A Comparison of Intimate Partner Homicide With Intimate Partner Homicide-Suicide: Evidence From a Norwegian National 22-Year Cohort

Abstract

Homicide-suicide incidents make up a relatively small proportion of homicides overall, but occur more frequently in certain subtypes of homicide, such as men who kill their female partners. This study investigates aspects of intimate partner homicide-suicide (IPHS) by comparing it with intimate partner homicide (IPH). All IPHs in Norway from 1990 to 2012 (N = 177) were included. Quantitative data were extracted from court documents. Qualitative data were collected by interviews with bereaved. Multivariate logistic regression analyses and systematic text condensation were conducted. Nearly one fourth of IPHs were identified as IPHS. Perpetrators of IPHS were less likely to have a previous criminal record, even having a history of disregard and violations of the law. Perpetrators of IPHS were mainly native-born citizens and were more educated than IPH perpetrators. The motive of IPHS was more often jealousy than a dispute, but the motive was most often recorded as “other” or “unknown.” IPHS was perceived as intentional, and the bereaved did not unambiguously support the interpretation that the IPHS had been triggered by stressful situations. The bereaved pointed to the loss of hope or loss of a future combined with an inability to cope with severe disappointments as an important risk factor. Within the framework of an interactional perspective, our findings indicate that IPHS shares more characteristics with IPH than it does with other categories of homicide and other violent deaths in general.

Keywords

intimate partner homicide homicide suicide homicide-suicide motives perpetrator distinction intimate partner violence

Homicide-suicide refers to a homicide of one or several individuals immediately followed by the perpetrator committing suicide (e.g., Dobash & Dobash, 2015; Knoll & Hatters-Friedman, 2015). Homicide-suicide incidents make up a relatively small proportion of homicides overall, but occur more frequently in certain subtypes of homicide, such as men who kill their female partners (Dobash & Dobash, 2015; Galta, Olsen, & Wik, 2010; Knoll & Hatters-Friedman, 2015; Liem, 2010; Liem, Barber, Markwalder, Killias, & Nieuwbeerta, 2011; Malphurs & Cohen, 2005; Marzuk, Tardiff, & Hirsch, 1992; Salari & Sillito, 2016). A homicide-suicide has far-reaching effects on surviving family as well as the local community. The impact might also be detrimental to the community at large, as the perpetrator cannot be prosecuted. The criminal justice system is then bypassed, missing the opportunity for formal closure through a public trial normally expected in serious crime cases (Dobash & Dobash, 2015; Knoll, 2016; Malphurs & Cohen, 2005; Panczak et al., 2013; Salari & Sillito, 2016).

Prevalence and Incidence

Men outnumber women as both perpetrators and victims of homicide (e.g., Stöckl et al., 2013). Intimate partner homicide (IPH) is the most common type of domestic homicide, and it is the only homicide category in which the principal victims are female (e.g., Liem & Roberts, 2009). The ratio between men and women as perpetrators of IPH is 6:1 (Stöckl et al., 2013). Women are far more likely to be killed by an intimate partner than by anyone else (e.g., Campbell, Webster, & Glass, 2009; Campbell, Glass, Sharps, Laughon, & Bloom, 2007). A systematic review of the global prevalence of IPH indicated that one in seven homicides is committed by an intimate partner (Stöckl et al., 2013).

As homicide-suicide cases are normally not subject to legal proceedings, there is no legal definition of this phenomenon (Liem, 2010; Salari & Sillito, 2016), and the scientific literature assigns varying time limits for the interval between the homicide and suicide (Liem, 2010). Commonly, cases are included in which the homicide(s) and the suicide are likely parts of the same action (Galta et al., 2010), as operationalized in terms of a time interval between homicide(s) and suicide ranging from the typical 24 hr (Dobash & Dobash, 2015; Knoll & Hatters-Friedman, 2015) to as long as 1 week (Eliason, 2009; Marzuk et al., 1992). Some experts consider the homicides in intimate partner homicide-suicides (IPHS) to be simply a side effect of the suicide, wherein the specific decision to kill oneself precipitates a perceived necessity to kill one’s partner or family (Eliason, 2009). Other experts claim that homicide-suicide cannot be categorized with either homicides or suicides but is actually a distinct behavior (Eliason, 2009; Knoll, 2016; Marzuk et al., 1992; Panczak et al., 2013). Suicide pacts and mercy killings appear to be very rare subcategories (Salari & Sillito, 2016).

Extensive literature reviews on homicide-suicide have shown that the incidence of homicide-suicide has been stable over time at less than 0.001% (Eliason, 2009; Knoll, 2016; Marzuk et al., 1992; Salari & Sillito, 2016). In the United States, the rate is 0.134 to 0.55 per 100,000 (Knoll & Hatters-Friedman, 2015; Liem, 2010; Liem et al., 2011; Panczak et al., 2013; Salari & Sillito, 2016). In European countries, the rate ranges from 0.20 in Finland to 0.05 in Great Britain and the Netherlands (Knoll & Hatters-Friedman, 2015; Liem, 2010; Liem et al., 2011; Panczak et al., 2013). In Scandinavia and other Western societies, IPH and IPHS have become issues of interest for policy makers and research (Caman, Kristiansson, Granath, & Sturup, 2017; Galta et al., 2010; Vatnar, Friestad, & Bjørkly, 2017a). Findings from Sweden showed that 24% of perpetrators of IPH with a female victim committed suicide after the homicide. This is about 4 times the rate of suicides committed after lethal violence under other circumstances (Belfrage & Rying, 2004). Studies from the United States have found that 27% to 32% of IPH was IPHS (Bossarte, Simon, & Barker, 2006; Campbell et al., 2007). Thus, the current empirical knowledge base does not unequivocally support Coid’s first law stating that in countries with high homicide rates, the percentage of homicide-suicides is lower than in places with low homicide rates (Eliason, 2009; Liem, 2010).

Characteristics of Escalation Process, Event, Perpetrator, and Victim

Systematic reviews and meta-analyses of the homicide-suicide literature have identified factors associated with the following: (a) relationship between the perpetrator and victim, (b) history of intimate partner violence (IPV), (c) sex of perpetrator and victim, (d) age of perpetrator and victim, (e) presence of divorce/separation, (f) use of weapon, (g) history of mental illness or attempted suicide, and (h) employment status (Campbell et al., 2007; Eliason, 2009; Liem, 2010; Panczak et al., 2013). Marzuk and coworkers (1992) proposed a typology for the classification of cases of IPHS by differentiating between pathological types of possessiveness and old age/illness. Both types of IPHS are dominated by male perpetrators (e.g., Knoll & Hatters-Friedman, 2015; Liem, 2010; Marzuk et al., 1992) (see Factor List Points 1, 3, 4, and 7). Although women commit one out of six IPHs, female perpetrators are rare in IPHS (Belfrage & Rying, 2004; Campbell et al., 2007; Eliason, 2009; Liem, 2010; Stöckl et al., 2013) (see Factor List Point 3). A more recently suggested classification scheme combines relationship and aspects of motive (Knoll, 2016) (see Factor List Points 1 and 5).

