Risk Posed by Different Intimate Partner Violence Offender Types: Findings From a Representative Police Sample

Abstract

Spanish intimate partner violence against women offender types (i.e., high instability/high antisociality, HiHa; low instability/high antisociality, LiHa; high instability/low antisociality, HiLa; low instability/low antisociality, LiLa) were matched with their police recidivism outcomes in a longitudinal study of 9,672 cases. Our goal was to examine whether these subtypes differed in (1) their recidivism rates, (2) the severity of the new violent episodes, and (3) the evolution of their risk levels. Results showed that individuals with high antisociality features where associated with the highest recidivism rates (26.5% HiHa; 22.6% LiHa), and higher likelihood of new severe violent episodes. HiHa offenders showed the highest risk over time, although the risk posed by all subtypes decreased during follow-up. Implications for police work are discussed.

Keywords

intimate partner violence recidivism risk typology VioGén System VPR

Introduction

According to the World Health Organization (2013), 30% of women worldwide have suffered physical and/or sexual intimate partner violence. The physical, psychological, and social consequences of intimate partner violence against women (IPVAW) on primary victims, their families and the wider community have been widely acknowledged (Campbell, 2002; Ellsberg et al., 2008; Guedes et al., 2016; Loxton et al., 2017; Trabold et al., 2020; Vilariño et al., 2018; Vives-Cases et al., 2011), and researchers, professionals, and public administrators are calling for more effective prevention strategies (Ellsberg et al., 2015; García-Moreno et al., 2015; Heise, 2011). In Spain, the country where this study was conducted, the last macro survey on Violence Against Women (n = 9,568 women, representative of the female population residing in Spain aged 16 years and older; Delegación del Gobierno contra la Violencia de Género, 2020) found 14.2% of women in the country had suffered physical and/or sexual intimate partner violence, and 31.9% psychological intimate partner violence, at some point in their lives. Nonetheless, the complexity of victim reporting decision-making and help-seeking in IPVAW cases (Blay Gil, 2014; Morgan et al., 2022; Xie & Baumer, 2019) makes it a largely unreported criminal phenomenon (78.3% of IPVAW victims in Spain did not report their aggressor; Delegación del Gobierno contra la Violencia de Género, 2020), and thus prevalence data might be much higher if all submerged IPVAW was accounted for (Gracia, 2004; Gracia et al., 2009).

One of the main concerns of law enforcement agencies is to ensure the protection of victims who report their aggressors and to prevent a violent episode from occurring again (González-Álvarez et al., 2018). In this regard, several standardized measures that facilitate risk classification according to the statistical likelihood of recidivism (new reported IPVAW incident) have been developed in the international context for IPVAW offenders (e.g., the Brief Spousal Assault Form for the Evaluation of Risk version 2 [B-SAFER], Kropp et al., 1995; the Ontario Domestic Assault Risk Assessment [ODARA], Hilton et al., 2008; the Spousal Abuse Risk Assessment-Version 3 [SARA-V3], Kropp & Hart, 2015; and the Violence Risk Screening−Police Version [V-RISK-POL], Roaldset et al., 2017). Recent meta-analytical evidence supports the predictive value of such tools, especially actuarial instruments (overall AUC = 0.657; van der Put et al., 2019).

The Spanish Comprehensive Monitoring System of Gender-Based Violence Cases

In Spain, IPVAW is legally referred to as gender-based violence (see Organic Law 1/2004 of 28 December on Integrated Protection Measures against Gender Violence), and is considered the manifestation of discrimination, the situation of inequality and the power relations of men over women, exercised over them by those who are or have been their spouses or who are or have been linked to them by similar relationships of affectivity, even without cohabitation. The Organic Law 1/2004 spurred the police and judicial protection for women exposed to IPVAW in the country and, in 2007, the Comprehensive Monitoring System of Gender-Based Violence Cases (VioGén System; a computer application which gathers information on all reported cases of gender-based violence in Spain) was created (González-Álvarez et al., 2018). Once a woman reports a man for gender-based violence, law enforcement investigate the situation and prepare a detailed report for the judge. First, police officers investigating a gender-based violence case in Spain use the data of their crime investigation to register and activate the case on the VioGén System. Next, the risk level of each offender is initially assessed with the Police Risk Assessment tool (Valoración Policial del Riesgo [VPR]; an actuarial protocol proven useful to predict and manage risk in cases of gender-based violence; López-Ossorio et al., 2019a, 2019b) and, according to the risk level detected (i.e., extreme, high, medium, low, or no risk), proportional police protection measures for the victim are adopted. In addition, the Spanish judges can use this police risk assessment to issue protective orders (Caballé-Pérez et al., 2020). During the time a case is active, the risk of IPVAW is reassessed with a second actuarial tool, the Police Assessment of Risk Evolution tool (Valoración Policial de la Evolución del Riesgo; VPER), every time a significant incident occurs (such as a new complaint), or in predetermined periods of time (López-Ossorio et al., 2019a). Cases remain active in the VioGén System until no recidivism risk is detected, which may take from a few days to a few years, depending on the case (González-Álvarez et al., 2018). The goal of the VioGén System dual risk assessment protocol (VPR and VPER) is to promptly identify reported individuals who pose a higher risk to their partners, in order to assign victim protection resources in the most efficient way; additionally, the risk management approach is oriented to the deactivation of the identified dynamic-relational risk indicators. Based on official reports, López-Ossorio et al. (2017) reported a police recidivism rate of 7.4% in a 6-month interval among arrested IPVAW offenders in Spain.

