Intimate partner violence (IPV) against women is one of our time’s greatest public health problems as well as a violation of human rights. Drawing on the definition provided by the Istanbul convention, IPV includes any form of physical, sexual, psychological, or economic harm or suffering perpetrated by a current or former partner (The Istanbul Convention, 2011). On a global scale, the most recent estimates show that 26–27% of women, aged 15 years or older, have been victimized by physical or sexual violence by a current or former partner (World Health Organization, 2021). Similar prevalence rates have been found in Sweden, where this study was conducted, where one in four women report ever experiencing physical or sexual IPV by a former or current partner (National Council for Crime Prevention [NCCP], 2014). Additionally, studies have also reported high rates of IPV re-victimization. To this end, in studies of criminal justice samples, such rates have been found to range between 15 and 60% (e.g., Klein & Tobin, 2008; Loinaz, 2014; Richards et al., 2014). However, most such studies rely on official records (e.g., arrest records or convictions) to measure re-victimization. In general, such studies tend to underestimate the actual rates of re-victimization as such crimes largely go unreported to the criminal justice system (e.g., NCCP, 2014; Sartin et al., 2006). As such, Bennett Cattaneo and Goodman (2005) found that IPV re-victimization rates in studies that relied on victim self-reports ranged between 22 and 74%. In Sweden, such rates have been found to range between 18 and 56% (Belfrage & Strand, 2012; NCCP, 2014; Petersson & Strand 2017; Svalin et al., 2018).

In their duty to prevent crimes, the police are one of the primary responders to IPV. The traditional police response to such calls, which best can be described as reactive in nature, has in many countries been replaced with a greater focus on violence prevention (Campbell et al., 2018; Kropp, 2008). This approach to IPV has been facilitated by using violence risk assessments, which play a key part in the prevention of such violence. In general, IPV risk assessment tools are based on the most prevalent and important perpetrator risk factors found in the literature, and by assessing these factors’ presence or absence the assessor makes an overall assessment of the risk for IPV re-victimization (Kropp et al., 2010). In Sweden, as in several other countries, the police use the Brief Spousal Assault Form for the Evaluation of Risk (B-SAFER: Kropp et al., 2010) to assess and prevent the risk of IPV re-victimization. The B-SAFER is based on the structured professional judgment (SPJ) approach to violence risk assessment and consists of ten perpetrator risk factors and five victim vulnerability factors (Kropp et al., 2010). When the risk for IPV re-victimization has been assessed with the B-SAFER, the assessor must try to mitigate the risk by suggesting a risk management strategy, which consists of various protective actions available to the assessor (e.g., a restraining order, alarm package, or arranging safe housing: Kropp et al., 2010).

Although the B-SAFER is one of many IPV risk assessment tools, it is one of the few tools that also includes victim vulnerability factors. To this end, research has shown how different victim characteristics may render the victim less able, or willing, to cooperate in the risk management strategy, which can increase the risk of being re-victimized (Belfrage & Strand, 2008; Bennett Cattaneo & Goodman, 2005; Kropp et al., 2010). However, it is important to stress that the assessment of victim vulnerability factors in no way is intended to blame the victims for their victimization. Instead, these factors help the police determine the likelihood that the victim will engage in self-protective actions and, thus, adhere to the recommended risk management strategy (Kropp et al., 2010).

The first vulnerability factor in the B-SAFER is a victim’s inconsistent behavior or attitude towards the perpetrator. This includes, for example, minimizing or denying the perpetrator’s violent behavior, or blaming oneself for being victimized (Kropp et al., 2010). There are several possible reasons for such behavior and attitudes from a victim, including feelings of fear, love, or sympathy for the perpetrator. Thus, the emotional attachment to a violent partner can result in an increased risk of being re-victimized as it may prevent the victim from leaving the abuser or from seeking help (Johnson, 2004; Kropp et al., 2010; Myhill, 2019).

The second vulnerability factor in the B-SAFER is an extreme fear of the perpetrator. This refers to fear so strong that the victim panics and becomes irrational, such as, for example, returning to the perpetrator due to a fear for violence escalation (Kropp et al., 2010). As such, this fear is likely to increase the risk for being re-victimized. The fear experienced by the victim may arise out of a fear for the victim’s own safety or for the safety and well-being of the victim’s children. Previous research has found a victim’s fear to be one of the most prevalent vulnerability factors among female victims of IPV reported to the Swedish police (Belfrage & Strand, 2008). Moreover, studies have also shown that a victim’s fear is positively correlated to a higher police-assessed risk for re-victimization (Belfrage & Strand, 2008; Robinson et al., 2018; Trujillo & Ross, 2008), as well as related to police recommendations of more protective actions (Trujillo & Ross, 2008). For police officers, it is possible that such fear may signal a higher legitimacy of the reported accusation, a higher severity of the IPV, an indicator of a history of abuse, or the perpetrator’s capacity for using severe forms of violence (Robinson et al., 2018; Trujillo & Ross, 2008).

