Gendered Entitlement or Generally Violent? Sociodemographic,Developmental,and Gender-Based Attitudinal Characteristics of Men Who Commit Homicide

Abstract

The study examined “gendered” and “general” factors associated with homicide using interviews with men convicted of murder or manslaughter in Australia. We compared men committing intimate partner femicide (IPF; n = 68) with men killing female non-intimate partners (MF; n = 44) and male non-intimate partners (MM; n = 135). They shared developmental and socio-economic characteristics. MM men reported extensive criminal histories and serious substance problems compared with IPF men. Many IPF men had perpetrated partner violence. Similarities existed across jealousy and marital role attitudes, though IPF men more likely condoned wife abuse and behaviorally controlled partners. Policies informed by complexity and diversity are important.

Keywords

intimate partner victim/offender relationship femicide gender criminal careers alcohol correlates drugs

One in seven homicides globally are perpetrated by intimate partners, with the likelihood of intimate partner homicide victimization higher for women than for men (Stöckl et al., 2013). In Australia, where the current study takes place, one in five homicides are perpetrated by intimate partners, with 75% of intimate partner homicide victims comprising women (Cussen & Bryant, 2015). Improving the knowledge base about intimate partner violence-related homicide is a key strategy for informing ongoing policy development aimed at better understanding risk factors and interventions, with the overarching goal of reducing lethal violence against women. Yet despite a growing body of research examining pathways to homicide, the literature remains divided on whether men who kill their female intimate partners should be considered a distinct “category” of individuals who commit homicide relative to those who kill, for example, friends/acquaintances or strangers (e.g., Decker, 1993; Dobash et al., 2004; Eriksson et al., 2019; Felson & Lane, 2010; Miethe & Regoeczi, 2004).

To inform prevention and intervention policies, it is crucial to understand the contributors to lethal violence. To date, however, empirical research into men’s lethal violence against female partners tends to draw on either a gendered perspective (emphasizing constructs such as entitlement, sexual jealousy, and gender role attitudes) or a general perspective of violence (emphasizing factors such as developmental experiences, low self-control, and criminal history). However, it is plausible that a combination of both gendered and general factors may be present, jointly contributing to men’s homicide offending.

Some scholars emphasize men’s violence against their intimate partners as a function of patriarchal gender structures, male entitlement, and proprietary attitudes (e.g., Daly & Chesney-Lind, 1988; Daly & Wilson, 1988; Dobash & Dobash, 1979; Heise & Kotsadam, 2015; Pence & Paymar, 1993; Taylor & Jasinski, 2011). Within this framework, both non-lethal and lethal intimate partner violence (IPV) against women have long been recognized as unquestionably gendered crimes. The underlying assumption is that men who kill their intimate partners differ to men who kill others in terms of certain attitudes and traits. Much empirical research supports this view. For example, data from the Murder in Britain study show significant psychosocial differences across the life course between men who kill intimate partners and men who kill other men (Dobash & Dobash, 2015; Dobash et al., 2004). These data reveal that men who kill other men are more likely to have problematic upbringings (e.g., a violent father), early contact with criminal justice agencies, and alcohol and drug problems, while men who kill intimate partners are more likely to display a history of violence perpetration in past relationships. Findings such as these support the view that men who kill their female intimate partners represent a subcategory or type of person who commits homicide.

Importantly though, while such factors as entitlement, sexual jealousy, and belief in traditional gender roles, which are known to be associated with intimate partner femicide (IPF), might be strongly “gendered,” a range of more “general” perpetrator characteristics, including a history of childhood abuse and/or neglect coupled with poor (or absent) parenting, compromised emotional functioning, negative/dysfunctional attachment style, socioeconomic disadvantage, and substance abuse, may also feature notably in the occurrence of IPF. In contrast to the gendered approach outlined above, these characteristics imply that men’s violence against female partners may not necessarily be an expression of gender and patriarchy, but rather may signal a propensity toward violence more generally. As argued by Felson (2006), the etiology of violence should remain the same irrespective of victim-offender relationship. There is some empirical support for this perspective. For example, using self-report data with inmates, Felson and Lane (2010) found few differences in terms of childhood abuse, official criminal history, and alcohol/drug use between those who had assaulted or killed their partners and those who had assaulted or killed others. The authors conclude that their findings appear more supportive of a generally violent perspective than a gendered perspective of violence.

The theoretical and empirical debate about how best to conceptualize men’s violence toward women has important policy implications, because each perspective suggests its own set of responses. For instance, the gendered perspective implies that prevention efforts centered around changing attitudes about gender and power structures may be beneficial in terms of reducing men’s violence against women. In contrast, the violence propensity perspective suggests that strategies to address a range of dysfunctional developmental and situational factors merit attention. Given that policy responses are typically constrained by limited resources and finite capacity, it is important to enhance knowledge in this field to guide effective decision-making and identify key areas for policy attention.

The current study begins to fill gaps in knowledge about gendered and general factors associated with men’s lethal violence by examining whether the combination of gendered and general factors present among men who commit homicide may vary depending on both the victim-offender relationship and the victim’s gender. Two research questions guide the study: (1) whether and how men who kill a female intimate partner differ from men who kill a female non-intimate (i.e., holding the victim’s gender constant, while the victim-offender relationship varies); and (2) whether and how men who kill a woman (irrespective of their relationship with that victim) differ from men who kill another man (in other words, considering differences that may be associated with the victim’s gender). Conceptualized this way, and considering the general perspective of violence, it would be predicted that all three groups (men who kill female intimates, men who kill female non-intimates, and men who kill other men) would share general characteristics associated with homicide and exhibit comparatively few differences on those factors. In terms of gendered factors, however, notable differences would be expected across gender and victim-offender groupings. Based on the gendered perspective, which emphasizes male/female power structures and dynamics, it may be argued that—irrespective of whether their relationship is intimate or not—gendered factors are inherently relevant to understanding any homicide involving a male perpetrator and female victim, but may simply be more prominent within intimate relationships. On that basis, it would be predicted that gendered factors would be the most relevant (in terms of prevalence or mean scores) among men who kill a female intimate, followed by men who kill female non-intimates, and that both groups should differ from men who kill other men.

