Introduction

Femicide is a persistent public health and social issue with cultural, social, psychological, and institutional factors playing a role in its perpetration [1, 2]. Despite international efforts to confront such violence, the gender-based killing of women and girls remains a global challenge [3]. To date, two sectors have most frequently examined intimate partner femicide (IPF)—the most prevalent form of femicide [3]. Public health researchers have framed IPF as an extreme form of intimate partner violence comparing femicide perpetrators to perpetrators of non-lethal intimate partner violence. Meanwhile, criminologists examine IPF as a form of homicide comparing femicides to other forms of homicide or serious crime. Yet, as discussed by feminist scholars, femicide cannot be reduced to an individual-level action and instead requires systems-level intervention [4–6].

IPF is often portrayed as being perpetrated by individual bad actors whose acts are rare, likely because the crime is perceived as being socially egregious [7–10]. Yet closer examination of homicide data reveals persistent statistical trends: women are disproportionately killed by their intimate partners. One in every three female homicides is perpetrated by a current or former intimate partner accounting for roughly 48,000 deaths per year [2, 3]. In Latin America, reported femicide rates have remained relatively constant throughout the last decade, in spite of the enactment of femicide-specific laws in several countries of the region [3, 11]. In this context, the interest in understanding if perpetrators have specific features has gained recent attention [12–14]. Here we focus on IPF, the gender-based killing of women or girls by their past or current intimate partners.

Because there is significant overlap in the risk factors for lethal and non-lethal intimate partner violence perpetration, it remains difficult to distinguish the distinctive features of IPF—a specific subtype of femicide—and by extension the men who perpetrate it. Most work has focused on the development and validation of lethality risk assessments based on the observations of female partners or law enforcement actors in the context of non-lethal intimate partner violence. Similarly, other scholars have tried to differentiate between IPF perpetrators and perpetrators of non-lethal intimate partner violence [15]. When compared to men who used non-lethal violence against an intimate partner, Dobash, Dobash and Medina-Ariz found that IPF perpetrators were more likely to use a weapon, strangle, or sexually assault their partners [16]. While they were also more likely to have been violent in a prior relationship, they were less likely to have previously used violence against the woman they killed. They were also less likely to have been drunk at the time of the femicide [16]. Goussinsky found that lethal violence is usually a planned or premeditated event, motivated by despair and a desire to destroy a partner [15]. The feeling of being threatened and belittled was a common emotion reported by IPF perpetrators across several Latin American countries [17]. These studies suggest that the behaviors of IPF perpetrators may be distinguished in some ways from men who use non-lethal intimate partner violence. Yet the data on the mental health status of these men is mixed.

Based on a comparison of IPF perpetrators and men who do not exercise violence against women in Turkey, Tosun Altınöz et al. found that perpetrators were not significantly different in terms of individual characteristics (i.e., childhood traumas, psychopathology, and gender attitudes), however, migration history was linked to femicide [18]. As understood by the authors, migration not only implies a change of location, but a radical social change and a defense of male-chauvinistic views, as found by other studies [19, 20]. This study also underscores that a unique psychopathology for femicide perpetration may not exist.

Others, most often criminologists, have compared femicides to non-partner homicides or perpetrators of other crimes. Much of the femicide perpetration research has focused on demographic characteristics and criminality, with medical pathologization occurring frequently [21]. While the demographic, socioeconomic, and criminogenic heterogeneity of male batterers has been highlighted, the degree to which IPF perpetrators suffer from poor mental health still needs further scrutiny especially given that many of these homicides are followed by suicide, potentially confounding the relationship between IPF and perpetrator mental health [22–24].

