Key messages

• IPH is a process that builds up over time.

• Risk factors for IPH should be contextualised, in order to determine which are pertinent at the time of the crime.

• The build-up to an IPH is complex, with several overlapping features. Some of these are clearly gendered and thus differ between male and female perpetrators.

Introduction

Intimate partner homicide (IPH) is often preceded by a long period of intimate partner violence (IPV). What are the signs that violent behaviour is about to turn lethal? And if the perpetrator has not been violent before, in which situations is s/he about to become a killer?

Several risk factors for IPH have been described in previous research. However, it may be difficult to put these risk factors into context, and to discern which are most pertinent at the time of the crime. Furthermore, risk factors are often described as personal rather than situational, which may lead to a static and simplified understanding of risk. In their pivotal work on men who murder women, Dobash and Dobash (2015) write about the project that changes: from the perpetrator trying to keep the woman in the relationship and/or under his control, to deciding to kill her as punishment when he realises that this is not possible (anymore). This hints at a process, which merits further investigation, in which the perpetrator moves from one agenda to another. Moreover, it is interesting to examine how women’s and men’s respective triggers to kill resemble or differ from one another.

Literature review

Women are more often victims of IPV and IPH than men (Garcia et al, 2007). One of seven homicides worldwide occur within an intimate partner relationship and more than a third (39%) of all femicides are committed by an intimate partner. In comparison, only 6% of homicides with male victims are committed by intimate partners (Stöckl et al, 2013). Among male IPH perpetrators, between 20% and 70% have a history of previous IPV perpetration (Kivisto, 2015). Thus there is an increased risk of a woman being killed by her male partner if he has been violent to her before, with strangulation as a particularly strong risk factor (Spencer and Stith, 2018). However, a violent man is also at greater risk of being killed by his partner (Campbell et al, 2003). In summary, previous IPV by a man against a woman has been described as ‘the number one risk factor for IPH’ (Campbell et al, 2003: 247), regardless of the IPH victim’s gender.

This means that women generally kill in self-defence or in retaliation of violence, which a thorough body of research supports (Swatt and He, 2006; Campbell et al, 2007; Caman et al, 2016). However, there are exceptions to this rule and there is perhaps a more complex picture to be drawn. Studies from Finland (Weizmann-Henelius et al, 2012), Norway (Vatnar et al (2018) and Sweden (Caman et al, 2016), indicate that quarrelling when drinking was another commonly reported motive, with or without previous IPV. Arguably, the latter makes a great difference to the interpretation of these findings and how the nature of the ‘quarrel’ should be understood. In contrast with previous research, the Norwegian study found no significant difference between male and female perpetrators regarding previous history of IPV, with half of both male and female perpetrators subjecting the IPH victim to IPV more than five times before the IPH. However, they also reported a difference in motives where court documents more commonly described fear and revenge as motives for female perpetrators, which points to gendered differences. This indicates complex patterns of female IPH perpetration that need to be further investigated.

Separation is a strong risk factor for male-to-female IPH, with approximately 50% of the homicides occurring within two months of the breakup (Kivisto, 2015). Jealousy is a common motive for killing (Kivisto, 2015), and the risk of IPH is five times higher if the victim has left the perpetrator for another partner (Campbell et al, 2003). Further, stalking is another major risk factor (Kivisto, 2015). Mental ill health is also an identified risk factor, although perpetrators who kill their partners have less mental illness in a lifetime perspective than those who kill other family members, such as children (Kivisto, 2015). Perpetrators who commit suicide after the IPH suffer from depression to a higher degree than perpetrators who do not Kivisto, 2015). Another risk factor is alcohol/drug abuse (Campbell et al, 2003), although perpetrators may not be intoxicated at the time of the crime (Kivisto, 2015). Firearms in the household also increase the risk of IPH, with a five-fold higher risk when a perpetrator has access to a firearm (Zeoli et al, 2017; see Spencer and Stith, 2018). The strongest sociodemographic risk factor is if the perpetrator is unemployed (Garcia et al, 2007). There is also a higher risk of IPH in families with children who are not the perpetrator’s biological offspring (Garcia et al, 2007).