IPVAW Offender Typologies Literature Review

Prior research has pointed toward the heterogeneous nature of IPVAW, both in terms of the psychological characteristics of the perpetrators and in relation to the type of violence that can occur (Delsol et al., 2003; Dixon & Browne, 2003; González-Álvarez et al., 2021; Johnson, 2008), which represents a challenge for the development of effective protection plans for victims. In this sense, typological approaches to classify IPVAW offenders have been discussed on the basis of their validity and applicability in professional practice (i.e., tailoring batterer intervention programs, risk management procedures, or police and judicial measures; Amor et al., 2009; Cavanaugh & Gelles, 2005; Holtzworth-Munroe & Stuart, 1994; Langhinrichsen-Rohling et al., 2000; Lohr et al., 2005; Weber & Bouman, 2020); although their practical use might not be comparable across countries and jurisdictions. Longitudinal studies on classic IPVAW offender typologies have explored the stability of such types over time (i.e., family only [FO]; borderline/dysphoric [BD]; low-level antisocial [LLA]; and generally violent/antisocial [GVA]; Holtzworth-Munroe et al., 2000). Holtzworth-Munroe et al. (2003) analyzed changes over time in a sample of 102 husband-violent couples (37 FO, 34 LLA, 16 GVA, and 15 BD offenders). After 1.5 and 3-year follow-up periods, the four subtypes continued to differ in their levels of IPVAW severity, with FO and LLA men displaying less severe violent episodes than BD and GVA men. The FO subgroup was the most stable, suggesting that IPVAW would not inevitably escalate over time; moreover, most LLA men who switched groups moved to the FO group as a result of decreasing levels of reported IPVAW and antisocial behavior. One-third remained in the LLA group, and almost one-fourth remained in the GVA over time (most moved to the FO or LLA groups), whereas only one man remained in the BD group 3 years later. Among those who had the opportunity for continued violence (they maintained contact with their partner), 93% of GVA, 86% of BD, 77% of LLA, and 60% of FO men engaged in new violent behavior against their partners. Similar results were found by Thijssen and de Ruiter (2011), who observed highest recidivism rates among GVA individuals and lowest rates among FO men (average 27-month follow-up period), as observed by Llor-Esteban et al. (2015), who labeled FO individuals as the lowest-risk group. In the same vein, after a 15-month follow-up (Clements et al., 2002), GVA and BD individuals were more likely than LLA and FO men to be rearrested for IPVAW (15% of GVA men, 22% of BD, 5.5% of LLA, and 5.5% of FO). Furthermore, GVA individuals showed higher violent (47%; including IPVAW) and general recidivism rates (i.e., any recidivism outcome; 62%) than the other IPVAW offender subtypes. This would support the implicit assumption that lower levels of IPVAW are due to dyadic, relationship, environmental, or interactional factors (situational aggression among FO individuals), while more severe IPVAW is primarily attributed to the man’s personal characteristics that reflect biological and intrapersonal factors (stable aggression among GVA men). In this sense, recent research points toward a strong association between antisocial personality traits and IPVAW recidivism (Hilton & Radatz, 2021).

In Spain, both theoretical (Amor et al., 2009; Torres et al., 2013) and empirical research on IPVAW offender typologies (Boira & Jodrá, 2013; García-Jiménez et al., 2014; Herrero et al., 2016; Loinaz, 2014; Loinaz et al., 2011) has been conducted in the last decade. Most recently, an empirical Spanish IPVAW offender typology was proposed using the VPR risk indicators as clustering variables in a large-scale representative national sample of gender-based violence cases (n = 9,731; González-Álvarez et al., 2021). Following this typology, IPVAW arrested offenders might be classified according to their levels of psychological instability (e.g., severe and very serious psychological violence, threats to harm the victim, death threat, aggression escalation, exaggerated jealousy or suspicion of infidelity, controlling behavior, and stalking behavior) and antisociality (i.e., indicators related to the aggressor’s criminal history), resulting in a four-group solution. High instability/low antisociality (HiLa) and high instability/high antisociality (HiHa) individuals shared most risk indicators related to the aggressor’s psychological instability; whereas HiHa and low instability/high antisociality (LiHa) men endorsed more antisociality indicators. The low instability/low antisociality (LiLa) group was characterized by the less presence of VPR risk indicators. If compared with Holtzworth-Munroe et al.’s (2000) classic typology, the LiLa subtype would fit the FO classic type, HiLa subtype would correspond to the classic BD type, LiHa would match the GVA classic group. The HiHa subtype did not resemble the classic LLA type, as these individuals presented with much higher antisocial levels than it was described for the LLA type but, instead, it could be considered a mixture of the BD and GVA offender types (González-Álvarez et al., 2021).

Little is known about the utility and applicability of IPVAW offender typologies in increasing the effectiveness of police work for the prevention of new IPVAW episodes. One of the few works that deals with this question is Petersson and Strand’s (2020) longitudinal study (28-month mean follow-up period), which describes and compares 628 arrested and non-arrested male IPVAW perpetrators in Sweden, in terms of their individual characteristics (i.e., generally violent [GV] and partner only [PO] violent), and variables related to the IPVAW incident. Results showed that arrested perpetrators were more likely to be reported for severe forms of IPVAW, being assessed by the police post-arrest with a higher recidivism risk, and being more likely to be prosecuted for the reported IPVAW incident. Furthermore, among those perpetrators who were arrested, GV perpetrators were four times more likely to recidivate than PO violent individuals.

Purpose of the Study

The main goal of the current study was to examine whether the four Spanish IPVAW offender subtypes (i.e., HiHa, HiLa, LiHa, and LiLa; González-Álvarez et al., 2021) had different recidivism rates, and if they differed in the severity of the new IPVAW episodes. To this end, the offender subtypes were matched with their recidivism outcomes (including violation of court protective orders) in a longitudinal study. The expectation was that the HiHa subgroup, with high levels of instability and antisociality (corresponding to the BD and GVA classic offender types), would present with higher recidivism rates and of a more severe nature than the other subgroups. Conversely, the LiLa subgroup (i.e., low instability and low antisociality, corresponding to the FO classic offender type), was expected to have the lowest recidivism rate, with the less severe new IPVAW episodes.

Moreover, we were interested in observing whether reoffending rates remained stable within IPVAW offender types over time, as measured by the VPER. Considering that the VioGén System allocates proportional police protection measures based on the risk level detected, it was hypothesized that the risk levels of all subtypes would decrease in the course of time, but perhaps at a different pace: the risk of HiHa individuals would take longer to decline than would be the case for the LiLa subgroup. This study would contribute to our knowledge of the risk posed over time by each offender type, which could be helpful in improving the police and judicial protection of IPVAW victims, by the adoption of early protective measures matched to the risk posed by each subgroup (e.g., low-risk IPVAW subtypes would need less surveillance/victim protection than high-risk offender subtypes); which in turn would optimize the scarce resources available.

