008, p. 2). However, due to the higher severity of male-to-female perpetrated IPV (e.g., Kropp et al., 2010), we chose to exclude female perpetrators of IPV for this study. Following the above definition of IPV, the types of perpetrator–victim relationship in this study consisted of being currently or formerly married, cohabitating, or in a boyfriend–girlfriend relationship while not living together.

All IPV-related crimes that are reported to the Swedish police are subjected to an initial risk assessment to identify those cases where the risk for future IPV is present (The Swedish National Police Board, 2010). This initial risk assessment is made by the individual police officer by the time the IPV crime is reported to the police and is unstructured in terms of being based on the police officer’s personal experience and knowledge of risk for future violence. The aim of this assessment is to provide an initial assessment of the extent to which the crime victim requires protective measures to prevent future victimization. If such measures are required, the police will perform a structured violence risk assessment with the B-SAFER. As a part of using this instrument, IPV perpetrators are, among other risk factors, assessed on “general criminality.” This item in the B-SAFER was used as a proxy for general violence to categorize the IPV perpetrators as either antisocial or family-only (see “Sample” section).

The data collection was carried out at the police headquarter in each district. Furthermore, we constructed a separate coding sheet to retrieve information from the B-SAFER assessments as well as from the police registers. Any personal information that could identify individuals in the sample was excluded.

Recidivism

In the present study, IPV recidivism was defined as any new police report of an IPV-related crime. Thus, this definition of IPV recidivism included both allegations as well as sentences for such criminality. Moreover, this definition of recidivism was broader compared with definitions used in previous studies (e.g., Eckhardt et al., 2008).

We categorized the type of IPV recidivism crime as physically or nonphysically violent. Physically violent crimes consisted of crimes such as attempted homicide, severe violation of a woman’s integrity, assault, and sexual assault. In Sweden, male perpetrators who are found guilty of repeated harassments (i.e., on more than one occasion and toward the same woman) aimed at violating an intimate female partner’s integrity (e.g., through assault, sexual violence, or threats) can be charged with “severe violation of a woman’s integrity.” To our knowledge, this crime is unique for Swedish crime legislation. Furthermore, nonphysically violent crimes consisted of crimes such as illegal threat, harassment, unlawfully entering someone’s home, interference in a judicial matter, arson, stalking, destruction of property, and breaching a restraining order. Moreover, to ascertain the identity of the victim in recidivism cases, supplementary police investigation registers were accessed.

Follow-Up

The follow-up time (i.e., time at risk) in this study consisted of the time between the date of a perpetrator’s police-reported IPV index crime and the final time when recidivism was examined in the police register (i.e., April 2015). As this was a prospective study where the follow-up was done every 6 months, the time at risk for IPV recidivism was not equal for all perpetrators in the sample. The mean follow-up time in this study was 28 months (SD = 11.0, range = 5-50 months). When analyzing the follow-up time between subtypes, it turned out that the antisocial perpetrators (M = 29, SD = 10.9 months) were followed up for a slightly longer time than the family-only perpetrators (M = 26, SD = 10.9 months) were t(628) = −3.26, p = .001.

Sample

A total of 729 IPV perpetrators were initially identified in the police material. However, to be included in the present study, the perpetrators had to meet a set of inclusion criteria: (a) male-to-female perpetrated IPV, (b) being categorized by the police as generally criminal or nongenerally criminal (as indicated in the B-SAFER), and (c) being accessible in the police registers containing recidivism data. In the initial sample of 729 perpetrators, 28 (3.8%) perpetrators were female and thus excluded. Another 44 (6.3%) perpetrators were not coded by the police on the item of general criminality in the B-SAFER and thereby excluded. Finally, 29 (4.4%) perpetrators were inaccessible in the police registers containing information of recidivism and were therefore excluded from the sample. As such, the final sample consisted of 628 male IPV perpetrators.

Moreover, the perpetrators in the sample were categorized as either antisocial or family-only based on each perpetrator’s coding of “general criminality” in the B-SAFER. Perpetrators who were assessed as generally criminal were designated as antisocial (n = 327). The remaining perpetrators (i.e., assessed as not being generally criminal) were classified as family-only (n = 301). The B-SAFER manual states that general criminality is a cursor for general violence (Kropp et al., 2010). Moreover, using general criminality as a proxy for general violence has been done in previous studies differentiating between family-only and antisocial perpetrators (Petersson et al., 2016; Serie et al., 2015; Thijssen & de Ruiter, 2011). Importantly, the resulting family-only and antisocial perpetrators identified in these studies have shown strong resemblance to the subtypes described by Holtzworth-Munroe and Stuart (1994) on the other suggested dimensions of violence severity and psychopathology (Petersson et al., 2016; Serie et al., 2015; Thijssen & de Ruiter, 2011). It should be noted that this is almost the same study sample that we used in our previous study where we compared risk factors for IPV between antisocial and family-only perpetrators (Petersson et al., 2016).