The third victim vulnerability factor in the B-SAFER is inadequate support or resources, which entails a victim’s lack of professional or social support (Kropp et al., 2010). For example, this includes access to legal aid, shelters, or the help and support offered by the victim’s social network (e.g., family and friends). Access to such support may be dependent on the victim’s knowledge, motivation, or ability to seek help. For instance, language barriers or living in isolated communities will likely complicate the access to help and support and can lead to a victim becoming more vulnerable and more likely to be re-victimized. To this end, women living in more sparsely populated areas have been found to be victimized by more severe forms of IPV compared to women living in urban areas (e.g., Edwards, 2015; Strand & Storey, 2019). This could be explained by that access to help and support is more restricted for victims who are more socially and physically isolated. The fourth vulnerability factor in the B-SAFER is a victim’s unsafe living situation. Victims may have to live, work, or travel in places where they risk encountering their perpetrator (Kropp et al., 2010). For instance, if the victim is unable to separate from the perpetrator due to having children together, the risk for IPV re-victimization will be higher since the perpetrator will have greater access to the victim. As such, it is common that victims choose to continue to live in an unsafe environment with their violent partner, due to love or fear of leaving the relationship (Kropp et al., 2010).

Finally, the fifth vulnerability factor is a victim’s health problems, which can include mental health problems, physical impairments, or substance abuse. However, health problems in the B-SAFER also include judicial problems or unemployment as this can create psychosocial problems such as depression and stress (Kropp et al., 2010). As such, the various forms of health problems among victims may render the victim dependent on the perpetrator, for example, in terms of financial support, alcohol, or drugs. Thus, this co-dependence on the perpetrator makes it harder for a victim to leave and, hence, the risk for IPV re-victimization increases (Capaldi & Kim, 2007; Kropp et al., 2010; Ørke et al., 2018). Moreover, research has also shown how co-dependency on substance abuse for both the victim and the perpetrator increases the risk for IPV re-victimization as substance abuse can increase the level of aggressive behaviors among both partners (Capaldi & Kim, 2007; Ørke et al., 2018).

The association between the B-SAFER victim vulnerability factors and risk for IPV re-victimization has previously been examined in a handful of studies. Drawing on these results, studies have found that the number of vulnerability factors assessed as present correlates positively with a higher police-assessed risk for IPV (Belfrage & Strand, 2008) and with a greater number of protective actions recommended by the police (Storey & Strand, 2017). Moreover, in a study analyzing the same sample as used in the present study, Strand and Storey (2019) found a higher frequency of victim vulnerability factors among victims living in rural and remote areas compared to victims living in urban areas. Assessing the presence of victim vulnerability factors can contribute to more accurate assessments of risk for IPV re-victimization (Belfrage & Strand, 2008). At the same time, this assessment also helps identify the specific needs of IPV victims, which could improve the matching of protective actions to a victim’s vulnerability profile (Graham et al., 2021). For example, some victims may require more support and services addressing their mental health problems, whereas others may need more help with practical issues such as their financial situation or help with housing. Thus, the heterogeneity among IPV victims in terms of their vulnerabilities is important to acknowledge and address, if the suggested risk management is to have a chance of being successful (Kropp et al., 2010).

A commonly used procedure to study the heterogeneity of any subject is to categorize the subject of interest into a typology. A typology can be defined as “a means of classifying or categorizing subject matter into groups” (Boxall et al., 2015, p. 1), where such groups are often labeled as “subtypes”. Typology research has been used to identify subtypes of IPV (Johnson, 2006), as well as subtypes of both male and female perpetrators of such violence (Ali et al., 2016; Holtzworth-Munroe & Stuart, 1994). However, such typology research has rarely focused on the victims. Thus, much less is known about the existence of victim subtypes and their vulnerability profiles, and most such studies have focused on the consequences of women subjected to various severity levels of IPV (e.g., Davies et al., 2015). Still, some knowledge can be found in the IPV perpetrator typology research, where comparisons of victim characteristics occasionally are included.