International comparisons of homicide-suicide and other violent deaths, systematic reviews, and recent original studies show that homicide-suicides were more likely than other types of lethal violence to involve a female victim, multiple victims, take place in a residential setting, and be carried out with a firearm by a male perpetrator (Banks, Crandall, Sklar, & Bauer, 2008; Galta et al., 2010; Knoll & Hatters-Friedman, 2015; Liem, 2010; Liem et al., 2011; Panczak et al., 2013; Salari & Sillito, 2016) (see Factor List Point 3 and 6). Studies also indicate that homicide-suicides have different sociodemographic characteristics and a unique etiology compared with suicide or homicide alone. Perpetrators tend to be older than other homicide perpetrators, as well as older than the victim (Banks et al., 2008; Belfrage & Rying, 2004; Campbell et al., 2007; Eliason, 2009; Galta et al., 2010; Liem et al., 2011; Salari & Sillito, 2016) (see Factor List Point 4). In Western countries, most IPHS involve Caucasian dyads (Bossarte et al., 2006; Campbell et al., 2007; Galta et al., 2010; Holland, Brown, Hall, & Logan, 2018; Liem, 2010; Malphurs & Cohen, 2005).

Comparisons of motives of homicide-suicide to homicide alone indicate that IPHS perpetrators have primarily suicidal motives including depression, failure or loss (Campbell et al., 2007; Eliason, 2009) (see Factor List Points 5 and 7). However, other studies assume perpetrators of IPHS have homicidal motives including anger, rage, or previous IPV, including threats of suicide and threats to kill (Bossarte et al., 2006; Campbell et al., 2007; Salari & Sillito, 2016) (see Factor List Points 2 and 7). In addition, results indicate that perpetrators motivated by jealousy, ill health, and other life stressors were more likely to commit suicide after the IPH than those triggered by other factors (Liem, 2010) (see Factor List Points 5, 7 and 8). Evidence also indicates differences in the primary intentions, with young perpetrators reflecting homicidal motives and older perpetrators more often being suicidal (Salari & Sillito, 2016) (see Factor List Point 4). Among the subgroup of ailing spouses, research has pointed to the role of economic strain, as well as changing health in one or both of the partners (Liem, 2010) (see Factor List Points 4, 7, and 8). Perpetrators of IPHS often suffer from mental illness, with depression being the most prevalent disorder (Campbell et al., 2007; Eliason, 2009; Knoll & Hatters-Friedman, 2015; Salari & Sillito, 2016) (see Factor List Point 7). Even though it has been found that perpetrators of IPHS were more likely than perpetrators of IPH to have been seen in health or mental health services due to depression or threats of suicide in the year prior to the incident (Campbell et al., 2007), female victims varied in their awareness of danger (Salari & Sillito, 2016).

A comparison of IPHS perpetrators by age found that a known history of IPV was most common in young dyads (Salari & Sillito, 2016) (see Factor List Point 2). In terms of personality characteristics, men who commit IPHS are described as overcontrolling and dependent (Liem, 2010; Marzuk et al., 1992) (see Factor List Points 5 and 7). It has also been suggested that when the continuation of a relationship is threatened, a breakthrough of aggression can take the shape of IPHS (Liem, 2010) (see Factor List Point 2, 5 and 7). These circumstances share jealousy as a motive—the trigger leading up to the event being the female partner’s rejection and an immediate threat of separation and estrangement. It has also noted that it could be assumed that perpetrators of IPHS are unable to cope with life’s disappointments, such as a terminated relationship, illness, and financial difficulties (Knoll & Hatters-Friedman, 2015; Liem, 2010; Salari, 2007) (see Factor List Points 5, 7, and 8). Others have suggested that suicidal men who commit IPHS may do so because they view their wives and children as part of an “extended self” (Bossarte et al., 2006; Knoll & Hatters-Friedman, 2015; Salari & Sillito, 2016) (see Factor List Points 5 and 7).

There is a widespread statement borne out by prior research that homicide-suicide perpetrators differ from the prototypical “killer.” Instead of emphasizing more general homicide risk factors, some researchers have argued that perpetrators of IPHS suffer from reactions to situational circumstances, such as distress over relationship termination (e.g., Eliason, 2009; Galta et al., 2010) (see Factor List Points 5 and 8). Research has described these perpetrators as less socially marginalized, better educated, and more often employed compared with other homicide perpetrators (Campbell et al., 2007; Dobash, Dobash, Cavanagh, & Lewis, 2004; Eliason, 2009; Galta et al., 2010; Panczak et al., 2013) (see Factor List Point 8). Common risk factors for homicide such as previous IPV, alcohol or substance use, family problems, mental illness, or a criminal record are less likely to apply to perpetrators of IPHS (Banks et al., 2008; Dobash et al., 2004; Eliason, 2009; Knoll & Hatters-Friedman, 2015; Panczak et al., 2013) (see Factor List Points 2, 5, and 7). Homicide-suicide studies have shown that substance involvement was about half of that found in homicide alone (Eliason, 2009). However, Panczak et al.’s (2013) meta-analysis indicated the influence of alcohol, history of IPV, and unemployment as common risk factors in both homicide and homicide-suicide (see Factor List Points 7 and 8). Campbell et al. (2007) also emphasized unemployment as the strongest demographic risk factor for both IPH and IPHS (see Factor List Point 8). Previous suicide attempts were more common among perpetrators of homicide-suicide than among homicide perpetrators (Panczak et al., 2013) (see Factor List Point 7). Others have reported a high prevalence of previous physical IPV by perpetrators of IPHS across cultures, which is one of the strongest risk factors for IPH (Eliason, 2009; Knoll & Hatters-Friedman, 2015; Liem, 2010; Malphurs & Cohen, 2005; Vatnar et al., 2017a) (see Factor List Point 2).

In brief, IPV researchers have made theoretical distinctions among types of IPH and other types of violence recognizing that characteristics, motivations, and outcome vary (Salari & Sillito, 2016). Due to shortcomings in existing IPV theories and homicide-suicide theories, new theoretical frameworks have been developed to enhance the comprehension of these extreme acts of violence (e.g., Bell & Naugle, 2008; Emery, 2011; Knoll, 2016; Liem, 2010; Winstok, 2007, 2011) Improvements to IPV and IPHS theory and research need to fully take into account the complexity and diversity of IPV and homicide-suicide (Arriaga & Capezza, 2005; Bell & Naugle, 2008; Briere & Jordan, 2004; Knoll, 2016; Liem, 2010; Winstok, 2007). A multidisciplinary, interactional perspective, which takes into consideration the characteristics, perspectives, and interplay of contexts and perpetrators and victims, has been proposed as a more comprehensive theoretical approach (e.g., Bell & Naugle, 2008; Emery, 2011; Knoll, 2016; Liem, 2010; Winstok, 2007, 2011). From this perspective, the traditional person–situation dichotomy has been questioned and replaced by an emphasis on the mutual impact of the two variables (Funder, 2006). The main idea is that violence involves an influential and continuous interaction between individuals and the various situations or contexts they encounter, such as the influence of substance use affecting an interaction at the time of a crime. The situation or context is defined as an actual situation as it is perceived, interpreted, and assigned meaning in the mind of a participant (Magnusson, 1981). Applying an interactional perspective to IPHS implies a focus on factors contributing to an escalation of the IPHS process. Thus, the continuous interaction between previous IPV incidents and contextual, clinical, and sociodemographic factors needs to be investigated to identify specific IPHS characteristics.

Aim of the Study

Previous studies on IPHS have been limited in at least three aspects. First, analyses have mainly been restricted to univariate or bivariate methods, which do not allow for control of multiple variables’ impact in multivariate designs (Vatnar et al., 2019). Second, homicide-suicide studies tend to include lethal events among nonintimate relationships, which may differ from findings based exclusively on current or ex-intimate partners (Salari & Sillito, 2016). Third, the majority of IPHS studies are based on small and selective samples studied within a qualitative design (Liem et al., 2011). This research aims to investigate the following aspects of IPHS incidents by comparing them with IPH and controlling for other group differences from a 22-year national cohort with the total number of IPHs in Norway included. Based on information from the national cohort of IPH court documents in Norway from 1990 to 2012, we investigated the following questions:

Research Question 1: What is the prevalence of IPHS among IPH incidents?

