Abstract
This systematic review is the first to examine treatment and recidivism outcomes in relation to typologies of perpetrators of intimate partner violence (IPV). Twenty-four (n = 24) studies presented treatment outcomes for different types of perpetrators, and n = 25 presented recidivism outcomes. Half of the studies that reported treatment (n = 12) and recidivism (n = 13) outcomes for different types classified men who used IPV using variations of the family-only/generally violent typology (including two, three, or four groups). Typology studies that examined treatment or recidivism outcomes also differentiated perpetrators by reactive or instrumental violence, severity and frequency of violence, personality type, or treatment responsivity, in addition to other unique typologies. Across various typologies, family-only perpetrators were significantly more likely to complete treatment (compared to generally violent perpetrators), and, overall, individuals deemed lower risk were more likely to complete treatment. Perpetrators classified as generally violent or antisocial consistently showed higher recidivism rates than those classified as family-only. Traits such as antisocial attitudes, anger, instability, and low readiness to change were also linked to a greater risk of reoffending, and, overall, individuals deemed higher risk were more likely to reoffend. This systematic review summarizes existing evidence indicating that typology impacts treatment and recidivism outcomes for different types of perpetrators; however, it also highlights the need for further investigation of recidivism and treatment outcomes across these groups.
Keywords
Introduction
It is clear from decades of research that there are different types of perpetrators 1 of intimate partner violence (IPV). Since 1974, researchers have created various typologies of perpetrators of IPV (Giesbrecht et al., 2026). Though this research has demonstrated that men who perpetrate IPV are not a homogenous group and that different types exist, a relatively small number of studies have examined treatment and recidivism outcomes in relation to typology. This systematic review is the first to summarize findings from all typology studies that included recidivism (n = 25) and treatment (n = 24) outcomes for types of perpetrators.
Typologies of Perpetrators of IPV
Although people of all genders can experience or perpetrate IPV, the majority of perpetrators of IPV and intimate partner homicide/femicide are men, and the majority of victims/survivors are women (Boyes, 2021; Dawson et al., 2021; Statistics Canada, 2023). While some studies have classified women who perpetrated IPV, the majority of IPV typology research has been conducted with men.
Over the past five decades, researchers have created typologies of IPV perpetrators using empirical (or inductive) methods, where types are not pre-set but are derived from the data, or rational (or deductive) methods, where perpetrators are classified into pre-set types, such as coding generality of violence (e.g., family-only or generally violent) or instrumental or reactive violence from police records. Theoretical or conceptual typologies have also been proposed, based on authors’ clinical experience or reviews of the extant literature.
Extant typology studies most commonly classify men who perpetrate IPV according to seven main types: family-only 2 and generally violent; family-only, generally violent, and borderline/dysphoric; 3 family-only, generally violent, low-level antisocial, and borderline/dysphoric; severity and frequency of violence; reactive and instrumental; situational couple violence and coercive control; and personality types. Typologies that classify IPV perpetrators based on factors such as treatment responsivity or physiological responses have also been developed. (A systematic review of 190 typology studies published between 1974 and 2024 is published in Giesbrecht et al., 2026).
Although typologies have been developed and tested by researchers and demonstrate the heterogeneity of perpetrators of IPV in terms of personality characteristics (e.g., antisocial traits), variation in forms of violence and control, severity of violence, and other factors, typologies are not widely used to inform treatment planning or supervision conditions (Cavanaugh & Gelles, 2005; Giesbrecht et al., 2023a, 2023b; Stewart et al., 2013, 2014). Two plausible reasons for this shortcoming include the lack of consensus on which typology to apply in practice and how to practically classify different types. Although no typology has emerged as the definitive or accepted model, the literature is clear that different types of perpetrators have distinct treatment and supervision needs (Hilton & Ennis, 2020; Radatz & Wright, 2016; Stewart et al., 2013).
Recidivism Among Perpetrators of IPV
Studies of IPV recidivism range in terms of sample sizes, length of follow-up periods, and measures of recidivism (e.g., rearrest, reconviction, victim reports). Generally, reported rates of IPV recidivism among perpetrators range from 15.3% to 60% (e.g., 15.3%, Kingsnorth, 2006; 27%, Hilton et al., 2010; 28.1%, Henning & Holford, 2006; 60%, Klein & Tobin, 2008). Klein and Tobin (2008) found that 32% of their sample had perpetrated IPV within a year; within 10 years, 60% committed IPV recidivism.
IPV perpetrators are also reported to have non-IPV recidivism. Klein and Tobin (2008) found that 43% had been rearrested for any crime within 1 year. Giesbrecht et al. (2025) found that 25.2% had some form of recidivism 4 within 1 year; 37.7% had recidivated within 3 years. The vast majority (90%) of Hilton and colleagues’ (2010) sample had a post-release charge (any offense), and 61% had charges relating to violence. Research indicates the presence of a subgroup of perpetrators who are chronic and repeat offenders who perpetrate severe violence and pose the greatest risk to their intimate partners. These individuals overlap with the general criminal population and engage in non-IPV crimes (Barnham et al., 2017; Sherman, 2007; University of New Brunswick & Canadian Association of Chiefs of Police, 2016). Findings from extant research show a relationship between the severity of IPV and recidivism (Goldstein et al., 2016; Hilton et al., 2001; Kingsnorth, 2006).
IPV and Risk-Need-Responsivity
The Risk-Need-Responsivity (RNR) Model of Offender Assessment and Treatment has been the dominant model of correctional intervention since the 1990s. Radatz and Wright (2016) described the Principles of Effective Intervention (PEI) for IPV, including risk, need, responsivity, treatment, and fidelity. The risk principle focuses on which individuals should receive intervention, stating that the intensity of intervention (e.g., duration, dosage) should be matched to the individual’s risk of reoffending. Although the majority of IPV treatment/intervention programs conduct risk assessment, not all treatment/intervention programs use risk assessment information to inform the provision of treatment tailored to participants’ level of risk (Giesbrecht et al., 2023b; Heslop et al., 2016).
The need principle emphasizes targeting criminogenic (or treatment) needs (i.e., dynamic risk factors) directly linked to criminal behavior. Depending on which forms and features of violence and abuse are present, treatment needs differ for different IPV perpetrators. For example, challenging a perpetrator’s attitudes of sexual proprietariness, ownership, and control of one’s partner (Stewart et al., 2013) is very different than addressing a lack of communication and conflict resolution skills.
