Sexual homicide: a descriptive analysis of demographics, behaviour, and body disposal in New Zealand

ABSTRACT

As a rare and complex phenomenon, the study of sexual homicide continues to be understudied in international contexts. This exploratory study sought to address this gap, utilising a dataset of 26 sexual homicide incidents which took place in New Zealand between 1968 and 2022, sourced from both judicial sentencing reports and a Police database. Descriptive analyses revealed men were the most common perpetrators, with females being the predominant victims. The act of moving or concealing the body emerged as a prevalent behaviour. Significant variation was observed between incidents, with each involving unique and distinctive sexual behaviours, further emphasising the challenges in understanding this complex phenomenon. By shedding light on both the commonalities and distinctive features across sexual homicides in a unique sample, this research highlights the heterogeneity among homicide offences of this nature, and provides insight to support the prevention, investigation and intervention with sexual homicide in both a local and global context.

PRACTICE IMPACT STATEMENT

This study sheds light on the dynamics of sexual homicide in New Zealand, highlighting common trends and unique behaviours among perpetrators and victims. By informing more targeted investigation strategies, it contributes to the collective aim of improving responses to sexual violence at both a local and global level.

KEYWORDS:

• Sexual violence
• homicide
• offender characteristics
• victim characteristics
• criminal behaviour

Sexual homicide is a rare yet interest-provoking phenomenon. Estimated to make up between 1% and 4% of all homicides (Abreu Minero et al., 2018; Chan & Heide, 2009; Kerr et al., 2013), the characteristics of those involved are diverse and the nature of this crime is shaped by the laws and cultures within which it takes place (Skott et al., 2021). As such, building a body of knowledge regarding best practice for working with an offence of this nature is hampered by the rarity of its occurrence, affecting both research and the accumulation of clinical or investigative experience (Beauregard, 2018, p. 437). In the study of rare phenomena, descriptive research with small samples has proven to be valuable (see Indrayan & Mishra, 2021). Gerring (2012) reasons that for events where little is known, “description must proceed independently of causal propositions” (p.733); causal arguments are instead more appropriately built upon prior descriptive arguments. However, despite examination of this area of offending increasing in recent years, its usefulness and generalisability is somewhat restricted beyond regions where the research has been conducted, due to legal and cultural variations (Skott et al., 2021). That said, some universal characteristics of sexual homicide offences and offenders have been identified that recur throughout sexual homicide literature (James et al., 2018). These common factors may act as theoretical points of interest for investigators and clinicians (Eichinger & Darjee, 2021).

Characteristics of sexual homicide and sexual homicide offenders

Sexual homicide refers to the presence of sexual behaviour or motivation during the killing of a person. Defining sexual homicide is complex, with definitions varying between studies (Kerr et al., 2013). Although sexual motivation is recognised as an important factor in many of these homicides, it can also be poorly understood or unidentified by both the investigators and the offender themselves and can add little value to the investigative process (Schlesinger, 2021, Practical Impediments to Research section). Consequently, a purely observational approach is typically taken in attributing a killing as sexual homicide, including evidence such as the victim's nudity, evidence of sexual assault or torture, or the perpetrator's use of sexual objects or implements as weapons (Kerr et al., 2013).

Sexual homicide has been recognised as a gendered crime, with both women and girls being more likely to be victims of sexual homicide than men and boys; estimates from international research suggest that approximately 80-90% of sexual homicide victims are female (Abrahams et al., 2017; Häkkänen-Nyholm et al., 2009; Schlesinger & Revitch, 1999; Van Patten & Delhauer, 2007). Sexual homicide offenders are most likely to be male (Clarkson et al., 2020; Karakasi et al., 2017), in their 20-30s (Myers et al., 2016; Van Patten & Delhauer, 2007), and unmarried (Beauregard & Martineau, 2013; Karakasi et al., 2017). Van Patten and Delhauer’s (2007) Los Angeles study found a positive age skew for both offenders (skew statistic = 0.74, SE = 0.21) and victims (skew statistic = 0.93, SE = 0.17), showing a preference for both younger offenders and victims. They noted however that around half of all offenders (49.7%) and victims (50.5%) fell between ages 26–59.