To this end, the two most consistently identified subtypes of perpetrators within the IPV literature are the partner only (PO) violent perpetrator and the generally violent (GV) perpetrator (Petersson & Strand, 2020). To this end, Petersson and Strand (2020) reviewed all papers on the topic of IPV perpetrator typologies published since the influential paper by Holtzworth-Munroe and Stuart (1994). As such, the PO violent perpetrator does not use violence against others outside of their relationship or family, whereas the GV perpetrator is violent also against others than their partner (e.g., friends or strangers). Furthermore, these two subtypes have been found to differ on several behavioral and psychosocial risk factors for IPV, as well as in terms of recidivism rates of such violence (Petersson & Strand, 2020). Compared to the PO violent subtype, the GV subtype use more severe and frequent forms of IPV, display more psychosocial risk factors for IPV (e.g., mental health problems, substance abuse, and unemployment), and are more likely to recidivate in IPV (Petersson & Strand, 2017, 2020). In contrast, the PO violent subtype has been found to be more socially well-adjusted outside of their relationship, displaying, for example, more pro-social attitudes and values, and shown higher levels of integration and participation in their communities (e.g., Herrero et al., 2016; Mach et al., 2020; Walsh et al., 2010). As such, PO violent perpetrators have been suggested to have a higher stake in conformity than GV perpetrators (Mach et al., 2020; Petersson & Strand 2020). However, few studies have included comparisons between the victims of these two perpetrator subtypes.

To date, only one previous study has focused primarily on examining victim vulnerability factors between victims of GV perpetrators and victims of PO violent perpetrators. As such, in a sample of IPV reported to the Spanish police, Aguilar Ruiz and González-Calderón (2020) found that victims of GV perpetrators were more often isolated from their family and social environments, more likely to minimize and justify their partner’s abusive behavior, more fearful of the perpetrator and for their own physical safety, displayed more substance abuse and physical or mental disabilities, and had experienced more abuse by previous partners, compared to victims of PO violent perpetrators. Moreover, victims of GV perpetrators were also found to be more likely to not be living together with the perpetrator and were more likely to have expressed intent of withdrawing previous complaints made to the police. Although not primarily examining victim vulnerabilities, Delsol et al. (2003) also found similar results pertaining to the fear of the perpetrator among the two subtypes of victims. However, Aguilar Ruiz and González-Calderón (2020) found no differences in relation to confronting the perpetrator, a victim’s mental disorder or suicidal behavior, pregnancy or recent childbirth, or having started an intimate relationship with a new partner.

Additionally, addressing the lack of knowledge on the importance of specific vulnerability factors related to an increased risk for suffering severe IPV victimization, Aguilar Ruiz and González-Calderón (2020) also found differences between the two groups of victims. Unique predictors for victims of GV perpetrators included living with the perpetrator, isolation, minimizing or justifying the abuse, and substance abuse. Corresponding vulnerability factors for victims of PO violent perpetrators included fear of suffering a lethal assault and having experienced abuse in a previous intimate relationship. Meanwhile, having filed prior complaints against the perpetrator was found to decrease the risk of having suffered severe IPV victimization among victims of GV perpetrators.

The Current Study

There is currently a wealth of research on partner violent men and several studies have found strong support for the heterogeneity among such perpetrators. To this end, a meaningful approach to studying these perpetrators has been to compare the two most consistently identified subtypes of such violence, namely, the GV and the PO violent perpetrators (see Petersson & Strand 2020 for a systematic review). This has generated knowledge about, for example, how such perpetrators differ in terms of IPV risk factors and rates of IPV recidivism (e.g., Cantos & O’Leary, 2014; Petersson & Strand, 2020). In contrast, the female victims of these two subtypes of male IPV perpetrators have not previously been given much attention and more knowledge about these victims’ vulnerabilities could further help to improve risk management strategies that, in turn, could reduce the rates of re-victimization. For example, as previously reported, victims of GV perpetrators present more psychosocial vulnerability factors for IPV re-victimization than victims of PO violent perpetrators (e.g., Aguilar Ruiz & González-Calderón 2020; Delsol et al., 2003), meaning that these victims require different forms of support and services. Currently, such variations among victims are not accounted for since vulnerability factors are often thought to be equally important for all victims (Aguilar Ruiz & González-Calderón, 2020). As previously mentioned, the sample used for this study has previously been used to study differences in vulnerability factors in terms of rurality (Strand & Storey, 2019). However, in the present study, we shifted focus from a contextual (i.e., rurality) to an individual level of analysis (i.e., the victim subtypes).

This study aimed to investigate how victims of PO violent and GV perpetrators differ in terms of vulnerability factors and assessed risk for IPV re-victimization. More specifically, the following research questions were addressed in this study:

  1. 1.

    How do victims of PO violent perpetrators and victims of GV perpetrators differ in terms of victim vulnerability factors?