The responsivity principle provides guidance for delivering treatment. This principle includes two facets—general and specific responsivity. The principle of general responsivity points to employing evidence-based strategies that are useful for a variety of offenders with diverse needs (Bonta & Andrews, 2024). The principle of specific responsivity emphasizes matching individual offenders to services and evidence-based treatment based on their needs, abilities, and demographic characteristics (Andrews et al., 2006; Bonta & Andrews, 2024). Although few researchers have examined incorporating the principles of RNR into interventions for IPV perpetrators, researchers have noted the relevance of typology (e.g., generality of violence, motivations for using violence [e.g., coercive control, impulsivity], or personality patterns [e.g., antisocial personality pattern]) to specific responsivity to improve treatment adherence and outcomes (Corvo & Spitzmueller, 2017; Hilton & Ennis, 2020; Radatz & Wright, 2016; Scott et al., 2013; Stewart et al., 2013, 2014).
Effectiveness of IPV Treatment
An early meta-analysis conducted by Babcock and colleagues (2004) reviewed 22 studies that evaluated treatment efficacy. Babcock et al. (2004) reported that, overall, treatment programs for perpetrators of IPV had only a small effect on recidivism. Wilson and colleagues (2021; 11 studies) found no conclusive evidence of the effectiveness of treatment programs for perpetrators of IPV. Fernández-Fernández et al. (2022; 26 studies) had a similar finding. Eckhardt et al. (2013; 30 studies) found outcomes were similar for perpetrators who were assigned to treatment interventions and those who were not. Another meta-analysis by Karakurt and colleagues (2019; 13 studies), however, found that participation in IPV intervention programs effectively reduced recidivism. Arias et al. (2013; 19 studies) reported that participation in IPV intervention programs had a nonsignificant positive effect on recidivism, with the recidivism rate reported by victims significantly higher than official reports. Similarly, a meta-analysis by Cheng and colleagues (2021; 17 studies) found that intervention programs appeared effective for reducing IPV and general recidivism, according to official reports, but not when reported by victims/survivors. Arce and colleagues (2020; 25 studies) also found significant differences between official reports and couples’ reports of recidivism. Travers et al. (2021; 31 studies) found that when follow-up periods of 2 years or more were used, the impact of treatment on recidivism was not statistically significant. They did, however, note that interventions based on the principles of RNR appeared to show the most promise.
One reason for the apparent lack of success in treating perpetrators of IPV may be that treatment programs are often standardized and/or only one program is available in a particular community, meaning that all perpetrators will participate in the same program (Giesbrecht et al., 2023a, 2023b). Researchers (Cavanaugh & Gelles, 2005; Cunha & Gonçalves, 2013; Hilton & Ennis, 2020; Lowenkamp & Latessa, 2004) have highlighted the risk of mismatching perpetrators of IPV to treatment. According to the RNR model and relating to criminal offending generally (not IPV specifically), the most effective approach is to direct more intensive programming and support to individuals deemed to be at a higher risk of reoffending (Andrews & Bonta, 2010; Bonta & Andrews, 2024; Lowenkamp & Latessa, 2004). Applying this approach necessitates deciding what, if anything, is appropriate treatment for low-risk offenders. Researchers have found that treatment is minimally effective for low-risk offenders, and they may be at an increased risk of reoffending after attending treatment (Hilton & Ennis, 2020; Lowenkamp & Latessa, 2004). This increased risk is likely due to a number of factors associated with intensive treatment programs that do not align with the needs of low-risk individuals. These factors may include associating with higher-risk offenders within the treatment group, program content that targets the criminogenic needs of higher-risk offenders (e.g., Cognitive Behavioral Therapy [CBT] that targets antisocial cognitions), and the potential disruption of prosocial supports when entering treatment (e.g., someone who works away from their home community may lose their job or have to take a leave of absence from work to attend treatment).
Babcock and colleagues (2016) made a case for making treatment available to low-risk perpetrators of IPV (or those at risk of perpetrating IPV) who self-refer without legal system involvement as a prerequisite. They stated that existing interventions, such as weekly group psychoeducational programming that targets the use of IPV and coercive controlling behavior, may work well for lower-risk perpetrators who do not have additional criminogenic needs that complicate treatment delivery. Babcock et al. (2016) also indicated that self-referred individuals, who are often lower risk, are more motivated to change at the beginning of treatment and can thus help create a constructive group environment. A caution, however, is that although lower-risk individuals may have a positive influence on higher-risk participants (Babcock et al., 2016), higher-risk individuals may have a negative influence on those who are lower-risk (Cunha & Gonçalves, 2013; Hilton & Ennis, 2020; Lowenkamp & Latessa, 2004). Stewart et al. (2013) wrote that, although it is unknown whether negative effects for low-risk program participants are caused by associating with antisocial co-participants, a potential solution is to have individuals participate in groups with others who share similar risk profiles. Therefore, treatment planning is likely to be improved when information on typology is available, given that typology is associated with risk and can provide a more comprehensive profile of individuals’ needs.
Attrition is a considerable problem for IPV treatment programs (Catlett et al., 2010; Cunha et al., 2024; Eckhardt et al., 2008b; Jewell & Wormith, 2010; Olver et al., 2011). A meta-analytic study by Olver and colleagues (2011) found that over one-third (37.8%) of individuals who attend IPV treatment never finish. Further, those who did not complete treatment had higher recidivism rates. This review also showed that those who failed to complete treatment were often the highest-risk individuals who would likely benefit the most from these interventions (Olver et al., 2011). Another meta-analysis by Cunha and colleagues (2024) found that drop-out rates among participants in IPV programs varied from 9% to 67%. Several variables, including substance abuse, personality disorder, criminal history, and perpetrator typology (specifically individuals other than those classified as family-only), were associated with higher rates of drop-out.
Although several typology studies have indicated that drop-out and completion of intervention programs (Cantos et al., 2019; Carbajosa et al., 2017a; Romero-Martínez et al., 2021; Stoops et al., 2010) and recidivism (Cantos et al., 2019; Giesbrecht et al., 2025; González-Álvarez et al., 2024; Johnson & Goodlin-Fahncke, 2015; Peters et al., 2023; Petersson & Strand, 2017) could be predicted by perpetrator type, the studies included in meta-analyses of treatment effectiveness do not examine outcomes in relation to typology. While results of these meta-analyses indicate that IPV treatment programs have a minimal effect on recidivism, it is likely that different types of perpetrators differ in terms of treatment participation, completion, and outcomes (e.g., recidivism after treatment). Therefore, the present systematic review examined studies that reported treatment and recidivism outcomes in relation to typologies of perpetrators of IPV, with the goal of clarifying how typology relates to outcomes for men who perpetrate IPV.
