Women Victims of Intimate Partner Violence and Intimate Partner Homicide: A Typology Based on Victimization Variables

Introduction

Intimate partner violence (IPV) refers to any behaviour that causes physical, psychological, or sexual harm to any member of an intimate partner relationship (World Health Organization [WHO, 2012a]). Intimate partner violence against women (IPVAW) (or gender-based violence, in Spain) is perpetrated by men against women and “encompasses all acts of physical and psychological violence, including assaults on sexual freedom, threats, coercion, or arbitrary deprivation of liberty” (Ley Orgánica 1/2004, p. 10).

The WHO (2021) notes that IPVAW continues to be a major social problem worldwide and that around one in three women have suffered physical and/or sexual intimate partner violence at some point in their lives. A recent systematic review provides worldwide prevalence data of up to 27% (Sardinha et al., 2022). In Spain, in particular, the prevalence of women who have suffered some type of intimate partner violence is around 13% (Bermúdez et al., 2020; Delegación del Gobierno contra la Violencia de Género de España, 2020; Gracia et al., 2019; Sanz-Barbero et al., 2019).

Femicide refers to the intentional killing of women because they are women (WHO, 2012b). WHO (2012b) contemplates four types of femicide: ‘honor’ killing (committed by a male or female family member for a real or alleged sexual or behavioral transgression), dowry-related femicide (occurs mainly in areas of the Indian subcontinent when newly married women are killed by in-laws for dowry-related conflicts), non-intimate femicide (committed by someone who is not intimate with the victim), and intimate femicide (committed by a current or former intimate partner). The latter corresponds to intimate partner homicide against women (IPHAW).

Globally, a systematic review concludes that it is estimated that around 38% of homicides of women have been perpetrated by their intimate partner (Stöckl et al., 2013). In Spain, despite the fact that in recent years the number of IPHAW has decreased (Torrecilla et al., 2019), the Spanish Government Delegation against Gender Violence (Delegación del Gobierno contra la Violencia de Género de España, 2022) has recorded a total of 1,133 female fatalities from 2003 to March 2022. Therefore, although Spain is one of the countries with lower rates of IPHAW and IPVAW (Bermúdez et al., 2020; Torrecilla et al., 2019), the rates are still alarming.

Risk Factors of Perpetrators and Victims of IPHAW and IPVAW

The interest of scientists has been especially in analysing the risk factors of both aggressors and victims in cases of IPHAW and IPVAW in order to identify the determinants of both phenomena (e.g., Aguilar Ruiz & Calderón, 2021; Esteves-Pereira et al., 2020; González-Álvarez et al., 2022; Hellmuth & McNulty, 2008; Matias et al., 2021; Pinto et al., 2021; Sabri et al., 2021; Sebire, 2017; Spencer & Stith, 2020; Vignola-Lévesque & Léveillée, 2021). However, although it previously seemed to be asserted that IPHAW is the most serious form of IPVAW, some studies conclude that in some cases they appear to be distinct phenomena to some extent, since in some cases of IPHAW no signs of IPVAW have been seen (there did not appear to be prior violence in the relationship), and the risk factors may be different. Thus, these are therefore two complex and multidimensional phenomena (Contreras-Taibo, 2014; López-Ossorio et al., 2018; Vignola-Lévesque & Léveillée, 2021).

On the one hand, studying the risk factors for male perpetration, both in cases of IPHAW and IPVAW, is crucial to prevent future cases and to design effective reintegration treatments (Crane & Easton, 2017; López-Ossorio et al., 2020; López-Ossorio et al., 2019; Vignola-Lévesque & Léveillée, 2021). Therefore, there are already quite a few studies that hae delved into the topic (e.g., Aguilar Ruiz & Calderón, 2021; González-Álvarez et al., 2022; Sebire, 2017; Spencer & Stith, 2020; Vignola-Lévesque & Léveillée, 2021) and some authors conclude that perpetrators do not appear to constitute a homogeneous group (López-Ossorio et al., 2018; Vignola-Lévesque & Léveillée, 2021).

For aggression in IPHAW cases, risk factors include access and threats with weapons, prior attempted strangulation, and sexual assault, controlling behaviors, threats to cause harm, abusing the pregnant victim, jealousy, stalking, substance abuse, having a primary school education and a psychopathological history, not having a job, being white and young, having children, being in a stable relationship with the victim, and having prior convictions. For aggression in IPVAW cases, risk factors would be the greater severity of the abuse, previous sexual assault or abuse, degrading treatment, aggression against other family members, threats of death and of harming the victim’s children, jealous or controlling behaviors, abuse of previous partners, other criminal records, and drug and/or alcohol abuse (e.g., Aguilar Ruiz & Calderón, 2021; Sebire, 2017; Spencer & Stith, 2020).On the other hand, for the same reason as in the case of aggressors, it is also scientifically and psychologically relevant to analyse the profile and risk factors of women victims of IPHAW and IPVAW, which would also facilitate police work by allowing a much more individualized risk prediction that, in turn, would facilitate the protection of reporting victims and the avoidance of a possible femicide (Aguilar Ruiz & Calderón, 2021; Esteves-Pereira et al., 2020; Matias et al., 2021; Pinto et al., 2021; Puente-Martínez et al., 2016; Sabri et al., 2021; Santos-Hermoso, González-Álvarez, Alcázar-Córcoles, et al., 2022; Sebire, 2017; Spencer & Stith, 2020; Taşkale & Soygüt, 2017).