Research Question 2: To what extent are perpetrators of IPHS different from other IPH perpetrators?

Research Question 3: Are IPHS incidents different from IPH not followed by suicide concerning escalation process and characteristics?

Method

The study was part of a mixed-methods study combining quantitative and qualitative data in a convergent parallel design (Vatnar et al., 2017a). The purpose of the convergent design is to obtain different but complementary data on the same topic (Morse, 1991 in Creswell & Plano Clark, 2011). The intent is to bring together the differing strengths of quantitative methods (large sample size, trends, generalization) with those of qualitative methods (details, in depth) (Patton, 1990 in Creswell & Plano Clark, 2011). In the convergent parallel design, the researcher uses concurrent timing to implement the quantitative and qualitative strands during the same phase of the research process, prioritizes the methods equally, keeps the strands independent during analysis, and then mixes the results during the overall interpretation (Creswell & Plano Clark, 2011). Our study is designed to investigate risk factors for IPH by combining structured risk assessment of violence and bereaved risk assessment in the analysis of court documents and information provided by bereaved. Transforming data from the text in court documents, into quantitative data drawn from the structured risk assessment instruments will involve reducing themes or risk factors to numeric information (Onwuegbuzie & Teddlie, 2003 in Creswell & Plano Clark, 2011). Because bereaved explanations of their risk assessment may provide important information about how risk factors influence perceived risk, we supplemented quantitative data with this type of qualitative data. To our knowledge, only two studies have taken a qualitative approach to understand this kind of risk assessments (Connor-Smith, Henning, Moore, & Holdford, 2011; Heckert & Gondolf, 2004).

This article is limited to data concerning IPHS, an issue that has not been addressed in previous publications arising from this IPH project. The study was approved by the Norwegian National Research Ethics Committee, and the Norwegian Higher Prosecuting Authority granted access to the court documents. The study includes all IPH cases in Norway within the time period 1990 to 2012. All cases are included, irrespective of socioeconomic status, race, ethnicity, language, nationality, sex, gender identity, sexual orientation, religion, geography, ability, and age.

Participants and Setting

Since 1990, IPH has been specified as an independent category of homicide (violation of §233/§275) in the Norwegian Criminal Investigation Service (NCIS). From 1990 to 2017, 941 homicides were committed in Norway (NCIS, 2018), representing a rate of 0.6 homicides per 100,000 inhabitants. Of these, 25% (233) were classified as IPH in the NCIS statistics. Only cases that had received a final judgment by December 31, 2013, when data collection was completed (N = 177), were included in the present study, including IPHS cases and cases with insane perpetrators and unfit to plead/incompetent to stand trial (12%). Data were extracted from the court documents pertaining to these 177 cases. These court documents contain all material and information collected and used during an investigation and a court trial (including police files, forensic expert witness reports, and court verdicts). In the majority of IPH cases from 1990 to 2012, the perpetrator was male and the victim was female (88.7%). There were no identified same-sex IPHs in Norway during this time. Mean age for perpetrators was 40.08 years (SD = 12.85) and for victims was 38.56 years (SD = 13.62).

The qualitative data stem from interviews with a stratified sample of the bereaved (n = 12): nine female and three male participants. The stratification was conducted to include a wider range of qualitative data from different categories of IPH: perpetrator male or female, IPH or IPHS, previous IPV or not, immigrants or native Norwegian, substance-related IPH or not, geography, time period, socioeconomic status, age, and so on. The bereaved may be relatives of the victim, relatives of the perpetrator, or friends of either or both, based on who was identified as core informants/witnesses in the court documents. Only one person from each selected case was interviewed. The sample of the bereaved was stratified to represent different types of IPH based on sociodemographic status, clinical and contextual factors, IPH characteristics, and help-seeking prior to the IPH incident. Two of the interviews represented IPHS.

Procedures

Quantitative data were collected by traveling to each police county or district in charge of a case, manually going through the set of court documents for each case, and coding the information into quantitative data according to a predefined codebook consisting of variables from NCIS homicide statistics and risk factor items drawn from three risk assessment tools. The reliability of this procedure was supported by results from an interrater reliability test—intraclass correlation coefficient (ICC), one-way random model, average measures = 0.835, confidence interval (CI) = [0.714, 0.923]—based on two independent raters’ coding of complete data sets from 20 randomly selected cases.

Any risk for IPH that the bereaved had observed was assessed by asking whether they could now, in retrospect, identify any risk or signals for IPH in the case that had caused their bereavement. Open-ended questions were then posed about these perceptions of possible risk factors and warning signals and how they now perceived their own risk assessment. The interviews were conducted by a specialist in clinical psychology. All interviews were audiotaped and later transcribed and saved in word processing files for analysis.

Measures

We defined IPHS as cases where the homicide(s) and the suicide were likely parts of the same action, typically occurring within 24 hr (Dobash & Dobash, 2015; Knoll & Hatters-Friedman, 2015). Cases where the perpetrator committed suicide after arrest or adjudication were not included as IPHS. Variables and measures of IPH characteristics and sociodemographic and contextual factors were drawn from NCIS statistics (for details, see Tables 2 -4). Only diagnoses (ICD-10 [International Classification of Diseases, 10th Revision]) set by health professionals qualified to diagnose mental illness (clinical psychologists and medical doctors) were registered in the variables perpetrator’s diagnosis and victim’s diagnosis. Criminal/legal problems measured as perpetrator previously convicted: no, yes, or omit. Perpetrators’ origin was measured as native citizen, naturalized citizen, or foreign citizen. Motive was measured in four main categories: dispute, jealousy, other, or unknown. Perpetrator in contact with police, health, or social service measured no, yes, or omit.

Risk factor items were taken from three validated risk assessment instruments. Most IPV risk assessment instruments aim at measuring risk of IPV, not IPH. The predictive validity of risk assessment for IPH is lower than for IPV due to the low base rates of IPH (see, for example, Messing & Thaller, 2013). In a preliminary review of the IPH literature, we found three validated risk assessment instruments with items on IPH such as the following: “Severe threats or threatening to kill in the past month,” “Victim’s perception of danger of death in the past month,” “Does he threaten to kill you?,” “Does he ever try to choke you?,” and “Do you believe he is capable of killing you?” (Vatnar, Friestad, & Bjørkly, 2017b). These risk assessment instruments were Danger Assessment Revised 20-item (DA-R20; Campbell et al., 2007; Campbell, Webster, & Glass, 2009), Spousal Assault Risk Assessment Guide (SARA; Kropp & Hart, 2000; Kropp, Hart, Webster, & Eaves, 1995), and Severe Intimate Violence Partner Risk Prediction Scale (SIVPAS; Echeburùa, Fernandez-Montalvo, de Corral, & Lopez-Goni, 2009). Together, these instruments cover a substantial number of possible risk factors of IPH (Vatnar et al., 2017a). The DA-R20 and SIVPAS are drawn from the actuarial tradition, and the SARA, from the structured professional judgment (SPJ) tradition. As the present study is conducted in the SPJ tradition, no total scores or weighing on measures were computed. All risk factors were measured as in the SARA: yes, partially, no, or unknown/omitted. SPJ is an approach that attempts to bridge the gap between the unstructured clinical judgment and actuarial decision-making approaches. It appears to be a viable approach for assessing IPV risk, suiting the requirements of criminal justice professionals (Kropp & Hart, 2015).