The Present Study
Systematic Review of Typologies of Men Who Perpetrate IPV
An earlier systematic review provided an overview of all studies to date of typologies of men who perpetrate IPV, including those that developed new typologies or classified individuals according to typologies derived in previous studies, published over a 50-year period, from 1974 to 2024 (Giesbrecht et al., 2026). That systematic review included 215 articles comprising 186 unique studies, which presented 190 typologies (new typologies or replications of typologies from previous research) of men who perpetrated IPV. That review was the first systematic review to include empirical, rational/deductive, and theoretically derived typologies. It was also the first to group typologies according to nine categories: (a) family-only and generally violent; (b) family-only, generally violent, and borderline/dysphoric; (c) family-only, generally violent, low-level antisocial, and borderline/dysphoric; (d) severity and frequency of violence; (e) reactive and instrumental; (f) situational couple violence and coercive control; (g) personality types; (h) other typologies (e.g., treatment responsivity, physiological reactivity); and (i) typologies of perpetrators of intimate partner femicide.
Method
The present systematic review includes all studies that examined treatment and recidivism outcomes in relation to typologies of perpetrators of IPV. The overall goal of the review was to summarize all existing studies that reported treatment and recidivism outcomes for different types of IPV perpetrators, to explore how perpetrator type is related to outcomes. The study protocol was registered with PROSPERO (registration number CRD42023452469) (Giesbrecht & Bruer, 2023).
Search Strategy
Studies included in the present review were identified through a systematic review of all studies of typologies of men who perpetrated IPV (Giesbrecht et al., 2026). Several articles relating to typologies of IPV perpetrators were obtained by the first author during an initial literature review. The reference lists of these articles were hand-searched to identify additional articles. Studies obtained by hand-searching (n = 436) were organized in Zotero (Corporation for Digital Scholarship, n.d.) and uploaded into Covidence systematic review software (Veritas Health Innovation, 2023).
Database searches (Criminal Justice Abstracts, PsycINFO, Social Services Abstracts, and Sociological Abstracts) to identify additional articles were conducted on August 13, 2023. Subsequent searches were conducted on March 31, 2024, and December 20, 2025. The systematic review includes articles published between 1974 and December 31, 2024. Search terms included IPV OR domestic violence AND perpetrat* OR batter* OR offend* AND typolog* OR subtype OR sub-type OR latent class OR cluster. Titles and abstracts were searched for these terms.
Results of the first database searches (August 13, 2023) and items obtained by hand-searching totaled 1,291 unique items, representing 1,261 studies. Using Covidence, 519 duplicate items were identified and removed, resulting in 742 items for further screening. Based on information included in the abstract, n = 271 studies were determined to be irrelevant to the systematic review topic and were screened out. The remainder (n = 471) were selected for full-text review. Seven of these articles could not be obtained, resulting in n = 464 studies for full-text review. During the full-text review, when multiple articles presenting results from the same study were identified, they were combined in Covidence to represent a single study. Of the n = 464 articles assessed for eligibility, n = 194 (comprising n = 166 studies) were included in the initial systematic review.
An additional search for new articles was conducted on March 31, 2024. Another search for articles published up to December 31, 2024, was conducted on December 20, 2025. Together, subsequent database searches identified five more articles. An additional 14 articles were identified by hand-searching, bringing the total to 19. In addition, the authors of the present review published a new typology of men who perpetrated IPV (Giesbrecht et al., 2025). Therefore, 215 articles (comprising 186 studies and presenting 190 typologies) were included in the full systematic review of studies of typologies of men who perpetrate IPV (Giesbrecht et al., 2026).
The 215 articles were read in full. All studies that reported treatment (n = 24) or recidivism (n = 25) outcomes by typology were included in the present review.
Inclusion and Exclusion Criteria
The population for the larger systematic review comprised adult men who perpetrated IPV against women. Typologies derived from primary studies with men who perpetrated IPV or from archival data were eligible for inclusion, as were studies conducted with couples that included perpetrators who were men. Typologies derived from women’s reports regarding their partners were eligible if the data provided related to the perpetrator. Typologies of perpetrators of all genders were included if data pertaining to men could be analyzed separately or if the majority (over 80%) of the sample were men.
As the study population was adult men who perpetrated IPV, studies of youth under the age of 18 were excluded. Typologies that only included women who perpetrated IPV were excluded. Typologies of relationship dynamics were excluded, as were typologies of victims.
The goal of the present study was to provide a comprehensive overview of all studies presenting treatment or recidivism outcomes by typology for men who perpetrated IPV. Therefore, no exclusion criteria were set relating to study quality. The systematic review only included studies published in English.
Data Extraction
Data collected for each article included reference(s), publication year, year data was collected, study country, typology category (e.g., reactive/instrumental, family-only/generally violent, situational couple violence/coercive control, personality), method (i.e., empirical, rational/deductive, or theoretical), sample size, sample population, sample recruitment (e.g., in treatment- mandated or self-referred, incarcerated, recruited from community), archival data or in-person participants, derived types, number of types, method of analysis, measures used to derive typology, measures used to compare derived types, and demographic variables for types.
Data that were collected pertaining to treatment outcomes included: treatment sample (size, population, recruitment), treatment intervention program (name, location), treatment approach (e.g., CBT, Duluth Model), program length, treatment outcome measure (e.g., attrition, completion, dropout, reduction of violent behavior, difference between pre-treatment and post-treatment scores, recidivism, stage of change after completing treatment [Levesqe et al., 2000], violent behavior while in treatment), and treatment outcomes for different perpetrator types.
Data that were collected pertaining to recidivism outcomes included: recidivism sample (size, population, recruitment), follow-up period, recidivism measure (i.e., new charge, rearrest, reconviction, reoffense [law enforcement record, self-reported, victim report], other), recidivism type (i.e., IPV, general [any] violence, general [any] offense, violation of protective order, non-IPV violence, other), and recidivism outcomes for different perpetrator types.
The present review presents findings from n = 24 typology studies that reported treatment outcomes and n = 25 typology studies that reported recidivism outcomes (Table 1).