In relation to women in fatal cases (IPHAW), a recent international meta-analysis (Spencer & Stith, 2020) concluded that substance abuse, having primary education, separation with the aggressor, and having children from a previous relationship increased the risk for female victimization. Other studies carried out in different countries that have expanded these investigations have found some additional factors in this type of women (victims of femicides) (Matias et al., 2021; Pinto et al., 2021; Sabri et al., 2021; Sebire, 2017). One of them highlights being young, not being pregnant, not being educated, not having white skin, having a disability, living in a rural or peri-urban area and in small municipalities, that the events do not occur in the home, that physical violence predominates (alone or together with other types of violence), and the use of weapons or sharp or blunt objects (as opposed to threats) (Pinto et al., 2021). Others add the absence (or very low percentages) of substance abuse and/or psychiatric history, low risk perception, having a profession, being in a formal relationship, predominating psychological violence, the presence of harassing and controlling behaviours, and filing a prior complaint (Matias et al., 2021; Sebire, 2017). In addition to most of these factors, some of these studies, however, indicate white skin as a risk factor (Sabri et al., 2021; Sebire, 2017).

Regarding the profile and characteristics of (non-fatal) victims of IPVAW, a recent international systematic review (Esteves-Pereira et al., 2020) concluded that having experienced violence during childhood, being economically dependent on the aggressor, lacking social support, and fearing for their lives are risk factors for victimization. In turn, these victims seem to present family and social isolation, low self-esteem, feelings of insecurity and inferiority, submissiveness, and pacification, in addition to having higher scores on the schizoid, schizotypal, avoidant, self-destructive, paranoid, and borderline personality scales.

Other studies (and systematic reviews of less recent studies) add having children, reciprocal violence toward the partner, depression, male cultures, and lack of democratization of the state, paternal parenting style and fathers with less than high school education, minimizing or justifying abuse, fear for their physical integrity, drug and/or alcohol abuse, and pregnancy or recent postpartum (Aguilar Ruiz & Calderón, 2021; Puente-Martínez et al., 2016; Taşkale & Soygüt, 2017; Yakubovich et al., 2018). Being in a new relationship, the existence of previous complaints, being older and being married seem to decrease the risk (Aguilar Ruiz & Calderón, 2021; Yakubovich et al., 2018).

Typology of Perpetrators and Victims of IPHAW and IPVAW

Holtzworth-Munroe and Stuart (1994) were among the first to propose a typology of aggressors using rational-deductive and empirical-inductive strategies. In their review, they proposed three subtypes of perpetrators: family only, dysphoric/borderline, and generally violent/antisocial, and concluded that perpetrators constitute a heterogeneous group. This study served as a theoretical frame of reference for other studies (e.g., González-Álvarez et al., 2022; Holtzworth-Munroe et al., 2000).

A recent study has delved further into these differences and sought to identify typologies of perpetrators (Vignola-Lévesque & Léveillée, 2021). To do so, it collected data from 67 male perpetrators of IPHAW and IPVAW and obtained a solution of four profiles: the homicidal abandoned partner (19.4%; most had committed IPHAW, experienced relationship breakdown, and had a history of self-destructive behaviours), the generally angry/aggressive partner (23.9%; most were IPVAW perpetrators, with alexithymia, and with a criminal history), the controlling violent partner (34.3%; committed IPHAW and had a criminal lifestyle), and the unstable dependent partner (22.4%; were IPVAW perpetrators and with alexithymia, but without criminal history).

Other studies focused on Spanish samples have also tried to identify typologies. One of them, focused on IPVAW offenders, obtained four groups: offenders with high instability and low antisociality, offenders with high instability and antisociality, offenders with low instability and high antisociality, and offenders with low instability and antisociality (González-Álvarez et al., 2022). The other study, focused on IPHAW cases, also obtained four groups: normalized, violent, pathological, and pathological/violent IPHAW (Santos-Hermoso, González-Álvarez, Alcázar-Córcoles, et al., 2022).

These studies emphasize that the importance of knowing these typologies lies in the fact that more specific and precise procedures and instruments can be designed for risk assessment in both IPHAW and IPVAW, in addition to helping in prevention and in the design of more individualized treatments (López-Ossorio et al., 2018; Vignola-Lévesque & Léveillée, 2021). However, some studies have concluded just the opposite in their analyses, noting that the characteristics that differentiate cases of IPHAW from cases of IPVAW are minimal (Jung & Stewart, 2019).