Because some studies compare femicide perpetrators to perpetrators of non-lethal intimate partner violence, and others compare them to perpetrators of other crimes, the data on the mental health status of IPF perpetrators is inconclusive. Depression, anxiety, posttraumatic stress disorder, antisocial personality disorder, and borderline personality disorder have all been found to be correlated with intimate partner violence perpetration [25]. While prior research has attempted to identify specific mental health diagnoses in IPF perpetrators [26], there is still no clear picture of the extent of mental health problems in the total population of perpetrators. Within the limited research on IPF perpetrators, scholars have found two types of perpetrators: those with and without antisocial features [27–29]. One study in Spain used the Psychopathy Checklist-Revised (PCL-R) to evaluate the degree of psychopathy of femicide perpetrators, finding low rates among their sample, with higher scores correlating with shorter romantic relationships as compared to relationships without lethal violence [30]. Likewise, Echeburúa & Fernández-Montalvo found that psychopathic batterers did not perpetrate IPF more often than non-psychopathic batterers, instead the crime might be attributable to other variables [31]. By contrast, in a study about intimate partner violence perpetration in Portugal, Cunha et al. found that PCL-R scores positively predicted the frequency of violence, more than other criminal variables [32]. To this end, some intimate partner violence recidivism assessment tools have included the PCL-R as part of their assessment of likely although these tools were not designed to assess femicide risk, again complicating the ability to differentiate between men likely to perpetrate lethal versus non-lethal violence.

Overall, few studies have focused on IPF perpetrator’s psychological or mental distress, and the results have varied. Belfrage and Rying compared the characteristics of IPF perpetrators to other homicide perpetrators in Sweden concluding that psychopathy is not overrepresented among IPF perpetrators; instead, depressive disorders were most common [33]. In comparing mental disorders among femicide perpetrators and male-male homicide perpetrators, Caman et al. found that, regardless of the type of homicide, one-third of the participants were diagnosed with mental disorders during their life span, yet low rates of serious or severe mental illness were found in both groups when aggregating lifetime diagnoses and diagnoses during the commission of the crime [23].

While the public and lay publications have taken for granted the fact that perpetrators have mental health problems, the evidence is far from settled. Femicide perpetration in general and particularly the mental health status and internal mental processes of femicide perpetrators remain under-researched [12]. We hypothesize that IPF perpetrators will not report experiencing higher rates of mental health problems, psychological distress, and psychopathy relative to other criminals. The hypothesis originated from our extensive qualitative fieldwork experience [17, 34, 35] and close readings of literature on femicide and homicide perpetration, including insights into the mental health, psychological distress, and psychopathy of such perpetrators. The purpose of this study was to compare levels of psychopathy, psychological distress, and treatment history among a sample of IPF perpetrators, male-male homicide perpetrators, and other offenders convicted of violent crimes in Buenos Aires, Argentina.

Methods

Results

The final sample included 205 prisoners including 68 IPF perpetrators, 73 homicide perpetrators, and 64 individuals convicted for other violent crimes. There were no significant differences across femicide, homicide, and perpetrators of other crimes based on their socio-demographic characteristics including age, educational attainment, ability to make ends meet before the crime, and prior criminal history. However, femicide perpetrators were less likely to have a prior criminal conviction relative to perpetrators of homicide and other crimes. Perpetrators of all types were in their mid-late 20s with over 90% being born in Argentina (Table 2).

The mean values of the PCL-R and LSRP were 17.6 (range = 3 to 30; SD = 6.5) and 34.1 (range = 3 to 67; SD = 14.9) respectively. The mean of GHQ-was 19.9 (range = 6 to31; 7.0 SD), and 180 (87.8%) respondents had a score above 11 for GHQ-12 likely indicating the presence of emotional disorders.

Participants did not differ in terms of their psychopathology across groups; however, femicide perpetrators had higher GHQ-12 scores meaning they were statistically more likely to experience psychological distress (Table 3). Across all participants, mental health treatment history was reported by 70 (34.1%), substance use treatment history by 80 (39.0%) participants. In addition, while femicide perpetrators self-reported previous episodes of mental (76.5%) and substance use (72.1%) treatments, these were much lower among homicide perpetrators (6.8%) or individuals convicted for other violent crimes (20.3%).

Table 3. Mental characteristics (psychopathy, psychological distress, and treatment history) of Argentinean perpetrators of femicide, homicide and other violent crimes.

https://doi.org/10.1371/journal.pmen.0000064.t003

In multivariate models, we explored whether type of crime was associated with psychopathology, psychological distress, and self-reported treatment history after accounting for relevant covariates.