Key Findings of the Systematic Review.
Note. IPV = Intimate partner violence.
Results
Descriptive Data for Included Studies
Studies of Treatment Outcomes by Typology
The n = 24 typology studies that reported treatment outcomes for types were reported in 27 articles (See Supplemental Material: Table 1). These articles comprised 19 journal articles, 6 doctoral dissertations, and 2 master’s theses. In addition, one study (Langhinrichsen-Rohling et al., 2000) classified perpetrators according to two different typological solutions (empirical and deductive); treatment outcomes for both typologies were reported. The first of these studies was published in 1996. The 24 studies were conducted in the United States (n = 18), Spain (n = 5), and South Africa (n = 1).
Sample sizes in the included studies ranged from n = 10 to n = 674 (M = 234.4, standard deviation [SD] = 184.8). All of the studies had samples of less than 1,000 participants; 22 had samples of less than 500. Sample sizes used in treatment analyses also ranged from n = 10 to n = 674 (M = 218.9, SD = 188.0). 5 The 24 typology studies were conducted with 23 different samples; 17 samples (approximately three-quarters) comprised in-person participants, and six employed archival datasets. Measures of treatment outcomes included attrition (measured by program completion, drop-out, or termination; n = 15), number of sessions attended (n = 9), compliance/non-compliance (n = 1), difference in scores on pre- and post-treatment measures (n = 6), reduction/continuation of violent behavior (n = 6), and recidivism (n = 10).
Studies in the systematic review that reported treatment outcomes for typologies of men who perpetrated IPV included 18 empirical (inductive or data-driven) typologies and six rational/deductive typologies. These 24 studies comprised six types: (a) family-only and generally violent (n = 3 deductive); (b) family-only, generally violent, and borderline/dysphoric (n = 6 empirical; n = 2 deductive); (c) family-only, generally violent, low-level antisocial, and borderline/dysphoric (n = 1 empirical); (d) severity and frequency of violence (n = 3 empirical); (e) reactive and instrumental (n = 1 empirical); and (f) personality types (n = 5 empirical). The review also included three typologies that did not fit the above categories, classified as (g) other: treatment responsivity/ failure and general violence/ IPV behavior (n = 2 empirical; n = 1 deductive). The systematic review of all IPV typologies (Giesbrecht et al., 2026) also included studies that classified perpetrators as using situational couple violence and coercive control; no studies of situational couple violence and coercive control reported treatment outcomes. A table summarizing details of the 24 studies on typology and treatment outcomes is available in Supplemental Material: Table 1.
The intervention programs in the reviewed studies included a variety of approaches and modalities. The intervention programs also varied in length (including number of sessions, duration in weeks, and length of sessions). Sixteen studies (n = 16) 6 in the review specified the number of sessions provided in the intervention program. The number of sessions ranged from 12 to 30 (M = 22.9, SD = 5.6). Fifteen studies (n = 15) 7 specified the length of the intervention program in weeks. The number of weeks ranged from 11 to 29 (M = 21.5, SD = 5.8). Twelve studies (n = 12) 8 specified the length of the intervention program in hours. The length in hours ranged from 18 to 63 hr (M = 43.2, SD = 13.1).
Sixteen studies (n = 16) in the review specified the treatment modality employed by the program. Eight (n = 8) 9 of these included employed CBT. Five studies (n = 5) 10 involved psychoeducational programs in line with the Duluth Model (Domestic Abuse Intervention Programs, n.d.). Two additional studies (n = 2) involved educational programs, and one involved a domestic violence and anger management program (n = 1). 11 Studies that reported differences in outcome between perpetrator groups included CBT (n = 5) and Duluth Model/psychoeducational groups (n = 3); the specific treatment modality was not reported for the remainder. In the reported studies, perpetrators participated in the same treatment program. It is not possible to discern from findings in published studies if different types of perpetrators would report different outcomes if they participated in separate treatment conditions (e.g., dosage, modality).
Studies of Recidivism Outcomes by Typology
The n = 25 studies in the systematic review that reported recidivism outcomes for typologies of men who perpetrated IPV were reported across 33 articles (See Supplemental Material: Table 2). These articles comprised 29 journal articles, 1 doctoral dissertation, 2 master’s theses, and 1 conference poster. The first of these studies was published in 2000. The 25 studies were conducted in the United States (n = 12), Spain (n = 7), Canada (n = 4), the Netherlands (n = 1), and Sweden (n = 1).
Sample sizes in the included studies ranged from n = 52 to n = 9,731 (M = 1,120.2, SD = 2,372.3). Twenty-one of the 25 studies had samples of less than 1,000 participants. Samples used in recidivism analyses ranged from n = 40 to n = 9,731 (M = 1,078.4, SD = 2,374.1). 12 Measures of recidivism included reoffense (self- [n = 4] and victim-reported [n = 3] and reported in police and correctional records; [n = 12]), rearrest (n = 8), new charge (n = 1), and reconviction (n = 4). Twenty-one (n = 21) studies reported IPV recidivism; several studies also reported general (any) violence and general (any) offending.
Studies in the systematic review that reported recidivism outcomes for typologies of men who perpetrated IPV included 18 empirical typologies and seven rational/deductive typologies. These 25 studies comprised 6 categories: (a) family-only and generally violent (n = 1 empirical; n = 4 deductive); (b) family-only, generally violent, and borderline/dysphoric (n = 4 empirical; n = 1 deductive); (c) family-only, generally violent, low-level antisocial, and borderline/dysphoric (n = 3 empirical); (d) severity and frequency of violence (n = 4 empirical); (e) reactive and instrumental (n = 1 empirical; n = 2 deductive); and (f) personality types (n = 2 empirical). The review also included three typologies that did not fit the above categories classified as (g) other: criminal history and antisociality, general violence/ IPV behavior, and readiness to change (n = 3 empirical). The systematic review of all IPV typologies (Giesbrecht et al., 2026) also included studies that classified perpetrators as using situational couple violence and coercive control; no studies of situational couple violence and coercive control reported recidivism outcomes. A table summarizing details for the 25 studies of typology and recidivism outcomes is available in Supplemental Material: Table 2.