Method

Context of the Study

The study consists of self-report data collected from homicide perpetrators in Australia through comprehensive face to face interviews. Australia has a relatively low homicide rate (0.9/100,000 in 2018–19; Bricknell & Doherty, 2021), and is on par with regions such as Western and Northern Europe, and much lower than the world average (6.1/100,000; United Nations Office on Drugs and Crime, 2019; see also Mazerolle et al., 2017). Although significant advances have been made in terms of gender equality in Australia over the last few decades, research still reveals disturbing patterns in terms of public attitudes toward gender equality and violence against women. For example, recent research shows that 34% of Australians believe it is natural for a man to want to appear in control of his partner in front of his male friends and 21% of Australians believe sometimes a woman can make a man so angry that he hits her even though he does not mean to do so (Webster et al., 2018). Such figures highlight the need to examine gender and violence in Australia further.

The set, setting, and contexts for homicide varies across geographic contexts, and yet many of the main correlates are consistent (see e.g., Daly & Wilson, 1988). The opportunity to collect meaningful information on various forms of homicide, and explore individual characteristics, pathways, precursors, and unique developmental and situational dimensions from a relatively large sample (for a study using self-report data collected from homicide offenders) illustrates the potential of this study to inform knowledge and raise understanding about homicide in a global context.

Data Source

The study used data from the Australian Homicide Project, which examines the causes of homicide and aims to improve understanding of pathways to homicide within the context of interventions by criminal justice, health, and social welfare agencies (Mazerolle et al., 2017). The project was approved by the Griffith University Human Research Ethics Committee. The full dataset consists of comprehensive interviews conducted between 2009 and 2013 with 302 individuals (262 men, 40 women) convicted of murder or manslaughter. The interviews were conducted at custodial and community correctional centers across Australia. Correctional staff provided eligible individuals with introduction packages a few weeks before data collection was due to commence, which contained information about the structure and aim of the research and expression of interest (EOI) forms. Individuals interested in the project put their EOIs into sealed envelopes, which were then returned to the research team. Research staff conducted the interviews after receiving extensive training in interview protocols and distress management procedures. Prior to commencement of the interview, participants were provided with an information sheet detailing the purpose of the study. The interviewers encouraged participants to ask questions about the interview and the nature of the study. Importantly, the interviewers made it clear to participants that some questions might be distressing, that their participation was voluntary, and that they had a right to withdraw their participation at any time without explanation. Those who wanted to proceed were provided with a consent form to sign. A modest amount of compensation was provided to respondents to compensate for their time in jurisdictions that allowed for it. The interviews were conducted face-to-face during which the interviewers read questions to the respondent and recorded their responses onto an interview schedule. Given the sensitive nature of the topic, a few of the respondents showed minor distress (mainly due to feelings of remorse). Actions taken by interviewers included skipping sections of the interview and, on limited occasions, notifying the prison psychologist (without revealing the trigger of the distress). The interviews lasted approximately 2 hours. On average, the participants were interviewed 12 years (SD = 8.2) after committing the homicide.

Variables

The interview schedule included a range of self-report measures to examine individual, sociodemographic, and developmental characteristics of individuals who commit homicide, as well as the extent and nature of their criminal history and attitudes about intimate relationships.

Type of homicide

Respondents self-reported their relationship with the person they killed, as well as gender of the victim. We used these data to refine the sample into three groups. Men who had committed intimate partner femicide (IPF) by killing a current or former female intimate partner (IPF; n = 68) were compared with men who had killed female non-intimate partners (MF; n = 44; such as a friend, neighbor, mother, stranger) and men who had killed a male non-intimate partner (MM; n = 135; such as a friend, employer, uncle, stranger). A few of the homicides involved more than one victim (IPF: n = 7; MF: n = 3; MM: n = 6). In these instances, we coded for the victim who was killed first, except for three cases of IPF where it was clear that the (-ex) intimate partner was the primary target, even though she was killed after other victims (e.g., new partner, children). The MF and MM groups excluded filicides, victims under the age of 15 and same-sex relationships. Although we recognize the importance of understanding the dynamics of these types of homicides, research suggests that filicides (e.g., Eriksson et al., 2016), child homicides (e.g., Dearden & Jones, 2008), and same-sex intimate partner homicide (e.g., Mize & Shackelford, 2008) display distinctive characteristics.

Of the IPF respondents, the majority (38%) were legally married to their victim, 29% were in a de facto or dating relationship with the victim, and 33% were separated from the victim at the time of the homicide. The relationship status for the MF and MM respondents differed extensively from that of the IPF respondents, with the majority in de facto or dating relationships (MF = 46%; MM = 52%), very few legally married (MF = 16%; MM = 7%), and the remainder not in a relationship at the time of the homicide (MF = 39%; MM = 42%).

Sociodemographic and developmental characteristics

Sociodemographics

A range of sociodemographic indicators were considered: (1) respondent age at the time of the homicide; (2) whether the respondent came from a non-English speaking background (NESB); (3) whether the respondent had completed high school; (4) whether the respondent received Commonwealth benefit (“welfare”) payments in the year prior to the homicide; (5) presence of financial stress (respondent unable to pay their bills in the year prior to the homicide); and (6) whether the respondent was unemployed at the time of the offense.