Recent research supports these findings in the Australasian region; assessing sexual homicide across both Australia and New Zealand, Eichinger and Darjee (2021) identified 130 victims, with 70.0% being female and having a mean age of 34 years old. Of the 118 offenders, 94.1% were typically male and had a mean age of 30 years old. They identified both offenders and victims as Caucasian in most cases (82.2% and 81.9%, respectively). Similar findings are seen internationally regarding victim and offender ethnicity (Chan & Heide, 2016; Smith et al., 2011). Research also tends to find most offenders do not have an existing psychiatric condition at the time of their offence (James & Proulx, 2014), although one previous study found that two-thirds of sexual homicide offenders had a history of depression (Meloy, 2000). However, personality disorders, including antisocial, borderline, and narcissistic, are more common among sexual homicide offenders (Beauregard & Martineau, 2017; Kerr et al., 2013).

Although there are some common patterns identified in the causes of death for sexual homicides, substantial variation also exists across cases. Asphyxiation (including both ligature or manual strangulation) has been identified as the most common cause of death, followed by beating and stabbing (Abrahams et al., 2017; Eichinger & Darjee, 2021; Kim et al., 2023; Stein et al., 2010; Van Patten & Delhauer, 2007). For example, Häkkänen-Nyholm and colleagues (2009) found strangulation occurred in 55.6% of sexual homicides in Finland. Comparatively, firearms are rarely used in sexual homicides. For example, in a study of 350 sexual homicide offenders in Canada, only 7.6% of offences involving a weapon used a firearm (Beauregard & Martineau, 2013). Instead, knives were the most commonly used weapon (37.1%), followed by bludgeoning (25.4%), a ligature (20.3%), other (6.1%), or an axe or hatchet (3.6%). However, the frequency of firearm use depends on the country in which the crime is committed (Abrahams et al., 2017; Chan et al., 2019), with higher rates of gun use being seen in countries with greater access to firearms (Altheimer & Boswell, 2012; Stolzenberg & D’Alessio, 2000).

Finally, findings related to the offender-victim relationship in sexual homicides are highly heterogenous. Some studies identify a stranger relationship in most cases (Beauregard & Proulx, 2002; Carter & Hollin, 2010), while others identify this as rare (Häkkänen-Nyholm et al., 2009) or about as likely to be a stranger as not (Carter et al., 2017; Darjee & Baron, 2013). One potential reason for this inconsistency is the complicating effect of staging in sexual homicide cases. Staging is a behaviour known to take place in homicide incidents that aim to mislead investigators and prevent the offender from becoming a suspect. For example, positioning the corpse in a manner which gives the appearance of suicide, or attempting to make the scene look like a burglary/robbery (Schlesinger et al., 2014). Although staging occurs relatively infrequently (Pettler, 2015), and even less so in sexual homicide offences (Schlesinger et al., 2014), it occurs most frequently when there is a known relationship between the offender and victim; some degree of familiarity between the two has been suggested as a near-requirement for staging to occur (Schlesinger et al., 2014). For example, Schlesinger and colleagues (2014) studied 946 US homicide incidents, and identified staging behaviours in 8.35% of their cases. Staging was evident in only 11 of 181 (6.07%) non-serial sexual homicide crime scenes, and none of the 195 crime-scenes created by 45 serial sexual homicide offenders. This association between staging behaviours and the relationship between offender and victim therefore means that staging may act as a confound when attempting to better understand relational patterns in sexual homicides.

Sexual behaviours in homicide offences

By nature, sexual homicide involves a variety of sexual behaviours which may hurt, offend, or dehumanise the victim. Cruelty, callousness and depravity are hallmark features of such offending (Chan, 2020; Meloy, 2000). In particular, degrading and humiliating behaviour endures as a frequent feature of sexual homicide. Eichinger and Darjee (2021) assessed the frequency of degrading or humiliating behaviour in their sample and found 20.9% of offenders performed these behaviours in some way. Mutilation can be considered an extreme form of degrading or callous behaviour, and has been associated with sexual homicide in previous research. For example, Hakkanen-Nyholm and colleagues (2009) examined the occurrence of mutilation in homicides, and found that in 38.5% of cases where this took place, sexual behaviour prior, during, or after the killing was also observed. This differed significantly from homicides where mutilation was not present (φ = .31, p < .001).

Perhaps surprisingly, non-consensual vaginal penetration (i.e. vaginal rape), while common, is not a defining characteristic of such offending. In Beauregard and Martineau’s (2013) examination of sexual homicides in Canada, they found this act took place in less than half (46.3%) of cases. Similarly, Darjee and Baron (2013; as cited in Beauregard & Martineau, 2017) identified this in 42.6% of their sample. Departing from this trend, Beauregard and Proulx (2002) found offenders had sexual intercourse with the victim in 78.6% of cases analysed, however they operationalised this to include oral, vaginal, and anal sex. This likely explains the increase in reported prevalence.