As noted above, there are several studies that have attempted to identify typologies of aggressors (with IPHAW and IPVAW samples, or with IPHAW or IPVAW samples only) (Dawson & Piscitelli, 2021; Dixon et al., 2008; González-Álvarez et al., 2022; Kivisto, 2015; Santos-Hermoso, González-Álvarez, Alcázar-Córcoles, et al., 2022; Vignola-Lévesque & Léveillée, 2021). In contrast, no studies have been identified that have attempted to identify victim typologies. Only studies with an approximate approach have been identified, which have tried to identify typologies of victims according to patterns of psychological abuse, emotional regulation, or symptoms of post-traumatic stress disorder (Hebenstreit et al., 2015; Marshall, 1996; Muñoz-Rivas et al., 2021).

Given the inconsistency in the literature regarding whether or not there are differences between perpetrators and victims of IPHAW and IPVAW cases (Jung & Stewart, 2019; López-Ossorio et al., 2018; Vignola-Lévesque & Léveillée, 2021), and taking into consideration that not all of the risk factors noted above between both types of victimization (i.e., IPHAW and IPVAW victims) are the same (Aguilar Ruiz & Calderón, 2021; Esteves-Pereira et al., 2020; Matias et al., 2021; Pinto et al., 2021; Puente-Martínez et al., 2016; Sabri et al., 2021; Sebire, 2017; Spencer & Stith, 2020; Taşkale & Soygüt, 2017), it is important to also explore victim typologies to design more accurate risk assessments and design more specific treatments (Esteves-Pereira et al., 2020; Spencer & Stith, 2020).

The Present Study

IPHAW and IPVAW continue to be problems with a great impact on society. For this reason, the authors have endeavoured to study the profiles and risk factors of both phenomena to design more precise instruments for assessment, prevent future cases, and design more individualized treatments (Crane & Easton, 2017; López-Ossorio et al., 2018; López-Ossorio et al., 2020; López-Ossorio et al., 2019; Vignola-Lévesque & Léveillée, 2021). In turn, the study of profiles and risk factors can also facilitate police work by allowing more accurate risk prediction and, therefore, greater protection for victims and the avoidance of possible femicide (Aguilar Ruiz & Calderón, 2021; Esteves-Pereira et al., 2020; Matias et al., 2021; Pinto et al., 2021; Puente-Martínez et al., 2016; Sabri et al., 2021; Santos-Hermoso, González-Álvarez, Alcázar-Córcoles, et al., 2022; Sebire, 2017; Spencer & Stith, 2020; Taşkale & Soygüt, 2017).

Therefore, given the importance of the analysis of profiles and risk factors for victimization, and considering the lack of studies that have tried to identify typologies of women, the main objective of this study was to identify typologies of Spanish victims of IPHAW and IPVAW. To this end, differences between the risk factors of the victims in both cases (i.e., between the sociodemographic variables, their personality and lifestyles, the psychosocial factors, stressors, and suicide factors involved in each case) were first analysed. Once the typology of victims was obtained, the differences and similarities between the profiles obtained and the determinants of aggression (perpetrator variables) were analysed, which in turn allowed us to interpret the main characteristics of the aggression. To achieve the objectives, the risk factors measured with the Valuation of Police Risk (VPR) tool, the Spanish protocol used to predict and manage the risk of IPVAW and IPHAW cases, were used (López-Ossorio et al., 2020).

Based on previous literature, we expect to find differences between the risk factors of victims of IPHAW and IPVAW cases (H1). However, since this is the first study that aims to identify victimization typologies of both cases (i.e., IPHAW and IPVAW), a hypothesis cannot be established based on previous research. In any case, taking the studies that have tried to identify typologies of aggressors (Spanish and from other countries; González-Álvarez et al., 2022; Santos-Hermoso, González-Álvarez, Alcázar-Córcoles, et al., 2022; Vignola-Lévesque & Léveillée, 2021) as a reference, we expect to obtain the same number of profiles, i.e., four (H2). Finally, once the different profiles have been obtained, we expect to find differences between them in terms of the variables that determine aggression in cases of IPHAW and IPVAW (the perpetrator variables), which in turn will allow us to interpret the main characteristics of aggression (H3).

Method

Participants

The sample consisted of 381 cases, of which 171 were IPHAW cases and 210 were IPVAW cases, all extracted from the Integral Monitoring System in Cases of Gender Violence (VioGén System, a computer application created in the Secretary of State for Security of the Spanish Ministry of Interior, which regulates data files, i.e., it contains and manages the country’s IPHAW and IPVAW cases; González-Álvarez, Ossorio, et al., 2018; López-Gutiérrez, 2021). Thus, 381 women with a mean age of 38 (SD = 13.52, range 13-84 years) and 381 men with a mean age of 41 (SD = 13.79, range 18-89 years) participated in the study. In cases of femicides, given that the victims could not participate directly, the information was extracted from people close to the victims (minimum 10 per case), such as friends, children, or relatives (especially from the latter).