Adverse childhood experiences

Respondents were asked about their adverse childhood experiences using individual items from Stewart et al.’s (1987) scale examining childhood experiences. We did not summate the full scale to create a total score. Instead, we examined the questions separately to assess various forms of abuse and neglect in childhood, including exposure to physical violence, sexual abuse, and physical neglect, as well as observing parental violence. All items were originally measured using a Likert scale ranging from 1 (Never) to 5 (Very often), but were dichotomized such that “never” indicated “no exposure to abuse/neglect” and all other responses (e.g., “sometimes”) indicated “exposure to abuse/neglect.” Attachment styles to caregivers in childhood were assessed using the brief self-report Parental Caregiving Style Questionnaire (Hazan & Shaver, 1987). The scale measures attachment styles to a respondent’s mother and father separately (or alternative caregivers if mother/father absent). The scale contains three descriptions of parental caregiving style that correspond with Ainsworth’s (1978) childhood attachment styles: (1) warm/responsive (secure), (2) cold/rejecting (avoidant), and (3) ambivalent/inconsistent (anxious-ambivalent). The avoidant and anxious-ambivalent styles were combined to create dichotomous variables measuring insecure attachment to maternal caregiver and insecure attachment to paternal caregiver.

Alcohol problems

Respondents’ alcohol problems in the year prior to the homicide were measured through the 10-item Alcohol Use Disorder Identification Test (AUDIT; Babor et al., 2001). Scores on this test range from 0 to 40. Cronbach’s alpha in the current study was .87. The continuous scores were recoded into a categorical measure as per AUDIT guidelines (0–7 = no problems; 8–15 = low/medium problems; and 16–40 = high problems).

Drug problems

Respondents’ drug problems in the year prior to the homicide were measured using the 10-item Drug Abuse Screening Test, Short Form (DAST-10; Skinner, 1982). Scores on this test range from 0 to 10. Cronbach’s alpha in the current study was .61. The continuous scores were recoded into a categorical measure as per DAST-10 guidelines (0 = no problems; 1–5 = low/moderate problems; 6–10 = substantial/severe problems).

Criminal history

Type of criminal history

Respondents were asked about lifetime engagement in four categories of offending (not including homicide incident): (1) assault (gang fights; assault incl. threatened and sexual; serious assault/serious injury intent); (2) robbery (strong-armed robbery and aggravated robbery with a weapon); (3) theft (theft of item(s) worth <A$50; theft of item(s) worth >A$50); illegal use of motor vehicle; vehicle theft; possession of stolen goods); and (4) illicit drug selling (trafficking/selling any drug incl., hard drugs).

Intimate partner violence

We created two composite variables to measure violence against (ex) partners in the year prior to the homicide and ever used intimate partner violence (against any partner). The first variable was created by combining the Physical Assault annual prevalence (CTAP) subscale of the Revised Conflict Tactics Scale (CTS2; Straus et al., 2003) with a separate question about whether the respondent had been violent toward their (ex) partner in the year prior to the homicide. The correlation between these two variables was r = .33 (p < .001). The second variable was created by combining the Physical Assault ever prevalence (CTAE) subscale of the CTS2 with three questions about whether respondents had ever been physically or sexually violent toward previous partners and a separate question about whether the respondent had been violent toward their (ex) partner in the year prior to the homicide (same as above). Cronbach’s alpha for this variable was .68. Both variables were dichotomized. Individuals who reported they had never been in a romantic relationship (22.7% of the MF group and 11.9% of the MM group) were categorized as not having engaged in violence against intimate partners.

Early offending onset

Respondents who indicated that they had a history of criminal behavior were asked how old they were when those behaviors commenced, with offending categorized in a dummy variable as “early” if onset occurred before 13 years of age and “late” if onset occurred after the age of 13.

Offending versatility

To measure versatility, we counted the number of different offense types in which the respondent had participated: assault, robbery, theft, illicit drugs, and domestic violence against any partner. The total score represents the versatility, ranging from 0 (respondents with no history of offending) to 5 (respondents who had participated in all five offense types at some point during their life).

Criminal justice system contact

Respondents were also asked to indicate whether, in the year prior to the homicide incident, they had been arrested or in prison (both dichotomous).

Attitudinal measures

Attitudes toward marital roles

We measured attitudes towards marital roles using the Gender-based Attitudes Toward Marital Roles scale (GATMR; Hoffman & Kloska, 1995). The six items (e.g., “Men should make the really important decisions in the family,” “A man should help in the house, but housework and childcare should mainly be a woman’s job”) were scored on a Likert scale ranging from strongly agree (1) to strongly disagree (4). Higher scores indicated lower levels of gender role stereotyping. Cronbach’s alpha indicated good reliability (.88).

Sexual jealousy

We used an adapted version of the eight-item Jealousy-provoking Situation Scale (Salovey & Rodin, 1988) to measure sexual jealousy, modified to assess how upset a range of situations (e.g., “Someone is flirting with your partner,” “Your partner talks about an old lover”) would make respondent feel. Responses were scored on a Likert scale ranging from not at all (1) to very (4). Higher scores indicate higher levels of jealousy. Cronbach’s alpha indicated good reliability (.84).