Method

Participants

For this study, we collected all gender-based violence cases registered in the VioGén System between October and December 2016 (first complaints made by female victims), provided by the Secretary of State for Security (Spanish Ministry of Home Affairs). In total, 10,623 cases were extracted in January 2018, resulting in a follow-up period ranging between 13 and 15 months (depending on the date of the initial complaint). All cases analyzed in our study included some sociodemographic information on victims and offenders (to describe the sample), as well as all the VPRs and VPERs conducted by the police officers and all the available data regarding compliance with the protective orders (in order to code risk levels and recidivism outcomes). Data extraction was refined, excluding those cases not including a VPR (cases recorded by a law enforcement agency that did not use this tool) and duplicated cases, resulting in a total sample of 9,672 valid cases.

In all cases, the offender was a man and the victim a woman. The mean age of the perpetrators was 39.1 years (SD = 12.4; range = 14–93), and they were mostly Spanish nationals (78.5%; i.e., individuals born in Spain and those who have legally acquired the nationality subsequently). Victims were also mostly Spanish nationals (75.6%) and their mean age was 36.1 years (SD = 11.7; range = 12–72). All the cases in this sample were included in the one used for the Spanish IPVAW typology development study (n = 9,731; González-Álvarez et al., 2021). The following distribution of IPVAW offender subtypes resulted in the current sample: LiLa (n = 3,926; 40.6%); HiLa (n = 2,672; 27.6%); HiHa (n = 2,068; 21.4%); and LiHa (n = 1,006; 10.4%). These percentages were equivalent to those of the development sample.

Instruments

The risk of recidivism of all cases included in our sample was assessed using version 4.0 of the VPR and the VPER (the most current version at the time of data extraction). Both are validated actuarial police IPVAW recidivism risk assessment tools integrated in the VioGén System and used by Spanish law enforcement agencies (see López-Ossorio et al., 2019a). The VPR is completed immediately by a police officer after the victim has filed the complaint, whereas the VPER is used to reassess the risk level for as long as the case is active in the VioGén System, every time a significant incident occurs (such as a new complaint or a violation of a protective order) or in predetermined periods of time according to the case risk level.¹ VPR_4.0 (area under the curve [AUC] = 0.658; 95% CI [0.625, 0.692]) includes 39 dichotomous risk indicators grouped in four dimensions (i.e., severity of the reported IPVAW episode, aggressor-related factors, victim vulnerability, and aggravating circumstances; López-Ossorio et al., 2017), resulting in five police recidivism risk levels (i.e., extreme, high, medium, low, or no risk).

VPER_4.0 includes 34 risk and 9 protective indicators, including variables related to the aggressor’s good prognosis, the adjustment and disposition of the victim toward protection measures, the degree of implication of the victim in her own protection and her own risk perception (López-Ossorio et al., 2017). There are two types of VPER_4.0 (López-Ossorio et al., 2019a): the VPER-S (AUC = 0.824; 95% CI [0.797, 0.851]), used to reassess risk in predetermined periods of time (as part of the risk management procedure); and the VPER-C (AUC = 0.823; 95% CI [0.796, 0.849]), adapted to reassess risk each time a new significant incident occurs between the perpetrator and his victim (i.e., each time the victim files a new complaint against her partner for an episode of intimate partner violence or a police officer becomes aware of a new incident involving the same two parties as the original report). Therefore, when a case includes a risk assessment with the VPER-C, it would be indicating that there has been a police recidivism event.

Measures

The allocation of cases to each of the IPVAW offender subtypes (i.e., LiLa, LiHa, HiLa, and HiHa) was retrieved from a previous study (González-Álvarez et al., 2021). To account for recidivism outcomes, two dichotomous variables (yes/no) were created: recidivism, coded positively when the case included (at least) one VPER-C risk assessment (reflecting a new reported IPVAW episode: physical, sexual, psychological violence or threats toward his victim,² including violations of court protective orders); and multi-recidivism, coded positively when more than one VPER-C risk assessment was conducted. Furthermore, a specific variable was created to account for the violation of court protective orders, along with five additional dichotomous variables detailing the characteristics of such violation (physical approach, virtual contact, with physical violence, with psychological violence, and stalking; Caballé-Pérez et al., 2020). Finally, the severity of the new IPVAW episode was coded dichotomously (severe/not severe) as indicated by the police officers in the first four factors of the first VPER-C of each case (i.e., the recidivism event was coded as severe when any of the indicators of physical, sexual, psychological violence, or threats was annotated as “severe”³ or “very severe”⁴).

Follow-up time was calculated as the difference between the date of the first complaint and the date when cases were extracted (January 2018). Time in custody (i.e., time spent in prison after the date of the first complaint) was subtracted, so follow-up time represented the individual’s opportunity to recidivate while residing in the community. We also calculated time at risk as the difference between the date of the first complaint and the date of the first police recidivism event, subtracting the time spent in prison (when applicable). The evolution of risk levels was analyzed taking into account the successive risk assessments (from the first to the last VPER) conducted for each case. Risk levels were categorized quantitatively (i.e., extreme = 5; high = 4; medium = 3; low = 2; and no risk = 1).

Data Analysis

First, we conducted a descriptive analysis of all variables for the total sample. Second, Chi-square tests were conducted to analyze the relationship between IPVAW offender types and recidivism outcomes. When χ² was significant, corrected standardized residuals (CSR; <−2; >2; a value above 2 means that in that cell values are higher than expected, whereas a value below −2 means that in that cell values are lower than expected) were used to determine which observed cells mainly contributed to this significance. Furthermore, Cramér’s V was used as the effect size measure (V values of 0.10, 0.30, and 0.50 were considered small, moderate, and large effect sizes, respectively; Cohen, 1988). Due to violations of the normality assumption, non-parametric Kruskal-Wallis and Mann-Whitney U tests were conducted for the analysis of quantitative variables (e.g., number of prison records). Bonferroni adjusted pairwise comparisons were used to investigate differences among IPVAW offender subtypes.

As there was no fixed follow-up time for the sample, differences in police recidivism rates among the four subgroups were examined employing the log-rank Kaplan-Meier method with log-rank (Mantel-Cox) test for significance, to explore the notion that the four groups recidivated at equal rates at the follow-up period. All statistical analyses were performed using SPSS Version 21 statistical software Package.