Analyses

Univariate and bivariate analyses and multivariate logistic regression analyses were used to test the association between risk factors and the dependent variable IPH followed by suicide or not. The stepwise options recommended for logistic regression for small samples were used (Altman, 1991; Pallant, 2010). As suggested by Altman and Pallant, initial comparisons of IPHS and IPH not followed by suicide were carried out by simple descriptive cross-tabulations (Step 1, Tables 2 -5). In the first multivariate logistic regression analyses (Step 2), variables with significant (p ≤ .05) or trend (p ≤ .10) univariate differences between the two dependent variables were adjusted for other significant differences within the following categories: (a) sociodemographic and clinical factors, (b) previous IPV, (c) IPH characteristics, and (d) previous help-seeking. Significant differences remaining in each of the four comparisons in Step 2 were forwarded to Step 3 (Table 6) where they were adjusted for all remaining group differences in Categories a, b, c, and d. Suitability for multivariate logistic regression analysis was investigated by the Hosmer and Lemeshow test. Cox and Snell R² and Nagelkerke R² were used to estimate the proportion of explained variance in the multivariate models that were tested. Values were estimated as model fit indices for the regression models (see notes in Table 6). Statistical analyses were performed using the statistical program package SPSS, version 23.0.

Transcribed text from the interviews was analyzed by systematic text condensation (STC; Malterud, 2001, 2012). The procedure consisted of the following steps: (a) total impression—from chaos to themes, (b) identifying and sorting meaning units—from themes to codes, (c) condensation—from code to meaning, and (d) synthesizing—from condensation to descriptions and concepts (Malterud, 2001, 2012).

Results

Prevalence and Incidence

According to information in the court case documents, 24.8% of the IPHs were identified as IPHS (Table 1), indicating a rate of 0.04 IPHS per 100,000 inhabitants. Nearly 6% (5.6%) of all IPHs—10 cases—included homicide of children in addition to the intimate partner (Table 1). All 10 familicide-suicide cases had a male perpetrator. The perpetrator was the biological father in nine out of 10 familicide-suicide cases.

Table 1.

Frequency Distributions Preceding Intimate Partner Homicide and Suicide (N = 177).

Variable	Measures	N	%
Categories of IPH	Intimate partner homicide (alone)	126	71.2
	Intimate partner and child victims homicide	4	2.3
	Intimate partner and other victims homicide	3	1.7
	Intimate partner homicide and suicide	34	19.2
	Intimate partner, child victims homicide and suicide	10	5.6

Note. Twenty-one children were <18 years when killed. Three children were ≥18 years when killed. Other victims were relatives or friends of intimate partner. Categories intimate partner and child victims, intimate partner and other victims, intimate partner, child victims and suicide also include cases where more than one child or one other person was killed in addition to the intimate partner. IPH = intimate partner homicide.

Characteristics of Escalation Process, Event, Perpetrator, and Victim

Several significant differences pertaining to perpetrators’ sociodemographic and clinical characteristics (gender, age, marital status, marital duration, origin, education, employment status/source of income, criminal record/legal problems, and substance misuse; Table 2) emerged from the bivariate analyses. The same goes for previous IPV in the current relationship (physical IPV, threats to kill; Table 3), IPH characteristics (modus operandi, motive, victim intent to break up, conflict of custody, economic problems; Table 4), and for perpetrators’ and victims’ previous help-seeking (perpetrator and victim in contact with police, health and social services, assessed risk, assessed risk forwarded, mention IPH plans; Table 5).

Table 2.

Differences Concerning Sociodemographic and Clinical Factors Preceding Perpetrators of IPH Compared With Perpetrators of IPH-Suicide (N = 177).

Independent Variable	Measures	IPH (n = 133)% (n)	IPH-Suicide (n = 44)% (n)	Total (N = 177)% (n)	p Value
Gender	Male	85.7 (114)	97.7 (43)	88.7 (157)	.029
Gender	Female	14.3 (19)	2.3 (1)	11.3 (20)
Age	Years (M)	38.5	44.9	40.1	.005
Marital status	Married	35.3 (47)	63.6 (28)	42.4 (75)	.006
	Cohabitant	37.6 (50)	18.2 (8)	32.8 (58)
	Separated	10.5 (14)	2.3 (1)	8.5 (15)
	Divorced	3.0 (4)	6.9 (3)	4.0 (7)
	Former cohabit.	13.5 (18)	9.1 (4)	12.4 (22)
Marital duration	Years (M)	8.8	14.2	10.1	.013
Origin	Native citizen	61.7 (82)	81.8 (36)	66.7 (118)	.033
	Naturalized citizen	12.0 (16)	2.3 (1)	9.6 (17)
	Foreign citizen	26.3 (35)	15.9 (7)	23.7 (42)
Education	Years (M)	9.9	12.8	10.2	≤.000
Source of income	Employed/student	36.8 (49)	63.6 (28)	43.5 (77)	.002
Source of income	Unemployed	63.2 (84)	36.4 (16)	56.5 (100)
Substance misuse	Yes/partly	14.3 (19)	0.0 (0)	10.7 (19)	.008
Substance misuse	No/omit	85.7 (114)	100.0 (44)	89.3 (158)
Criminal/legal problems	Yes	54.9 (73)	18.2 (8)	45.7 (81)	≤.000
	No	39.8 (53)	72.7 (32)	48.0 (85)
	Omit	5.3 (7)	9.1 (4)	6.2 (11)

Note. Mann–Whitney U tests were estimated to test for possible independent group differences for variables with nonparametric distributions. The chi-square test was used for nominal data. Criminal/legal problems measured as perpetrator previously convicted. Perpetrator having children, mutual children in the intimate partnership, perpetrator’s diagnosis and victim’s diagnosis, symptoms of mental illness, suicidal ideation/behavior/threats, year of crime were tested with nonsignificant results. IPH = intimate partner homicide.

Table 3.

Previous IPV for IPH Compared With IPH-Suicide (N = 175) and (N = 126).

Independent Variable	IPH(n = 133)% (n)	IPH-Suicide(n = 44)% (n)	Total(N = 177)% (n)	p Value
Threats to kill
Yes/partly	60.3 (79)	36.4 (16)	54.3 (95)	.006
No/omit	39.7 (52)	63.6 (28)	45.7 (80)
Previous physical IPV
Yes/partly	89.9 (98)	64.7 (11)	86.5 (109)	.005
No/omit	10.1 (11)	35.3 (6)	13.5 (17)

Note. Analyses of characteristics of IPV previous IPH only includes cases with previous IPV (n = 126). The chi-square test was used for nominal data. Previous IPV in general, previous psychological IPV, previous sexual IPV, mutual IPV, IPV registered by authority, forwarded IPV information to authority were tested with nonsignificant results. Two cases were missing previous IPV, n = 175. For physical IPV, only previous IPV cases are included, n = 126. IPV = intimate partner violence; IPH = intimate partner homicide.

Table 4.

Characteristics of IPH Cases Compared With IPH-Suicide Cases (N = 177).