Relational entitlement

The 28-item Relational Entitlement and Proprietariness Scale (REPS; Hannawa et al., 2006) was used to measure relational entitlement. The REPS contains three subscales; Behavioral Control (e.g., “If I can’t have my partner, no one can”), Social Control (e.g., “I insist on knowing where my partner is at all times”), and Information Control (e.g., “I look through my partner’s drawers, handbag, or pockets”). Items were scored on a Likert scale ranging from strongly disagree (1) to strongly agree (7). Higher scores indicate higher relational entitlement. Cronbach’s alpha indicated good reliability for the overall scale (.96) and the three sub-scales (Behavioral Control = .89, Social Control = .93, and Information Control = .89).

Attitudes toward wife abuse

Respondents’ attitudes toward using violence within intimate relationships were assessed using the 14-item Revised Attitudes toward Wife Abuse scale (RAWA; Yoshioka et al., 2001). The RAWA contains three subscales; Endorsement of Male Privileges (e.g., “A husband should have the right to discipline his wife”), Perceived Lack of Alternatives (e.g., “A wife should move out of the home if her husband hits her”) (reverse coded), and Situation-Specific Approval of Violence (e.g., “A husband has the right to hit his wife if she nags him too much”). Responses are scored on a Likert scale ranging from strongly disagree (1) to strongly agree (6). Higher scores indicate higher violence-supportive attitudes. Cronbach’s alpha indicated good reliability for the overall scale (.83) and two of the sub-scales (Perceived Lack of Alternatives = .70, Situation-Specific Approval of Violence = .87), while the other sub-scale (Endorsement of Male Privileges displayed a lower Cronbach’s alpha (.52).¹

Analyses

Homicide in Australia, as is the case in many developed nations, is a rare event. Further, not all homicide offenders are convicted, and, of those who are, fewer still agree to participate in self-report studies such as this research project. Consequently, the analytical decisions were determined on the basis of the sample size available for the study. Chiefly, this dictated adopting a descriptive focus, along with the use of non-parametric bivariate analyses (given that achieving sufficient statistical power for reliable multivariate analyses would have required a larger sample size, particularly in the MF group). These decisions reflect methods adopted in prior published research on homicide (e.g., Dobash et al., 2009) as well as widely accepted principles for analysis of modest sample sizes and non-parametric data (e.g., Mackridge & Rowe, 2018).

Given these considerations, non-parametric bivariate tests were used to identify group differences across the variables of interest. Either a Chi-Square test or Kruskal-Wallis test was conducted, and this was determined by the attribute (categorical or continuous) of the dependent variable (e.g., sociodemographics, substance problems, offending history, and attitudes) but not the grouping variable (IPH, MF, MM; always categorical). As the non-parametric equivalent of the one-way independent ANOVA, Kruskal-Wallis is based on ranked scores (rather than the actual scores) and calculates whether the mean rank of a given continuous variable differs across the grouping variable (in this case, homicide type) (Kraska-Miller, 2014). Missing data were below 15% for all variables except observing parental violence (18.6%), insecure attachment (maternal = 23.1%; paternal = 26.7%), sexual jealousy (17.0%), relational entitlement (overall = 21.5%; behavioral control = 21.5%; social control = 21.9%; information control = 21.5%), and attitudes toward wife abuse situation-specific approval sub-scale (15.4%). Missing values were not imputed. Pair-wise differences between groups (i.e., IPF vs. MF, MF vs. MM, and IPF vs. MM) were assessed using column proportion comparison and post-hoc multiple comparison (Bonferroni corrected) follow up tests for Chi-Square and Kruskal-Wallis analyses, respectively.

Results

Individual, Sociodemographic, and Developmental Characteristics

The groups were generally similar across a range of individual, demographic, and developmental characteristics, including education and employment status and childhood experiences (see Table 1). The men in the IPF group were more likely to be older at time of the offense and to come from a non-English speaking background. However, they were less likely to have been on welfare and to report severe substance problems compared with the MM group in the year preceding the homicide. On these indicators, the men in the MF group were situated at the midway point between the IPF and MM groups, either displaying similarities with the MM group (age, NESB), the IPF group (welfare payments), or both (substance problems). Table 1 provides further detail.

Table 1.

Individual, Sociodemographic, and Developmental Characteristics Among Intimate Partner Femicide (IPF), Male-on-Female (MF) Homicide, and Male-on-Male (MM) Homicide.

	Full sample (n = 247)	IPF (n = 68)	MF (n = 44)	MM (n = 135)	Test statistic	p
	% or M (SD)	% or M (SD)	% or M (SD)	% or M (SD)	Test statistic	p
Sociodemographics
Age at homicide	30.6 (10.3)	38.3 (11.0)^a	28.6 (10.4)^b	27.5 (7.7)^b	44.54	.000
NESB	14.2	25.0^a	6.8^b	11.1^b	9.55	.008
Did not complete high school	67.9	62.5	61.4	72.7	3.13	.209
On welfare payments	46.6	29.8^a	29.3^a	60.2^b	20.49	.000
Financial stress	22.0	28.1	29.3	16.8	4.46	.108
Unemployed	27.5	19.1	29.5	31.1	3.37	.186
Adverse childhood experiences
Physical violence	58.9	48.3	65.1	61.7	3.87	.144
Sexual abuse	18.9	27.6	13.6	16.8	4.00	.135
Physical neglect	35.3	34.5	43.2	32.8	1.55	.460
Observing parental violence	47.3	42.0	47.4	49.6	0.80	.672
Insecure attachment (maternal)	31.6	23.9	31.4	34.9	1.80	.408
Insecure attachment (paternal)	50.8	44.4	65.6	49.0	3.67	.160
Alcohol problems
None	36.5	55.4^a	32.5^a,b	29.3^b	15.49	.004
Low-medium	25.1	25.0^a	30.0^a	23.6^a
High	38.4	19.6^a	37.5^a,b	47.2^b
Drug problems
None	38.8	59.3^a	50.0^a	25.0^b	22.87	.000
Medium	30.8	22.0^a	25.0^a	37.1^a
Substantial	30.4	18.6^a	25.0^a,b	37.9^b

Note. Pair-wise differences between groups were assessed using column proportion comparison and post-hoc multiple comparison (Bonferroni corrected) follow up tests. Each superscript letter denotes a subset of homicide categories (IPF, MF, MM) whose column proportions do not differ significantly from each other at the .05 level. Percentages are presented for the categorical variables. Mean and standard deviation is presented for the continuous variable (respondent age at time of homicide).