Regarding bondage, several studies have found sexual homicide offenders will often use physical restraints, although this has not been a consistent finding. Beauregard and Proulx (2002) identified that just under one-third (30%) of all offenders used physical restraints, whereas Beauregard and Martineau (2017) found only 10.9% of sexual homicide offenders used restraints. Similarly, Salfati and Taylor (2006) identified physical restraints being used in 8% of offences. Notably however, both Beauregard and Martineau (2017) and Salfati and Taylor (2006) assessed blindfolds and gags separately to restraints, seeing this in 7.1% and 5% of their samples, respectively. The overlap between blindfold/gags and physical restraint in these samples is unknown.

While sexual homicide typically occurs after the commission of rape or to prevent victim reporting, sexual behaviour can take place after death. Homicide involving necrophilia is highly infrequent, however it does occur in a modest proportion of sexual homicides (Stein et al., 2010); necrophilia for the primary purpose of sexual intercourse with a corpse is even rarer, instead more frequently being an opportunistic or sadistic secondary outcome (Chopin & Beauregard, 2021). Swart and Mellor (2017) argued that in spite of its scarcity, in addition to obvious post-mortem sexual activity, post-mortem mutilation should lead to a strong consideration of necrophilic activity due to its statistical rarity. Excised body parts may be touched or viewed by the offender as they masturbate, or may come into contact with the offender’s genitals (Mellor, 2016a). Where investigators are concerned, Mellor (2016b, p. 110) stated “the overarching paraphilia of necrophilia should always take precedence over the individual differences present in necrophilia-spectrum behaviours.” Therefore, evidence of mutilation/dismemberment may indicate the need to examine for sexual abuse, and vice versa.

Body disposal in sexual homicide

Previous research on sexual homicide has been particularly concerned with the means of body disposal that occurs within these incidents. “Body disposal” refers to any interference with the body post-mortem that aims to conceal criminal offending or avoid detection. This may include body transportation, concealment, or destruction. This research can provide clues around the identity of the offender, the method used to kill the victim, the motive for the crime, and, for homicides in which the body remains unfound, potential locations of interest. International literature reveals body disposal occurs in around one-third of sexual homicides (Beauregard & Martineau, 2013). In an Australian-New Zealand study, Eichinger and Darjee (2021) found that 34.6% of offenders moved the body and 21.21% concealed it following the offence. Additionally, in their sample of Finnish homicides, Häkkänen-Nyholm et al. (2009) observed that the bodies of 38.9% of sexual homicide victims were located at an alternative location to the scene of the killing.

Utilising small samples of sexual homicides, previous research has found certain characteristics and behaviours are more strongly associated with the victim’s body being moved than others. For unclear reasons, demographics of victims or offenders are rarely assessed in these studies (e.g. Beauregard & Martineau, 2016; Chai et al., 2021). However, those that have included offender age typically find this to be unrelated to body disposal patterns (e.g. Beauregard & Field, 2008). By contrast, the victim’s age has been found to be related to body disposal patterns, with the bodies of older victims being more likely to be left at the scene (d = 0.96; Beauregard & Field, 2008). Previous studies have identified further victim characteristics that are associated with the moving of the body post-offence in sexual homicide. For example, Chai et al. (2021) found that in solved homicides, if the victim was a sex worker or homeless, the body was more likely to be moved. Conversely, utilisation of a con approach (i.e. the use of deception or manipulation) or insertion of a foreign object into the victim was associated with a reduction in the likelihood of body transportation.

Post-crime, concealment of the body, being found lying face down, having removed or destroyed evidence, recovering the body in the victim’s residence, and finding the body outdoors were all associated with reduced odds of the body being transported (Chai et al., 2021). Notably, in Chopin and Beauregard’s (2024) comparison of solved sexual homicides to unsolved, perpetrators who attempted to removed or destroy evidence were significantly more likely to be caught as it resulted in more evidence being created. Stein et al.’s (2010) study of necrophilia in the United States found 11 of 16 victims were left at the crime scene unmoved and with no effort to conceal.

Current study

Considering the niche subject area, the existing literature examines a wide variety of crime scene behaviours, yet rarely examines these in different geographical, socioeconomic, or cultural contexts. Expanding research into different settings aids in understanding the generalisability of findings and helps to account for sociocultural differences between populations. This gap in the existing evidence base has led to the identification of need for the current study. Furthermore, there is a lack of research on the association between body disposal patterns and offender/victim characteristics to help guide practice/investigations in New Zealand.