Ten studies in the systematic review examined both treatment and recidivism outcomes (Berbary et al., 2018; Cantos et al., 2019; Carbajosa et al., 2017a; Eckhardt et al., 2008a, 2008b; Graña et al., 2014; Graña Gómez et al., 2017; Huss & Ralston, 2008; Redondo et al., 2019; Romero-Martínez et al., 2021; Stoops et al., 2010).
Outcome Data in Included Studies
Treatment Outcomes by Typology
Three of the typology studies that examined treatment outcomes classified perpetrators according to the family-only and generally violent dichotomy, using rational/deductive methods (Cantos et al., 2019; Mach, 2015; Mach et al., 2020). All three studies reported that generally violent men were less likely to complete treatment than family-only perpetrators. In addition, Mach et al. (2020) reported that family-only individuals attended significantly more sessions than generally violent men.
Eight studies classified perpetrators according to the tripartite family-only, generally violent, and borderline/dysphoric typology (n = 6 empirical: Adcox Meade, 2006; Bartels, 1996; Carbajosa et al., 2017a; Huss & Ralston, 2008; Romero-Martínez et al., 2021; Sartin, 2004; and n = 2 rational/deductive: Brashear, 2005; Langhinrichsen-Rohling et al., 2000 [Study 1]). Two studies reported that family-only perpetrators were most likely to complete treatment programming and generally violent individuals were least likely, with borderline/dysphoric perpetrators falling in the middle (Langhinrichsen-Rohling et al., 2000 [Study 1]; (Romero-Martínez et al., 2021)). Family-only perpetrators also completed significantly more sessions (Huss & Ralston, 2008), and generally violent men attended fewer (Carbajosa et al., 2017a; Romero-Martínez et al., 2021). Generally violent individuals were also more likely to drop out (Carbajosa et al., 2017a). Other findings included that borderline/dysphoric perpetrators were two to three times more likely to drop out after the first session (Adcox Meade, 2006), and generally violent and antisocial individuals who stayed in treatment reported the greatest reduction in self-reported anger and IPV (Huss & Ralston, 2008). Other studies reported that although typology did not significantly predict completion, family-only had the highest rates of completion (Adcox Meade, 2006; Sartin, 2004).
One empirical study (Eckhardt et al., 2008a) examined the four-class family-only, generally violent, low-level antisocial, and borderline/dysphoric typology in relation to treatment outcomes. This study also reported that family-only men were most likely to complete treatment programming and generally violent men were least likely, with borderline/dysphoric perpetrators falling in the middle. Another empirical study classified perpetrators based on IPV behavior (Stoops et al., 2010). This study found that those who perpetrated more general violence were more likely to require multiple referrals before completing treatment and were also significantly less likely to complete treatment.
Five empirical studies that derived typologies based on personality reported treatment outcomes (Belaga, 2004; Eckhardt et al., 2008b; Freeman, 1999; Langhinrichsen-Rohling et al., 2000 [Study 2]; Murphy et al., 2007). Individuals with the lowest levels of anger were most likely to complete treatment (Eckhardt et al., 2008b) and attend more sessions (Murphy et al., 2007). Belaga (2004) reported that individuals with the highest levels of violence, the least secure parental attachment, and the lowest physiological arousability improved the least in treatment.
Three empirical studies classified individuals on severity and frequency of IPV (Berbary et al., 2018; Gibbs et al., 2020a; Graña Gómez et al., 2017). Berbary et al. (2018) reported that participants deemed higher risk perpetrated more violence after treatment. Gibbs et al. (2020a) reported that the higher-risk group showed the most significant improvements. Graña Gómez et al. (2017) also reported that individuals with moderate and high levels of violence and psychopathology showed a significant reduction between pre- and post-treatment scores on the Revised Conflict Tactics Scales.
One empirical study that classified perpetrators as reactive or instrumental found that both groups had significant reductions in levels of violence following treatment (Redondo et al., 2019). Two studies (empirical: Stalans & Seng, 2007; rational/deductive: Carbajosa et al., 2017b) classified perpetrators on treatment responsivity.
Across various typologies, family-only perpetrators were significantly more likely to complete treatment (compared to generally violent perpetrators) and, overall, individuals deemed lower risk were more likely to complete treatment.
Recidivism Outcomes by Typology
Four studies in the systematic review that reported recidivism outcomes classified perpetrators according to the family-only and generally violent dichotomy, using rational/deductive methods (Cantos et al., 2019; Loinaz, 2014; Petersson & Strand, 2017; Stalans et al., 2004). In addition, an empirical study by Peters et al. (2023) described three types: IPV-specific, generally violent/antisocial, and IPV and generally violent/antisocial. In these studies, generally violent/antisocial perpetrators also had a higher rate of recidivism than family-only perpetrators (Cantos et al., 2019; Peters et al., 2023; Petersson & Strand, 2017, 2020; Stalans et al., 2004). An additional study (Loinaz, 2014) did not find a statistically significant difference between the two groups; however, generally violent/antisocial perpetrators had a higher rate of recidivism.
Five studies classified perpetrators according to the tripartite family-only, generally violent, and borderline/dysphoric typology (n = 4 empirical: Carbajosa et al., 2017a; Huss & Ralston, 2008; Romero-Martínez et al., 2021; Walsh et al., 2010; and n = 1 rational/deductive: Wexler, 2000). Two studies reported that generally violent/antisocial individuals had the most recidivism, followed by borderline/dysphoric; family-only had the least (Huss & Ralston, 2008; Romero-Martínez et al., 2021). Wexler (2000) found this for non-Indigenous perpetrators; however, a significant difference was not observed between types of Indigenous perpetrators. Two studies did not find significant differences between the groups (Carbajosa et al., 2017a; Walsh et al., 2010).
Three empirical studies (Eckhardt et al., 2008a; Johnson & Goodlin-Fahncke, 2015; Thijssen & de Ruiter, 2011) found the four-class family-only, generally violent, low-level antisocial, and borderline/dysphoric typology. Again, two studies reported that generally violent/antisocial had the most recidivism, whereas family-only had the least, with borderline/dysphoric and low-level antisocial in the middle (Eckhardt et al., 2008a; Johnson & Goodlin-Fahncke, 2015). Thijssen and de Ruiter (2011) did not find a significant difference between the groups; however, generally violent/antisocial individuals had the most recidivism, and family-only individuals had the least.