Most of the risk factors described above pertain to male perpetrators. Risk factors particularly pertinent to female perpetrators are previous victimisation, unemployment, alcohol/drug abuse, and having no mutual children with the IPH victim (Caman et al, 2016, Vatnar et al, 2018). In sum, female perpetrators seem to display both similarities and differences to male perpetrators and while some authors stress the similarities, others stress the differences.

In Sweden, where this study was conducted, most enquiries and research concerning IPH are undertaken by various governmental agencies. Sweden has excellent registry data which it is possible to connect to via an individual’s personal identification number, which forms the basis of this research. Thus, while a fairly robust body of knowledge on IPH exists in Sweden, only a minor part is accessible to non-Swedish speakers. Swedish research results mainly concur with those of international research (Nybergh, 2016). However, regarding risk factors, Caman (2017) found, in a longitudinal study, that alcohol had become decreasingly involved, over time, in male-perpetrated IPH cases and that perpetrators were increasingly less likely to have been previously reported for IPV. In general, according to Caman, male IPH perpetrators ‘seem to become more conventional’ (p. 49), and thus more difficult to distinguish from a normal male population.

As evident from the above, much of the literature covers risk factors for becoming a perpetrator. However, Caman’s research (2017) points out how difficult it may be to actually identify any valid risk factors. Furthermore, risk factors are often described as ‘static’, without taking contextual factors into account (see Baldry and Ferraro, 2010). This indicates a need for another way to approach risk in relation to IPH, that is, as the outcome of a process proceeding over time. This has recently been proposed by Monckton Smith (2019) who describes eight stages of progression in male-to-female IPH. Drawing on Johnson’s (2008) typology of IPV as either intimate terrorism, situation couple violence or violent resistance, and on Stark’s (2007) conception of coercive control, she describes this pathway as a form of increasing intimate terrorism with a temporal sequencing of increasing coercive control. Similar to her we want to explore what the build-up to an IPH entails, but in both male and female perpetrated cases. The term IPH will be used for all these cases, as common within the field, but we are keenly aware of that the term is problematic in female-to-male perpetrated cases of violent resistance, that is, when the perpetrator kills in the context of and/or direct response to violence directed at her.

Aim

The aim of this article is to explore and describe the build-up preceding IPH as the first phase of a threefold process surrounding the homicide. The additional two phases of this process, ‘changing the project’ and ‘the aftermath’, will be explored in a future article.

Theoretical framework

Our point of departure is a feminist understanding of IPH, connecting it to violence against women and bearing in mind that men’s violence against women is an important precursor to IPH, including when men are killed. As Hearn (1998) puts it, feminist researchers interpret violence against women as ‘structured oppression’ (p. 31). Although few feminists would claim that the oppression of women rests solely on the threat of force, violence is regarded as an important ingredient of this oppression. Within non-feminist frameworks, male violence is often interpreted as a less common phenomenon and an outcome of some other primary factor, such as individual pathology or the social system of class (Hearn, 1998). Feminist researchers, on the other hand, tend to place the violence itself, and its connection to gender, in the analytical focus.

Like Baldry (Baldry and Ferraro, 2010), however, we also acknowledge that perspectives derived from psychiatry and criminology are valid when examining IPH cases. Regarding male perpetration, many men exert violence against women, but very few to a lethal extent. According to our framework of understanding, those who do exert lethal violence probably differ from those who do not by having poorer mental health and facing graver social problems. Regarding female perpetration, few women exert violence against men, and even fewer kill a partner. Female perpetrators are thus both rare and, as shown by Caman et al (2016), in some respects more psychosocially vulnerable than male perpetrators.

One rationale for our choice of theoretical framework and focus of analysis is that our findings point to a very complex picture of female perpetration, demanding multiple lenses. We only partly found what we expected: that women kill in the context of being victimised, which was clearly true in half (5) of the cases. One case was ambiguous in the sense that there was no indication of abuse other than a verbal attack preceding the killing. In the remaining four cases we found no indication of any form of abuse. Previous IPV may of course have been under-reported in the court files, but the fact remains that other situations and motives stood out as being pertinent.