Similarly, only three studies are known to have examined sexual homicide in the Australasian region, one utilising an Australian sample (Kocsis, 1999) and two using the same Australia-New Zealand combined sample (Clarkson et al., 2020; Eichinger & Darjee, 2021). While similarities undoubtedly exist between the two Western countries, differences may also be present due to their individual cultural climates (McPhedran et al., 2011). Eichinger and Darjee (2021) do not state how many cases in their sample originate from New Zealand and so the extent of crossover between their study and the current research is unclear. However, as some of the current research’s cases occurred since Eichinger and Darjee’s (2021) publication, deviations between the samples are certain.

The current study aimed to better understand the occurrence of sexual homicide in New Zealand, and to explore victim/offender/crime characteristics and body disposal features. In particular, the present study aimed to examine the circumstances, means and manner in which a victim’s body may be manipulated during and after death to avoid apprehension. Beauregard and Field’s (2008) research on the relationship between victim and offender ages prompted a more comprehensive evaluation in the current study, particularly in relation to body disposal. As noted by Beauregard and Martineau (2017), there is substantial heterogeneity within the features observed across various forms of homicide, meaning it is important not only to compare features or characteristics between homicide types, but also to conduct research that comprehensively evaluates each type within itself. This type of research allows variation between cases to become more transparent, and aids in developing useful profiling tools which better adapt to inconsistencies between cases of each homicide type.

Three research questions were central to this study:

1. What are the common characteristics of sexual homicide offenders and victims?

2. What behaviours (sexual and non-sexual) are typically observed in sexual homicide incidents?

3. What sexual homicide behaviours/characteristics are associated with body disposal methods?

Method

This research received ethics approval from the University of Canterbury Human Research Ethics Committee (HREC 2021/39) prior to data collection. This study endeavoured to succinctly outline the methodology, including sample size determination, inclusion/exclusion criteria, and analytical measures, to enhance transparency of our findings.

Sampling procedure

Sexual homicide cases in New Zealand were identified using a combination of the New Zealand Police Violent Crime Linkage Analysis System (ViCLAS) and Ministry of Justice sentencing reports; these sources were also used to collect the desired information about each case (see further details below). The ViCLAS is a database which systematically records data regarding violent and sexual incidents. Its primary purpose is crime linkage, and the exploration of similarities between cases. Sentencing reports produced by the Ministry of Justice are publicly-available and outline the sentencing process, case summary, and offender information. These are available for sentencings which take place in the High Court or above. Information collected from these sources was supplemented by media articles obtained through a comprehensive search of national and local news websites. This supplemental approach is common in the study of homicide (Clarkson et al., 2020).

Sexual homicide classification was based on the FBI’s criteria set out by Ressler et al. (1988). This required at least one of the following criteria to be present: victim’s attire or lack of attire; exposure of sexual parts of the victim’s body; sexual positioning of the victim’s body; insertion of foreign objects into the victim’s body cavities; evidence of sexual intercourse; or evidence of substitute sexual activity, interest, or sadistic fantasy. Fact summaries of homicide cases with the databases were assessed to determine if any criteria were met.

Because sexual homicide is an uncommon phenomenon, no timeframe was specified as to when the homicide occurred, in order to gain the largest sample possible. The final sample included offences that occurred between 1968 and 2022.

Cases/Participants

This study identified 26 incidents involving 30 offenders and 26 victims. One case involved an additional victim, however because they did not perish they were excluded from the dataset.

All but three cases involved only one offender. Two cases involved three offenders, and the third involved two. Notably, offenders were male in all incidents except for one triple-offender case, where two of the offenders were female (including the primary perpetrator). One offender appeared twice as he was responsible for two separate incidents. In the first incident, the victim was the offender's neighbour, while in the second incident, the victim was his wife. Some similarities were evident between these cases; both victims were manually suffocated and had their mouths gagged. Their bodies were subsequently placed in similar locations. The first victim was discovered only after the offender came under investigation following the disappearance of his wife. Although these cases shared similarities, both were included as they differed in the circumstances of each offence. We were interested in analysing all relevant incidents investigated by the police to gain insights into their characteristics as they were found during investigation.

Variables

Given the incorporation of information from a variety of sources, a codebook was developed to ensure consistent recording of information between cases. Variables included canvassed demographics, motivations, crime scene behaviour, and coronial evidence.