Sample characteristics are presented in Table 1. Antisocial perpetrators were younger than the family-only perpetrators. Likewise, victims of antisocial perpetrators were younger than victims of family-only perpetrators. Moreover, antisocial perpetrators were less likely to be of immigrant background and were to a larger extent reported for a nonphysically violent index crime. Although more antisocial perpetrators were sentenced for their index crime, there was no difference between the antisocial and family-only perpetrators in terms of being imprisoned for this crime.

Table 1:

Differences in Demographic and Legal Characteristics Between 628 Antisocial and Family-Only IPV Perpetrators and Their Victims

Variable	Antisocial perpetrators(n = 327),M (SD)	Family-only perpetrators (n = 301),M (SD)	t test	df	p
Age
Perpetrator	38 (12.4)	41 (13.4)	2.75	626	.006
Victim	34 (12.2)	38 (13.2)	3.70	622	.001
	n (%)	n (%)	χ²	df	p
Relationship status
Cohabitating	250 (76.7)	248 (82.7)	3.43	1	.064
Children (<18 years) living at home	144 (44.6)	152 (51.4)	2.56	1	.109
Immigrant background
Perpetrator	58 (17.7)	86 (28.7)	10.56	1	.001
Victim	39 (11.9)	70 (23.3)	14.17	1	.001
Index crime
Physical violence	241 (75.5)	254 (84.7)	0.63	1	.425
Sexual violence	5 (1.6)	10 (3.3)	2.04	1	.153
Nonphysical violence	64 (20.1)	33 (11.0)	9.61	1	.002
Convicted for index crime	106 (32.4)	74 (24.6)	4.70	1	.030
Imprisoned^a	48 (14.7)	24 (8.0)	3.00	1	.083

Note. Physical violence included attempted murder or manslaughter, severe violation of a woman’s integrity, assault, and unlawful deprivation of liberty. Nonphysical violence included illegal threats, harassment, arbitrary conduct, slander, home invasion, interference in a judicial matter, arson, stalking, destruction of property, and nonphysical violations of a restraining order. Sexual violence consisted of rape. N varies from 628 due to missing data for some variables. IPV = intimate partner violence.

Included perpetrators sentenced to forensic psychiatric care.

Statistical Analyses

Statistical comparisons were carried out with chi-square tests, using odds ratios (OR) with 95% confidence intervals (CI) as measures of effect size. Furthermore, survival analyses were performed to compare the temporal aspects of IPV recidivism among antisocial and family-only perpetrators. Cox proportional hazards model (Cox regression) was performed to analyze whether predictors had a significant impact on the survival time of study participants (Tabachnick & Fidell, 2013). In this study, perpetrator type (i.e., antisocial or family-only) was used as the predictor. Moreover, in the Cox regression analysis, hazard ratios (HR) were used. Rather than merely examining occurrence of events, HR also accounts for time between events (Tabachnick & Fidell, 2013). The Cox regression analysis’ assumption of hazard functions as proportional for any levels of a predictor during the follow-up time (Tabachnick & Fidell, 2013) was not violated.

Survival analysis using the Kaplan–Meier product-limit method was administered to illustrate the survival distributions of study participants as a graphical function (Tabachnick & Fidell, 2013). Moreover, the Kaplan–Meier survival analysis allows for statistical comparisons between groups in survival time. Perpetrators who did not recidivate during their follow-up time (i.e., censored cases) were included in the analyses. Furthermore, differences between survival distributions among antisocial and family-only perpetrators were carried out using the log rank test. According to Stansfield and Williams (2014), “survival analysis is particularly useful in recidivism studies where perpetrators enter the at-risk period at different points along a timeline and censored before the study period ends” (p. 168).

The significance level used in this study was p < .05. Statistical analyses were carried out using IBM SPSS (version 22.0). This study received ethical approval by the Swedish Ethical Review Board.

Results

Recidivism Rates Among Antisocial and Family-Only Perpetrators

Overall, more antisocial perpetrators recidivated in any new criminality compared with the family-only perpetrators (32.2% vs. 16.1%), χ²(1, 628) = 49.98, p < .001 (OR = 3.2; 95% CI = [2.3, 4.4]). Moreover, there were also more antisocial perpetrators who recidivated in both general criminality and IPV compared with family-only perpetrators (11.3% vs. 1.7%), χ²(1, 628) = 23.41, p < .001 (OR = 7.5; 95% CI = [2.9, 10.9]). However, considering the focus of this study, the following results will only concern IPV recidivism among the antisocial and the family-only perpetrators.

The antisocial perpetrators recidivated to a larger extent in IPV compared with the family-only perpetrators (27.2% vs. 12.9%), χ²(1, 628) = 19.64, p < .001 (OR = 2.5; 95% CI = [1.7, 3.8]). Figure 1 displays the amount of IPV recidivism crimes committed by antisocial and family-only perpetrators.