Criminal History

As seen in Table 2, all criminal history variables showed significant differences across the three groups except for one (early onset of offending). Here the MF group was generally positioned midway between the IPF and MM groups on most criminal history indicators (except for IPV where they showed the lowest prevalence rates). The MM group showed higher prevalence rates of engaging in assault, robbery, theft, and illicit drug selling compared with the IPF and/or MF groups. The MM group had also engaged in a higher number of offense types (versatility) compared with the IPF and MF groups. Over one-third of the MM group had been arrested in the year leading up to the homicide, a statistically significantly higher proportion compared with the one-tenth arrested within the IPF group. The only criminal history indicator where the IPF group reported the highest prevalence rate was partner violence (statistically higher than the MM group). Nearly half of the men in the IPF group reported IPV against the victim they killed.

Table 2.

Criminal History and Past Violence Within Intimate Relationships Among Intimate Partner Femicide (IPF), Male-on-Female (MF) Homicide, and Male-on-Male (MM) Homicide.

	Full sample (n = 247)	IPF (n = 68)	MF (n = 44)	MM (n = 135)	Test statistic	p
	% or M (SD)	% or M (SD)	% or M (SD)	% or M (SD)	Test statistic	p
Early offending onset (<13)	56.7	43.8	63.4	59.5	4.42	.110
Offending history (ever)
Assault	80.5	65.6^a	72.7^a	90.2^b	18.73	.000
Robbery	25.3	8.1^a	20.5^a,b	35.1^b	16.96	.000
Theft	77.6	53.1^a	86.4^b	86.5^b	30.01	.000
Illicit drug selling	47.5	25.0^a	36.4^a	62.1^b	26.50	.000
Partner violence
Violent to (ex-)partner (year prior)	34.2	44.6^a	18.4^b	34.4^a,b	6.92	.032
Ever violent (past or current partner)	45.3	60.3^a	30.0^b	43.2^a,b	9.30	.010
Versatility (# offense categories)	2.8 (1.3)	2.2 (1.4)^a	2.4 (1.2)^a	3.2 (1.2)^b	13.64	.000
CJS contact (year prior)
Arrested	29.6	10.5^a	28.6^a,b	38.6^b	14.88	.001
In prison	15.2	6.8^a	11.9^a	20.2^a	6.05	.049

Attitudinal Measures

No significant differences were observed between the three groups on attitudes toward marital roles and sexual jealousy (see Table 3). In terms of relational entitlement, only one of the subscales showed significant differences. Here the IPF and MF groups both showed higher levels of relational entitlement compared with the MM group. The scale with the most differences was that measuring attitudes toward wife abuse. Here the IPF and MF groups both scored statistically higher compared with the MM group on the overall scale. The IPF group also scored statistically higher than the MM group in terms of expressing lack of alternatives (e.g., not agreeing that wife beating is grounds for divorce) and situation-specific approvals of wife beating (e.g., agreeing that a man has a right to hit his wife if she refuses to have sex with him). The MF group scored statistically higher than the MM group (but not the IPF group) on the endorsement of male privilege subscale (which includes items such as “a man is the ruler of his home”).

Table 3.

Scores on Attitudinal Measures Among Intimate Partner Femicide (IPF), Male-on-Female (MF) Homicide, and Male-on-Male (MM) Homicide.

	Full sample (n = 247)	IPF (n = 68)	MF (n = 44)	MM (n = 135)	Test statistic	p
	M (SD)	M (SD)	M (SD)	M (SD)	Test statistic	p
Attitudes toward marital roles	2.0 (0.6)	2.1 (0.5)	2.1 (0.6)	2.0 (0.6)	2.04	.360
Sexual jealousy	2.1 (0.7)	2.3 (0.7)	2.1 (0.7)	2.1 (0.6)	1.67	.433
Relational entitlement
Overall	1.9 (0.9)	2.1 (1.0)	2.1 (1.1)	1.8 (0.8)	4.89	.087
Behavioral control	1.7 (0.8)	1.8 (0.9)^a	1.9 (1.1)^a	1.5 (0.6)^b	11.26	.004
Social control	2.2 (1.3)	2.4 (1.5)	2.4 (1.3)	2.1 (1.2)	2.33	.312
Information control	2.0 (1.0)	2.0 (1.0)	2.2 (1.1)	1.9 (0.9)	3.54	.171
Attitudes toward wife abuse
Overall	1.7 (0.6)	1.9 (0.5)^a	1.9 (0.6)^a	1.6 (0.5)^b	13.24	.001
Endorsement of male privilege	1.9 (0.7)	1.9 (0.7)^a,b	2.1 (0.7)^a	1.8 (0.8)^b	10.10	.006
Lack of alternatives	1.8 (0.8)	2.0 (0.8)^a	1.9 (0.7)^a,b	1.7 (0.8)^b	7.94	.019
Situation-specific approval	1.6 (0.6)	1.9 (0.6)^a	1.7 (0.6)^a,b	1.5 (0.5)^b	18.42	.000

Discussion

Gender represents a focal concern in the debate about causes of violence. Some scholars argue that men who are violent toward intimate partners share characteristics with men who are violent in other settings (e.g., Felson, 2006). This perspective may be contrasted with the argument that men who perpetrate violence in intimate relationships are distinctly different to other violent men, particularly in areas such as patriarchal power structures, male feelings of entitlement, proprietary attitudes, and a need to control aspects of their partner’s life (e.g., Daly & Wilson, 1988; Dobash & Dobash, 2015; Heise & Kotsadam, 2015). To date, much empirical research into men’s lethal violence against female partners has tended to draw on either a gendered perspective (emphasizing constructs such as entitlement, sexual jealousy, and gender role attitudes) or a general perspective of violence (emphasizing factors such as developmental experiences and criminal history).