Stoops et al. (2010) also found that general violence was associated with recidivism. Eckhardt et al. (2008b) found that higher anger expression was associated with recidivism, and González-Álvarez et al. (2022) found that antisociality and instability were associated with recidivism. The studies that classified perpetrators according to risk level (Berbary et al., 2018; Klostermann et al., 2009) found that high-risk individuals had more recidivism. Graña Gómez and colleagues (2017) classified perpetrators according to levels of violence and psychopathology; however, they did not find a significant difference between the derived types and the control group in terms of recidivism.
Three studies that assessed recidivism classified perpetrators as reactive or instrumental (n = 1 empirical: Redondo et al., 2019; n = 2 rational/deductive: Ennis et al., 2017; Lee et al., 2018). Two studies (Ennis et al., 2017; Lee et al., 2018) did not find a statistically significant difference between the two groups. The third study (Redondo et al., 2019) found that recidivism was predicted by individuals’ level of pre-treatment hostility, with undercontrolled individuals displaying more hostility than overcontrolled. Another study (Redondo et al., 2023) found that low readiness to change predicted short-term recidivism at 1 year. In addition, the authors of the present review created a tripartite typology that differentiated types based on criminal history and antisocial attitudes and compared these types on measures of general recidivism (Giesbrecht et al., 2025). The three groups differed significantly on all reoffending outcomes, with criminal history and antisocial attitudes linked to the highest rates of four measures of general recidivism.
Discussion
Typologies have limited utility if only employed for descriptive purposes (Holtzworth-Munroe & Stuart, 1994); the goal of classifying perpetrators of IPV must be not only to demonstrate that perpetrators are a heterogenous population in terms of characteristics and behavior but also to use information regarding differences to inform case management, including more accurately predicting risk of recidivism and implementing strategies to reduce and manage risk, including assignment to treatment/intervention programs. Therefore, it is essential for researchers to examine outcome data by typology (whether in studies that solely use data from police, court, or correctional records or by linking outcome data to other data sources [e.g., participants’ self-report measures, data from intervention programs]).
Treatment
Half (n = 12 of 24) of the studies that reported treatment outcomes for different types were a variation of the family-only/generally violent typology (including 2, 3, or 4 groups); the other half included typologies that differentiated perpetrators by personality type, perpetration of violence, or treatment responsivity.
Measures of treatment outcomes included attrition (measured by program completion, drop-out, or termination), number of sessions attended, compliance/non-compliance, difference in scores on pre- and post-treatment measures, reduction/continuation of violent behavior, and recidivism. This variety of measures elucidates not only post-treatment outcomes (e.g., cessation of IPV) but also differences between types in terms of attrition and compliance. While more research is needed, findings from this systematic review indicate that additional conditions should be put in place to encourage and support IPV treatment program involvement for those who are least likely to complete (i.e., generally violent IPV perpetrators).
Typology studies that examined treatment outcomes included in this review had relatively small sample sizes (M = 218.9). Most commonly, studies examining treatment outcomes are conducted with in-person participants in treatment programs, as treatment outcomes are rarely available in archival datasets from official sources (e.g., law enforcement or correctional databases).
Twelve studies in this review identified family-only and generally violent types, including the dichotomous typology, as well as the tripartite family-only, generally violent, and borderline/dysphoric typology, and the quadripartite family-only, generally violent, low-level antisocial, and borderline/dysphoric typology; other included studies identified types that had similarities to the family-only and generally violent types. Taken together, this extant research demonstrates that family-only perpetrators are more likely to complete treatment and less likely to drop out, whereas generally violent perpetrators are more likely to drop out and less likely to complete (Cantos et al., 2019; Carbajosa et al., 2017a; Eckhardt et al., 2008a; Langhinrichsen-Rohling et al., 2000 [Study 1]; Mach, 2015; Mach et al., 2020; Romero-Martínez et al., 2021; Stalans & Seng, 2007).
In addition, family-only perpetrators attend more sessions than generally violent men (Carbajosa et al., 2017a; Huss & Ralston, 2008; Mach et al., 2020; Romero-Martínez et al., 2021; Sartin, 2004). Another study also reported that family-only individuals were more compliant with treatment (Brashear, 2005). This is unsurprising, given that family-only perpetrators have been described in previous research as having more stability factors, including a higher stake in conformity, whereas generally violent individuals are distinguished by antisocial traits and behavior and display more lifestyle instability (Giesbrecht et al., 2026). The studies that include a borderline/dysphoric type report that these individuals fall median to family-only and generally violent on measures of dropout and number of sessions attended (Carbajosa et al., 2017a; Langhinrichsen-Rohling et al., 2000 [Study 1]; Romero-Martínez et al., 2021).
Similarities can also be seen between the family-only and generally violent types and other perpetrator types reported in reviewed studies, in terms of their characteristics and treatment outcomes. In a study that identified very high-, moderate-, and low-risk groups, with multiple subgroups comprising each, Stalans and Seng (2007) highlighted the importance of separating family-only and generally violent types for treatment delivery. Stalans and Seng (2007) identified three subgroups with the highest probability of treatment failure: unemployed generalized aggressors, high school dropouts ordered into substance abuse treatment, and unemployed perpetrators ordered into substance abuse treatment. In their study, generally violent men were divided into unemployed and employed generalized aggressors, as those who were unemployed had a much higher likelihood of treatment failure (Stalans & Seng, 2007). The results of the systematic review of all typology studies illustrate that generally violent perpetrators are more likely than family-only individuals to have challenges with employment, education, and substance misuse (Giesbrecht et al., 2026).
Stoops and colleagues’ (2010) low-level criminality, dysphoric general violence, and dysphoric volatile behavior typology has similarities with the family-only, generally violent, and borderline/dysphoric typology. Stoops et al. (2010) found that dysphoric general violence men were more likely to require multiple referrals before participating in treatment and were less likely to complete treatment. Freeman’s (1999) conforming and overcontrolled types were most likely to complete treatment. These individuals are characterized by high social status, low psychopathology, minimal criminal history, and lower use of IPV (Freeman, 1999); characteristics that are similar to family-only perpetrators (Giesbrecht et al., 2026). Freeman’s (1999) impulsive-borderline perpetrators were less likely to complete treatment, followed by chronic criminal men. Impulsive-borderline individuals have characteristics similar to the borderline/dysphoric type and have the most extensive criminal history. Chronic criminals have more substance misuse issues, more anger, and displayed attitudes that are more supportive of IPV; traits similar to generally violent perpetrators in other studies (Giesbrecht et al., 2026). Belaga (2004) found that individuals who perpetrated the highest levels of violence and displayed the least secure parental attachment improved least in treatment (i.e., reduced IPV and improved interpersonal skills); again, similarities can be noted between this group and generally violent perpetrators.