In summary, when analysing IPH cases in connection to gender, we find it valuable to try to connect strands of research often regarded as conflicting. In our analysis, we regard gender both as static, that is, what men do to women and what women do to men, but we also consider gender to be something being ‘done’ in interaction (West and Zimmerman, 1987). By so doing, we aim for a deeper understanding of IPH as a gendered crime, whether perpetrated by a man or a woman.

Methodology

The ongoing IPH-STOP study aims at identifying and analysing all IPH cases perpetrated in the Västra Götaland Region¹ in western Sweden during 2000–2016. The major aim of the project is to identify risk factors for IPH, which could possibly contribute to prevention work. Via police records, 59 IPH cases were identified. Although a few cases, not included in the study, were ambiguous and may in fact have been IPH, we are confident that we, through thoroughly going through all relevant police data on suspicious deaths, have identified all cases that police investigations have shown to be clear instances of IPH. In eight of these cases the perpetrators, all male, committed suicide before the case could be taken to court. We obtained district court records for the remaining cases, including 51 victims and 50 perpetrators, since one man killed two female ex-partners. Of these 50 IPH perpetrators, the subject of this article, 40 were men and ten were women. It is noteworthy that another male perpetrator had also killed two female victims, but one was not included in the study since this IPH was committed before the year 2000. There was one same-sex couple among the cases. Thirty-nine perpetrators were convicted of murder, four of manslaughter and four of involuntary manslaughter. In three cases, the perpetrator was acquitted by the district court. In one of these, the woman was found to have killed her male partner but was acquitted on grounds of self-defence. In the other two cases the (male) perpetrators were later convicted in a higher court of murder and manslaughter, respectively. The analysed court records all recount and summarise the court proceedings, but they vary in length (from nine to almost 60 pages), style and level of detail.

The aim of the study underlying this and a forthcoming article is to explore and describe IPH as a process. Our analysis departs from Dobash’s and Dobash’s (2015) concept of a ‘changed project’ that triggers the killing. The research questions guiding our analysis were: What does the ‘build-up’ to the killing consist of?; When does the project change (that is, at what point does the perpetrator decide to kill)?; and, finally, What does the aftermath of the deadly violence look like (in terms of what the perpetrator does and says at the scene, adjacent to the killing, and in the courtroom)? We have let these three research questions guide us through a thematic analysis, according to Braun and Clarke (2006). The method allows the use of pre-established themes, as well themes created while working with the material. A quote or text extract may cover and exemplify several themes in the material, rather than having to be forced into a single one. Furthermore, themes are allowed to, and often do, overlap.

The thematic analysis was undertaken by the first and the last author and proceeded in the following steps: 1) All pertinent court report data were coded according to three pre-established overarching themes: ‘build-up’, ‘changing the project’ and ‘the aftermath’. In order to establish consensus, coding was done independently by each author, and then compared and harmonised; 2) All coded text extracts were literally cut out and put into three piles; 3) Themes were developed jointly by the authors, again by piling quotes touching on the same themes together. This hands-on method of developing themes allowed for continuous development of consensus around coding and substantial familiarity with the material, but it also meant that quotes were mainly put under one theme, although they could touch on several; 4) Themes were further elaborated on and some were re-coded as subthemes; 5) The result of steps 1–4 were presented to the rest of the research team, whose input further refined the analysis.

We encountered some difficulties in establishing at what point the project had transitioned from build-up into changing the project. In some cases, it was simply not clear, while it was more a matter of interpretation in others. However, these difficulties helped us to discern IPH as a process that is only partly linear; the deed itself obviously constitutes a clear demarcation between before and after. But the decision to kill builds up over time and that stage of the process may be less linear.

Using court records as a source of information in IPH cases has some obvious limitations; the foremost is that the victim is dead and her/his voice can no longer be represented. This must be kept in mind throughout the analysis, so that we as researchers do not risk colluding with the perpetrators’ world view and description of events. This is especially pertinent since previous research shows that denial and victim-blaming is common among male IPH perpetrators (for example, Dobash and Dobash, 2011; 2015). Second, the courts’ task is to establish guilt, not to undertake an in-depth description of the case. In cases where the perpetrator had confessed and there were no other probable perpetrators, the information may be very sparse. On the other hand, many court records are more prolific and detailed. It is, however, important to remember that an analysis of court records only reflects what the court has found of interest to register.