It is reasonable to suggest that the combination of gendered and general factors present among those who commit homicide may vary depending on both the victim-offender relationship and the victim’s gender. This study therefore focused on three distinct groups: men who commit intimate partner femicide (IPF), men who kill female non-intimates (MF), and men who kill other men (MM). A general violence perspective would predict that all three groups would share characteristics associated with homicide and exhibit comparatively few differences on those factors. Based on a gendered perspective, it may be argued that gendered factors are inherently relevant to understanding any homicide involving a male perpetrator and female victim, but may simply be more prominent within intimate relationships.

After examining a diverse range of factors, it appears that both the gendered and the general perspectives of intimate partner femicide offending offer valuable insights. In support of a general perspective were the findings revealing similarities in developmental factors such as childhood experiences and attachment style to parents. Here the three groups tended to be quite similar, reporting high levels of abusive and neglectful experiences. The groups also showed similarities across a number of socio-economic factors, such as limited education and financial stress in the year prior to the homicide.

In contrast, and in line with a more gendered perspective, are the findings that the MM group reported extensive problems across a variety of criminal history indicators and presence of serious problems with drugs and alcohol compared with the IPF group (with the MF group located in between). Indeed, quite high proportions of the MF and MM groups had engaged in violent acts such as assault (broadly speaking, not specifically against intimate partners) and robbery prior to committing homicide. Thus, they “arrive” at the homicide scene (so to speak) with extensive “experiences” committing violent acts. Importantly, and consistent with substantial past research (e.g., Dobash & Dobash, 2015; Dobash et al., 2004), a considerable proportion of the IPF group reported using violence against an intimate partner in the year before the homicide, as well as lifetime use of violence against intimate partners. Their prevalence rates, however, did not differ statistically from the MM group (a substantial proportion of whom also reported partner violence), but only from the MF group. Although the IPF group reported the highest prevalence rates of engaging in past intimate partner violence, this was also a relatively common pattern among MF and MM (30.0% and 43.2% had ever committed intimate partner violence). Together, these findings align with other research, which suggests that homicide offenders, especially those who commit homicide outside of family/partner relationships, have extensive criminal histories and previous contact with criminal justice agencies prior to committing the homicide (Dobash et al., 2004; Eriksson et al., 2019; Kivivuori & Lehti, 2012; Thomas et al., 2011).

Nevertheless, in contrast to what gendered perspectives of IPF would predict, the groups did not differ in terms of levels of sexual jealousy and attitudes around marital roles. While the measure of sexual jealousy was developed specifically for the current study, prior research on the GATMR scale (used to measure attitudes toward marital roles in the current study) reveals that the three groups of men who kill examined in the current study displayed much higher levels of gender-role traditionalism (in other words, less egalitarian attitudes) compared with a normative sample of fathers in the United States (Hoffman & Kloska, 1995). This suggests that while no differences were observed within the current sample, men who commit lethal violence potentially display much greater levels of traditional attitudes compared with the general population (irrespective of who they kill).

Other attitudinal measures revealed differences between the groups in line with a gendered perspective. On measures of attitudes to wife abuse, the results support the hypothesis that gendered factors would be more noticeable among men who commit IPF compared with men who kill other men (MM), particularly on questions measuring perceived lack of alternatives to violence (e.g., a husband should not be arrested for hitting his wife) and beliefs about male violence being justified in certain situations (e.g., when a partner refuses sexual advances). In terms of relational entitlement, both the IPF and MF groups scored significantly higher than MM on measures of behavioral control (e.g., feeling entitled to treat their partner however they choose and wanting the final say in disagreements). Such overt control tactics are used to monitor victims and compel their obedience, and are central to the concept of coercive control (Stark, 2007). Nevertheless, the groups did not differ in terms of social and informational control. It is not clear why behavioral aspects of control emerged as significant, but social and information aspects did not. Potentially, it may simply be due to the sample sizes involved. However, the result may also suggest that some of the “subtler” behaviors that characterize coercive control—which are picked up in the social and informational control measures—may not be as specifically associated with IPF as previously assumed. The result may, instead, indicate that overt behavioral control measures may be especially relevant in understanding the phenomenon of men killing women (irrespective of whether an intimate relationship exists). Given the exploratory nature of the present study, we caution against drawing any strong conclusions at this stage, and highlight the need for undertaking more detailed research into this specific finding. Interestingly, however, comparisons with the mean scores of a normative sample of undergraduate students from a completely different study reveal that the current sample of men convicted of murder or manslaughter scored higher (higher scores indicate more controlling behavior) in terms of behavioral and social control, while the undergraduate sample actually scored higher on information control, which includes behavior such as paying their partner a surprise visit to see with whom they might be socializing (Hannawa et al., 2006). Further studies are required to ascertain the relative normalcy of such behavior within the general population.