IPVAW victims’ group was made up of 102 cases in which the perpetrator was in a Social Integration Centre (CIS) (cases classified as less severe), 99 cases in which the perpetrator was serving a sentence in a penitentiary centre (cases classified as serious), and 9 cases in which there was an attempt homicide and serving in a penitentiary centre too.

Measures

The instrument used to obtain the data of interest was a semi-structured interview designed by the National Team for the Detailed Review of Gender Homicides of the Secretary of State for Security of Spanish Ministry of the Interior (González-Álvarez, Garrido, et al., 2018; Santos-Hermoso, González-Álvarez, Alcázar-Córcoles, et al., 2022; Santos-Hermoso, González-Álvarez, López-Ossorio, et al., 2022; Santos-Hermoso et al., 2021). Two interview models were used, but with the same type of questions, one for women victims of IPVAW and another for people close to the victims of IPHAW to elaborate a psychological autopsy (research methodology to improve knowledge in cases of IPHAW; McPhedran et al., 2022).

With the aim of ensure inter-judge agreement (i.e., concordance between reviewers when filtering the documentation needed for the study), the team also designed a dictionary of variables (after an exhaustive review of the literature) and a template that facilitated the coding of the different variables and avoided discordance between experts when it came to understanding each variable. In total, the interview collected 45 variables of interest for this study (from the VPR tool; López-Ossorio et al., 2020), referring to socio-demographic (12 variables; e.g., “At the time of the events, what was your work situation?”), personality, and lifestyle factors (8 variables; e.g., “How do you usually react to conflicts, unforeseen events, frustrations?” or “What do you like to do in your free time?”), psychosocial risk (13 variables; e.g., “At the time, did you have someone to support you?”), stressors (10 variables; e.g., “How and when do relationship problems start?”), and suicide factors (2 variables; e.g., “At the time, did you identify reasons for living?” “Which ones?”).

All variables were coded as 0 when not present in the person (e.g., no alcohol consumption) and as 1 when present (e.g., feelings of loneliness), except for the three personality traits (extraversion, neuroticism, and psychoticism) which were coded as 0 = low and 1 = high, and risk perception which was coded as 0 = low, 1 = medium, and 2 = high. Of the sociodemographic variables, age was coded as 0 = older and 1 = younger, nationality as 0 = foreign and 1 = Spanish, socioeconomic level as 0 = medium high, 1 = low, and 2 = very low, and educational level as 0 = high school or university, 1 = secondary school or vocational training, 2 = high school and 3 = no studies.

Procedure

Firstly, a national working group was formed, made up of different criminal analysts from different Spanish universities, with which both the Secretary of State for Security (who approved the project as meeting the ethical requirements) and the General Secretariat of Penitentiary Institutions of the Ministry of the Interior have collaboration agreements.

Secondly, for each case a detailed review was carried out of the police, judicial and penitentiary documentation, and of that stored in the VioGén System, to gather all the available information that would be useful for the project.

Thirdly, interviews were conducted with the perpetrators and victims of IPHAW and IPVAW cases in different Spanish territories between 2015 and 2021 (all participants gave written informed consent). In cases where it was not possible to interview the victim (cases of intimate partner homicide), between 10 and 15 people close to the victim, aggressor, and both were interviewed for each case. The experts then transferred the information collected in the interviews to the automated template, and a group of field monitors (Coordination and Studies Office of the Secretary of State for Security) checked that data from the interviews and templates coincided. All reviewers were previously trained in indirect profiling to ensure the reliability of the information collected in the templates (Muñoz-Espinosa & Santos-Hermoso, 2020; Sotoca et al., 2019). For this purpose, the profiling was based on Eysenck’s PEN Model of the Big Three personality traits (Espinosa & García-Rodríguez, 2004; Eysenck & Eysenck, 1985) (psychoticism, extraversion, and neuroticism), the model that has been used in all the studies carried out with samples from the VioGén system to measure personality. Finally, all the variables collected in the template were anonymized and transferred to a database for statistical analysis.

Data Analysis

First, to obtain a detailed description of the characteristics of the sample and to find the main differences between victims of IPHAW and victims of IPVAW (H1), descriptive statistics were calculated using contingency tables. For this purpose, the different sample sizes (and their percentages) in both samples were found for the 45 different variables of interest: sociodemographic, personality and lifestyle factors, psychosocial risk factors, stress factors, and suicide factors. To find statistical differences, chi-square, phi coefficient (in the case of 2 x 2 data matrices) or Cramer’s V (in the case of 2 x 3 data matrices), and odds ratios (to find the risk determinants for the IPHAW group and the IPVAW group) were calculated.