Eckhardt et al.’s (2008b) typology of anger types found that low anger men were more likely than the moderate anger–inexpressive or high anger–expressive types to complete treatment. Another typology of anger types by Murphy et al. (2007) reported that the pathological anger group attended significantly fewer total sessions than low and normal anger types. The normal anger group had less physical aggression, psychological aggression, and dominance–intimidation than the other two groups following treatment. The results of the larger systematic review illustrate that generally violent perpetrators display more anger than the other types (Giesbrecht et al., 2026).
Eight of the reviewed studies examined outcomes from a CBT intervention program, and five presented findings from a Duluth Model intervention. Although meta-analyses demonstrate that, overall, participation in IPV treatment/intervention programs has a minimal effect on outcomes for perpetrators of violence (as discussed above), some meta-analyses found positive and significant effects for CBT and Duluth Model programs, with little difference between the two models (Arias et al., 2013; Babcock et al., 2004; Eckhardt et al., 2013). Some programs deliver a combination of the two models (Cunha et al., 2024; Wilson et al., 2021).
Bonta and Andrews (2024) stated that cognitive-behavioral and cognitive social learning treatment approaches, in line with the RNR model, are the most effective, no matter the targeted behavior. Radatz and Wright (2016) indicated that the most effective treatment strategies for IPV are cognitive-behavioral and employ social learning techniques. The meta-analysis by Travers et al. (2021) reported that interventions grounded in the principles of RNR yielded the greatest positive results.
It appears that some intervention programs in the reviewed studies reported greater positive changes among higher-risk perpetrators who remained in treatment (Huss & Ralston, 2008; Sartin, 2004; Gibbs et al., 2020a). In a study that differentiated perpetrators according to reactive or instrumental use of violence, Redondo et al. (2019) found that both groups had significant reductions in levels of violence following treatment; it may be possible that this treatment program was effective for perpetrators of both types, whereas other programs are more effective for one type than the other. Other studies also did not find significant differences between groups on treatment outcomes; again, this could be related to the typology or treatment program employed in the study, or both.
Recidivism
Approximately half (n = 13 of 25) of the studies that reported recidivism outcomes by typology used a variation of the family-only/generally violent typology (including two, three, or four types); the other half included severity and frequency of violence, reactive and instrumental, and personality typologies as well as three typologies that did not fit the above categories (criminal history and antisociality, general violence/IPV behavior, and readiness to change). Overall, findings of this systematic review indicate that generally violent/antisocial perpetrators had the most recidivism and family-only had the least. Low-level antisocial and borderline/dysphoric perpetrators fell intermediate to generally violent and family-only perpetrators on measures of recidivism. In addition, criminal history, anger, hostility, antisociality, instability, and low readiness to change were associated with recidivism.
Sources of recidivism data (including IPV and general violence reoffenses) in the reviewed studies included any reoffense (identified in law enforcement or correctional records, self-reports, or victim reports), rearrests, new charges, and reconvictions. The majority (n = 19) used only official data (e.g., law enforcement records of rearrest). The majority of incidents of IPV are not reported (Conroy, 2021); therefore, the full extent of IPV recidivism is not captured in official data sources. Two studies assessed recidivism using perpetrators’ self-reports, and one study used only survivors’ reports of repeated violence. Three studies used multiple sources of reoffense data: Walsh et al. (2010) used self-reports, collateral informants, and official records to assess recidivism; 13 Johnson and Goodlin-Fahncke (2015) used victim reports and official reports; 14 and Eckhardt and colleagues (2008a) 15 assessed recidivism information collected from victims, perpetrator self-reports, and official data. One study in the review (Bartels, 1996) indicated that partner-only perpetrators self-reported more IPV despite fewer official reports.
Meta-analytic studies have identified differences in rates of recidivism when comparing victim reports with official reports and perpetrator self-reports (Arce et al., 2020: Arias et al., 2013; Cheng et al., 2021). Perpetrators of IPV may be reluctant to self-report their level of IPV recidivism, and not all repeat IPV incidents come to the attention of law enforcement. Although there are challenges inherent in collecting recidivism data from survivors (it may not be possible to contact survivors, survivors who remain in the relationship may be reluctant to provide data, and those who no longer remain in the relationship may have limited information to provide), verification from current or former partners can assist in improving the accuracy of recidivism data.
Implications for Practice and Research
Multiple studies have developed new typologies or classified IPV perpetrators according to typologies derived from previous studies (Giesbrecht et al., 2026), though a relatively small body of literature has examined recidivism and treatment outcomes across these types. This systematic review was the first to summarize findings from all typology studies that included recidivism (n = 25) and treatment (n = 24) outcomes for types of perpetrators. This systematic review summarizes evidence that typology impacts treatment outcomes (including attrition, attendance, compliance, and behavior change) and recidivism outcomes for different types of perpetrators; however, the findings also highlight the need for further investigation of recidivism and treatment outcomes for different types of perpetrators (Table 2).
Implications for Research and Practice.
Note. IPV = Intimate partner violence; RNR = Risk-need-responsivity; PEI = Principles of effective intervention; CBT = Cognitive behavioral therapy.
In the reviewed studies, all perpetrators participated in the same treatment program; therefore, it is not possible to determine whether different types of perpetrators would report different outcomes if they participated in different treatment programs (e.g., outcomes for family-only perpetrators in CBT versus Duluth Model programming). It is possible that some perpetrator types would experience better outcomes following different treatment modalities. Research that compares perpetrators of the same type on outcomes after different types of programs is needed to answer this question. Researchers have also highlighted the potential for detrimental outcomes when lower-risk individuals participate in treatment alongside higher-risk perpetrators (Cunha & Gonçalves, 2013; Hilton & Ennis, 2020; Lowenkamp & Latessa, 2004). In all the studies included in the review, perpetrators of different types participated in group programming together.
Half of the typology studies that reported treatment and recidivism outcomes classified perpetrators according to the family-only/generally violent typology (including two, three, or four groups). Few studies of treatment and recidivism employed the reactive and instrumental typology and other studies presented different typologies relating to severity and frequency of violence, personality types, or other typologies (e.g., criminal history and antisociality); therefore, more evidence is needed to confirm the applicability of these typologies for the prediction of treatment and recidivism outcomes and, therefore, case management strategies for perpetrators of these different types. No studies of treatment and recidivism outcomes differentiated perpetrators of situational couple violence and coercive control.