  2. 2.

    How do victims of PO violent perpetrators and victims of GV perpetrators differ in terms of assessed risk for IPV re-victimization?

  3. 3.

    Which victim vulnerability factors are most strongly associated with an elevated assessed risk for IPV re-victimization for victims of PO violent perpetrators and victims of GV perpetrators?

Method

Sample

The sample in this study was drawn from an eight-year (2009–2016) prospective research project conducted in collaboration with the Swedish police. The aim of the project was to implement and evaluate the use of structured violence risk assessments for various forms of family violence in four Swedish police districts (Strand et al., 2016). These four districts included urban, rural, and remote areas in Sweden and have been described more closely in the previous study using the same sample (Strand & Storey, 2019). The police officers conducting the risk assessments were recruited from family violence units within the police. Thus, these police officers had received specialized training in IPV and risk assessment.

The total sample consisted of 1479 women who had reported being victimized by IPV, by a current or former male partner, to the Swedish police. Information about age was only available for a subset of the sample (n = 980), where the mean age was 35.8 years (SD = 12.6, Range = 14–83). The following demographical information was based on the final sample of 1479 women. As such, 21.0% of the victims had an immigrant background and 52.7% had children younger than 18 years living at home. The most common index crimes in the police reports were physical assault (n = 723, 51.8%), followed by gross violation of a woman’s integrity (n = 369, 26.4%), threats (n = 196, 14.0%), other crimes (e.g., violation of restraining orders, harassment, or stalking: n = 76, 5.4%), sexual assault (n = 22, 1.6%), and attempted murder/manslaughter (n = 8, 0.6%). Eighty-three cases (5.4%) were missing information about index crime reported. The crime of gross violation of a woman’s integrity is internationally unique for Swedish legislation, where male perpetrators who are found guilty of repeatedly committing crimes against the same female partner (i.e., on more than one occasion), aimed at violating her integrity, can be charged with this crime. The purpose of this crime is that it considers repeated crimes towards a female partner to be more severe than each crime separately, and it allows the court to sentence the perpetrator to a stricter sentence compared to if the perpetrator was to be sentenced for several crimes separately, such as physical assault or illegal threats.

Materials and Measures

The material in this study mainly consisted of police risk assessments of IPV, based on the Swedish version of the B-SAFER (Kropp et al., 2010), as well as additional case-related information. As such, alleged perpetrators who were reported to any of the four police districts between 2009 and 2014 for an IPV-related crime were subjected to a B-SAFER assessment.

Victim vulnerability factors. The victim vulnerability factors in this study were retrieved from the B-SAFER, which is a SPJ tool specifically intended to be used by the police, and the Swedish version was developed together with the Swedish police (Belfrage & Strand, 2008; Kropp et al., 2010). Although the B-SAFER consists of ten risk factors divided into the nature of the IPV (i.e., violent acts, violent threats or thoughts, escalation, violation of court orders, and violent attitudes), the psychosocial adjustment of the perpetrator (i.e., general criminality, intimate relationship problems, employment problems, substance use problems, and mental health problems), given the aim of this study only the five victim vulnerability factors described in the introduction section (inconsistent attitudes or behavior, extreme fear of perpetrator, inadequate support or resources, unsafe living situation, and health problems) were analyzed. These risk and victim vulnerability factors are assessed on a three-point Likert scale as either present (yes), partially present (possibly), or absent (no). This decision should be based on several sources of information, including hearings with the victim, perpetrator, and any witnesses, as well as on information retrieved from police registers (e.g., previous convictions or access to firearms). If the assessor lacks information about a risk or victim vulnerability factor, the factor is omitted. As in other risk assessment tools based on the SPJ approach, the risk and vulnerability factors are drawn from systematic reviews of the IPV literature and included in the B-SAFER (Kropp et al., 2010). Moreover, the assessor can also include other case-specific risk and victim vulnerability factors if deemed relevant to the risk of future violence.

Risk for IPV re-victimization. Based on the presence of the ten risk factors and the five victim vulnerability factors in the B-SAFER, the assessor produces two summary risk ratings for IPV re-victimization: one pertaining to the risk for imminent IPV, and one pertaining to the risk for severe or deadly IPV (Kropp et al., 2010). However, only the latter risk rating was analyzed in this study as few IPV risk assessment instruments assess the risk for imminent IPV, thus, making the comparability of our results to previous studies easier (e.g., Aguilar Ruiz & González-Calderón 2020). In the B-SAFER, the risk ratings are scored as either low, moderate, or high risk. In contrast to other methods of violence risk assessment, risk assessment tools based on the SPJ approach do not sum the scores from the risk and vulnerability factors in each case algorithmically. Rather, the summary risk rating is achieved by the assessor drawing on his or her own professional knowledge and discretion of IPV in weighing the risk and vulnerability factors together into an assessed, summarized, risk rating (Kropp et al., 2010). Finally, the assessor should attempt to mitigate the identified risk by suggesting risk management strategies. However, since the risk management strategies implemented in each case were not systematically documented by the police, this was not analyzed in this study.