Discussion

Our objective was to contribute to the academic literature regarding the individual characteristics of femicide perpetrators, by comparing their levels of psychopathy, psychological distress, and treatment history relative to perpetrators of homicide and other serious crimes in Buenos Aires, Argentina. These data may contribute towards disentangling questions about the risk factors—including mental health status—for lethal violence and upstream prevention strategies.

First, among our sample of demographically similar Latin American perpetrators of femicide, homicide, and other violent crimes we did not find significant differences in terms of their psychopathology, antisocial behavior or emotional detachment. The average values for psychopathology (PCL-R) were similar to Santos-Hermoso [30]. The lack of significant differences between the compared groups coincides with several studies in other regions [18, 30, 31, 33]. Dobash, Dobash and Medina-Ariza found IPF perpetrators do not report significantly different socio-demographic or mental health characteristics when compared to perpetrators of non-lethal intimate partner violence [16]; this comparison group differs from our comparison with homicide and perpetrators of other crime, underscoring the challenges in comparing IPF perpetrators to others (either intimate partner violence or other homicide and other crime perpetrators). As the first study to analyze these variables in Latin America, we interpret these findings as further supporting the notion that there are no psychopathology diagnoses for perpetrators of IPF. Thinking of femicide perpetrators as men with a singular distinguishable mental health diagnosis or pathology likely undermines our efforts at identifying them, and thus our ability to prevent femicide.

Second, we found higher psychological distress (GHQ-12 scores) among femicide perpetrators, although the reason for this distress is unclear. The fact that this group is statistically more likely to experience psychological distress was similarly found in other studies [33]. The elevated scores in this aspect may be explained by the increased perceived contextual and interpersonal pressures experienced by perpetrators, such as economic strain, fear of losing their partner, and general social anxiety [17, 23, 34]. Qualitative analysis of IPF perpetrator’s rationality aligns with this interpretation [12]. We also found considerably higher self-reported prior mental health and substance use treatment among femicide perpetrators. The fact that femicide perpetrators are more than three times as like as individuals convicted for other crimes to self-report mental health treatment is in line with many studies on risk factors of intimate partner violence and femicide specifically [3, 58, 59]. The findings of increased psychological distress and prior mental health and substance use treatment among femicide perpetrators suggest that there may be missed opportunities for femicide prevention in the fields of mental health and substance use disorders.

Based on a meta-analysis, Spencer et al. argue that timely and comprehensive violence screening is crucial for preventing various forms of intimate partner violence [60]. Professionals in mental health and substance use disorder fields can play a pivotal role in identifying and intervening with men seeking support. However, men are generally less inclined to seek mental health care, indicating the necessity for proactive efforts beyond these sectors. Given that IPF perpetrators reported higher treatment levels, it’s crucial to identify potential intervention points within social systems such as healthcare facilities, psychologists, and counselors. Screening, risk assessment, and intervention strategies should involve a broad range of professionals equipped to address the immediate mental, social, and legal aspects of this situation to detect potential femicide perpetration risks. Tools that differentiate between men likely to commit lethal versus non-lethal violence, along with corresponding interventions, are essential in this context.

Thirdly, femicide perpetrators reported fewer prior convictions. While not statistically significant—or included as a confounder for the other variables—this aspect is fundamental for the purposes of identifying potential perpetrators, as it would imply the possibility of early intervention, for instance regarding gender norms [61]. Since both male-male homicide and the other violent crimes clustered in this study are more associated with criminal careers (i.e., property crimes, drug dealing, etc.) and previous contact with penal institutions [29], it seems likely that that men who commit IPF do not have the same institutional pathways or systems-level engagement, perhaps even if they are also perpetrators of intimate partner violence. This finding underscores difficulties in detecting men at risk of femicide. IPF perpetrators may possess social skills, such as manipulating their partners to avoid legal action, and have networks of peers who support their coercive control before the femicide [29, 62]. These factors contribute to their ability to avoid engagement with law enforcement. As a result, upstream approaches for prevention are necessary since these men may evade or otherwise not be engaged with the criminal justice system. Femicide perpetrators likely require distinctive interventions—including self-assessments and harm mitigation tactics to prevent their potential for femicide perpetration.