Although no studies in this review compared perpetrators of situational couple violence and coercive control in terms of treatment outcomes, it is possible that Duluth model programs, which “[believe] that battering is a pattern of actions used to intentionally control or dominate an intimate partner and actively [work] to change societal conditions that support men’s use of tactics of power and control over women,” (Domestic Abuse Intervention Programs, n.d.), would be more effective for targeting coercive controlling behavior rather than situational couple violence which is not rooted in the dynamics of domination and control. It is also possible that perpetrators of coercive control and situational couple violence may differ in terms of treatment compliance and attrition, as seen with generally violent and family-only perpetrators. More studies of treatment outcomes that take typology into account are needed overall, and research relating to the coercive control/situational couple violence typology is explicitly needed.
Effective case management is essential for reducing the risk of recidivism among men who have perpetrated IPV and for enhancing the safety of victims/survivors (intimate partners and children). Case management can be informed by incorporating typology, thereby enabling professionals to deliver interventions and provide supervision in line with risk factors, criminogenic needs, and specific responsivity factors prevalent within each type. Researchers have indicated that different types of perpetrators have different treatment and supervision needs (Hilton & Ennis, 2020; Radatz & Wright, 2016; Stewart et al., 2013). Given that the principles of RNR and PEI state that perpetrators should be assigned to a level of treatment intensity corresponding to their level of risk, results of this review indicate that generally violent perpetrators should receive the most intense programming (in terms of dosage and duration) (Andrews & Bonta, 2010; Bonta & Andrews, 2024; Hilton & Radatz, 2021; Radatz & Wright, 2016; Stewart et al., 2014). The principle of general responsivity points to the need to use strategies that are broadly effective for a variety of perpetrators, specifically, treatment programs that incorporate cognitive-behavioral and cognitive-social learning approaches (Andrews et al., 2006; Arce et al., 2020; Bonta & Andrews, 2024; Fernández-Fernández et al., 2022; Radatz & Wright, 2016). The principle of specific responsivity states that individuals should be assigned to the most appropriate evidence-based treatment given their needs, abilities, demographic characteristics, motivation to participate, and readiness to change (Andrews et al., 2006). Individual sessions are also useful for addressing individuals’ specific risks and needs, or factors relating to typology that are not addressed in group treatment. It will also be necessary to put conditions in place to encourage and support program involvement (and institute consequences for non-attendance) for those who are least likely to attend and complete treatment (i.e., generally violent perpetrators).
Limitations
As the goal of this systematic review was to provide a comprehensive review of all studies that examined treatment and recidivism outcomes by perpetrator typology, no exclusion criteria were set relating to study quality. Therefore, the included studies differ in methodological quality. This review only included studies published in English. It is possible that relevant studies have been published in other languages. As in all systematic reviews, relevant studies may have been missed and therefore not included, despite hand-searching and multiple database searches.
Thirteen of 24 studies of treatment outcomes and 11 of 25 studies of recidivism outcomes reported demographic data for derived types. Reported demographic data varied among the studies but included, for example, age, ethnicity, immigrant status, employment, and marital status. Given that demographic data were not available for half of the included studies and that reported demographic variables differed among the remaining studies, this review did not examine differences in treatment and recidivism outcomes for different types of perpetrators of IPV based on demographic variables (e.g., age, Indigenous identity). Future research may investigate the relevance of various IPV typologies for diverse groups of perpetrators and the implications of typology on treatment and recidivism outcomes for these groups. Research that takes diversity characteristics, beyond typology, into account is expected to enhance the applicability of IPV typologies to effective case management strategies.
Conclusion
The present systematic review was the first to summarize findings from typology studies that included treatment (n = 24) and recidivism outcomes (n = 25) for different types of perpetrators of IPV. Measures of treatment outcomes included attrition (program completion, drop-out, or termination), attendance, compliance, scores on pre- and post-treatment measures, reduction of violent behavior, and recidivism. Measures of recidivism included reoffense (self- and victim-reported and reported in police and correctional records), rearrest, new charge, and reconviction. Studies reported IPV recidivism and general (any) violence and offending.
Studies in the review categorized perpetrators as: (a) family-only and generally violent; (b) family-only, generally violent, and borderline/dysphoric; (c) family-only, generally violent, low-level antisocial, and borderline/dysphoric; (d) severity and frequency of violence; (e) reactive and instrumental; (f) personality types; and (g) other (criminal history and antisociality, general violence/IPV behavior, treatment responsivity/failure, and readiness to change). No studies of treatment and recidivism outcomes categorized perpetrators according to the situational couple violence and coercive control typology.
Overall, generally violent men were less likely to complete treatment, more likely to drop out, and attended fewer sessions. Individuals deemed higher risk were less likely to complete treatment. General violence, antisocial attitudes, anger, instability, and low readiness to change were associated with recidivism. Individuals deemed higher risk were more likely to reoffend. Evidence from this systematic review illustrates that typology impacts treatment and recidivism outcomes for different types of perpetrators; however, it also highlights the need for further investigation of recidivism and treatment outcomes across these groups.
Supplemental Material
sj-pdf-1-tva-10.1177_15248380261451842 – Supplemental material for Treatment and Recidivism Outcomes for Perpetrators of Intimate Partner Violence: A Systematic Review of Typology Research
Supplemental material, sj-pdf-1-tva-10.1177_15248380261451842 for Treatment and Recidivism Outcomes for Perpetrators of Intimate Partner Violence: A Systematic Review of Typology Research by Crystal J. Giesbrecht, Leslie Anne Keown and Kaila C. Bruer in Trauma, Violence, & Abuse
Footnotes
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This article is based on Crystal J. Giesbrecht’s doctoral dissertation titled, Classifying Men Who Perpetrate Intimate Partner Violence: A 50-year Systematic Review and a New Typology Applicable to Case Management.
Crystal J. Giesbrecht gratefully acknowledges the support of a Vanier Canada Graduate Scholarship (Social Sciences & Humanities Research Council; SSHRC).
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability Statement
References for all studies included in the systematic review, as well as tables summarizing the findings of the review, are available in Supplemental Material.
Supplemental Material
Supplemental material for this article is available online.
Notes
Author Biographies
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