Independent Variable	IPH(n = 133)% (n)	IPH-Suicide(n = 44)% (n)	Total(N = 177)% (n)	p Value
Modus operandi
Knife/axe	48.9 (65)	18.2 (8)	41.2 (73)	≤.000
Firearm	15.8 (21)	52.3 (23)	24.9 (44)
Blunt force	10.5 (14)	9.1 (4)	10.2 (18)
Choking	20.3 (27)	18.2 (8)	19.8 (35)
Other	4.5 (6)	19.8 (35)	4.0 (7)
Motive
Dispute	30.1 (40)	4.5 (2)	23.7 (42)	≤.000
Jealousy	39.8 (53)	40.9 (18)	40.1 (71)
Other	19.6 (26)	20.5 (9)	19.9 (35)
Unknown	10.5 (14)	34.1 (15)	16.4 (29)
Victim intent to break up/separate
No/omit	37.1 (43)	57.1 (24)	42.4 (67)	.024
Yes/partly	62.9 (73)	42.9 (18)	57.6 (91)
Conflict of custody
No/omit	78.6 (103)	95.5 (42)	82.9 (145)	.010
Yes/partly	21.4 (28)	4.5 (2)	17.1 (30)
Economic problems
No/omit	25.2 (33)	68.2 (30)	36.0 (63)	≤.000
Yes/partly	74.8 (98)	31.8 (14)	64.0 (112)

Note. The chi-square test was used for nominal data. Police county, year of crime, scene of crime/location of the incident, perpetrator’s and victim’s substance influence at time of crime were tested with nonsignificant results. IPH = intimate partner homicide.

Table 5.

Differences in Help-Seeking Previous to the Incident for IPH Compared With IPH-Suicide (N = 177).

Independent Variable	IPH(n = 133)% (n)	IPH-Suicide(n = 44)% (n)	Total(N = 177)% (n)	p Value
Perpetrator in contact with police, health, and social services				≤.000
No	12.2 (16)	14.0 (6)	12.6 (22)
Yes	84.7 (111)	60.5 (26)	78.7 (137)
Omit	3.1 (4)	25.6 (11)	8.6 (15)
Assessed risk in contact with perpetrator				≤.000
No	49.2 (62)	33.3 (14)	45.2 (76)
Yes	40.5 (51)	14.3 (6)	33.9 (57)
Omit	10.3 (13)	52.4 (22)	20.8 (35)
IPH plans mentioned to nonprofessionals				.012
No	51.1 (67)	34.1 (15)	46.9 (82)
Partly	12.2 (16)	11.4 (5)	12.0 (21)
Yes	28.2 (37)	27.3 (12)	28.0 (49)
Omit	8.4 (11)	27.3 (12)	31.1 (23)
Victim in contact with police, health, and social services				≤.000
No	3.8 (5)	18.2 (8)	7.4 (13)
Yes	84.0 (110)	36.4 (16)	72.0 (126)
Omit	12.2 (16)	45.5 (20)	20.6 (36)
Assessed risk in contact with victim				≤.000
No	29.5 (38)	27.9 (12)	29.1 (50)
Yes	51.2 (66)	9.3 (4)	40.7 (70)
Omit	19.4 (25)	62.8 (27)	20.2 (52)
Assessed risk forwarded after contact with victim				.022
No	65.7 (69)	23.8 (12)	67.6 (96)
Yes	23.8 (25)	10.5 (11)	19.0 (27)
Omit	10.5 (11)	73.0 (27)	13.4 (19)

Note. The chi-square test was used for nominal data. Homicide plans identified by professionals, victim awareness of homicide risk, risk assessment forwarded to other professionals after contact with perpetrator were tested with nonsignificant results. IPH = intimate partner homicide.

Few of the bivariate differences in Tables 2 to 5 remained significant in Step 2 when adjusted for other significant group differences within the same aim categories: sociodemographic and clinical factors, previous IPV, IPH characteristics, and previous help-seeking.

After controlling for all remaining group differences, the following differences between IPHS and IPH remained significant in the multivariate models in Step 3 (Table 6). (a) Perpetrators of IPHS were 80% less likely to have a previous criminal record; (b) they were nearly 95% more likely to be a native citizen compared with a naturalized citizen or nearly 80% more likely to be a native citizen compared with a foreign citizen; (c) their motive for the crime was 6 times more likely to be ascribed to jealousy than to dispute, or 14 times more likely to be ascribed to “other,” or 29 times more likely to be ascribed to “unknown”; (d) information of IPHS perpetrators’ previous contact with police, health, and social services had 8 times increased risk of being lacking in the court case documents; and (e) IPHS perpetrators were higher educated compared with other perpetrators of IPH.

Table 6.

IPH-Suicide Compared With IPH Without Suicide (Baseline) Multivariate Logistic Regression.

Independent Variables	Adj. Odds Ratio (OR)	95% CI	p Value
Model 1 (N = 177)
Origin			.015
Native citizen (baseline)
Naturalized citizen	0.061	0.006, 0.633	.019
Foreign citizen	0.225	0.061, 0.827	.025
Criminal/legal problems			.002
No (baseline)
Yes	0.198	0.072, 0.547	.002
Omit	4.979	0.928, 26.720	.061
Gender			ns
Age			ns
Marital status			ns
Marital duration			ns
Source of income			ns
Substance misuse			ns
Model 2 (n = 174)
Criminal/legal problems			.009
No (baseline)
Yes	0.220	0.077, 0.629	.005
Omit			ns
Motive			.013
Dispute (baseline)
Jealousy	6.205	1.114, 33.661	.034
Other	14.289	1.804, 113.187	.012
Unknown	28.711	3.685, 233.631	.001
Perpetrator in contact with police, health, and social services			.026
No
Yes			ns
Omit	7.668	1.160, 50.684	.035
Modus operandi			ns
Origin			ns
Model 3 (n = 137)
Education (years)	1.269	1.023, 1.524	.030
Origin			ns
Criminal record			ns
Motive			ns
Modus operandi			ns

Note. Multivariate binary logistic regression. Adj. odds ratio = adjusted odds ratio. Model 1, Cox and Snell R² = .255 and Hosmer and Lemeshow test = .756. Model 2, Cox and Snell R² = .381 and Hosmer and Lemeshow test = .581. Model 3, Cox and Snell R² = .236 and Hosmer and Lemeshow test = .270. Model 4 including previous IPV was tested with nonsignificant results. IPH = intimate partner homicide; OR = odds ratio. CI = confidence interval.

The qualitative interviews also showed that when the perpetrator had no criminal record, the bereaved still expressed concerns about the perpetrator’s history of disregard and violations of the law:

But he had done some things that I would never do. And in a very stressful situation, you might do the wrong thing . . . As he clearly did . . . Yes, he pushed the limits, changed and bent the rules to fit himself . . . But he would always push the limits on what he was allowed to do . . . He was clear that he had smoked marijuana and that he had an illegal radar detector in his car, he paid speeding fines, but not speeding fines from abroad, he misused prescription medications and so on.

[Interviewer] He was not previously convicted, right?

Yes, he had managed to get away. He certainly got a few speeding fines in Norway, but nothing of a serious nature. And as I said about his upbringing, that his father forced him to cut down trees owned by the neighbor, things like that. It shapes attitudes, what your dad tells you to do.

However, the idea of IPHS perpetrators as different from both the prototypical “killer” as well as “batterer” was supported by the bereaved:

No, it was not his style to shoot someone—it was quite unthinkable. Plus he was very calm, kind, helpful, nice, a good friend. And the first to stand up for someone, and there was no aggression in him in any way.

Still, the qualitative interviews with the bereaved also showed that the IPHSs were interpreted as intentional, not as accidental reactions in a dispute:

It wasn’t a case like you throw a punch because you get so annoyed or because someone hits you, you hit back and the other falls—it was not like that. There was something that was very wrong and sounded planned—ice cold planned. And that made it even worse, you know.