Body disposal factors (the dependent variables in this study) were dichotomously coded as follows: Body moved (0 = body left, 1 = body moved); Body hidden (0 = body uncovered, 1 = body concealed); Body destroyed (e.g. set alight, dissolve with acid or feed body to animals; 0 = body intact, 1 = body destroyed/attempted destruction).

Sexual variables were coded according to the following definitions:

Humiliation. The performance of degrading behaviour, such as exposure of breasts or genitals for the purposes of embarrassment or degradation (0 = no humiliation, 1 = humiliation identified).

Bondage (non-consensual). Restraint of a victim using ties, rope, or other material (0 = no bondage, 1 = bondage identified).

Non-genital torture. Intentional acts causing physical pain or psychological torment in excess of cause of death (0 = no non-genital torture, 1 = non-genital torture identified).¹

Non-insertional genital torture. Intentional acts causing physical pain to breasts or genitals, such as cutting or burning (0 = no non-insertional genital torture, 1 = non-insertional genital torture identified).

Rape/Sodomy/Foreign object insertion (FOI). Intentional sexual assault to living victim including vaginal rape, sodomy, and/or foreign object insertion (0 = no rape/sodomy/FOI, 1 = rape/sodomy/FOI identified).

Necrophilia. Sexual assault known to have taken place post-mortem (0 = no necrophilia, 1 = necrophilia identified).

Sexual behaviours were coded with the acknowledgement that the initial degree of consent by the victim could not always be determined pre-homicide (e.g. consensual rope/knife play, sex work). Resultantly, coding was applied according to what the investigating officer might find at the crime scene.

Data analysis

It was not feasible to have multiple coders for data due to the restrictions on data access, however international research has generally demonstrated good internal reliability regarding ViCLAS data (Martineau & Corey, 2008; cf. Snook et al., 2012).

A descriptive assessment of victim and offender demographics, crime-scene characteristics, and sexual behaviours, and the overlap amongst these, was conducted. Qualitative descriptions were also included where this added value to the descriptive results.

Results

Demographics

As noted, 26 incidents of sexual homicide in New Zealand were identified through the New Zealand Police ViCLAS system and Ministry of Justice sentencing reports.

Victim characteristics

Victim characteristics and their association with each body disposal factor are presented in Table 1.

Table 1. Victim demographics by body disposal factor^a.

		Overall			Body Moved			Body Hidden
		(n = 26)		No (n = 11)		Yes (n = 15)		No (n = 14)		Yes (n = 12)
		n	%	n	%	n	%	n	%	n	%
Sex	Female	26	100%	11	100.0%	15	100.0%	14	100.0%	12	100.0%
	Male	0	0.0%	0	0.0%	0	0.0%	0	0.0%	0	0.0%
Ethnicity	Asian	2	7.7%	1	9.1%	1	6.7%	1	7.1%	1	8.3%
	Caucasian	16	61.5%	5	45.5%	11	73.3%	8	57.1%	8	66.7%
	Māori	3	11.5%	1	9.1%	2	13.3%	1	7.1%	2	16.7%
	MELAA	1	3.8%	1	9.1%	0	0.0%	1	7.1%	0	0.0%
	Pacific Peoples	0	0.0%	0	0.0%	0	0.0%	0	0.0%	0	0.0%
	Unknown	4	15.4%	3	27.3%	1	6.7%	3	21.4%	1	8.3%
Marital Status	Single	4	15.4%	3	27.3%	1	6.7%	3	21.4%	1	8.3%
	Partner	10	38.5%	2	18.2%	8	53.3%	4	28.6%	6	50.0%
	Unknown	12	46.2%	6	54.5%	6	40.0%	7	50.0%	5	41.7%
Employed	No	8	30.8%	4	36.4%	4	26.7%	3	21.4%	5	41.7%
	Yes	14	53.8%	4	36.4%	10	66.7%	8	57.1%	6	50.0%
	Unknown	4	15.4%	3	27.3%	1	6.7%	3	21.4%	1	8.3%
Residence	House	21	80.8%	10	90.9%	11	73.3%	11	78.6%	10	83.3%
	Homeless	1	3.8%	0	0.0%	1	6.7%	0	0.0%	1	8.3%
	Unknown	4	15.4%	1	9.1%	3	20.0%	3	21.4%	1	8.3%
Living Situation	Alone	7	26.9%	5	45.5%	2	13.3%	6	42.9%	1	8.3%
	Family	12	46.2%	4	36.4%	8	53.3%	5	35.7%	7	58.3%
	Other	7	26.9%	2	18.2%	5	33.3%	3	21.4%	4	33.3%

Note. ^aBody destroyed excluded due to low cell counts.