Overall, these results appear to indicate that a combination of gendered and general factors offer valuable insights that can enrich our understanding of IPF. This inference implies that a variety of factors and circumstances would be relevant for policy and program development. For instance, the results show that while men who commit IPF may be particularly likely to justify the use of intimate partner violence, they also share attitudes about gender-role traditionalism, as well as adverse experiences in childhood and low socio-economic status, with other men who commit homicide. A particularly fruitful direction for future research would be to examine the possibility of specific subtypes within different victim-offender relationship and victim gender categories. For example, it is possible that different offender typologies or clusters of characteristics may exist within each type of victim-offender relationship. Such work is much more common within non-lethal intimate partner violence research (e.g., Holtzworth-Munroe & Stuart, 1994; Johnson, 2008) but is gaining traction within the homicide literature (e.g., Dixon et al., 2008; Elisha et al., 2010). Such efforts may reveal certain typologies or clusters of men who commit homicide that emerge independently of the relationship between victim and offender or of victim gender, knowledge which would, in turn, inform the development of tailored and nuanced policy responses to homicide (whether between intimate partners, or otherwise).

A number of limitations must be acknowledged within the present work. As is the case with many published studies of homicide using self-report data, the sample size available for this work was relatively modest, which constrains the likelihood of detecting significant effects (in other words, it reduces statistical power). As such, the sample sizes for each group (particularly, the MF group) should be considered when interpreting, and/or seeking to generalize, the results. Further, the sample sizes, in conjunction with the novel focus of this work, dictated the decision to undertake bivariate analyses only. Performing multivariate analyses—ideally, on a larger sample—represents a fruitful area for future examination. For example, it would be beneficial to extend the current findings by undertaking multivariate analysis (controlling for factors such as age, non-English speaking background, and employment status) to determine whether differences observed between the three groups persisted when controlling for demographic characteristics. Considering factors such as different groups’ past involvement in intimate relationships, as well as the intensity and duration of those relationships, would also be valuable. Multivariate analyses would also reveal which variables uniquely contribute to the model, and which predictors are the strongest, whilst controlling for all other variables. This would allow for a more direct comparison of the general and gendered perspectives. In addition, as this work did not directly compare the sample of men who had committed homicide with either a non-lethal violent offending sample or a normal population sample, inferences cannot be drawn about “risk” factors for intimate partner (or, indeed, other) homicide offending. As such, due interpretive caution should be applied.

In summary, the findings from this study provide partial support for both the gendered and the general perspectives of IPF. Overall, this work indicates the value of seeking comprehensive, multi-dimensional theoretical and empirical approaches to understanding IPF. The results also suggest that theories of homicide which place the victim-offender relationship at their foundation may require far more nuance in order to capture diversity within victim-offender relationship/victim gender categories, as well as similarities across different relationship/gender categories. While considerably more empirical work is required to provide a truly comprehensive understanding of IPF, the present research indicates that it would be prudent to develop policy responses that are informed by complexity and multiple different perspectives and understandings of offending.

Footnotes

Acknowledgements

We wish to acknowledge the assistance of the various correctional departments across Australia for their significant support in the conduct of this research. The views expressed in this report are those of the authors and do not represent the policies or views of the correctional departments.

Authors’ Note

The views expressed in this report are those of the authors and do not represent the policies or views of the correctional departments.

Samara McPhedran is now affiliated to University of Queensland, St Lucia, Australia.

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) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported under the Australian Research Council’s Discovery Projects funding scheme (DP0878364).

ORCID iD

Li Eriksson

Notes

Author Biographies

Dr Li Eriksson is a Senior Lecturer at the School of Criminology and Criminal Justice at Griffith University and member of the Griffith Criminology Institute. Her research examines developmental and contextual predictors of lethal and non-lethal violence, including the role of adverse childhood experiences, offending history, and substance use.

Dr Samara McPhedran is Honorary Associate Professor at the TC Beirne School of Law, University of Queensland. Her research focuses on mental health, gender-based and other violence, homicide, suicide, and gun control. She has worked extensively in legislative and policy evaluation in criminology and public health.

Professor Paul Mazerolle is President and Vice Chancellor of the University of New Brunswick. His research examines processes that shape offending behaviour across the life-course. His primary focus is in building knowledge on violence to inform theories, advance understanding, and improve policy and practices to reduce or prevent violence.

Professor Richard Wortley is Director of the Jill Dando Institute of Security and Crime Science at University College London. His research interests centre on the role immediate environments play in criminal and antisocial behaviour, and the implications this has for situational crime prevention.

References

Ainsworth

M. D.

(1978). Patterns of attachment: A psychological study of the strange situation. Lawrence Erlbaum Associates.

Babor

T. F.

Higgins-Biddle

J. C.

Saunders

J. B.

Monteiro

M. G.

(2001). AUDIT: The alcohol use disorder identification test: Guidelines for use in primary care (2nd ed.). World Health Organization.

Bricknell

Doherty

(2021). Homicide in Australia 2018-19: Statistical report 34. Australian Institute of Criminology.

Cussen

Bryant

(2015). Domestic/family homicide in Australia. Research in Practice, 38, 1–7.

Daly

Chesney-Lind

(1988). Feminism and criminology. Justice Quarterly, 5(4), 497–538.

Daly

Wilson

(1988). Homicide. Aldine De Gruyter.

Dearden

Jones

(2008). Homicide in Australia: 2006-2007 National Homicide Monitoring Program (NHMP) annual report (AIC reports: Monitoring reports, Issue). Australian Institute of Criminology.

Decker

S. H.

(1993). Exploring victim-offender relationships in homicide: The role of individual and event characteristics. Justice Quarterly, 10(4), 585–612.