Figure 1:

Number of Recidivism Crimes Among IPV Perpetrators Who Recidivated (n = 128), Divided for Antisocial and Family-Only Perpetrators

As many antisocial and family-only perpetrators recidivated in several crimes (as can be seen in Figure 1), the most severe IPV recidivism crime was selected as the recidivism index crime and included in further analyses. As such, if a perpetrator committed both physically violent and nonphysically violent IPV recidivism crimes, physically violent crimes were deemed as more severe. Crimes of sexual violence were considered as physically violent IPV crimes.

In total, 80 (62.5%) recidivism index crimes were characterized as physically violent, where one recidivism crime consisted of sexual violence. Moreover, 47 (36.7%) recidivism index crimes were nonphysically violent. In one (0.8%) case of recidivism, information concerning the specific crime committed was missing. More antisocial perpetrators than family-only perpetrators (15.9% vs. 9.3%) recidivated in a physically violent index crime, χ²(1, 627) = 6.06, p = .014 (OR = 1.8; 95% CI = [1.1, 3.0]). This difference also held true for nonphysically violent recidivism where more antisocial perpetrators were reported for such an index crime compared with the family-only perpetrators (11.3% vs. 3.3%), χ²(1, 627) = 14.38, p < .001 (OR = 3.7; 95% CI = [1.8, 7.6]).

Time to IPV Recidivism Among Antisocial and Family-Only Perpetrators

First, a Cox regression survival analysis was carried out to estimate the impact of perpetrator type (i.e., antisocial or family-only) on time to recidivism. The Cox regression model was significant, χ²(1, 628) = 6.49, p = .011, and predicted that antisocial perpetrators were more likely to recidivate faster compared with family-only perpetrators (β = .49, p = .012, HR = 1.6; 95% CI = [1.1, 2.4]). In terms of HR, the antisocial perpetrators were 1.6 times more likely to recidivate faster compared with family-only perpetrators.

Second, a Kaplan–Meier survival analysis was performed to confirm the results of the Cox regression analysis. The Kaplan–Meier analysis graphically illustrated and determined whether there were significant differences between antisocial and family-only perpetrators’ survival distributions regarding time to recidivism. As indicated by Figure 2 and the log rank test, antisocial perpetrators recidivated significantly faster compared with the family-only perpetrators, χ²(1, 628) = 6.62, p = .010.

Figure 2:

Kaplan–Meier Survival Analysis of IPV Recidivism Among Family-Only and Antisocial Perpetrators (N = 628)

Table 2 offers a more detailed account of IPV recidivism among antisocial and family-only perpetrators during the first year after the reported index crime. As can be seen in Table 2, antisocial and family-only perpetrators recidivated to the same extent during the first 8 months. However, antisocial perpetrators were more likely to recidivate between 9 to 10 months. There was no difference between the antisocial and the family-only perpetrators in recidivism between 11 and 12 months after the reported index crime. Finally, antisocial perpetrators were more than 3 times as likely to recidivate after the first year compared with family-only perpetrators. Further inspection of Table 2 also reveals that, of the perpetrators who recidivated, the majority of family-only (n = 30, 76.9%) and antisocial (n = 60, 67.4%) perpetrators recidivated within the first year after their reported index crime.

Table 2:

Proportion of Antisocial and Family-Only IPV Perpetrators at Risk and Recidivating in IPV, Divided in 2-Months Intervals and Later Than 12 Months (N = 628)

Time interval (months)	Perpetrators recidivating in IPV
	Antisocial perpetrators(n = 327)				Family-only perpetrators(n = 301)				Differences in recidivism between subtypes
	At risk (n)	n	%	Cumulative %	At risk (n)	n	%	Cumulative %	χ²	df	p	OR	95% CI
0-2	327	14	4.3	4.3	301	9	3.0	3.0	0.74	1	.388	—	—
3-4	327	15	4.6	8.9	301	6	1.9	4.9	3.26	1	.059	—	—
5-6	327	8	2.4	11.3	301	5	1.7	6.6	0.48	1	.538	—	—
7-8	320	8	2.4	13.7	290	7	2.3	8.9	0.01	1	.934	—	—
9-10	311	10	3.1	16.8	287	1	0.3	9.2	6.77	1	.033	9.5	[1.2, 74.4]
11-12	304	5	1.5	18.3	267	2	0.7	9.9	1.06	1	.341	—	—
Later	294^a–1^b	29	8.9	27.2	256^c–2^d	9	3.0	12.9	9.53	1	.003	3.2	[1.5, 6.7]

Note. Later = 13 to 47 months after the initial index IPV crime. IPV = intimate partner violence; OR = odds ratio; CI = confidence interval.

13 months follow-up. ^b50 months follow-up. ^c13 months follow-up. ^d48 months follow-up.

As a final examination of the temporal aspects of recidivism among antisocial and family-only perpetrators, we compared time to nonphysically and physically violent IPV recidivism. Thus, we performed additional Kaplan–Meier analyses where time to recidivism for antisocial and family-only perpetrators was examined in relation to type of recidivism index crime. The results are demonstrated in Figure 3.