All 26 victims were female, with a mean age of 32.52 (SD = 14.4) and ranging between 15 and 69 years old. The majority of victims were Caucasian (61.5%). Where known, victims were more likely to be married or partnered (38.5%), with comparatively fewer being single (15.4%). Most victims’ place of residence was a house (80.8%), where nearly half lived with family (46.2%). More than half of victims were in steady employment (53.8%).

Victims who were moved had a mean age of 28.13 (SD = 10.84, n = 15), ranging between 17 and 56 years old. Those who were not had a mean age of 39.10 (SD = 17.02, n = 10) and ranged between 16 and 69 years old. Three-quarters of moved victims were Caucasian (73.3%, n = 11) yet less than half of those not moved were Caucasian (45.5%, n = 5). Ten (66.7%) moved victims were known to be employed compared to four (36.4%) not moved. Notably, it was more common for those not moved to live alone (45.5% vs. 13.3%).

Victims who were hidden had a mean age of 27.67 (SD = 12.32, n = 12), ranging between 15 and 56 years old. Those who were not had a mean age of 37.00 (SD = 15.16, n = 13) and ranged between 22 and 69 years old. It was more common for hidden victims to have a partner (50.0% vs. 28.6%), however those not hidden were more frequently employed (57.1% vs. 50.0%). Those hidden lived alone less often than those not hidden (8.3% vs. 42.9%).

Offender characteristics

Offender characteristics and their association with each body disposal factor are presented in Table 2.

Table 2. Offender demographics by body disposal factor^a.

		Overall		Body Moved				Body Hidden
		(n = 30)		N (n = 14)		Y (n = 16)		N (n = 17)		Y (n = 13)
		n	%	n	%	n	%	n	%	n	%
Sex	Female	2	6.7%	0	0.0%	2	12.5%	0	0.0%	2	15.4%
	Male	28	93.3%	14	100.0%	14	87.5%	17	100.0%	11	84.6%
Ethnicity	Asian	0	0.0%	0	0.0%	0	0.0%	0	0.0%	0	0.0%
	Caucasian	14	46.7%	5	35.7%	9	56.3%	7	41.2%	7	53.8%
	Māori	10	33.3%	6	42.9%	4	25.0%	6	35.3%	4	30.8%
	MELAA	1	3.3%	0	0.0%	1	6.3%	1	5.9%	0	0.0%
	Pacific Peoples	2	6.7%	1	7.1%	1	6.3%	1	5.9%	1	7.7%
	Unknown	3	10.0%	2	14.3%	1	6.3%	2	11.8%	1	7.7%
Marital Status	Single	13	43.3%	5	35.7%	8	50.0%	7	41.2%	6	46.2%
	Partner	12	40.0%	4	28.6%	8	50.0%	5	29.4%	7	53.8%
	Unknown	5	16.7%	5	35.7%	0	0.0%	5	29.4%	0	0.0%
Employed	No	9	30.0%	3	21.4%	6	37.5%	4	23.5%	5	38.5%
	Yes	9	30.0%	3	21.4%	6	37.5%	5	29.4%	4	30.8%
	Unknown	12	40.0%	8	57.1%	4	25.0%	8	47.1%	4	30.8%
Residence	House/Boarding home	21	70.0%	12	85.7%	9	56.3%	14	82.4%	7	53.8%
	Homeless	3	10.0%	0	0.0%	3	18.8%	1	5.9%	2	15.4%
	Motor camp	1	3.3%	0	0.0%	1	6.3%	0	0.0%	1	7.7%
	Unknown	5	16.7%	2	14.3%	3	18.8%	2	11.8%	3	23.1%
Living Situation	Alone	9	30.0%	4	28.6%	5	31.3%	5	29.4%	4	30.8%
	Family	12	40.0%	6	42.9%	6	37.5%	5	29.4%	7	53.8%
	Other	3	10.0%	1	7.1%	2	12.5%	3	17.6%	0	0.0%
	Unknown	6	20.0%	3	21.4%	3	21.4%	4	23.5%	2	15.4%
Build	Larger than victim	17	56.7%	8	57.1%	9	56.3%	10	58.8%	7	53.8%
	Similar to victim	1	3.3%	0	0.0%	1	6.3%	1	5.9%	0	0.0%
	Smaller than victim	0	0.0%	0	0.0%	0	0.0%	0	0.0%	0	0.0%
	Unknown	12	40.0%	6	42.9%	6	37.5%	6	35.3%	6	46.2%
Prior Criminal Record	No	4	13.3%	1	7.1%	3	18.8%	2	11.8%	2	15.4%
	Yes	19	63.3%	7	50.0%	12	75.0%	8	47.1%	11	84.6%
	Unknown	7	23.3%	6	42.9%	1	6.3%	7	41.2%	0	0.0%
Prior Violent Record	No	7	23.3%	2	14.3%	5	31.3%	2	11.8%	5	38.5%
	Yes	14	46.7%	4	28.6%	10	62.5%	6	35.3%	8	61.5%
	Unknown	9	30.0%	8	57.1%	1	6.3%	9	52.9%	0	0.0%
Prior Sexual Record	No	12	40.0%	3	21.4%	9	56.3%	6	35.3%	6	46.2%
	Yes	9	30.0%	4	28.6%	5	31.3%	3	17.6%	6	46.2%
	Unknown	9	30.0%	7	50.0%	2	12.5%	8	47.1%	1	7.7%