Findings and analysis

Build-up is the overarching theme, covering the situational precursors to IPH that constitute the first stage of the IPH process (see Figure 1). All themes presented later in the article are underlying themes to this overarching theme. They are often overlapping rather than discrete, and it is our impression that they, together, encompass all the data we found pertaining to the build-up phase. However, the cases are very different and the build-up phase of each individual case differs, in content and sequence, from the others in ways not described later in the article. What is most important is that we have not separated male and female perpetrated cases in the model and themes presented below. However, it should be obvious, we hope, that some themes pertain more to male than female perpetrators and/or may imply different things for male and female perpetrators, which we indeed will try to illustrate.

View Full Size

The IPH process

Citation: Journal of Gender-Based Violence 5, 1; 10.1332/239868020X15922355479497

• Download Figure
• Download figure as PowerPoint slide

Separation

According to the prosecutor, it had been the victim calling to tell the perpetrator that their relationship was over. During the following five days, he frequently phoned and texted her, as well as physically stalking her. When she agreed to meet him one last time, she tried to make it clear to him (again) that it was over. After this meeting, she told her sister that she thought the perpetrator would still not leave her alone. Indeed, he did not; some hours after that last meeting he killed her.

In some relationships there seem to have been several previous breakups, or threats of breaking up, making it more difficult to discern whether an impending separation was part of the build-up phase in the IPH process. In the court records, this is often described as the partners having a ‘stormy relationship’, a term that has various implied meanings. Sometimes it refers to both partners having a drinking problem and to high levels of conflict, while the pattern of dominance, control and violence constituting the core of male-to-female IPV is referred to in other cases. Judging from previous research (see Ptacek, 1999), it is also possible that the courts may have interpreted control as conflict.

Possessiveness

It was that sleepless night that the perpetrator killed the victim.

In some cases, the perpetrators’ controlling behaviour had been noted years before the killing. It can thus perhaps be more adequately described as a part of the general context of the homicide. In other cases, it was more obvious that the perpetrators had increased these controlling behaviours prior to the killing, for instance when they started stalking the victim.

We found possessiveness to be a precursor of male-to-female IPH, in line with previous research (for example, Dobash and Dobash, 2015), which we, similar to Monckton Smith (2019), connect to male dominance and coercive control, the core feature of abusive relationships (Stark, 2007). Female perpetrators may also have been jealous, but did not display controlling and stalking behaviours, with the exception of one woman who at a previous point in time, according to the court records, had stalked and threatened her ex-partner and his new family. However, there was no indication that she had displayed such behaviours towards her present partner, whom she killed. Thus even in this case, possessiveness was not seen as a precursor to the killing.

Being violent and abusive

The victim was killed two years later, shortly after separating from the perpetrator. Ten days before the murder, the perpetrator had, according to a note written by the victim, repeatedly threatened her over the phone: pertinent subthemes to this theme were aggressive communication, escalation of violence and a desire for revenge.

In this case, the victim was stabbed by the perpetrator after his having verbally abused her for hours. In four further cases, we interpret the female perpetrators actions as violent resistance, although only one female perpetrator was acquitted on the legal grounds of self-defence. In practice, these translate as that the violence employed by the person defending herself from an attack must be in reasonable proportion to the violence employed by the attacker. This is a strictly situational evaluation that does not take the wider context of previous and ongoing victimisation into account. However, taking violent resistance into account and concurring with Campbell et al (2003), our analysis suggests that male abusiveness was a pertinent precursor to the killing in both male-to-female and female-to-male IPH cases.