When you speak about a person who you are in such a hateful relationship with, you might say you wish someone could shoot her, right? And that was the case, that NN said such things, which was understood as a joke, of course. That’s what people say without meaning it. And so, in retrospect, maybe NN meant it? Such extreme statements, there was a message behind it. . . to say it aloud and test his friends’ reactions. And I actually think that’s what he did.

It has been argued that perpetrators of IPHS are prone to react to situational circumstances, and the bereaved added some nuances and concerns to this interpretation of motive:

Something was written in the newspaper about his workload and so on. But I thought it was completely wrong to start linking the pressure at work to the incident—the next step from workload is to kill three people? You can’t make such connections. Children at school [where the murdered children had been pupils] were terrified when their fathers were exhausted after work.

Rather, the bereaved perceived loss of hope or lack of future prospects, combined with an inability to cope with severe disappointments to be an important risk factor for IPHS.

One of the reasons then was NN’s despair, for it was a deep despair, because he saw everything he had built up throughout his life was lost . . . So what he lost then, it was completely lost. And then there was no future for him. And I think that if you have no future, then you get into a slightly different situation than younger people who can look forward and who have the future ahead.

I don’t think it was helpful, that he was educated from this prestigious university. At that university, they were all taught that you are a winner; you will be the top executive. And if you do not, then what happens? There will be some failures, and there will be a gap between your expectations and where you actually end up.

Discussion

Main Findings

We applied an interactional perspective to understand and investigate possible differences between IPH and IPHS. To accomplish this, we analyzed the context and interaction before, during, and after an incident. Nearly one fourth of IPH were identified as IPHS. Perpetrators of IPHS were less likely to have a previous criminal record, even if data indicated previous undetected misdemeanors. Perpetrators of IPHS were mainly native Norwegian citizens and had more education than IPH perpetrators. The motive of IPHS was more often jealousy than a dispute; however, most often, the motive was recorded as “other” or “unknown.” IPHS was perceived as intentional, and the bereaved did not unambiguously support the interpretation of IPHS as triggered by stressful situations. The bereaved pointed to the loss of hope or of a future, combined with an inability to cope with severe disappointments, as an important risk factor.

Prevalence and Incidence

The court case documents showed that nearly one fourth of IPHs in Norway were identified as IPHS, paralleling findings from Sweden (Belfrage & Rying, 2004), but being somewhat lower than U.S. estimates (Bossarte et al., 2006; Campbell et al., 2007). The rate of IPHS per 100,000 inhabitants was 0.04, placing Norway among the European countries with the lowest rate of IPHS (Knoll & Hatters-Friedman, 2015; Liem, 2010; Liem et al., 2011; Panczak et al., 2013). As the Norwegian homicide rate (0.6) is also low in international comparisons, the findings on IPHS do not support Coid’s first law’s hypothesis/prediction (Eliason, 2009; Liem, 2010).

Characteristics of Escalation Process, Event, Perpetrator, and Victim

According to previous research, perpetrators of IPHS are mainly male native citizens (Bossarte et al., 2006; Campbell et al., 2007; Galta et al., 2010; Holland et al., 2018; Liem, 2010; Malphurs & Cohen, 2005). In our 22-year cohort of IPH, only one IPHS case involved a female perpetrator. All familicide-suicide cases had a male perpetrator, and he was the biological father in nine of the 10 cases. This concurs with previous studies that showed that homicide-suicides were more likely than other types of lethal violence to involve a female victim, multiple victims, take place in a residential setting, and be committed by a male perpetrator (Banks et al., 2008; Galta et al., 2010; Knoll & Hatters-Friedman, 2015; Liem, 2010; Liem et al., 2011; Panczak et al., 2013; Salari & Sillito, 2016).

In accordance with previous research, perpetrators of IPHS were less socially marginalized and more often employed, compared with perpetrators of homicide generally (Campbell et al., 2007; Dobash et al., 2004; Eliason, 2009; Galta et al., 2010; Panczak et al., 2013). Still, IPHS perpetrators emerged as more similar to, than different from, IPH perpetrators in findings from multivariate analysis comparing other sociodemographic characteristics. Previous research has found that perpetrators of IPHS were more likely than perpetrators of IPH to have used health or mental health services due to depression or threats of suicide in the year prior to the incident (Campbell et al., 2007). We found that compared with that of perpetrators of IPH, information on IPHS perpetrators’ previous contact with police, health, and social services was more often absent in the court case documents.

The widespread preconception that homicide-suicide perpetrators are different from the typical profile of a “killer” was supported by interviews with the bereaved. However, even if perpetrators of IPHS were less likely than perpetrators of IPH to have a previous criminal record, qualitative data indicated previous violations of the law that were undetected or that did not incur criminal consequences. In addition, there was no significant difference pertaining history of previous IPV when comparing IPHS with other IPH.

As well, some, instead of emphasizing more general homicide risk factors, have argued that perpetrators of IPHS suffer from reactions to situational circumstances (e.g., Eliason, 2009; Galta et al., 2010). The bereaved voiced some concerns about interpretations that link IPHS to situational circumstances. Some pointed out that linking the homicides to ordinary stressful situations such as a heavy workload or the termination of a partnership caused anxiety and concern, at least in the local community. As well, the motive for IPHS was more often ascribed to jealousy than to a dispute. However, the motive for the crime was most often officially recorded as “other” or “unknown,” as both perpetrator and victim(s) were dead and access to information regarding motives was therefore limited. At the same time, qualitative data indicated that the IPHS was perceived as intentional and planned, not an accidental reaction to situational circumstances. In addition, the bereaved felt that the loss of hope or loss of a future in combination with an inability to cope with severe life disappointments was an important risk factor. This finding adds some perspective to previous findings that homicide-suicide perpetrators are more commonly depressed and their motives are failure and loss compared with homicide perpetrators (Campbell et al., 2007; Eliason, 2009; Holland et al., 2018; Knoll & Hatters-Friedman, 2015; Liem, 2010; Salari, 2007). It has also been speculated that suicidal men who commit IPHS may do so because they view their wives and children as part of an “extended self” (Bossarte et al., 2006; Knoll & Hatters-Friedman, 2015; Salari & Sillito, 2016). To speculate, our findings concerning sociodemographic and clinical factors, previous IPV, IPH characteristics, and previous help-seeking behavior could be perceived and interpreted as indicating that IPHS is the Caucasian male’s version of honor-based homicide. However, our findings concur with previous studies claiming that homicide-suicide cannot be categorized either as homicide or as suicide, but is actually a distinct behavior (Eliason, 2009; Knoll, 2016; Marzuk et al., 1992; Panczak et al., 2013). Nevertheless, it would be naive to assume that homicide-suicide is unique and shares no characteristics with other forms of fatal violence or that its typology is exhaustive (Marzuk et al., 1992).

Conclusion

We found few differences and many similarities between IPHS and IPH without suicide within the framework of an interactional perspective, addressing the escalation process, the context, previous IPV, clinical and sociodemographic factors, and IPH characteristics. To put it more simply, IPHS differs more from other categories of homicide and other violent death, in general, than it does from IPH. IPH and IPHS have several characteristics in common that distinguish them from other homicides. Future research may address the following findings from our study: (a) an absence of criminal records for perpetrators even when previous physical IPV had been found in 65% of the IPHS cases and previous threats to kill, in 36%; (b) an increased likelihood that IPHS perpetrators belonged to the native-born rather than the immigrant population; (c) an absence of discernible (to us) dominant motives for IPHS; and (d) a difficulty finding information concerning perpetrators’ previous help-seeking.