Results

Police Recidivism

The mean duration of follow-up for this longitudinal study was 1.2 years (SD = 0.11; range = 0–1.33). A total of 144 individuals (1.5% of the whole sample) served time in one of the prisons managed by the Spanish General Secretary of Penitentiary Institutions during the follow-up period: 53.5% were HiHa offenders, 24.3% LiHa, 16.7% LiLa, and 5.6% HiLa (χ²[3, N = 9,672] = 143.963, p = .000, V = 0.12). Of these, seven individuals (five HiHa and two LiHa) were in custody during the whole follow-up period, and thus they did not have the opportunity to recidivate while residing in the community. Therefore, all subsequent analyses of recidivism variables were performed excluding such individuals (n = 9,665).

First, 25.3% of cases (n = 2,444) did not include any VPER risk assessment (neither VPER-S nor VPER-C). This was due to an immediate deactivation of the case in the VioGén System, as a result of cessation of subsequent police or judicial measures (i.e., risk of recidivism was not estimated and/or the suspect was acquitted).⁵ The absence of VPER risk assessments was statistically related to the offender type (χ²[3, N = 9,665] = 946.11, p = .000, V = 0.31): LiLa cases were less likely to include a VPER than expected (59.3%), whereas HiHa (90.6%) and HiLa cases (85.8%) included more than expected. No statistically significant differences were found for the LiHa subgroup (72.8%).

Second, 1,717 cases included (at least) one VPER-C risk assessment, meaning the IPVAW recidivism rate for an average follow-up of 1.2 years (in days: M = 438.9; SD = 38.6; range = 1–486) was 17.8%. IPVAW repeat offenders took, on average, 133.92 days to reoffend (Mdn = 105; SD = 110.67; range = 0–458; those individuals with 0 days at risk were reported again the same day of their index arrest). Among these cases, 551 (5.7% of the total sample) included more than one VPER-C risk assessment, and they were labeled as multi-recidivism cases. In 465 cases (4.8%), the new reported incident was labeled as “severe.”

IPVAW recidivism rates for each offender type and differences in the severity of the new violent episodes are reported in Table 1. Significant (but small) differences in recidivism outcomes were observed between IPVAW offender subtypes. HiHa and LiHa individuals were more likely to be reported for new IPV episodes (single and multiple) against their victims, and to use severe violence (more likely among LiHa than among HiHa offenders). In contrast, the LiLa subgroup showed the smallest percentage of recidivism variables, followed by the HiLa subgroup.

Table 1.

Recidivism Rates for Each Offender Typology and Severity of the New Violent Episodes (N = 9,665).

Recidivism variables	Offender typology				χ²	Cramer’s V
	LiLa	HiLa	HiHa	LiHa
	(n = 3,926)	(n = 2,672)	(n = 2,063)	(n = 1,004)
Recidivism
Presence (%)	11.9	17.8	26.5	22.6	214.042***	0.149
CSR	−12.4	0	11.7	4.2	214.042***	0.149
Multi-recidivism
Presence (%)	2.7	5.7	10.2	8.1	152.087***	0.125
CSR	−10.4	0.1	9.9	3.4	152.087***	0.125
Multi-recidivism
Presence (%)	2.7	5.7	10.2	8.1	152.087***	0.125
CSR	−10.4	0.1	9.9	3.4	152.087***	0.125
Severe
Presence (%)	3.2	3.8	7.7	8.1	88.681***	0.096
CSR	−6.2	−2.9	6.8	5.1	88.681***	0.096

Note. CSR = corrected standardized residuals.

p < .05. **p < .01. ***p < .001.

The follow-up period for each case was significantly different depending on the offender type (H[3, 9672] = 12.545, p = .006): HiLa cases were the ones with the longest follow-up period (M = 441.70; SD = 29.85; range = 1–486); conversely, HiHa cases had the shortest follow-up time (M = 435.26; SD = 50.64; range = 1–486; Bonferroni-adjusted alpha level of .008). Nonetheless, the log-rank Kaplan-Meier survival estimator confirmed that, controlling for individual differences in follow-up time, the survival rates of the four IPVAW offender subtypes differed significantly for any new reported IPVAW incident (χ²[3, N = 9,665] = 220.897, p = .000; see Figure 1). No statistically significant differences were detected among IPVAW offender subtypes for the amount of time repeat offenders (n = 1,717) took to reoffend.

Figure 1.

Kaplan-Meier survival plot for police recidivism distinguishing among IPVAW offender subtypes.

Protective Orders

In our sample, 49% of victims (n = 4,732) were issued a court order of protection (i.e., restraining orders and/or prohibition of communication with the victim). Such protective orders were distributed differently across groups (χ²[3, N = 9,672] = 582.9, p = .000; V = 0.245): HiHa (67%; CSR = 18.5) and HiLa cases (55.1%; CSR = 7.4) accumulated the highest percentages, in contrast with LiHa (47.2%; CSR = −1.2) and LiLa cases (35.8%; CSR = −21.4; Table 2). In relation to recidivism, details on the percentage of protective order violations and the type of such violations are shown in Table 3. In this regard, statistically significant differences were detected among IPVAW offender types, although with small effect sizes associated. The highest proportion of protective order violations occurred in HiHa and LiHa cases; furthermore, these offenders were more likely to be arrested for approaching the victim. HiHa offenders were more likely to virtually contact their victims than the other subtypes, whereas LiHa individuals stood out for displaying physical violence after the breach of the order.

Table 2.

Violation of Protective Orders for Each Offender Typology (N = 4,732).

Type of protective order violation	Offender typology				χ²	Cramer’s V
	LiLa	HiLa	HiHa	LiHa
	(n = 1,406)	(n = 1,472)	(n = 1,381)	(n = 472)
Protective order violation
Presence (%)	16.6	22.8	28.8	28.1	64.800***	0.117
CSR	−7	−0.5	5.8	2.6
Protective order violation
Presence (%)	16.6	22.8	28.8	28.1	64.800***	0.117
CSR	−7	−0.5	5.8	2.6
Physical approach
Presence (%)	14.1	18.7	24.3	24.3	50.894***	0.106
CSR	−6	−0.9	5.2	2.7
Virtual contact
Presence (%)	3.4	6.3	9.8	5	41.902***	0.104
CSR	−5	0	5.8	−1
With physical violence
Presence (%)	12.3	10.4	15.4	20.3	9.366*	0.092
CSR	−0.8	−2.2	1.2	2.3
With psychological violence
Presence (%)	17.1	18.8	23.2	24.1	5.071	0.068
Stalking
Presence (%)	14.8	21.1	20.6	24.8	6.203	0.075

Note. CSR = corrected standardized residuals.

p < .05. **p < .01. ***p < .001.