‘Arguing’

At this point, according to witnesses, the perpetrator furiously stabbed her partner to death, finally telling him, ‘Nobody speaks to me like that’. Since the quotes describe apparently similar situations, they merit a further examination by comparing the cases. In the male-to-female case, the perpetrator had previously been repeatedly violent to the victim. Judging from the document instituting the court proceedings, he had severely assaulted her and attempted to strangle her six months before the murder. He assaulted her again and possibly kept her locked up in his apartment on an occasion four months before the killing, and was yet again violent just a month before the deed. Thus, there was an evident escalation of violence, peaking in the ‘argument’ described by the perpetrator. In the female-to-male case, the perpetrator – although she had previously been violent towards other people – seems not to have been violent or in other ways abusive towards the victim. Thus, there was no gradual escalation in violence, but rather a more rapid and situational escalation, in fact triggered by verbal abuse from the victim, calling her a ‘fat whore’ and ‘bad mother’. Furthermore, it is interesting to note that the statements that apparently triggered the perpetrator are related to gendered norms of being a good woman. Hence, since women are not socialised to be violent, on the one hand the perpetrator transgresses norms of femininity by exerting severe and lethal violence, but on the other hand she seems to kill in response to being criticised as a woman.

In this particular case there is no indication that the perpetrator had previously been subjected to violence by the victim, apart from what triggered her to kill, and thus it is ambiguous whether her actions should be viewed as violent resistance. In other cases, women who killed during an ‘argument’ were clearly acting in the context of and in response to being repeatedly victimised by the IPH victim.

Escalating alcohol/drug abuse

In some cases, there was an evident escalation of alcohol/drug abuse, such as when a couple who periodically drank together seemed to have more frequent and longer drinking periods. However, their drinking problem came as a surprise to their neighbour, whose help the victim sought, showing her her bruises. When the neighbour accompanied the victim home, she was shocked to see that the couple’s normally tidy home was in complete disarray and full of empty bottles. The perpetrator angrily asked her to leave. In court, she later recounted that she had never seen them so intoxicated or the perpetrator so aggressive as during the summer that preceded the killing.

Of the 40 male perpetrators, 17 had an alcohol/drug addiction and 21 were intoxicated at the time of the homicide. Of the ten female perpetrators, six had an alcohol/drug addiction and five were intoxicated at the time of the homicide. Even if the numbers are small, it is interesting to note that female perpetrators had addiction problems to at least the same extent as male perpetrators. Thus, while as Caman (2016) notes, male perpetrators seem to become more ‘conventional’ over time, female perpetrators are, at least in our small sample, quite unconventional. This is, however, in line with previous research underlining that the contexts of female perpetration, victimisation and alcohol/drug abuse overlap Caman (2016).

Escalating mental health problems

By coding this as suicidal communication, we want to draw attention to the fact that the perpetrators had often said or otherwise indicated that they were contemplating suicide, which opens up for possible intervention. It is also noteworthy, in the example above, that the perpetrator told his boss that he had loaded the gun with two shots and had placed it close to his wife.

It is also important to note that only male perpetrators seemed to contemplate suicide, which must be considered in light of the fact that perpetrators first killing their partners and then themselves is a predominantly male phenomenon (Campbell et al, 2007; Balica and Stöckl, 2016). This is also the case in the IPH-STOP study; as previously mentioned, we identified eight such cases during the study period.

Financial issues

Financial issues were sometimes precursors to the killings. These assumed various forms, making this a rather wide theme. In some cases, the perpetrator had experienced a series of financial setbacks that seemed to trigger the decision to kill the partner. Accessing the victim’s money emerged as one of several motives for killing one’s partner in both male and female perpetrated cases. Conflicts regarding finances were a substantial part of the-build-up to the killing in a few cases, as when the conflicts regarding a jointly owned business became aggravated when the female victim wanted a divorce.

In one case, an elderly male perpetrator recounted that his wife had told him that all their financial assets belonged to another person. This worried him, the perpetrator stated, and triggered him to kill her.

Moreover, she told the court that she had suggested to her husband that she could move out to the garden shed, while he could stay in the main house with his new partner. A friend of the perpetrator confirmed in court that the perpetrator had had serious concerns regarding what the divorce would entail for her economic and living situation and that she had said that the best thing would be if her husband died, making her his heir. Eventually, she hit him with a hammer and killed him.