Figure 3:

Kaplan–Meier Survival Analyses of Time to Physically (n = 80), As Well As Nonphysically (n = 47), Violent IPV Recidivism Among Family-Only and Antisocial Perpetrators

In terms of time to physically violent recidivism, the Kaplan–Meier survival analysis (see Figure 3) and the log rank test demonstrated that there was no difference between the antisocial and the family-only perpetrators, χ²(1, 628) = 1.37, p = .240. However, there were differences between antisocial and family-only perpetrators in terms of time to nonphysically violent IPV recidivism, χ²(1, 628) = 8.04, p = .005.

Discussion

The aim of this study was to compare antisocial and family-only perpetrators in terms of IPV recidivism rates, type of recidivism crime, and time to recidivism. The results demonstrated that antisocial perpetrators were more than twice as likely to recidivate in IPV compared with the family-only perpetrators (27.2% vs. 12.9%). Moreover, antisocial perpetrators were more likely to recidivate in both physically and nonphysically violent IPV. In terms of time to IPV recidivism, two major results were found. First, of those perpetrators who recidivated by perpetrating another act of IPV, the majority recidivated within the first year after their reported index crime. Second, antisocial perpetrators were found to recidivate faster than family-only perpetrators in nonphysical IPV. However, no difference was evident in terms of time to physical IPV recidivism among the subtypes.

Unlike the majority of previous studies comparing IPV recidivism rates among antisocial and family-only perpetrators (Eckhardt et al., 2008; Loinaz, 2014; Stoops et al., 2010; Thijssen & de Ruiter, 2011), we were able to report significant differences between these two subtypes. In line with our hypothesis, as well as the results of Huss and Ralston (2008), we found that antisocial perpetrators were more likely to recidivate in any new IPV-related crime compared with family-only perpetrators (27.2% vs. 12.9%). This could be explained with our use of a larger sample size and a longer follow-up period than used in previous studies of criminal justice samples (e.g., Eckhardt et al., 2008; Loinaz, 2014; Thijssen & de Ruiter, 2011). The antisocial perpetrators’ higher propensity to recidivate may likely be ascribed to their characterization as generally criminal. As noted by Kropp et al. (2010), generally criminal perpetrators are described as engaging in frequent and persistent antisocial behavior. Moreover, antisocial perpetrators have demonstrated more risk factors associated with IPV recidivism than family-only perpetrators (e.g., Cavanaugh & Gelles, 2005; Holtzworth-Munroe et al., 2003). Thus, their higher recidivism rate is rather unsurprising and their IPV may likely reflect an overall pattern of violent behavior and part of a criminal lifestyle. This assumption is consistent with the results of our study on risk factors for IPV in this sample (Petersson et al., 2016).

The difference in recidivism rates between antisocial and family-only perpetrators could also be explained using the stake in conformity hypothesis. Toby (1957) argued that compared with individuals who have weak ties to conventional society, individuals who have strong such ties (e.g., being married, employed, and having a high level of education) have more to lose in committing a crime. Such individuals not only risk losing their job and their marriage, but perhaps more importantly, they risk the loss of social approval and social status among friends and neighbors (Goldstein et al., 2015; Klein, Wilson, Crowe, & DeMichele, 2008; Toby, 1957). On the contrary, individuals with weaker ties to conventional society (e.g., unmarried, unemployed, and low levels of education) already consider themselves as outsiders and have less to lose in terms of the social reactions of others (Toby, 1957). Thus, such individuals are more inclined to commit crimes. In support of this hypothesis, Sherman, Smith, Schmidt, and Rogan (1992) demonstrated that perpetrators with a high stake in conformity were to a larger extent deterred from recidivating in IPV following an arrest, compared with perpetrators with a low stake in conformity.

Arguably, family-only perpetrators can be seen as having higher stakes in conformity compared with antisocial perpetrators (Cantos & O’Leary, 2014; Goldstein et al., 2015). The notion of considering the family-only and antisocial perpetrators in this study as having a high and a low stake in conformity, respectively, is largely based on past research. In our previous study of this sample, we demonstrated that family-only perpetrators were more psychosocially well-adjusted, in terms of having less risk factors for IPV and being assessed with lower risk for IPV recidivism (Petersson et al., 2016). In contrast, antisocial perpetrators evidenced more risk factors for IPV (e.g., problems related to employment, intimate relationships, substance use, and mental health) and also received higher risk ratings for IPV recidivism. These results were consistent with previous findings (e.g., Serie et al., 2015; Thijssen & de Ruiter, 2011). Moreover, Walsh et al. (2010) found that family-only perpetrators demonstrated a higher degree of positive social functioning compared with antisocial perpetrators. In combination with the results of IPV recidivism rates in this study, these findings support the notion of family-only perpetrators as having a higher stake in conformity than antisocial perpetrators (Petersson et al., 2016; Serie et al., 2015; Thijssen & de Ruiter, 2011; Walsh et al., 2010). This seems to be a plausible explanation as to why antisocial perpetrators recidivated to a larger extent compared with family-only perpetrators. However, future research is encouraged to more thoroughly test the stake in conformity theory in relation to family-only and antisocial perpetrators’ characteristics.