Dixon

Hamilton-Giachritsis

Browne

(2008). Classifying partner femicide. Journal of Interpersonal Violence, 23(1), 74–93.

10.

Dobash

R. E.

Dobash

(2015). When men murder women. Oxford University Press.

11.

Dobash

R. E.

Dobash

R. P.

(1979). Violence against wives: A case against the patriarchy. Free Press.

12.

Dobash

R. E.

Dobash

R. P.

Cavanagh

(2009). “Out of the blue”: Men who murder an intimate partner. Feminist Criminology, 4(3), 194–225.

13.

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(6), 577–605.

14.

Elisha

Idisis

Timor

Addad

(2010). Typology of intimate partner homicide: Personal, interpersonal, and environmental characteristics of men who murdered their female intimate partner. International Journal of Offender Therapy and Comparative Criminology, 54(4), 494–516.

15.

Eriksson

Mazerolle

Wortley

Johnson

(2016). Maternal and paternal filicide: Case studies from the Australian Homicide Project. Child Abuse Review, 25(1), 17–30.

16.

Eriksson

Mazerolle

Wortley

Johnson

McPhedran

(2019). The offending histories of homicide offenders: Are men who kill intimate partners distinct from men who kill other men? Psychology of Violence, 9(4), 471–480.

17.

Felson

R. B.

(2006). Is violence against women about women or about violence? Contexts, 5(2), 21–25.

18.

Felson

R. B.

Lane

K. J.

(2010). Does violence involving women and intimate partners have a special etiology? Criminology, 48(1), 321–338.

19.

Hannawa

A. F.

Spitzberg

B. H.

Wiering

Teranishi

(2006). “If I can’t have you, no one can”: Development of a Relational Entitlement and Proprietariness Scale (REPS). Violence and Victims, 21(5), 539–560.

20.

Hazan

Shaver

(1987). Romantic love conceptualized as an attachment process. Journal of Personality and Social Psychology, 52(3), 511–524.

21.

Heise

L. L.

Kotsadam

(2015). Cross-national and multilevel correlates of partner violence: An analysis of data from population-based surveys. The Lancet Global Health, 3(6), e332–e340.

22.

Hoffman

L. W.

Kloska

D. D.

(1995). Parents’ gender-based attitudes toward marital roles and child rearing: Development and validation of new measures. Sex Roles, 32(5–6), 273–295.

23.

Holtzworth-Munroe

Stuart

G. L.

(1994). Typologies of male batterers: Three subtypes and the differences among them. Psychological Bulletin, 116(3), 476–497.

24.

Johnson

M. P.

(2008). A typology of domestic violence: Intimate terrorism, violent resistance, and situational couple violence. Northeastern University Press.

25.

Kivivuori

Lehti

(2012). Social correlates of intimate partner homicide in Finland: Distinct or shared with other homicide types? Homicide Studies, 16(1), 60–77.

26.

Kraska-Miller

(2014). Nonparametric statistics for social and behavioral sciences. CRC Press.

27.

Mackridge

Rowe

(2018). A practical approach to using statistics in health research: From planning to reporting. John Wiley & Sons.

28.

Mazerolle

Eriksson

Wortley

Johnson

(2017). Homicide in Australia and New Zealand: Precursors and prevention. In Brookman

Maguire

E. R.

Maguire

(Eds.), International handbook on homicide (pp. 412–431). John Wiley & Sons.

29.

Miethe

T. D.

Regoeczi

W. C.

(2004). Rethinking homicide: Exploring the structure and process underlying deadly situations. Cambridge University Press.

30.

Mize

K. D.

Shackelford

T. K.

(2008). Intimate partner homicide methods in heterosexual, gay, and lesbian relationships. Violence and Victims, 23(1), 98–114.

31.

Pence

Paymar

(1993). Education groups for men who batter: The Duluth model. Springer.

32.

Salovey

Rodin

(1988). Coping with envy and jealousy. Journal of Social and Clinical Psychology, 7(1), 15–33.

33.

Skinner

H. A.

(1982). The drug abuse screening test. Addictive Behaviors, 7(4), 363–371.

34.

Stark

(2007). Coercive control: How men entrap women in personal life. Oxford University Press.

35.

Stewart

Senger

M. M.

Kallen

Scheurer

(1987). Family violence in stable middle-class homes. Social Work, 32(6), 529–531.

36.

Stöckl

Devries

Rotstein

Abrahams

Campbell

Watts

Moreno

C. G.

(2013). The global prevalence of intimate partner homicide: A systematic review. Lancet, 382(9895), 859–865.

37.

Straus

M. A.

Hamby

S. L.

Warren

W. L.

(2003). Conflict tactic scales handbook. Western Psychological Services.

38.

Taylor

Jasinski

J. L.

(2011). Femicide and the feminist perspective. Homicide Studies, 15(4), 341–362.

39.

Thomas

K. A.

Dichter

M. E.

Matejkowski

(2011). Intimate versus nonintimate partner murder. Homicide Studies, 15(3), 291–311.

40.

United Nations Office on Drugs and Crime. (2019). Global study on homicide. United Nations Office on Drugs and Crime.

41.

Webster

Diemer

Honey

Mannix

Mickle

Morgan

Parkes

Politoff

Powell

Stubbs

Ward

(2018). Australians’ attitudes to violence against women and gender equality: Findings from the 2017 National community attitudes towards violence against women survey (NCAS). Australia’s National Research Organisation for Women’s Safety (ANROWS).

42.

Yoshioka

M. R.

DiNoia

Ullah

(2001). Attitudes toward marital violence: An examination of four Asian communities. Violence Against Women, 7(8), 900–926.