In this case, gender enters into the analysis in the form of an unequal economic relationship, in which the perpetrator seems to have been economically dependent on the victim. As with other aspects discussed in our analysis, economic dependence should obviously not be regarded as a general precipitator of women killing. But it stands out as important in the analysis of the court records in this particular case, which is one of the female perpetrated cases where there was no indication that the woman had killed in response to any kind of violence, including economic abuse.

Impending catastrophe

In this case, also quoted under the theme of escalating mental health problems, repeated help-seeking behaviour also seems to have preceded the killing. According to the prosecutor, the perpetrator called in sick ten days before the murder and told his employer that he was contemplating suicide. However, the employer did not think that there was an acute risk of suicide, but recommended that he see a doctor. During the nine days preceding the murder, the perpetrator talked to a general practitioner on the phone and saw two different psychiatrists. According to the prosecutor, the perpetrator did not express any suicide plan to these psychiatrists. According to the perpetrator himself, this was because he found it difficult to communicate with them and he felt that he had not been taken seriously at his last visit. He also stated that it was after this encounter that his thoughts of suicide expanded to include his wife and children. Ultimately, however, he shot neither himself nor the children, but his wife.

In another case, the perpetrator had, according to the court records, tried to make the victim move out of his apartment ‘for some time’, since ‘he was afraid that something might happen’, a concern that he also expressed to a friend shortly before the murder. What eventually happened was that he beat his partner to death with a hammer.

Strategic planning

Judging from the court records, the deed was planned or at least prepared in some cases. In one such case, the perpetrator seemed to have undertaken several practical preparations. He rented a car and placed a gun in it, put a cap on (something he never did otherwise), probably to disguise himself, and then located the victim via her job. The court interpreted these actions as planning the deed, although the perpetrator denied this. Strategic planning may indicate that the perpetrator had in fact decided to kill the victim, that is, that the build-up phase had turned into a ‘changed project’. However, our analysis of the court records indicates that some ‘planning and preparing’ perpetrators had refrained from carrying out the deed at previous opportunities, before finally killing the victim. This indicates blurred and shifting boundaries between the first two stages of the IPH process. Regarding gender, strategic planning was mainly evident in male-to-female cases, with one female perpetrator who prepared the poisoning of her partner as an exception to the rule. In male to-female cases, strategic planning often went hand-in-hand with stalking and was closely connected to possessiveness and separation.

Conclusions

This article puts IPH in a process perspective. Our results show that the first, build-up, phase of an IPH is complex, and encompasses several different features. Some of these, like possessiveness, are clearly gendered, differing between male and female perpetrators, with no female perpetrators displaying possessiveness. Others appear more ambiguous. Several of the presented themes indicate an escalation: of possessiveness and violent behaviour in male-to-female cases, of alcohol/drug abuse, of mental health problems and/or of fears for the future, often connected to a real or imagined threat of separation. An escalation of violence over time was part of the build-up to IPH for some male perpetrators, while female perpetrators displayed a more rapid and situational escalation, and/or violent resistance. Male perpetrators also seemed to plan more. Escalating fears for the future could be part of the build-up for both female and male perpetrators, but suicide plans were only conspicuous in the accounts of male perpetrators. The practical implications of these findings are that practitioners should be (given the tools to be) particularly attentive to escalation of known risk factors, especially male possessiveness, and be aware that (the victim wanting) a separation may initiate escalation. Furthermore, in some cases, asking men who express that they are contemplating suicide if they have ever have thought of ‘taking someone with them’, may possibly open up opportunities for intervention.

Finally, concurrent with previous research we found that women often kill in the context of their own victimisation. However, other situations and motives also stand out as pertinent, which corresponds with recent research from several Nordic countries. If and why the picture of female IPH perpetration may look different in the Nordic countries is thus an important question for future research.

Acknowledgements

The authors wish to acknowledge General Counsel Barbro Jönsson and Chief Archivist Henrik Svensson, Police Region Väst, Swedish Police, for giving us access to crucial data and kind help along the way. Further, the input and support from all our colleagues at VKV has been much appreciated. We also want to thank the anonymous reviewers for insightful comments. Finally, the first author would like to thank Professor Liz Kelly for an important conversation over coffee.

Conflict of interest

The authors declare that there is no conflict of interest.