Limitations

The variables used in the present analysis did not cover all possible risk factors of IPHS. Criminal case documents relating to each of the 177 IPHs were the only quantitative data source in this study. These documents are produced for purposes other than research and consequently did not provide exhaustive data related to our research questions. Accordingly, there may be a risk of false negatives—for example, failure to identify diagnoses of mental illness, previous IPV, and help-seeking. Nonetheless, the risk for false positives from this source of data is low. At the same time, the NCIS database on homicide in Norway can provide a clear and unique insight based on a representative population-based sample, where the perpetrators who commit suicide in the commission of an offense are included. Some findings may be underestimated due to a lack of information from the victim and, in IPHS cases, from both victim and perpetrator, particularly in relation to previous IPV.

Some of the independent variables have few response options, for example, motivation is limited to four categories (including “other”). This was done due to limitations in the court documents and to strengthen statistical power. By limiting the number of potential responses, it further limits the likelihood of variation between IPH and IPHS. However, the risk for statistical Type II errors would increase by differentiating into higher numbers of attributes and categories. In our research design, the qualitative data add valuable and more nuanced information. Even though this was a 22-year cohort study, the number of IPHS (n = 44) was small, increasing the risk for statistical Type II errors. The odds ratios were large for some associations. Still, wide CIs call for caution in interpretation. However, the investigation covered the total prevalence of IPH in Norway within the actual time period. This strengthens the external validity of the findings, at least for IPH in Norway. The Hosmer and Lemeshow tests confirmed that the data were suitable for multivariate logistic regression. The model fit indices were high for the models in the multivariate analyses. This enhanced the internal validity of the study.

It is reasonable to assume that attitudes and perceptions regarding the relationship may vary significantly depending on the relationship of the bereaved to the victim and offender. However, some of the bereaved were friends or grown-up children of both perpetrator and victim. Thus, the interviews with IPHS bereaved reflected the experiences of both perpetrators and victims.

Clinical and Policy Implications

The typical social sanctions against homicide—that is, a criminal justice response—are not possible after IPHS, and prevention of the crime appears to be a huge challenge if limited only to the potential IPHS perpetrator population. When individuals perceive they have lost everything, there is a greater danger of IPHS (Salari & Sillito, 2016). Because IPHS is such a rare event, to screen for it is not feasible without many false positives. However, the literature and our results reveal some characteristics that may be helpful when performing a risk assessment and management. As IPH is very rare and IPHS is even rarer compared with other IPV, it is important to emphasize the difference between structured risk judgments (SPJ) and the more limited approach of risk predictions. The scope of risk prediction is to identify a context-independent risk for future violence for an actual person. In contrast to this, the SPJ tradition has two important aims: (a) to identify violence risk as an interactional or situational phenomenon and (b) to develop measures that can mitigate this risk. Instigation of preventive measures is an integral and core aspect of this approach. Hence, further research on IPHS may want to integrate two strongly associated issues: risk identification and prevention.

Research Implications

Because IPHS has an extremely low base rate, studies must be largely retrospective in design and cover either a large geographic area or a long time span. News reports were often the material in IPHS research, and obtaining accurate representation of the events as they are reported in news media has been crucial (e.g., Salari & Sillito, 2016). IPHs in Norway follow a socially biased pattern, which is consistent with research on recorded crime in general (Aldridge & Browne, 2003; Dobash, Dobash, & Cavanagh, 2009; Garcia & Hurwitz, 2007; Vatnar et al., 2017a). Thus, the news media’s well-documented tendency to present IPH and especially IPHS as happening out of the blue is at odds with current evidence. Consequently, news reports may not be reliable as sources for IPHS research. Research on IPHs needs to focus on deepening the understanding of the circumstances that precede IPHS and under what circumstances IPHS interacts with other contextual and individual variables. In such studies, sample selection is an urgent issue. Comparing IPHS with IPH and not to all kinds of homicide or suicide is essential. Finally, the great majority of IPHS studies include only univariate and bivariate analyses. It is of paramount importance for future studies to develop and test multivariate models to better grasp the complexity of IPHS.

Footnotes

Declaration of Conflicting Interests

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

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Solveig K. B. Vatnar

Author Biographies

Solveig K. B. Vatnar is a professor of clinical psychology at the Molde University College, Faculty of Health and Social Sciences. She also currently holds position as PhD, specialist in clinical psychology at the Centre for Research and Education in Forensic Psychiatry, Oslo University Hospital, Norway. She was educated at University of Oslo, Norway. Her expertise is in the field of intimate partner violence, with a special focus on the interactional perspective of IPV, risk assessment of IPV, and intimate partner homicide. She has authored more than 20 international publications in journals and books in the fields of clinical psychology/psychiatry.

Christine Friestad (psychologist, PhD) is a researcher at the Centre for Research and Education Forensic Psychiatry at the Oslo University Hospital and also holds a position as associate professor at the University College of Norwegian Correctional Service. Her research mainly focus on prisoners, mental health, and living conditions, with a special emphasis on persons convicted of sexual and violent crimes, and the correctional management and health services provided to this group.

Stål Bjørkly is a professor of clinical psychology at the Molde University College, Faculty of Health and Social Sciences, and a clinical psychologist. He is also research consultant at the Centre for Research and Education in Forensic Psychiatry, Oslo University Hospital, Norway. He was educated at the University of Bergen. His main work concerns the psychology of aggression, violence among mentally ill patients, and risk analysis. He has published several books on the psychology of aggression and more than 100 international publications in journals and books in the fields of clinical psychology/psychiatry.

References

Aldridge

M. L.

Browne

(2003). Perpetrators of spousal homicide. A review. Trauma, Violence, & Abuse, 4, 265-276.

Altman

D. G.

(1991). Practical statistics for medical research. London, England: Chapman & Hall/CRC.

Arriaga

W. B.

Capezza

N. M.

(2005). Targets of partner violence: The importance of understanding coping trajectories. Journal of Interpersonal Violence, 20, 89-99.

Banks

Crandall

Sklar

Bauer

(2008). A comparison of intimate partner homicide to intimate partner homicide-suicide: One hundred and twenty-four New Mexico cases. Violence Against Women, 14, 1065-1078.

Belfrage

Rying

(2004). Characteristics of spousal homicide perpetrators: A study of all cases of spousal homicide in Sweden 1990-1999. Criminal Behaviour and Mental Health, 14, 121-133.

Bell

K. M.

Naugle

A. E.

(2008). Intimate partner violence theoretical considerations: Moving towards a contextual framework. Clinical Psychology Review, 28, 1096-1107.

Bossarte

R. M.

Simon

T. R.

Barker

(2006). Characteristics of homicide followed by suicide incidents in multiple states, 2003-04. Injury Prevention, 12(Suppl. 2), ii33-ii38. doi:10.1136/ip.2006.012807

Briere

Jordan

C. E.

(2004). Violence against women: Outcome complexity and implications for assessment and treatment. Journal of Interpersonal Violence, 19, 1252-1276.

Caman

Kristiansson

Granath

Sturup

(2017). Trends in rates and characteristics of intimate partner homicides between 1990 and 2013. Journal of Criminal Justice, 49, 14-21. doi:10.1016/j.jcrimjus.2017.01.002

10.

Campbell

J. C.

Glass

Sharps

P. W.

Laughon

Bloom

(2007). Intimate partner homicide: Review and implications of research and policy. Trauma, Violence, & Abuse, 8, 246-269.

11.

Campbell

J. C.

Webster

D. W.

Glass

(2009). The danger assessment: Validation of a lethality risk assessment instrument for intimate partner femicide. Journal of Interpersonal Violence, 24, 653-674.