The B-SAFER has been reported to be a useful tool for the Swedish police (e.g., Belfrage & Strand, 2008, 2012; Storey et al., 2014), although with some limitations (e.g., Svalin et al., 2017), Furthermore, the tool is reported to have good-to-excellent interrater reliability and predictive validity (e.g., Serie et al., 2017; Storey et al., 2014; Thijssen & de Ruiter, 2011), as well as shown good concurrent and construct validity (Au et al., 2008).

Subtype categorization. The victims in the sample were categorized as either victims of PO violent perpetrators (n = 728, 49.2%) or victims of GV perpetrators (n = 751, 50.8%), based on whether the perpetrators were rated by the police as generally violent/criminal or not. This rating was done using the item ‘general criminality’ in the B-SAFER (Kropp et al., 2010). Information about this item was retrieved from police registers (e.g., previous convictions) and from hearings with the victims and/or the perpetrator, and included any criminal acts committed outside of the relationship or family, regardless of whether the perpetrator was convicted for the offence or not. Thus, this could include physical assaults against strangers or friends, or drug related offences. As such, those perpetrators who were assessed by the police as being generally violent/criminal according to the procedure described above were labeled as GV perpetrators. Those perpetrators who were found to be violent or using criminal acts that were confined to the relationship or family were labeled as PO perpetrators. This procedure has previously been adopted when using risk assessment instruments to categorize both perpetrators (Serie et al., 2017; Thijssen & de Ruiter, 2011) and victims as either GV or PO violent (e.g., Aguilar Ruiz & González-Calderón 2020).

Procedure

In Sweden, when a police report is made or registered in their systems, the police officer indicates via a tick-box whether the crime is IPV-related. Thus, any crimes reported where the suspect was either a current or former intimate partner to the victim was identified as IPV. Subsequently, all crimes in the four districts that were categorized as IPV-related were subjected to an initial screening assessment of the risk of IPV re-victimization (Swedish National Police Board, 2019). If such a risk was identified, the case was referred to the family violence unit, which is responsible for carrying out structured violence risk assessments using the B-SAFER (Kropp et al., 2010), as well as suggesting and implementing adequate risk management strategies. In general, the police followed the five steps required when conducting a B-SAFER assessment. First, background information about the case was collected (e.g., demographics of the victim and perpetrator), including a brief description of the IPV incident. Second, the perpetrator risk factors in the B-SAFER were assessed. Third, the victim vulnerability factors in the B-SAFER were assessed. Fourth, based on the second and third steps, the summary risk ratings for the risk of IPV re-victimization were made. Fifth, based on the outcome of the B-SAFER assessment, the police officer suggested various risk management strategies to be initiated in each case.

Statistical Analyses

Before conducting the statistical analyses, the victim vulnerability factors were collapsed into dichotomous variables, measuring the presence (i.e., combining ratings of present and partially present) or absence of such factors. Furthermore, the summary risk rating of severe or lethal IPV was also collapsed into a dichotomous variable, indicating low risk or elevated risk (i.e., combining ratings of moderate and high risk). The rationale for this conversion of data was to facilitate the interpretations of the results and this procedure has also been used previously with similar data (Petersson et al., 2019). Comparisons between the victims of PO violent and GV perpetrators on categorical data were carried out using chi-square tests of independence (χ2) with phi as measure of effect size. Odds ratios (OR) with 95% confidence intervals were also used to further explore the magnitude of differences. OR greater than 1 indicate an increase in odds of the outcome, whereas OR less than 1 indicate a decrease in odds of the outcome measured. Moreover, two binary logistic regression analyses were conducted to examine the relative importance of the vulnerability factors in relation to assessed risk for IPV re-victimization. As such, for each victim subtype, the vulnerability factors were entered as predictors and the summary risk rating (i.e., the risk for severe or lethal IPV) was used as the dependent variable. No issues with multicollinearity between the predictors were identified before running the logistic regression analyses. All statistical analyses were conducted using the IBM SPSS version 26. This study received ethical approval from the Swedish Ethical Review Authority.