In terms of types of recidivism crime, the results demonstrated that antisocial perpetrators were more likely to recidivate in physically as well as nonphysically violent IPV. This is consistent with prior findings (Holtzworth-Munroe et al., 2003) and further establishes the antisocial perpetrator as the most violent subtype, compared with the family-only perpetrator. Moreover, physically violent IPV recidivism in this study was deemed as more severe compared with nonphysically violent IPV recidivism. As such, the potential consequences of physical IPV (e.g., homicide or severe physical injury) were deemed as more severe than the consequences of nonphysical IPV. Moreover, the criminal justice system considers physical violence as more severe compared with psychological violence, as well as imposes harsher punishments for physical violence in general. However, while making this distinction, we did not intend to minimize the effects of psychological abuse. Noteworthy, the use of severe violence has been reported to be predictive of future such violence among antisocial perpetrators, but not among family-only perpetrators (Goldstein et al., 2015). Therefore, victims of severe forms of IPV committed by antisocial perpetrators should therefore be prioritized in terms of being offered more intense protective measures (e.g., protected living) by the criminal justice system. At the same time, the antisocial perpetrators are more likely to require higher levels of interventions to minimize the risk for future severe IPV.

Another key finding in this study relates to the identification of a critical time period for IPV recidivism among antisocial and family-only perpetrators. Although the majority of antisocial (67.4%) and family-only (76.9%) perpetrators recidivated within the first year, their recidivism rates began to differ significantly after a year of follow-up. Antisocial perpetrators recidivated to a larger extent compared with family-only perpetrators after the first year following their index crime. However, the significant result in Table 2 for 9 to 10 months can most likely be explained in terms of methodological issues. The low amount of recidivists suggests that the difference between the two subtypes is a result of a lack of statistical power rather than a true difference. Thus, our interpretation is that antisocial and family-only perpetrators diverge in their recidivism rates after the first year following their index crime. This suggests that the critical time period for antisocial perpetrators to recidivate is more long-term in nature. Several authors have argued that the existence of a critical time period for IPV recidivism is important from a victim safety perspective and should guide the response from the criminal justice system to prevent future IPV (Klein et al., 2008; Richards et al., 2014). Drawing on the results of this study, interventions and protective measures offered by the criminal justice system should be in effect as soon as possible following a police-reported IPV crime regardless of perpetrator subtype. Importantly, such interventions need to be in effect for at least the first year and even for a longer period of time for the antisocial perpetrators, to reduce IPV recidivism.

More detailed examination of time to recidivism also revealed that antisocial perpetrators recidivated faster in nonphysical IPV compared with the family-only perpetrators (see Figure 3). This was partially in line with our hypothesis that antisocial perpetrators would recidivate faster compared with family-only perpetrators. These results indicate that nonphysical IPV (e.g., threats and harassments) is a frequent and stable component to antisocial perpetrators’ violent behavior toward their partners. The notion that nonphysical IPV is a stable behavior among antisocial perpetrators in this study is supported by the sample characteristics (see Table 1) demonstrating that such perpetrators were more likely to have committed a nonphysically violent IPV index crime. In further support of this conclusion, Holtzworth-Munroe et al. (2003) demonstrated that antisocial perpetrators, compared with family-only perpetrators, were more consistently psychologically abusive over time toward their female partner. Previous research (see O’Leary, 1999, for a review) has reported that psychological abuse has as deleterious and negative impact on women’s health as physical violence. Psychological abuse has also been demonstrated as a precursor to physical violence (Kropp et al., 2010; O’Leary, 1999), as well as co-occurring with physical violence (e.g., NCCP, 2014). Thus, it is important to consider previous nonphysically violent IPV (e.g., threats of violence) as equally important as previous physical IPV when assessing the risk for IPV recidivism, especially among antisocial perpetrators.

The nonsignificant results concerning time to physical IPV were partially in contrast to our hypothesis. We initially expected that antisocial perpetrators would recidivate faster than family-only perpetrators in any IPV (i.e., both physical and nonphysical). To our knowledge, no previous study has examined time to recidivism among antisocial and family-only perpetrators in such a detailed manner. The only previous study examining time to IPV recidivism (Huss & Ralston, 2008) did not differentiate between time to physical and nonphysical IPV. This complicates comparisons between the results in this study and previous findings. However, a possible explanation to the nonsignificant result concerning time to physical IPV recidivism could be that victims of family-only perpetrators might be more prone to contact the police when they are being victimized of physical violence, compared with when being victimized by nonphysical violence. As family-only perpetrators generally use low levels of violence (e.g., Holtzworth-Munroe et al., 2003), victims’ of such perpetrators might wait to contact the police until an act of more severe violence (i.e., physical violence) occurs. However, this hypothesized tendency among victims of family-only perpetrators not to contact the police until an act of physical violence occurs needs to be examined in future research.