Table 3.

VPR Police Recidivism Risk Levels for Each Offender Typology (N = 9,672).

VPR risk level	Offender typology				χ²	Cramer’s V
	LiLa	HiLa	HiHa	LiHa
	(n = 3,926)	(n = 2,672)	(n = 2,068)	(n = 1,006)
Extreme					6,031.005***	456
Presence (%)	0	1	6	0
CSR	−10.2	−2.5	18.2	−4.2
High
Presence (%)	0	7.5	20.1	0.8
CSR	−21.3	2.7	28.4	−7.7
Medium
Presence (%)	1.3	36.7	45.1	11.7
CSR	−40	22.5	29.4	−8
Low
Presence (%)	28.1	49.3	28.3	52.7
CSR	−14.3	16.1	−8.8	11.2
No risk
Presence (%)	70.6	5.4	0.6	34.9
CSR	63	−36.6	−36.1	0.7

Note. CSR = corrected standardized residuals.

p < .05. **p < .01. ***p < .001.

Risk Levels and Evolution of Cases

At the time of a first complaint, all cases in the VioGén System are assessed with the VPR tool, resulting in five police recidivism risk levels, which were distributed significantly differently (medium effect size) across the IPVAW offender subtypes (see Table 3). The HiHa cluster accumulated the higher proportion of cases with the three highest risk levels (6% of the HiHa individuals were labeled as extreme risk offenders, 20.1% as high, and 45.1% as medium risk); followed by the HiLa subgroup, with a majority of individuals classified in the medium (36.7%) and low risk levels (49.3%). LiHa individuals stood out for their low risk level (52.7%), and no risk was assessed in 70.6% of LiLa offenders. No cases were coded as extreme in the LiHa and LiLa subgroups, and no high-risk cases were detected among the LiLa offenders.

During the time a case is active in the VioGén System, police officers use the VPER tool to follow the impact of the protection measures over the recidivism risk. Cases are only deactivated when no risk is estimated and/or the suspect is acquitted. In this sense, the total number of VPER risk assessments and their resulting risk levels give us a sense of the evolution of risk in the cases analyzed: a large number of VPER risk assessments and high risk levels over time indicate that the offender continues to pose a risk to his victim and that the police protection measures cannot be withdrawn in a short period of time. In our sample, during a follow-up period ranging between 13 and 15 months, the average number of VPER risk assessments conducted per case was 4.03 (Mdn = 3; range = 0–51). In the specific case of VPER-C risk assessments (conducted after a new relevant incident), the average number was 0.27 (Mdn = 0; range = 0–14). Taking into account each IPVAW offender type, the higher amount of VPER risk assessments was observed among HiHa cases (M = 5.91; Mdn = 6; range = 0–32), followed by HiLa (M = 4.58; Mdn = 4; range = 0–51), LiHa (M = 4.06; Mdn = 2; range = 0–30), and, finally, LiLa offenders (M = 2.66; Mdn = 1; range = 0–23). These differences were statistically significant (H [3] = 1,078.546, p = .000), as confirmed by all post hoc pairwise comparisons (Bonferroni-adjusted alpha level of .008).

Resulting risk levels were coded quantitatively (from extreme = 5, to no risk = 1), and the mean risk level for each IPVAW offender type was observed after a new VPER risk assessment was conducted. As shown in Figure 2, the mean risk level of all offender types decreased over time. Furthermore, differences in average risk levels over time were detected among offender types: HiHa cases were coded as those with the highest risk over time, followed by LiHa and HiLa cases; conversely, at all times, LiLa cases were associated to the lowest risk levels.

Figure 2.

Evolution of recidivism risk levels over time for each IPVAW offender subtype.

Discussion

The main purpose of this study was to match Spanish IPVAW offender types (i.e., LiLa, LiHa, HiHa, and HiLa; González-Álvarez et al., 2021) with their risk of recidivism. Reported results showed that, during the follow-up period (M = 1.2 years), the police IPVAW recidivism rate in our sample (n = 9,665) was 17.8% (higher than the recidivism rate reported by López-Ossorio et al., 2017); furthermore, the multi-recidivism rate was 5.7%. New IPVAW incidents occurred, on average, between 4 and 5 months after the first complaint, which would highlight the importance of long-term police protection of the victims. As expected, statistically significant differences were found among IPVAW offenders in terms of their risk of recidivism, when they were classified according to the Spanish IPVAW offender typology (González-Álvarez et al., 2021).

First, results from the analysis of IPVAW police recidivism outcomes confirmed our hypothesis: IPVAW offenders with high levels of instability and antisociality (i.e., HiHa) showed the highest recidivism (and multi-recidivism) rates (26.5%), followed by those with low levels of instability and high levels of antisociality (i.e., LiHa; 22.6%). Likewise, such IPVAW offender subtypes were more likely to be reported for new severe IPVAW episodes against their victims (8.1% LiHa; 7.7% HiHa) than the other two subtypes. Furthermore, these offenders were also more likely to violate the protective orders imposed (28.8% HiHa; 28.1% LiHa), and in such cases, the highest proportion of physical violence toward their victim occurred in LiHa cases (20.3%). Conversely, IPVAW offenders with low levels of instability and antisociality (i.e., LiLa) showed the lowest recidivism rates (11.9%). These results align with prior findings on Holtzworth-Munroe et al.’s classic typology (i.e., highest recidivism rates among GVA individuals and lowest among FO men; Clements et al., 2002; Holtzworth-Munroe et al., 2003; Llor-Esteban et al., 2015; Petersson & Strand, 2020; Thijssen & de Ruiter, 2011) and highlight the importance of individual level factors of IPVAW offending (e.g., male’s antisocial tendencies, extensive criminal history, psychological instability) on recidivism (Hanson & Wallace-Capretta, 2004; Hilton & Radatz, 2021).