Results

Differences in Vulnerability Factors

The first research question aimed to examine differences between the victims of PO violent perpetrators and the victims of GV perpetrators in terms of the B-SAFER vulnerability factors. Results showed that the proportions of victim vulnerability factors present were high in the overall sample of victims. Nevertheless, as can be seen in Table 1, significant differences were found between the two groups on all victim vulnerability factors, where victims of GV perpetrators displayed a higher proportion of all such factors compared to victims of PO violent perpetrators. However, the effect sizes were overall small. Albeit, except for unsafe living situation, the odds of displaying the B-SAFER victim vulnerability factors were almost twice as high among victims of GV perpetrators compared to among victims of PO violent perpetrators.

Table 1 Presence of victim vulnerability factors among PO victims and GV victims (N = 1479)
Full size table

Differences in Assessed Risk for IPV Re-Victimization

The second research question aimed to examine differences between the two groups of victims in terms of their assessed risk of IPV re-victimization. To this end, the proportions of victims of GV perpetrators (48.7%) and victims of PO violent perpetrators (26.3%) assessed with an elevated risk for severe or lethal IPV re-victimization differed significantly, χ2(1, 1436) = 76.62, p = .000, phi = 0.23 (OR = 2.7; 95% CI = [2.1, 3.2]). As such, victims of GV perpetrators were generally assessed with a higher risk of being re-victimized by severe or lethal forms of IPV. More specifically, the odds of being assessed with an elevated risk for such victimization were nearly three times as high for victims of GV perpetrators compared to victims of PO violent perpetrators.

Vulnerability Factors Associated with an Elevated Assessed Risk for IPV Re-Victimization

Finally, logistic regression analyses were conducted to examine the vulnerability factors’ association with an elevated assessed risk for IPV re-victimization among the two groups of victims separately. For victims of PO violent perpetrators, extreme fear of the perpetrator and an unsafe living situation emerged as significantly related to an elevated assessed risk for IPV re-victimization (see Table 2). Likewise, the same two vulnerability factors were also significantly related to an elevated assessed risk for IPV re-victimization for victims of GV perpetrators. For both groups of victims, extreme fear of the perpetrator demonstrated the strongest association with an elevated assessed risk for IPV.

Table 2 Logistic regression models of the B-SAFER victim vulnerability factors’ predictability of elevated assessed risk for IPV re-victimization among PO victims and GV victims
Full size table

Discussion

The aim of this study was to compare victim vulnerability factors and the assessed risk for IPV re-victimization among the victims of the two most consistently identified subtypes of IPV perpetrators: the Partner Only (PO) violent and the Generally Violent (GV: Petersson & Strand 2020). To this end, using a sample of IPV reported to the Swedish police, the presence of victim vulnerability factors was high in the overall sample (above 40% for almost all factors). However, the results showed that these factors were significantly more common among victims of GV perpetrators. In general, victims of GV perpetrators demonstrated more inconsistent attitudes or behaviors, extreme fear of the perpetrator, inadequate support or resources, an unsafe living situation, and health problems. Moreover, victims of GV perpetrators were generally assessed by the police with a significantly higher risk for IPV re-victimization. Finally, in relation to the victim vulnerability factors most strongly associated with an elevated assessed risk for IPV re-victimization, the presence of extreme fear of the perpetrator and having an unsafe living situation were significantly related to such an outcome for both groups of victims.

The frequency of victim vulnerability factors was high in the overall sample of victims in our study. This is perhaps not surprising as this was a sample drawn from police data, where the severity of IPV is usually higher compared to, for example, community samples (Johnson, 2006; NCCP, 2014). This highlights the importance of assessing vulnerability factors among victims of IPV reported to the police when assessing the risk for re-victimization of such violence (Belfrage & Strand, 2008), as well as the need to offer these victims support and services that target their psychological and behavioral vulnerabilities (e.g., Kuijpers et al., 2011).

Nevertheless, the proportion of vulnerability factors present was generally higher for victims of GV perpetrators than victims of PO violent perpetrators. As such, our findings echo those reported in previous studies (Aguilar Ruiz & González-Calderón, 2020; Delsol et al., 2003), which also found that victims of GV perpetrators were more likely to display minimization or justification of the abuse (i.e., inconsistent attitudes or behavior in the B-SAFER), fear of the perpetrator, isolation (i.e., inadequate support or resources in the B-SAFER), as well as substance abuse and disabilities (i.e., health problems in the B-SAFER). The higher degree of vulnerability among these victims may be explained by the risk profile and characteristics displayed by their GV perpetrators. In general, GV perpetrators use more severe and frequent forms of IPV, more controlling and coercive behaviors, as well as display more violent attitudes and psychosocial risk factors, including substance abuse and mental health problems (Aguilar Ruiz & González-Calderón, 2020; Petersson & Strand, 2020). For example, the higher prevalence of violent attitudes found among GV perpetrators may also have led their victims to minimize and justify their partner’s use of violence (e.g., Myhill 2019). Likewise, the greater prevalence of substance abuse among GV perpetrators (e.g., Petersson & Strand 2020) may have caused the victims to also resort to such abuse to cope with their victimization (Aguilar Ruiz & González-Calderón, 2020). Problematically, previous studies have found that such co-dependence increases the risk of re-victimization even further (Capaldi & Kim, 2007; Ørke et al., 2018).