Strengths and Limitations

Previous studies of time to IPV recidivism among antisocial and family-only perpetrators are scarce. Only one study (Huss & Ralston, 2008) was identified and the authors did not provide detailed information of the time to, nor type of, recidivism among such perpetrators. Thus, the current study is considered an improvement in terms of outlining the temporal aspects of IPV recidivism with greater precision, as well as novel in terms of considering type of IPV recidivism crime for the antisocial and family-only perpetrators. We also used a longer follow-up time as well as a larger sample size, allowing for stronger power in statistical analyses compared with previous studies. In addition, by relying on police records for recidivism data, the present study were attentive to previous calls advising this source of information (Loinaz, 2014; Stoops et al., 2010). However, the use of official records measuring IPV recidivism has its limitations. Several studies have shown that relying on such data greatly underestimates the amount of actual recidivism (e.g., Goldstein et al., 2015). Although not possible in this study, the use of victim reports to measure recidivism would likely demonstrate higher recidivism rates, as well as more accurate information concerning the time to recidivism. Usually, only the more severe forms of IPV come to the criminal justice system’s attention (Sartin, Hansen, & Huss, 2006), indicating that the less severe forms of violence largely go unreported. Moreover, in this study we used any new police report of an IPV-related crime to measure recidivism. As we did not solely rely on convictions, we do not know if all the alleged perpetrators in this study actually committed the crime.

Moreover, it can be seen as a limitation that the antisocial perpetrators had a longer follow-up time compared with the family-only perpetrators. However, as illustrated in Table 2, the majority of all perpetrators recidivated within the first year. As such, the 3 months longer mean follow-up time for the antisocial perpetrators was not considered a limitation in this study. Finally, using general criminality as a corresponding measure of generally violent and antisocial behavior could also be seen as a limitation. General criminality is, however, described in the B-SAFER manual as an indicator of antisocial behavior (including general violence: Kropp et al., 2010). Furthermore, this procedure has been adopted in previous studies when categorizing family-only and antisocial perpetrators (Petersson et al., 2016; Serie et al., 2015; Thijssen & de Ruiter, 2011).

Practical Implications

Several results in this study have direct practical implications for various actors within the criminal justice system. A first step for criminal justice professionals assigned to perform IPV risk assessments should be to categorize the perpetrator as antisocial or family-only, based on general criminality or similar measures related to generality of violence. Drawing on the results of this study, this procedure would facilitate the identification of perpetrators most at risk for recidivism (i.e., antisocial perpetrators). Thus, interventions aimed at preventing future IPV should primarily be directed toward antisocial perpetrators. Subsequently, to minimize the risk for IPV recidivism, the next step would be to match perpetrator subtype with adequate interventions (Cavanaugh & Gelles, 2005). Importantly, the frequently employed strategy of court-mandating perpetrators to treatment has been deemed nonbeneficial for antisocial perpetrators (Cantos & O’Leary, 2014). Due to their high rates of treatment attrition (e.g., Huss & Ralston, 2008), pretreatment motivational interventions have been suggested as a first step focusing interventions on raising their stake in conformity (e.g., targeting substance abuse, unemployment, low income, and education; Cantos & O’Leary, 2014). However, there is support that IPV treatment is effective for family-only perpetrators (Cantos & O’Leary, 2014; Huss & Ralston, 2008).

Perhaps more importantly, the above-mentioned perpetrator–intervention matching needs to be done quickly. Regardless of perpetrator subtype, the first year following the reported index crime was a critical time period for recidivism in this study, with several perpetrators recidivating during the first few months. This further supports the recommendations of implementing “fast track” policies for IPV perpetrators to prevent them from recidivating (Klein et al., 2008; Richards et al., 2014). From a temporal perspective, this could include striving to avoid delays when working with IPV cases in police investigations, trials, and sentencing of perpetrators, for the criminal justice system to intervene against the perpetrators as soon as possible. Moreover, any sentences or sanctions directed toward IPV perpetrators, as well as protective measures directed toward their victims, should be in effect for the first year following the reported index crime. However, our results indicate that compared with family-only perpetrators, interventions for antisocial perpetrators, as well as protective measures for their victims, need to be in effect beyond the first year. Finally, when an IPV case goes through the criminal justice system, it is important to understand that both antisocial and family-only perpetrators can pose a high risk for recidivism. Our results show that it is a misconception that all family-only perpetrators are judged as low risk, while antisocial perpetrators are all high risk (e.g., Cavanaugh & Gelles, 2005).

Conclusion

Antisocial perpetrators were more than twice as likely as family-only perpetrators to recidivate in any type of IPV, although recidivism occurred for family-only perpetrators as well. This supports the assertion of antisocial perpetrators as constituting a high-risk subtype for IPV recidivism (e.g., Cavanaugh & Gelles, 2005; Petersson et al., 2016). Moreover, the critical time period for IPV recidivism for antisocial perpetrators were longer compared with family-only perpetrators, indicating that antisocial perpetrators require long-term interventions and monitoring by the criminal justice system. The present study highlights the usefulness of identifying perpetrators as antisocial or family-only in a risk assessment context. Such identification can provide the risk assessor with valuable information regarding the level and timing of interventions necessary to prevent future IPV.