Given that the Spanish IPVAW offender typology was created incorporating some IPVAW risk indicators related to the recidivism of the offenders, correspondence was found among the risk levels estimated by the VPR risk assessment and the offender types. HiHa offenders were labeled as those with the highest recidivism risk, followed by HiLa, LiHa, and LiLa offenders. In relation to the stability of the risk posed by IPVAW offenders, it was predicted that all types would reduce their risk levels as a consequence of the adoption of police protection measures. According to the hypothesis of the study by Holtzworth-Munroe et al. (2003), FO offenders would present a greater decrease in their risk levels than BD and GVA subtypes. According to our results, all the offender types showed a lower risk level over time (follow-up period ranging between 13 and 15 months). Specifically, HiHa offenders (corresponding to BD and GVA subtypes) were coded as those with the highest risk over time and had a longer police follow-up (highest number of VPER risk assessments), followed in descending order by the HiLa (i.e., BD), LiHa (i.e., GVA), and LiLa (i.e., FO) typologies. Despite observing that some offenders recidivated during the follow-up period, the fact that none of the offender types showed an increase in their risk levels confirms the usefulness of the early adoption of protection measures by Spanish law enforcement agencies in IPVAW cases, and, especially, the long-term maintenance of this type of measures in the cases that are more resistant to change (i.e., HiHa cases).

Limitations and Implications for Practice

A major concern for research based on police data is that the information gleaned from law enforcement agencies may not represent the full extent of any offending (i.e., the police cannot assume that no incriminating information has been hidden, deleted, or remains otherwise undetected), especially when investigating IPVAW, where lack of reporting does not necessarily mean reduced IPVAW (individual, relationship, and abuse characteristics may shape reporting rates; Blay Gil, 2014; Morgan et al., 2022; Xie & Baumer, 2019). Considering the high percentage of unreported IPVAW in Spain (Delegación del Gobierno contra la Violencia de Género, 2020), studies including data from IPVAW cases registered in the VioGén System can only offer a detailed analysis of the characteristics of reported cases. On the other hand, taking into account the presumption of innocence of the individuals arrested by the police, some of the cases incorporated into the VioGén System may not end in a conviction (according to the Spanish State Attorney General’s Office report, 28.7% of IPVAW defendants tried in 2019 were not convicted; Fiscalía General del Estado, 2020), and thus, would no longer be considered as IPVAW cases. However, and unlike most of the published studies on IPVAW offenders that inform as limitations the use of relatively small, non-representative samples; the current sample represents, to date, the largest and most representative sample used in typological studies.

Taking into account potential intercultural differences in the profile of IPVAW offenders, we must highlight that our study was carried out with a Spanish sample, classified according to the Spanish IPVAW offender typology (González-Álvarez et al., 2021). In this sense, we encourage new large-scale empirical typological studies in geographically and culturally different samples, which might provide new evidence about the cross-cultural validity of this suggested classification.

This study offers empirical evidence of the potential utility of the Spanish IPVAW offender typology (González-Álvarez et al., 2021) in police settings. From an applied perspective, it could be useful to provide training to police officers in charge of victim protection on the characteristics of IPVAW offender types and warn victims of the profile of their aggressors, in order to allocate more police protection measures to the most dangerous cases and improve the self-protection plans (especially in those cases where the perpetrator is labeled as HiHa or LiHa). Furthermore, taking into account that in most cases resources are scarce in day-to-day police practice, operational decision-making must be optimized. After conducting the individual risk assessment of each offender using the VPR tool, police officers often face the difficult situation of deciding which case to prioritize given an identical estimated risk level (e.g., which victim to provide a telematic device that detects the geolocation of the aggressor or which victim needs to move to a shelter more urgently). With the individual risk being equal, knowledge of the type of IPVAW offender we are facing in each investigated case could allow us to introduce a qualitative nuance that could improve risk management operational decision-making.

Footnotes

Acknowledgements

The authors would like to thank all members of the Spanish Law Enforcement Agencies, users of the Viogén System, for their meticulous work.

Declaration of Conflicting Interests

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

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was partially supported by the University Research Institute of Criminology and Criminal Science of the University of Valencia.

ORCID iD

Virginia Soldino

Notes

Author Biographies

José Luis González-Álvarez (jlga@interior.es), PhD, is a Colonel Psychologist of Guardia Civil and founder of the Criminal Behaviour Analysis Section (SACD; 1995) and the Woman-Minor Teams (EMUMEs; 1996). He holds the position of Head of the Cabinet for Coordination and Studies of the Spanish Secretary of State for Security, where he is in charge of the coordination of criminological research projects.

Virginia Soldino (virginia.soldino@uv.es), PhD, is an Assistant Professor in the Department of Criminal Law and a researcher at the University Research Institute of Criminology and Criminal Science of the University of Valencia. Her research is focused on the study of offending behavior from a criminological perspective, with a special emphasis on the development of cross-cultural validation studies in Spain.

Jorge Santos-Hermoso (jorge.santos@icfs-uam.es). BA in Sociology; PhD student in Psychology. Researcher at the Institute for Forensic and Security Sciences (ICFS; Autonomous University of Madrid). Specialized in criminal—particularly, violent—behavior. His research interests include homicide, intimate partner violence, sexual aggression, and terrorism.

Enrique J. Carbonell-Vayá (enrique.carbonell@uv.es), PhD, is an Associate Professor at the University of Valencia. He is an experienced Forensic Psychologist in Spain (development of forensic assessments and consultation services in all jurisdictions), and the academic coordinator in the Valencian Community of the National Project for the Detailed Review of Intimate Partner Homicides.

References

Amor

P. J.

Echeburua

Loinaz

(2009). ¿Se puede establecer una clasificación tipológica de los hombres violentos contra su pareja? [Is it possible to establish a typological classification of violent men against their partner?] International Journal of Clinical and Health Psychology, 9(3), 519–539.

Blay Gil

(2014). “Voy o no voy”: el recurso a la policía en el caso de la violencia de género. Perspectivas de las víctimas. [“I go or I don’t go”: recourse to the police in the case of gender violence. Victims’ perspectives]. Estudios Penales y Criminológicos, 33, 369–400.