Thus, the consequences of being victimized by GV perpetrators are bound to be more severe and diverse. For example, Davies et al. (2015) reported that victimization of several forms of violence (e.g., physical, psychological, and sexual), as well as controlling and coercive behaviors, had a cumulative negative effect on victims of IPV, who presented worse health outcomes than victims who were exposed to less various forms of IPV. Noteworthy, the vulnerability factor that differed the most between victims of GV and PO violent perpetrators in this study was extreme fear of the perpetrator (56.5% vs. 40.3%). This result corroborates the results of previous studies that examined such fear among victims of GV and PO violent perpetrators (Aguilar Ruiz & González-Calderón, 2020; Delsol et al., 2003). Thus, being victimized by frequent, diverse, and severe forms of IPV is more likely to instill fear in the victims.

This study also found that victims of GV perpetrators were, in general, more likely to be assessed with an elevated assessed risk for IPV re-victimization. This result was expected, drawing on previous findings of GV perpetrators as posing an overall higher risk for recidivating in IPV (e.g., Cantos et al., 2015). Given the GV perpetrators’ propensity to use more severe and frequent IPV in general, the police may be more likely to assess the risk for future IPV as higher for these victims than for victims of PO violent perpetrators. Furthermore, these results are in line with previous research on the Swedish police’s use of the B-SAFER, which found a positive correlation between the presence of vulnerability factors and an elevated police-assessed risk for IPV re-victimization (Belfrage & Strand, 2008). Thus, it is unsurprising that the combination of the GV perpetrators’ risk profile and the heightened degree of vulnerability among their victims results in a higher assessed risk for IPV re-victimization.

This study also examined the specific importance of victim vulnerability factors in relation to an elevated assessed risk for IPV re-victimization among both subtypes of victims. Thus, we addressed an area previously identified as in need of more research (Aguilar Ruiz & González-Calderón, 2020). The results showed that, in general, the same victim vulnerability factors were related to an elevated assessed risk for IPV re-victimization for both types of victims. To this end, the presence of the factors extreme fear of the perpetrator and an unsafe living situation were significantly related to such an outcome. These results indicate that the Swedish police in our sample find these two vulnerability factors in the B-SAFER to be most important for an elevated risk for IPV re-victimization and, thus, do not differentiate between victims of PO violent perpetrators and victims of GV perpetrators when assessing risk for such violence. These results contrast those reported by Aguilar Ruiz and González-Calderón (2020), who found unique predictors for suffering severe IPV victimization for victims of GV perpetrators (i.e., living with the aggressor, minimizing or justifying the abuse, isolation, and substance abuse) and victims of PO violent perpetrators (i.e., fear of suffering a lethal assault and being abused by a previous partner). However, it is important to bear in mind that Aguilar Ruiz and González-Calderón (2020) examined predictors among victims who had already suffered severe forms of IPV victimization, whereas in our study we examined the risk factors the police assessed as most important for victims being re-victimized by severe forms of IPV. It is important to bear in mind the current lack of studies and knowledge within both these areas and stress the need for further research.

Nonetheless, in this study, extreme fear of the perpetrator and having an unsafe living situation were regarded as contributing to an elevated assessed risk for IPV re-victimization for victims overall. To this end, the former vulnerability factor has in previous research been reported as one of the most important factors for an elevated police-assessed risk for IPV re-victimization, as it may signal a higher legitimacy of the reported accusation and is indicative of a perpetrator’s capacity for using severe IPV (Belfrage & Strand, 2008; Robinson et al., 2018; Trujillo & Ross, 2008). Regarding an unsafe living situation, this usually means that the victim continues to live together with the perpetrator or that the perpetrator knows about the victim’s whereabouts (e.g., where the victim lives and his or her daily routines: Kropp et al., 2010). Thus, this vulnerability is an indicator of the perpetrator’s access to the victim, which increases the risk for violence.