Footnotes

Acknowledgements

The authors would like to thank the police officers in the counties of Västernorrland and Jämtland for their participation in this research.

This research was funded by a grant from the Swedish Crime Victim Compensation and Support Authority.

Joakim Petersson received his MSc in criminology from Mid Sweden University, Sweden, in 2013. He is currently a lecturer in criminology at Örebro University in Sweden and a researcher at CAPS (Center for Criminological and Psychosocial Research) at Örebro University. He is also a PhD student at Mid Sweden University.

Susanne Strand is an associate professor of criminology at CAPS at Örebro University in Sweden and at the Centre for Forensic Behavioural Science at Swinburne University of Technology in Melbourne, Australia. The focus of her research is risk assessment and risk management in three different areas: intimate partner violence, stalking, and honor-based violence.

References

Belfrage

Strand

(2012). Measuring the outcome of structured spousal violence risk assessments using the B-SAFER: Risk in relation to recidivism and intervention. Behavioral Sciences & the Law, 30, 420-430.

Bowen

Gilchrist

E. A.

Beech

A. R.

(2005). An examination of the impact of community-based rehabilitation on the offending behaviour of male domestic violence offenders and the characteristics associated with recidivism. Legal and Criminological Psychology, 10, 189-209.

Boyle

D. J.

O’Leary

K. D.

Rosenbaum

Hassett-Walker

(2008). Differentiating between generally and partner-only violent subgroups: Lifetime antisocial behavior, family of origin violence, and impulsivity. Journal of Family Violence, 23, 47-55.

Cantos

A. L.

Goldstein

D. A.

Brenner

O’Leary

K. D.

Verborg

(2015). Correlates and program completion of family-only and generally violent perpetrators of intimate partner violence. Behavioral Psychology, 23, 549-569.

Cantos

A. L.

O’Leary

D. K.

(2014). One size does not fit all in treatment of intimate partner violence. Partner Abuse, 5, 204-236.

Cavanaugh

M. M.

Gelles

R. J.

(2005). The utility of male domestic violence offender typologies. Journal of Interpersonal Violence, 20, 155-166.

Dixon

Browne

(2003). The heterogeneity of spouse abuse: A review. Aggression and Violent Behavior, 8, 107-130.

Eckhardt

Holtzworth-Munroe

Norlander

Sibley

Cahill

(2008). Readiness to change, partner violence subtypes, and treatment outcomes among men in treatment for partner assault. Violence and Victims, 23, 446-475.

Garcia-Moreno

Pallitto

Devries

Stöckl

Watts

Abrahams

Petzold

(2013). Global and regional estimates of violence against women: Prevalence and health effects of intimate partner violence and non-partner sexual violence. Geneve, Switzerland: World Health Organization.

10.

Goldstein

D. A.

Cantos

A. L.

Brenner

L. H.

Verborg

R. J.

Kosson

D. S.

(2015). Perpetrator type moderates the relationship between severity of intimate partner violence and recidivism. Criminal Justice and Behavior, 43, 879-898.

11.

Hilton

N. Z.

Harris

G. T.

Popham

Lang

(2010). Risk assessment among incarcerated male domestic violence offenders. Criminal Justice and Behavior, 37, 815-832.

12.

Holtzworth-Munroe

Meehan

J. C.

Herron

Rehman

Stuart

G. L.

(2000). Testing the Holtzworth-Munroe and Stuart (1994) batterer typology. Journal of Consulting and Clinical Psychology, 68, 1000-1019.

13.

Holtzworth-Munroe

Meehan

J. C.

Herron

Rehman

Stuart

G. L.

(2003). Do subtypes of maritally violent men continue to differ over time? Journal of Consulting and Clinical Psychology, 71, 728-740.

14.

Holtzworth-Munroe

Stuart

G. L.

(1994). Typologies of male batterers: Three subtypes and the differences among them. Psychological Bulletin, 116, 476-497.

15.

Huss

M. T.

Ralston

(2008). Do batterer subtypes actually matter? Treatment completion, treatment response, and recidivism across a batterer typology. Criminal Justice and Behavior, 35, 710-724.

16.

Klein

A. R.

Crowe

(2008). Findings from an outcome examination of Rhode Island’s specialized domestic violence probation supervision program: Do specialized supervision programs of batterers reduce reabuse? Violence Against Women, 14, 226-246.

17.

Klein

A. R.

Tobin

(2008). A longitudinal study of arrested batterers, 1995-2005: Career criminals. Violence Against Women, 14, 136-157.

18.

Klein

A. R.

Wilson

Crowe

A. H.

DeMichele

(2008). Evaluation of the Rhode Island probation specialized domestic violence supervision unit (American Probation and Parole Association and BOTEC Analysis Corporation). Retrieved from https://www.ncjrs.gov/pdffiles1/nij/grants/222912.pdf

19.