Boira

Jodrá

(2013). Tipología de hombres condenados por violencia de género en un contexto de intervención psicológica en la comunidad [Typology of men convicted of gender-based violence in a context of psychological intervention in the community]. Revista Latinoamericana de Psicologia, 45(2), 289–303. https://doi.org/10.14349/rlp.v45i2.806

Caballé-Pérez

Villalba-García

Santos-Hermoso

López-Ossorio

J. J.

González-Álvarez

J. L.

(2020). El quebrantamiento de las órdenes de protección en violencia de género: análisis de los indicadores de riesgo mediante el formulario VPR4.0 [Breach of protection orders in gender-based violence: Analysis of risk indicators using the VPR4.0 form]. Anuario de Psicología Jurídica, 30, 63–72. https://doi.org/10.5093/apj2019a17

Campbell

J. C.

(2002). Health consequences of intimate partner violence. The Lancet, 359(9314), 1331–1336. https://doi.org/10.1016/S0140-6736(02)08336-8

Cavanaugh

M. M.

Gelles

R. J.

(2005). The utility of male domestic violence offender typologies. Journal of Interpersonal Violence, 20(2), 155–166. https://doi.org/10.1177/0886260504268763

Clements

Holtzworth-Munroe

Gondolf

Meehan

(2002, November). Testing the Holtzworth-Munroe et al. (2000) batterer typology among court-referred maritally violent men [Poster presentation]. Annual Meeting of the Association for the Advancement of Behavior Therapy.

Cohen

. (1988). Statistical Power Analysis for the Behavioral Sciences (2nd ed.). Lawrence Erlbaum Associates.

Delegación del Gobierno contra la Violencia de Género. (2020). Macroencuesta de violencia contra la mujer 2019 [Macro-survey of violence against women 2019]. https://www.lamoncloa.gob.es/serviciosdeprensa/notasprensa/igualdad/Documents/2020/110920-Macroencuesta2019_Estudio.pdf

10.

Delsol

Margolin

John

R. S.

(2003). A typology of maritally violent men and correlates of violence in a community sample. Journal of Marriage and Family, 65(3), 635–651. https://doi.org/10.1111/j.1741-3737.2003.00635.x

11.

Dixon

Browne

(2003). The heterogeneity of spouse abuse: A review. Aggression and Violent Behavior, 8(1), 107–130. https://doi.org/10.1016/S1359-1789(02)00104-0

12.

Ellsberg

Arango

D. J.

Morton

Gennari

Kiplesund

Contreras

Watts

(2015). Prevention of violence against women and girls: What does the evidence say? The Lancet, 385(9977), 1555–1566. https://doi.org/10.1016/S0140-6736(14)61703-7

13.

Ellsberg

Jansen

H. A.

Heise

Watts

C. H.

Garcia-Moreno

(2008). Intimate partner violence and women’s physical and mental health in the WHO multi-country study on women’s health and domestic violence: An observational study. The Lancet, 371(9619), 1165–1172. https://doi.org/10.1016/S0140-6736(08)60522-X

14.

Fiscalía General del Estado. (2020). Memoria de la Fiscalía General del Estado (State Attorney General’s Office report). Fiscalía General del Estado. Ministerio de Justicia.

15.

García-Jiménez

J. J.

Godoy-Fernández

Llor-Esteban

Ruiz-Hernández

J. A.

(2014). Differential profile in partner aggressors: Prison vs. mandatory community intervention programs. European Journal of Psychology Applied to Legal Context, 6(2), 69–77. https://doi.org/10.1016/j.ejpal.2014.06.003

16.

García-Moreno

Zimmerman

Morris-Gehring

Heise

Amin

Abrahams

Montoya

Bhate-Deosthali

Kilonzo

Watts

(2015). Addressing violence against women: A call to action. The Lancet, 385(9978), 1685–1695. https://doi.org/10.1016/S0140-6736(14)61830-4

17.

González-Álvarez

J. L.

López-Ossorio

J. J.

Urruela

Rodríguez-Díaz

(2018). Integral monitoring system in cases of gender violence VioGén System. Behavior & Law Journal, 4(1), 29–40.

18.

González-Álvarez

J. L.

Santos-Hermoso

Soldino

Carbonell-Vayá

E. J.

(2021). Male perpetrators of intimate partner violence against women: A Spanish typology. Journal of Interpersonal Violence, Advance online publication. https://doi.org/10.1177/0886260521997442

19.

Gracia

(2004). Unreported cases of domestic violence against women: Towards an epidemiology of social silence, tolerance, and inhibition. Journal of Epidemiology & Community Health, 58, 536–537. https://doi.org/10.1136/jech.2003.019604

20.

Gracia

García

Lila

(2009). Public responses to intimate partner violence against women: The influence of perceived severity and personal responsibility. The Spanish Journal of Psychology, 12(2), 648–656. https://doi.org/10.1017/s1138741600002018

21.

Guedes

Bott

Garcia-Moreno

Colombini

(2016). Bridging the gaps: A global review of intersections of violence against women and violence against children. Global Health Action, 9(1), 31516. https://doi.org/10.3402/gha.v9.31516

22.

Hanson

R. K.

Wallace-Capretta

(2004). Predictors of criminal recidivism among male batterers. Psychology, Crime & Law,10(4), 413–427. https://doi.org/10.1080/10683160310001629283

23.

Heise

(2011). What works to prevent partner violence? An evidence overview. https://researchonline.lshtm.ac.uk/id/eprint/21062/1/Heise_Partner_Violence_evidence_overview.pdf

24.

Herrero

Torres

Fernández-Suárez

Rodríguez-Díaz

F. J.

(2016). Generalists versus specialists: Toward a typology of batterers in prison. European Journal of Psychology Applied to Legal Context, 8(1), 19–26. https://doi.org/10.1016/j.ejpal.2015.09.002

25.

Hilton

N. Z.

Harris

G. T.

Rice

M. E.

Houghton

R. E.

Eke

A. W.

(2008). An indepth actuarial assessment for wife assault recidivism: The domestic violence risk appraisal guide. Law and Human Behavior, 32(2), 150–163. https://doi.org/10.1007/s10979-007-9088-6

26.

Hilton

N. Z.

Radatz

D. L.

(2021). Criminogenic needs and intimate partner violence: Association with recidivism and implications for treatment. Psychological Services, 18(4), 566–573. https://doi.org/10.1037/ser0000450

27.

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(6), 1000–1019. https://doi.org/10.1037/0022-006X.68.6.1000

28.

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(4), 728–740. https://doi.org/10.1037/0022-006X.71.4.728

29.

Holtzworth-Munroe

Stuart

G. L.

(1994). Typologies of male batterers: Three subtypes and the differences among them. Psychological Bulletin, 116(3), 476–497. https://doi.org/10.1037/0033-2909.116.3.476

30.

Johnson

M. P.

(2008). A typology of domestic violence: Intimate terrorism, violent resistance, and situational couple violence. Northeastern University Press.

31.