Note: Body destroyed excluded due to low cell counts.

The majority of offenders were male (93.3%, n = 28), with an age range from 17 to 54 years old (M = 30.13, SD = 9.41). Most offenders were either Caucasian or Māori, and a relatively even split was identified between offenders being single or partnered, and employed or unemployed. More information was available on the place of residence and living situation of each offender than for victims. Most offenders lived in a house or in a boarding home (70%), either with their family (40%) or alone (30%).

Two-thirds (63.3%, n = 19) of offenders had a known criminal record, however this dropped to below half (46.7%, n = 14) when looking at prior violent offending, and just under a third (30%, n = 9) for prior sexual offending. The offender was of a bigger build than the victim in 58.1% (n = 18) of incidents and similar to the victim in one incident (3.2%). No incidents were recognised as involving an offender with a smaller build than the victim.

Offenders who moved the body had a mean age of 29.50 (SD = 8.02, n = 16), ranging between 16 and 45 years old. Those who left the body at the crime scene had a mean age of 30.97 (SD = 11.32, n = 12), ranging between 19 and 54 years old. The body was more frequently moved by Caucasian individuals (56.3%, n = 9), and as likely by those who were single (50.0%, n = 8) or with a partner (50.0%, n = 8). Regarding offenders who left the body, all with a known residence lived in a house or boarding house compared to just over half (56.3%) of those who moved it. Offenders who moved the body were more likely to have a criminal record (75.0% vs 50.0%), particularly for violence (62.5% vs 28.6%), than those who did not.

Offenders who hid the body had a mean age of 28.84 (SD = 8.51, n = 13), ranging between 16 and 45 years old. Those who did not had a mean age of 31.25 (SD = 10.28, n = 15) and ranged between 19 and 54 years old. Half of those who hid the body were Caucasian (53.8%, n = 7), had a partner (53.8%, n = 7), and/or lived in a house or boarding home (53.8%, n = 7). Eleven offenders who hid the body had a criminal record (84.6%), while eight involved violent offending (61.5%) and six were sexual (46.2%).

Offender-Victim relationship

The offender-victim relationship tended to be relatively distant, with more than half of incidents being perpetrated against stranger victims (54.8%, n = 17). The victim was known to the offender (e.g. friends, neighbours, acquaintance) in eleven cases (35.5%), and was a partner in only two cases (6.5%); these two cases were committed by the same offender. The offender-victim relationship was unknown in one case (3.2%). Caucasian offenders more frequently targeted intra-racial victims, while minority populations targeted both intra-racially and inter-racially. Except for offenders from a Pacific background where the victim’s ethnicity was unknown, Caucasian women were most targeted.

Motivation and premeditation

Variation in the motivators of sexual homicide are common. Sexual satisfaction was the prominent motivator in the current sample, identified in over half of cases (53.8%, n = 14). Anger (19.2%, n = 5), burglary (11.5%, n = 3), jealousy (3.8%, n = 1), and pleasure (3.8%, n = 1) were also identified as drivers of offender behaviour. Notably, at least three of the five anger-motivated killings were propelled by rejected sexual advances. The motivations behind two killings were undetermined (7.7%).

In assessing premeditation, a four-level approach was used. Nine offenders (29%) were found to have spontaneously committed their crime. None resulted from unstructured premeditation (i.e. a reactionary response to perceived provocation or injustice). Four (12.9%) offenders began their assault intending violence alone and not foreseeing death, and six (19.4%) intended death to occur. The degree of premeditation was unable to be determined in 12 cases (38.7%).