Limitations

The present study is not without limitations. First and foremost, the challenges when conducting first-hand quantitative research in prison settings inherently imply institutional barriers, necessary agreement with penal institutions and sometimes being subjected to informal regulations within each facility (access blocks to certain areas, social dynamics influencing participant’s behavior). While we aimed to implement a recruitment strategy that prioritized the safety, trust, and comfort of potential participants, we acknowledge that we are unaware of factors influencing their decision to participate, such as concerns about potential retaliation from prison staff (e.g., informal punishments for “speaking out of turn”). However, the high enrollment rates (87%, 94%, and 79%) may suggest otherwise. Moreover, institutional permissions have an impact in the feasibly of using randomized samples, calculating the power of statistical tests, as mentioned in other studies [23, 30, 63]. In our prior systematic review of femicide perpetrators, we found few studies that used randomized samples [12].

Second, the sampling strategy limits the potential generalization of the results. The fact that the first stage of the sampling included a purposeful selection of institutions has potential implications for the data collected. However, since the institutions where fieldwork was conducted house a large number of imprisoned people, we believe this limitation was partially overcome. Likewise given that femicide is a relatively rare event the inclusion of comparably sized groups of perpetrators of femicide, homicide, and other crimes is not unimportant.

Third, the fact that most of the variables are self-reported (with the exception of the PCL-R) might imply certain biases, such as that answers may be exaggerated, or participants may be embarrassed or unwilling to talk about certain private topics. However, the high response rates could be interpreted as an interest in participating, which could be explained by dominant ideas in people in prison about talking about their lives [54].

Fourth, the lack of prior studies focusing on the mental health of imprisoned population and research using the scales applied here to a largely Argentinean population make it challenging to establish comparisons. Our results, however, show lower rates of psychological distress in comparison to average scores in other studies focused on imprisoned populations [51, 64]. Unfortunately, due to the lack of studies applying GHQ-12 on general Argentinean population, no further comparison can be made. Furthermore, the dearth of these data for the general population limits a possible understanding of the national frame for mental health measurements.

Fifth, due to the cross-sectional design of this project, this study cannot establish causal conclusions. Specifically, this means we cannot state whether the mental health problems we examined were already present before committing the crime, emerged after the event, or even developed during imprisonment.

Sixth, it is worth mentioning that a proportion of the femicide cases are unsolved or wrongly classified in Argentina, as in other countries [3, 65]. Consequently, some femicide perpetrators could have either been wrongly classified (i.e. as homicide perpetrators) or not imprisoned at all meaning that a subsample of femicide perpetrators may not be represented.

Lastly, the high rates of IPF-suicide (calculated to be 30% of the cases overall) could be considered a limitation [3]. Particularly, because suicide is linked to mental health status, this could potentially confound the relationship between IPF and the variables measured [23, 24].

Despite these limitations, studies such as the one presented here are useful as they acknowledge that there is not a completely ideal methodological scenario to analyze femicide perpetrators in the current institutional and legal scenario, in which this crime has been incorporated into the penal frameworks in relatively recent years. Furthermore, in spite of the obstacles, the present study identifies data on how common, or rare, various mental disorders are in a comparative sample of perpetrators of femicide, homicide, and other crimes providing windows of opportunity for prevention.

Conclusion

Deciphering the extent to which mental health characteristics have a role in the perpetration of femicide is crucial to understand not only the risk factors, root causes, and triggers of this crime, but also to plan adequate prevention strategies. Our study indicated that there are no significant differences in terms of sociodemographic characteristics or psychopathology when comparing perpetrators of IPF, homicide, and other serious violent crimes. Femicide perpetrators report higher levels of prior mental health and substance use treatment, and psychological distress—although the temporal nature of this distress is unclear. These findings suggest potential missed opportunities for screening, risk assessment, and interventions prior to the crime within the fields of mental health and substance use disorders. The study suggests that femicide perpetrators likely need tailored interventions, including self-assessments and harm mitigation tactics, to prevent their potential for femicide perpetration.