Second, to identify the typologies of IPHAW and IPVAW victims (H2), a latent class analysis (LCA) was performed considering the variables of interest (i.e., personality and lifestyle factors, psychosocial risk factors, stress factors, and suicide factors) those with a statistical significance level of p < .05 obtained in the previous analyses (i.e., with a phi coefficient or a Cramer’s V with significance of p < .05).

In this sense, 12 variables were introduced to obtain the profiles, but after finding the graphs with the different solutions (from 1 to 4 profiles) it was observed that two of the variables hardly contributed to the model, so it was considered appropriate not to consider them in the analysis (these were the separation process with the perpetrator and being pregnant). Therefore, 10 variables were used to identify the victim typologies: neuroticism, psychoticism, alcohol abuse, isolation, feelings of loneliness, previous reconciliations with the perpetrator and/or withdrawal of complaints, high risk perception, loss of a loved one, carer stressor, and suicidal ideation.

Once all the variables of interest were identified, different fit indices were calculated from one class to four classes to determine the most optimal profile model. These indices were the Akaike information criterion (AIC), the Bayesian information criterion (BIC), the sample size adjusted BIC (SSA-BIC), the entropy, the Vuong-Lo-Mendel-Rubin test (VLMR), the adjusted likelihood ratio test (adjusted LRT), and the parametric bootstrapped LRT. The combination of all these indices determines the most optimal model by considering the significance of the p-values of the VLMR, the adjusted LRT and the parametric bootstrapped LRT, a value as close to 1 for entropy, and small values of AIC, BIC, and SSA-BIC with the largest number of profiles. In turn, the elbow graph also serves to determine the best solution, so it was calculated considering the values extracted from the AIC, BIC, and SSA-BIC indices. The statistical program Mplus (version 8.7) was used to perform all these analyses.

After obtaining the most optimal number of profiles, descriptive statistics were calculated using contingency tables to obtain the differences between the profiles found and the type of case, i.e., IPHAW group or IPVAW group. The main purpose was to determine the number of IPHAW and IPVAW victims in each of the profiles. A linear regression analysis was also carried out to analyse whether the type of case (IPHAW and IPVAW) predicted the profiles obtained.

Finally, to test H3 and to be able to compare the profiles found with the determinants of aggression, descriptive statistics were calculated using contingency tables. All the variables that determine aggression in cases of IPHAW and IPVAW were introduced, which are the same as those that determine victimization, used to test H1 (except for being pregnant, risk perception, and previous reconciliations). Only those that were significant are presented in the results tables. For this purpose, IBM SPSS statistical software (version 23) was used again.

Results

Determinants of Victimization in Cases of IPHAW and IPVAW

The difference between the determinants of victimization in cases of IPHAW and in cases of IPVAW are presented in Table 1. First, in relation to sociodemographic determinants, the models showed that female IPVAW victims are more likely to be young and Spanish compared to female IPHAW victims (OR = 0.45 and 0.43, p < .001 and p < .01), while female IPHAW victims are more likely to have a very low socioeconomic and educational level (OR = 1.50 and 1.53, p < .01), with effect sizes below .001 and .01 (phi coefficient and Cramer’s V).

Second, in relation to personality and lifestyle, the models showed that having neurotic traits was related to higher odds of IPVAW (OR = 0.63, p < .05), whereas having traits of psychoticism and abusing alcohol were related to higher odds of IPHAW (OR = 2.04 and 1.65, p < .05), with effect sizes less than .05 (phi coefficient).

Third, with respect to psychosocial risk factors and with effect sizes less than .001, .01, and .05 (phi coefficient), female victims of IPVAW were more likely to feel isolated, have feelings of loneliness, be pregnant, have previous reconciliations with the offender or withdrawals of allegations, and have a perception of high risk compared to female victims of IPHAW (OR = 0.19-0.62, p < .001, p < .01, and p < .05).

Fourth, in relation to stressors factors, losing a loved one and having to take care of someone were related to higher odds of IPVAW (OR = 0.31 and 0.33, p < .01), whereas being in a separation process with the perpetrator was more related to higher odds of IPHAW (OR = 1.76, p < .05), with effect sizes less than .01 and .05 (phi coefficient).

Finally, with respect to the suicide factors and with effects sizes less than .05 (phi coefficient), IPVAW victims were more likely to have suicidal ideation (OR = 0.37, p < .05).

Latent Class Analysis with IPHAW and IPVAW

The results of the LCA from the ten variables of interest (the victimization risk factors, i.e., neuroticism, psychoticism, alcohol abuse, isolation, feelings of loneliness, previous reconciliations with the perpetrator and/or withdrawal of complaints, high risk perception, loss of a loved one, carer stressor, and suicidal ideation) for the solutions from one to four profiles are presented in Table 2. After analysing the best combination of these, the different fit indices, and the Elbow Graph (Figure 1), the 3-profile solution was considered the most optimal.