Kropp

Hart

Webster

C. D.

Eaves

(1995). Manual for the spousal assault risk assessment guide (2nd ed.). The British Columbia Institute Against Family Violence.

32.

Kropp

P. R.

Hart

S. D.

(2015). The spousal assault risk assessment guide Version 3 (SARA-V3). ProActive ReSolutions Inc.

33.

Morgan

Boxall

Payne

J. L.

(2022). Reporting to police by intimate partner violence victim-survivors during the COVID-19 pandemic. Journal of Criminology. Advance online publication. https://doi.org/10.1177/26338076221094845

34.

Langhinrichsen-Rohling

Huss

M. T.

Ramsey

(2000). The clinical utility of batterer typologies. Journal of Family Violence, 15, 37–53. https://doi.org/10.1023/A:1007597319826

35.

Llor-Esteban

García-Jiménez

J. J.

Ruiz-Hernández

J. A.

Godoy-Fernández

(2015). Profile of partner aggressors as a function of risk of recidivism. International Journal of Clinical and Health Psychology, 16, 39–46. https://dx-doi-org.web.bisu.edu.cn/10.1016/j.ijchp.2015.05.004

36.

Lohr

J. M.

Bonge

Witte

T. H.

Hamberger

L. K.

Langhinrichsen-Rohling

(2005). Consistency and accuracy of batterer typology identification. Journal of Family Violence, 20, 253–258. https://doi.org/10.1007/s10896-005-5989-7

37.

Loinaz

(2014). Typologies, risk and recidivism in partner-violent men with the B-SAFER: A pilot study. Psychology, Crime & Law, 20(2), 183–198. https://doi.org/10.1080/1068316X.2013.770854

38.

Loinaz

Ortiz-Tallo

Sánchez

L. M.

Ferragut

(2011). Clasificación multiaxial de agresores de pareja en centros penitenciarios [Multiaxial classification of partner-violent men in prisons]. International Journal of Clinical and Health Psychology, 11(2), 249–268.

39.

López-Ossorio

J. J.

González-Álvarez

J. L.

Buquerín Pascual

García

L. F.

Buela-Casal

(2017). Risk factors related to intimate partner violence police recidivism in Spain. International Journal of Clinical and Health Psychology, 17, 107–119. https://dx-doi-org.web.bisu.edu.cn/10.1016/j.ijchp.2016.12.001

40.

López-Ossorio

J. J.

González-Álvarez

J. L.

Muñoz Vicente

J. M.

Urruela Cortés

Andrés-Pueyo

(2019a). Validation and calibration of the Spanish police intimate partner violence risk assessment system (VioGén). Journal of Police and Criminal Psychology, 34(4), 439–449. https://doi.org/10.1007/s11896-019-09322-9

41.

López-Ossorio

J. J.

Loinaz

González-Álvarez

J. L.

(2019b). Protocolo para la valoración policial del riesgo de violencia de género (VPR4.0): revisión de su funcionamiento. [Protocol for the police gender violence risk assessment (VPR4.0): Review of its performance]. Revista Espanola de Medicina Legal, 45(2), 52–58. https://doi.org/10.1016/j.reml.2019.01.002

42.

Loxton

Dolja-Gore

Anderson

A. E.

Townsend

(2017). Intimate partner violence adversely impacts health over 16 years and across generations: A longitudinal cohort study. PLoS One, 12(6). https://doi.org/10.1371/journal.pone.0178138

43.

Petersson

Strand

(2020). Characteristics and recidivism in relation to arrest: Differentiating between partner violent perpetrator subtypes. Nordic Journal of Criminology, 21(2), 203–222, https://doi.org/10.1080/2578983X.2020.1727119

44.

Roaldset

J. O.

Hartvig

Bjørkly

(2017). Psychometric properties and predictive validity of a police version of a violence risk screen – A pilot study. International Journal of Law and Psychiatry, 54, 133–139. https://doi.org/10.1016/j.ijlp.2017.06.007

45.

Thijssen

de Ruiter

(2011). Identifying subtypes of spousal assaulters using the B-SAFER. Journal of Interpersonal Violence, 26(7), 1307–1321. https://doi.org/10.1177/0886260510369129

46.

Torres

Lemos-Giráldez

Herrero

(2013). Violencia hacia la mujer: Características psicológicas y de personalidad de los hombres que maltratan a su pareja [Violence against woman: Psychological and personality characteristics of men who abuse their female partner]. Anales de Psicologia, 29(1), 9–18. https://doi.org/10.6018/analesps.29.1.130621

47.

Trabold

McMahon

Alsobrooks

Whitney

Mittal

(2020). A systematic review of intimate partner violence interventions: State of the field and implications for practitioners. Trauma, Violence, & Abuse, 21(2), 311–325. https://doi.org/10.1177/1524838018767934

48.

van der Put

C. E.

Gubbels

Assink

(2019). Predicting domestic violence: A meta-analysis on the predictive validity of risk assessment tools. Aggression and Violent Behavior, 47, 100–116. https://doi.org/10.1016/j.avb.2019.03.008

49.

Vilariño

Amado

B. G.

Vázquez

M. J.

Arce

(2018). Psychological harm in women victims of intimate partner violence: Epidemiology and quantification of injury in mental health markers. Psychosocial Intervention, 27(3), 145–152. https://doi.org/10.5093/pi2018a23

50.

Vives-Cases

Ruiz-Cantero

M. T.

Escribà-Agüir

Miralles

J. J.

(2011). The effect of intimate partner violence and other forms of violence against women on health. Journal of Public Health, 33(1), 15–21. https://doi.org/10.1093/pubmed/fdq101

51.

Weber

Bouman

Y. H. A.

(2020). Intimate partner violence: Variations in perpetrators and treatment allocation. Journal of Interpersonal Violence, 35(3–4), 1012–1034. https://doi.org/10.1177/0886260517692994

52.

World Health Organization. (2013). Global and regional estimates of violence against women: prevalence and health effects of intimate partner violence and non-partner sexual violence. https://apps.who.int/iris/bitstream/handle/10665/85239/9789241564625_eng.pdf;jsessionid=0D801E4E8F709E508151EE1BA7C60F21?sequence=1

53.

Xie

Baumer

E. P.

(2019). Crime victims’ decisions to call the police: Past research and new directions. Annual Review of Criminology, 2, 217–240. https://doi.org/10.1146/annurev-criminol011518-024748