12.

Connor-Smith

J. K.

Henning

Moore

Holdford

(2011). Risk assessments by female victims of intimate partner violence: Predictors of risk perceptions and comparison to an actuarial measure. Journal of Interpersonal Violence, 26, 2517-2550.

13.

Creswell

J. W.

Plano Clark

V. L.

(2011). Designing and constructing mixed methods research (2nd ed.). Thousand Oaks, CA: Sage.

14.

Dobash

R. E.

Dobash

R. P.

(2015). When men murder women. Oxford, UK: Oxford University Press.

15.

Dobash

R. E.

Dobash

R. P.

Cavanagh

(2009). “Out of the blue”: Men who murder an intimate partner. Feminist Criminology, 4, 194-225.

16.

Dobash

R. E.

Dobash

R. P.

Cavanagh

Lewis

(2004). Not an ordinary killer—Just an ordinary guy: When men murder an intimate woman partner. Violence Against Women, 10, 577-605.

17.

Echeburùa

Fernandez-Montalvo

de Corral

Lopez-Goni

J. J.

(2009). Assessing risk markers in intimate partner femicide and severe violence: A new assessment instrument. Journal of Interpersonal Violence, 24, 925-939.

18.

Eliason

(2009). Murder-suicide: A review of the recent literature. Journal of the American Academy of Psychiatry and the Law, 37, 371-376.

19.

Emery

C. R.

(2011). Disorder or deviant order? Re-theorizing domestic violence in terms of order, power and legitimacy: A typology. Aggression and Violent Behavior, 16, 525-540.

20.

Funder

D. C.

(2006). Towards a resolution of the personality triad: Persons, situations, and behaviors. Journal of Research in Personality, 40, 21-34.

21.

Galta

Olsen

S. L.

Wik

(2010). Murder followed by suicide: Norwegian data and international literature. Nordic Journal of Psychiatry, 64, 397-401.

22.

Garcia

Hurwitz

E. L.

(2007). Homicides and intimate partner violence: A literature review. Trauma, Violence, & Abuse, 8, 370-383.

23.

Heckert

D. A.

Gondolf

E. W.

(2004). Battered women’s perceptions of risk versus risk factors and instruments in predicting repeat reassault. Journal of Interpersonal Violence, 19, 778-800.

24.

Holland

K. M.

Brown

S. V.

Hall

J. E.

Logan

J. E.

(2018). Circumstances preceding homicide-suicides involving child victims: A qualitative analysis. Journal of Interpersonal Violence, 33, 379-401. doi:10.1177/0886260515605124

25.

Knoll

J. L.

Hatters-Friedman

(2015). The homicide-suicide phenomenon: Findings of psychological autopsies. Journal of Forensic Sciences, 60, 1253-1257. doi:10.1111/1556-4029.12819

26.

Knoll

J. L. T.

(2016). Understanding homicide-suicide. Psychiatric Clinics of North America, 39, 633-647. doi:10.1016/j.psc.2016.07.009

27.

Kropp

P. R.

Hart

S. D.

(2000). The Spousal Assault Risk Assessment (SARA) Guide: Reliability and validity in adult male offenders. Law and Human Behavior, 24, 101-118.

28.

Kropp

P. R.

Hart

S. D.

(2015). Manual for the Spousal Assault Risk Assessment Guide, 3rd ed. (SARA-V3). Sydney, Australia: Proactive Resolutions.

29.

Kropp

P. R.

Hart

S. D.

Webster

C. D.

Eaves

(1995). Manual for the Spousal Assault Risk Assessment Guide, 2nd ed. Vancouver, Canada: British Columbia Institute on Family Violence.

30.

Liem

(2010). Homicide followed by suicide: A review. Aggression and Violent Behavior, 15, 153-161.

31.

Liem

Barber

Markwalder

Killias

Nieuwbeerta

(2011). Homicide-suicide and other violent deaths: An international comparison. Forensic Science International, 207(1-3), 70-76. doi:10.1016/j.forsciint.2010.09.003

32.

Liem

Roberts

D. W

. (2009). Intimate partner homicide by presence or absence of a self-destructive act. Homicide Studies, 13, 339-354.

33.

Magnusson

(1981). Towards a psychology of situations: An interactional perspective. Hillsdale, NJ: Lawrence Erlbaum.

34.

Malphurs

J. E.

Cohen

(2005). A statewide case-control study of spousal homicide-suicide in older persons. American Journal of Geriatric Psychiatry, 13, 211-217.

35.

Malterud

(2001). Qualitative research: standards, challenges, and guidelines. The Lancet, 358, 483-488.

36.

Malterud

(2012). Systemic text condensation: A strategy for qualitative analysis. Scandinavian Journal of Public Health, 40, 795-805.

37.

Marzuk

P. M.

Tardiff

Hirsch

C. S.

(1992). The epidemiology of murder-suicide. Journal of the American Medical Association, 267, 3179-3183.

38.

Messing

J. T.

Thaller

(2013). The average predictive validity of intimate partner violence risk assessment instruments. Journal of Interpersonal Violence, 28, 1537-1558.

39.

Norwegian Criminal Investigation Service. (2018). National statistics of murder (Norway). Drapsoversikt. Retrieved from https://www.politiet.no/globalassets/04-aktuelt-tall-og-fakta/drap/drapsoversikt-2018.pdf

40.

Pallant

(2010) SPSS Survival manual: A step by step guide to data analysis using SPSS (4th ed.). Berkshire, UK: Open University Press.

41.

Panczak

Geissbuhler

Zwahlen

Killias

Tal

Egger

(2013). Homicide-suicides compared to homicides and suicides: Systematic review and meta-analysis. Forensic Science International, 233(1-3), 28-36. doi:10.1016/j.forsciint.2013.08.017

42.

Salari

(2007). Patterns of intimate partner homicide suicide in later life: Strategies for prevention. Clinical Interventions in Aging, 2, 441-452.

43.

Salari

Sillito

C. L.

(2016). Intimate partner homicide-suicide: Perpetrator primary intent across young, middle, and elder adult age categories. Aggression and Violent Behavior, 26, 26-34. doi:10.1016/j.avb.2015.11.004

44.

Stöckl

Devries

Rotstein

Abrahams

Campbell

Watts

Moreno

C. G.

(2013). The global prevalence of intimate partner homicide: A systematic review. The Lancet, 382, 859-865. doi:10.1016/S0140-6736(13)61030-2

45.

Vatnar

S. K. B.

Friestad

Bjørkly

(2017a). Intimate partner homicide in Norway 1990–2012: Identifying risk factors through structured risk assessment, court documents, and interviews with bereaved. Psychology of Violence, 17, 395-405.

46.

Vatnar

S. K. B.

Friestad

Bjørkly

(2017b). Intimate partner homicide, immigration, and citizenship: Evidence from Norway 1990-2012. Journal of Scandinavian Studies in Criminology and Crime Prevention, 18(2), 103-122.

47.

Vatnar

S. K. B.

Friestad

Bjørkly

(2019). The influence of substance use on intimate partner homicide: Evidence from a Norwegian national 22-year cohort. International Journal of Forensic Mental Health. doi:10.1080/14999013.2018.1525777

48.

Winstok

(2007). Toward an interactional perspective on intimate partner violence. Aggression and Violent Behavior, 12, 348-363.

49.

Winstok

(2011). The paradigmatic cleavage on gender differences in partner violence perpetration and victimization. Aggression and Violent Behavior, 16, 303-311.