Figure 2 shows the profiles of the solution considered most optimal, i.e., the one that offers three latent classes, as well as the percentages of people in each of the profiles. As can be seen, Class 1 represents victims characterized by low neuroticism, low isolation and feelings of loneliness, less reconciliation with the offender, lower risk perception, and low suicidal ideation. In general, it is the class with the fewest characteristics of all those analysed. Class 2 represents victims characterized by presenting the loss of a loved one and the role of caregiver as stressors, low psychoticism and alcohol abuse, high feelings of loneliness, high risk perception and suicidal ideation. Finally, Class 3 represents victims characterized by high neuroticism and psychoticism, alcohol abuse, isolation, and higher reconciliations with the offender, and by the absence of bereavement and caregiving as stressors.

Table 3 shows the comparison between the three classes obtained in the LCA and the group of IPHAW and the group of IPVAW. As can be seen, Class 1 (fatal victims, hereinafter) is mostly represented by IPHAW victims (58.70%) while Class 2 (non-fatal victims, hereinafter) is mostly represented by IPVAW victims (75.50%). Class 3 (mixed, hereinafter) appears to be represented by both groups. In addition, to analyze if indeed the obtained classes were predicted by the type of case (IPHAW and IPVAW), the regression analysis showed significant results, with an R² of .04 (p < .001) and a β of -.20, confirming such an association.

Differences between Profiles and the Determinants ofAggression in Cases of IPHAW and IPVAW

Table 4 shows the comparison between the three classes obtained in the LCA and the significant determinants of aggression. As can be seen, there are significant differences (p < .05) between fatal victims’ profile and non-fatal victims’ profile in terms of socioeconomic level, drug use, and optimism (identifying reasons to live), which could be interpreted as it is more likely that when the victim is not murdered (non-fatal victims’ profile) the aggressor will have a low socioeconomic status, use drugs, and identify reasons to live.

Significant differences (p < .01 and p < .05) are also observed between fatal victims’ profile and mixed profile in terms of cohabitation, alcohol consumption, criminal history, criminal subculture, separation process with the victim, problems with crime and drugs and, again, socioeconomic level. In this case, it could be interpreted as meaning that it is more likely that when the victim is murdered (fatal victims’ profile) the aggressor will not live with her or will be in the process of separation, will not have a criminal record or live with a criminal subculture, will not have problems with crime or drugs, and will have a medium-high socioeconomic level.

Between non-fatal victims’ profile and mixed profile, significant differences (p < .01) were only observed in terms of alcohol consumption, which could indicate that it is more likely that when the victim is not murdered (non-fatal victims’ profile), the aggressor will not consume alcohol. Finally, there are also differences (p < .05) in age within non-fatal victims’ profile, which could indicate that it is more likely that when victims are not murdered the aggressor will be younger, and in the recent loss of employment of the perpetrator within fatal victims’ profile, which could indicate that it is more likely that when the victim is murdered the aggressor will not have this loss.

Discussion

Given the inconsistency in the literature regarding whether or not there are differences between perpetrators and victims of IPHAW and IPVAW cases (e.g., Jung & Stewart, 2019; López-Ossorio et al., 2018), the main objective of this study was to identify typologies of Spanish victims of IPHAW and IPVAW. The first step was to analyse the differences between the victims in the two cases in terms of their sociodemographic characteristics, their personality and lifestyles, and the psychosocial risk factors, stressors, and suicide factors involved in each case.

On the one hand, the results obtained in this study show that women victims of IPVAW are more likely to be young, Spanish, have traits of neuroticism, feel isolated, and present feelings of loneliness, be pregnant, present previous reconciliations with the aggressor or withdrawals of complaints, have a perception of high risk, have lost a loved one, have to take care of someone, and present suicidal ideation. On the other hand, female victims of IPHAW are more likely to have a very low socioeconomic and educational level, present traits of psychoticism and alcohol abuse, and be in the process of separating from the aggressor.

Therefore, H1 is confirmed, since different risk factors have been found depending on the case (IPHAW or IPVAW), which coincides with the results found in previous systematic reviews, meta-analyses, and primary studies (Aguilar Ruiz & Calderón, 2021; Esteves-Pereira et al., 2020; Matias et al., 2021; Pinto et al., 2021; Puente-Martínez et al., 2016; Sebire, 2017; Spencer & Stith, 2020; Taşkale & Soygüt, 2017; Yakubovich et al., 2018). Knowing all these characteristics that differentiate IPHAW and IPVAW victims can help distinguish which cases need greater protection. In other words, being able to distinguish between the risk factors that differentiate IPHAW victimization from IPVAW victimization can facilitate the work of preventing future cases, since when a woman reports a case all of these characteristics can be considered and protective resources can be targeted to a greater or lesser extent (Santos-Hermoso, González-Álvarez, Alcázar-Córcoles, et al., 2022).