Kropp

P. R.

Hart

S. D.

Belfrage

(2008). Bedömning av risk för upprepat partnervåld (SARA:SV) Version 2: Version 2, Användarmanual [Assessing the risk for spousal assault recidivism], Swedish translation of the Brief spousal assault form for the evaluation of risk [B-SAFER], User manual. Vancouver: Proactive Resolutions. Sundsvall, Sweden: Rättspsykiatriska Regionkliniken.

20.

Kropp

P. R.

Hart

S. D.

Belfrage

(2010). Brief spousal assault form for the evaluation of risk (B-SAFER, User manual, 2nd ed.). Vancouver, British Columbia, Canada: Proactive Resolutions.

21.

Lin

S.-C.

C.-Y.

Chou

H.-C.

Chen

S.-P.

Huang

J.-J.

T.-E.

. . . Chen

C.-C.

(2009). Domestic violence recidivism in high-risk Taiwanese offenders after the completion of violence treatment programs. The Journal of Forensic Psychiatry & Psychology, 20, 458-472.

22.

Loinaz

(2014). Typologies, risk and recidivism in partner-violent men with the B-SAFER: A pilot study. Psychology, Crime & Law, 20, 183-198.

23.

National Council for Crime Prevention. (2014). Brott i nära relationer: En nationell kartläggning (Brå- rapport 2014: 8) [Offences in close relationships: A national survey]. Stockholm, Sweden: Fritzes.

24.

O’Leary

D. K.

(1999). Psychological abuse: A variable deserving critical attention in domestic violence. Violence and Victims, 14, 3-23.

25.

Petersson

Strand

Selenius

(2016). Risk factors for intimate partner violence: A comparison of antisocial and family-only perpetrators. Journal of Interpersonal Violence. Advance Online Publication. doi:10.1177/0886260516640547

26.

Petrucci

C. J.

(2010). A descriptive study of a California domestic violence court: Program completion and recidivism. Victims & Offenders, 5, 130-160.

27.

Richards

T. N.

Jennings

W. G.

Tomsich

Gover

(2014). A 10-year analysis of rearrests among a cohort of domestic violence offenders. Violence and Victims, 29, 887-906.

28.

Sartin

R. M.

Hansen

D. J.

Huss

M. T.

(2006). Domestic violence treatment response and recidivism: A review an implications for the study of family violence. Aggression and Violent Behavior, 11, 425-440.

29.

Serie

C. M. B.

van Tilburg

C. A.

van Dam

de Ruiter

(2015). Spousal assaulters in outpatient mental health care: The relevance of structured risk assessment. Journal of Interpersonal Violence, 1, 1-20.

30.

Sherman

L. W.

Smith

D. A.

Schmidt

J. D.

Rogan

D. P.

(1992). Crime, punishment, and stake in conformity: Legal and informal control of domestic violence. American Sociological Review, 57, 680-690.

31.

Stalans

L. J.

Yarnold

P. R.

Seng

Olson

D. E.

Repp

(2004). Identifying three types of violent offenders and predicting violent recidivism while on probation: A classification tree analysis. Law and Human Behavior, 28, 253-271.

32.

Stansfield

Williams

K. R.

(2014). Predicting family violence recidivism using the DVSI-R: Integrating survival analysis and perpetrator characteristics. Criminal Justice and Behavior, 41, 163-180.

33.

Stoops

Bennett

Vincent

(2010). Development and predictive ability of a behavior-based typology of men who batter. Journal of Family Violence, 25, 325-335.

34.

Svalin

Mellgren

Torstensson Levander

Levander

(2014). Utvärdering av SARA:SV: Uppsalapolisens arbete med riskbedömningar för upprepade brott i nära relationer [Evaluation of the B-SAFER: Working with risk assessments for repeat intimate partner violence in the police county of Uppsala] (The Swedish National Police Board’s report, Vol. 2014, No. 6). Malmö, Sweden: Institutionen för kriminologi, Malmö högskola.

35.

The Swedish National Police Board. (2010). Riktlinjer för polisära riskanalyser vid våld på individnivå [Guidelines for police risk assessments for interpersonal violence]. Stockholm: The Swedish National Police Board.

36.

Tabachnick

B. G.

Fidell

L. S.

(2013). Using multivariate statistics (6th ed.). Boston, MA: Pearson Education.

37.

Thijssen

de Ruiter

(2011). Identifying subtypes of spousal assaulters using the B-SAFER. Journal of Interpersonal Violence, 26, 1307-1321.

38.

Toby

(1957). Social disorganization and stake in conformity: Complementary factors in the predatory behavior of hoodlums. Journal of Criminal Law and Criminology, 48, 12-17.

39.

Walsh

Swogger

M. T.

O’Connor

B. P.

Chatav Schonbrun

Shea

M. T.

Stuart

G. L.

(2010). Subtypes of partner violence perpetrators among male and female psychiatric patients. Journal of Abnormal Psychology, 119, 563-574.