Limitations

Compared to the study by Aguilar Ruiz and González-Calderón (2020), the victim vulnerability factors examined in this study were more broadly defined, thus, making it impossible to know the specific vulnerabilities displayed by the victims. For example, health problems in the B-SAFER can vary from mental health problems to unemployment, which entails a broad range of problems that affect victims in very different ways. Therefore, more detailed information about the vulnerability factors would have been desired. Another limitation concerns the internal validity of victims categorized as victims of the PO violent subtype. To this end, it may be the case that several perpetrators were misclassified as PO violent because their general criminality and violence had not yet been detected by the police, or that these perpetrators began displaying a generally criminal and violent behavior after the risk assessment was conducted. Although possible, previous research has shown that the subtypes tend to be stable over time (Holtzworth-Munroe et al., 2003). Therefore, the potential misclassification of victims as victims of PO violent perpetrators instead of victims of GV perpetrators is not believed to have had a substantial impact on our results.

Finally, drawing on the sample used in this study, the generalizability of the results is limited to male-to-female perpetrated IPV, reported to the Swedish police. Although the sample was recruited from four Swedish police districts, the findings are generally similar to previous studies from other police districts in Sweden (e.g., Belfrage & Strand, 2008; Svalin et al., 2018). These similarities include the rates of vulnerability factors present, rates of missing values, as well as sample characteristics (e.g., type of crimes reported and age of the victims ). Albeit, IPV committed by men against women, as well as IPV reported to the police, is generally more severe in nature (NCCP, 2014). Therefore, generalizations to other forms of IPV, such as female-to-male, or to other samples (e.g., community samples) cannot be made. Nonetheless, it is important to bear in mind the conclusion by Sartin et al. (2006), stating that “individuals who have come to the attention of the legal system…are likely to represent the more severe end of the continuum for domestic violence perpetration. Here is an easily identified population for intervention, and failure of the intervention is likely to have serious consequences for the victim” (p. 428). Thus, although IPV reported to the police only provides a glimpse of all forms of such violence, any intervention aiming to prevent this from being repeated may have potentially life-saving consequences for the victims.

Practical Implications

There are some important practical implications that can be drawn from the results of this study. First, overall, we found that the presence of victim vulnerability factors was high among the victims in our sample. As these vulnerability factors can complicate a victim’s possibility to engage in self-protective actions and the proposed risk management, thus, increasing the risk for IPV-revictimization (Kropp et al., 2010), such factors need to be included in IPV risk assessments (e.g., Aguilar Ruiz & González-Calderón 2020; Belfrage & Strand, 2008). Moreover, the high frequency of vulnerability factors also highlights the victims’ need for support and services that target their psychological and behavioral vulnerabilities (e.g., Kuijpers et al., 2011).

Perhaps the most important implication of the results in this study, however, relates to the demonstrable heterogeneity among IPV victims. As such, studying victims of IPV through the same typological lens as applied to IPV perpetrators (i.e., the PO violent and GV categorization) provided useful information about, among other things, the various vulnerabilities and needs of different victims. Thus, as advocated by Aguilar Ruiz and González-Calderón (2020), victims of GV perpetrators may require more extensive protective actions targeting their overall more vulnerable situation. This could include helping with safer housing options (e.g., women’s shelters), restraining orders, and various alarm solutions. Meanwhile, the higher degree of vulnerability among victims of GV may make them more reluctant to, at first, receive help from the police or crime victim support organizations. To this end, it is important for those offering help and support to motivate and empower the victim to participate in the police investigation as well as to adhere to the protective actions suggested. For practitioners who are working with risk assessment and management of IPV, categorizing such cases as either involving victims of a PO violent perpetrator or victims of a GV perpetrator would therefore be an important first step towards providing accurate ratings of risk for re-victimization as well as implementing adequate and tailored protective actions. As previously suggested (Aguilar Ruiz & González-Calderón, 2020), the police need more training in terms of typologies and their implications for the risk of re-victimization.

Conclusions and Recommendations for Future Research

The current study supports the findings of the only previously conducted study on this topic (Aguilar Ruiz & González-Calderón, 2020). As such, victims of GV perpetrators are more vulnerable and more at risk of being re-victimized by severe forms of IPV. To this end, it is important that those who work with risk assessment and management of IPV make the distinction between victims of PO violent perpetrators and victims of GV perpetrators, as this could help inform about the level of risk as well as both level and type of protective actions necessary to mitigate this risk. However, it should also be remembered that this study g used police risk ratings as the outcome measure. Thus, future research should attempt to validate the findings of this study using actual re-victimization data. It is also important that future research continue to examine potential differences between victims of PO violent perpetrators and victims of GV perpetrators to increase the body of knowledge on the heterogeneity among victims of IPV.