These results also allow us to compare the main characteristics of Spanish IPHAW and IPVAW victims with those of victims from other countries. As for IPHAW victims, some of the risk factors obtained in this study also seem to be present in victims from other countries, i.e., having a low level of education, presenting substance abuse, and being in the process of separation with the perpetrator. On the contrary, low socioeconomic status and psychotic traits do not seem to be risk factors for victimization in other countries (Matias et al., 2021; Pinto et al., 2021; Sabri et al., 2021; Sebire, 2017; Spencer & Stith, 2020). Regarding IPVAW victims, there are also some risk factors obtained in this study that coincide with those obtained in studies with victims from other countries, that is, being young, presenting isolation and feelings of loneliness, being pregnant, and having a high perception of risk. Other factors, such as neuroticism, previous reconciliations and withdrawal of complaints, and suicidal ideation, seem to be risk factors in Spanish victims, but other risk factors along the same lines, such as schizoid, schizotypal, avoidant, self-destructive, paranoid, or borderline personality, and justification of the abuse, are present in victims from other countries (Esteves-Pereira et al., 2020; Puente-Martínez et al., 2016; Taşkale & Soygüt, 2017; Yakubovich et al., 2018).

It is also interesting to be able to compare the factors associated with the victims, obtained in this study, with the factors associated with the aggressors, obtained in previous studies. Thus, having a low level of education, having problems with substances, and being in the process of separation are characteristics that aggressor and victim seem to share in cases of IPHAW. However, previous studies conclude that other risk factors, such as having children and having a psychopathological and criminal history, are common among IPHAW offenders, and factors, such as substance abuse and criminal history, are common among IPVAW offenders. In contrast, this study has not concluded that these factors are characteristic of either victims (IPHAW or IPVAW) (Aguilar Ruiz & Calderón, 2021; Dobash et al., 2009; Eriksson et al., 2022; Sebire, 2017; Spencer & Stith, 2020).

Once the differences between the factors most strongly associated with some cases or others were found, the interest lay in identifying the IPHAW and IPVAW typologies. LCA makes it possible to identify qualitatively different subgroups. Thus, unlike other more classical statistical methods, LCA analysis makes it possible to group individuals into latent classes and to analyze the differences between the classes according to the variables of interest with more powerful analyses. Knowing this, this study set out to analyze IPHAW and IPVAW profiles and to analyze possible differences with this statistical method, which allows us to obtain more adjusted results (Nylund-Gibson & Choi, 2018; Weller et al., 2020).

The LCA showed a solution of three victim profiles, different from what was expected (H2 is rejected; Vignola-Lévesque & Léveillée, 2021): one profile represented mostly by women victims of IPHAW (fatal victims) and characterized by presenting low neuroticism, low isolation and feeling of loneliness, lower reconciliations with the aggressor, lower risk perception, and low suicidal ideation; another by women victims of IPVAW (non-fatal victims) and characterized by presenting loss of a loved one and the role of caregiver as stressors, low psychoticism and alcohol abuse, high feeling of loneliness, high risk perception and suicidal ideation; and another by women from both groups (mixed profile) and characterized by presenting high neuroticism and psychoticism, alcohol abuse, isolation, and higher reconciliations with the offender, and by the absence of grief and caregiving as stressors.

There are no previous studies that have attempted to identify typologies of victims of IPHAW and IPVAW cases, but these results indicate, as has been reported in other studies with different approaches and different statistical analyses, that there may be differences in the characteristics and risk factors of these victims (Aguilar Ruiz & Calderón, 2021; Esteves-Pereira et al., 2020; Matias et al., 2021; Pinto et al., 2021; Puente-Martínez et al., 2016; Sabri et al., 2021; Sebire, 2017; Spencer & Stith, 2020; Taşkale & Soygüt, 2017; Yakubovich et al., 2018).

These differences are also reinforced by the results found in the last analysis made in this work, since differences were found between the three victim profiles obtained and the determinants of aggression. Therefore, H3 is confirmed, agreeing with the results of previous literature that have stated that there are also differences between aggressors in IPHAW and IPVAW cases (Aguilar Ruiz & Calderón, 2021; González-Álvarez et al., 2022; López-Ossorio et al., 2018; López-Ossorio et al., 2020; Sebire, 2017; Spencer & Stith, 2020; Vignola-Lévesque & Léveillée, 2021). It is also interesting to note the similarities between previous findings and the results obtained in this study in terms of the differences between the profile of fatal and non-fatal victims in terms of the determinants of aggression. Thus, previous research has indicated that substance use is a risk factor for assault in IPVWA and IPHAW cases and the results of this study have exposed that the offender is more likely to use drugs when the victim is not killed than when the victim is killed (Aguilar Ruiz & Calderón, 2021; Sebire, 2017; Spencer & Stith, 2020).

Attending again to the victimization classes obtained in this study, mixed profile may be somewhat more confusing when classifying victims in this group, as there are both IPHAW and IPVAW victims. However, in the case of the profile of fatal victims and non-fatal victims, knowing that the former is mostly represented by female IPHAW victims and the latter by female IPVAW victims (hence the choice of names), it can be very useful to be able to classify victims according to the characteristics and risk factors they present, as it will be easier to predict when a woman will be more at risk of being killed by her partner and, therefore, to allocate the relevant resources and to be able to prevent future cases (González-Álvarez et al., 2022; Koppa & Messing, 2021; López-Ossorio et al., 2018; Vignola-Lévesque & Léveillée, 2021).

In addition, considering the determinants of aggression analyzed in this study, it may also be easier to identify these victims according to these variables of aggression. From the differences found in this study, attending to those found between fatal victims’ profile and non-fatal victims’ profile, i.e., socioeconomic status, drug use, and the identification of motives for living by the perpetrator, may again guide the classification of the victim into one typology or another. Specifically, if the victim belongs to non-fatal victims profile, it is more likely that the aggressor has a low socioeconomic level, uses drugs, and identifies motives for living, which again affirms that the characteristics of the aggressors may also be different depending on whether it is a case of IPHAW or IPVAW (González-Álvarez et al., 2022; López-Ossorio et al., 2019; Vignola-Lévesque & Léveillée, 2021). In short, all of this could facilitate police and judicial work in preventive work (Crane & Easton, 2017; López-Ossorio et al., 2020; López-Ossorio et al., 2019; Vignola-Lévesque & Léveillée, 2021). Therefore, there are already quite a few studies that have delved into the topic and have also tried to identify the differences between IPHAW and IPVAW offenders (e.g., Aguilar Ruiz & Calderón, 2021; González-Álvarez et al., 2022; Sebire, 2017; Spencer & Stith, 2020; Vignola-Lévesque & Léveillée, 2021).

Limitations and Future Research Directions

Among the main limitations of the study is the difficulty in generalizing the results found in this study due to the sample size. Also noteworthy is the difficulty in generalizing the results to an international level, since although the similarities between the risk factors of Spanish victims and victims from other countries have been presented, some factors were not coincidental and cross-cultural differences in the profile of IPHAW and IPVAW victims could result in different victim typologies. Future research should replicate the results with a larger sample to determine whether the results are consistent. It would also be interesting to replicate this study in other countries and analyze cross-cultural differences.

Another limitation related to the sample is the fact that women victims of IPVAW who participated in this study were women who had suffered violence by men who were or had been serving a sentence in the CIS or in a penitentiary center, so that in the distribution of the groups there were both victims of more severe cases and victims of milder cases (in addition to the fatal victims). Therefore, another interesting future line of research would be to analyze the differences between the non-fatal victims of more severe cases and milder cases and, in turn, to analyze these differences with fatal victims.

A final limitation to be highlighted is related to the method of data collection in the case of fatal victims. Unlike in the case of non-fatal victims, the information was collected by means of psychological autopsy, which, although it has proven to be a reliable method, does not allow data to be collected directly from the target persons (McPhedran et al., 2022).

Conclusions

All the results found in this study highlight the importance of considering that both IPHAW and IPVAW are complex and multidimensional phenomena, and knowing the differences between both cases, however minimal they may be, allows for the design of more specific and precise procedures and instruments for risk assessment in both IPHAW and IPVAW, in addition to helping in prevention and in the design of more individualized treatments (Koppa & Messing, 2021; López-Ossorio et al., 2018; Vignola-Lévesque & Léveillée, 2021).

All of this also facilitates police work in identifying IPHAW and IPVAW victims and being able to immediately deploy the most intense police and judicial protection measures, as in the case of aggressors. The main advantage of knowing these differences is to be able to distinguish the victims as a preventive method and, with this, to be able to target the most appropriate measures and resources (González-Álvarez et al., 2022). But all this not only facilitates the detection of risk factors from the police point of view, since other agents (such as health, social services or even people close to the victims and aggressors) may also be able to detect them and bring them to the attention of the authorities when the victims have not yet reported the aggressor (Santos-Hermoso, González-Álvarez, Alcázar-Córcoles, et al., 2022).

In sum, knowing that one group is represented mostly by women victims of IPHAW and another group by women victims of IPVAW, the results of this study can enable the prevention of future cases by differentiating the different risk factors in both victimization cases, which will allow the design of much more accurate risk assessments. In other words, when new reports arrive it may be easier to classify the victim after analyzing her characteristics and risk factors for her case, which in turn will make it possible to allocate greater resources when there is suspicion that a femicide could occur (González-Álvarez et al., 2022; Santos-Hermoso, González-Álvarez, Alcázar-Córcoles, et al., 2022).

In any case, given the complexity of the phenomenon (IPHAW and IPHVW), these results should be interpreted as merely indicative when it comes to facilitating police work, with analyzing each case individually and particularly being important because, as has been repeated throughout this study, victims and aggressors of IPHAW and IPVAW do not represent fully homogeneous groups.