Offence-related behaviours

While crime staging was rarely observed (7.7%), the crime scene was cleaned in nearly a quarter of cases (23.1%). Being unknown in only one case (3.8%), behaviours relating to the hiding or destruction of evidence were present in a minority of cases; the offender hid evidence (such as a weapon) in ten (38.5%) cases and destroyed evidence (such as offender or victim clothing, a weapon, or material used in the commission of the offence) in two (7.7%) cases. Items were identified as stolen from the victim or their property in nearly half (42.3%) of cases, again being unknown in one case (3.8%).

Cause of death and weapon involvement

The main cause of death and weapons used are presented in Table 3. The most common cause of death was asphyxiation (38.5%, n = 10). A quarter of victims died through blunt force trauma (23.1%, n = 6); four of these (15.4%) were caused by direct head injuries. One victim died from a combination of blunt force trauma and stab wounds (3.8%); it was indeterminable which caused the fatal injury. Two (7.7%) victims’ bodies were decomposed to the extent that cause of death could not be determined.

Table 3. Cause of death and weaponry.

		Overall (n = 26)
		n	%
Cause of Death	Asphyxiation	10	38.5%
	Blunt Force Trauma	6	23.1%
	Stab Wounds	6	23.1%
	Shot	1	3.8%
	Combination Trauma	1	3.8%
	Indeterminable	2	7.7%
Weapon Involvement
Primary Weapon	Airway-restrictive Device	5	19.2%
	Firearm	1	3.8%
	Knife/Cutting Tool	6	23.1%
	Melee	11	42.3%
	Percussive Instrument	2	7.7%
	Unknown	1	3.8%
Secondary Weapon (n = 15)	Airway-restrictive Device	3	20.0%
	Firearm	0	0.0%
	Knife/Cutting Tool	1	6.7%
	Melee	5	33.3%
	Percussive Instrument	3	20.0%
	Car	1	6.7%
	Unknown	2	13.3

Across the 26 offences, half (50%, n = 13) were known to have involved more than one weapon. Eleven (42.3%) did not use a secondary weapon, and it was not known in two (7.7%) cases. Regarding the primary weapon, 42.3% (n = 11) predominantly involved melee during the attack (i.e. the use of hands, feet, and body to assault the victim), and nearly a quarter (23.1%, n = 6) involved a knife or cutting tool. Regarding the 13 cases where a secondary weapon was used, melee was the most frequent (33.3%, n = 5).

Sexual behaviours

At least one form of sexual behaviour was known to have occurred in each case. Almost half (46.2%, n = 12) of all cases were known to involve only one sexual behaviour. Five (19.2%, n = 5) cases involved two sexual behaviours, and another five (19.2%) involved three or more. Four (15.4%) cases were unclear regarding the presence of two or more sexual behaviours.

Rape, sodomy, or foreign object insertion was known to have occurred in 18 cases (69.2%). It was unclear if sexual activity was consensual prior to killing in one (3.8%) case, and the evidence was undeterminable due to decomposition in one (3.8%) other. Vaginal penetration specifically was known to have occurred in 65.4% (n = 17) of cases, and believed not to have taken place in 15.4% (n = 4). This was unknown in five (19.2%) cases. Necrophilia was found to have occurred in five (19.2%) cases, and suspected to have taken place in two (7.7%) more but not confirmed. It was unknown whether necrophilia occurred in three cases (11.5%).

Non-consensual bondage was known to have occurred in six (23.1%) cases. A further four (15.4%) victims had their mouth stuffed or blocked with material, however were otherwise unrestrained. It was unknown in one (3.8%) case if bondage took place.

Humiliation took place in six (23.1%) incidents, but not in 18 (69.2%). It was unclear in two (7.7%) cases. Evidence of non-genital torture was available in five (19.2%) cases, however non-insertional genital torture was rare (3.8%, n = 1). Non-genital torture did not occur in three-quarters of cases (76.9%, n = 20), and was unknown in one (3.8%). Non-insertional genital torture did not occur in 84.6% (n = 22) of cases, however it was unclear in three (11.5%) cases.

There were some differences identified in the presence of sexual behaviours depending on the motive of the homicide. In cases where the motivation was known, two-thirds (66.7%, n = 8) of sexually motivated killings involved only one sexual behaviour, compared to only one-third (33.3%, n = 3) of cases with alternative motivations. Conversely, two-thirds of cases motivated by anger, jealousy, joy, or other crime (n = 6) involved multiple sexual behaviours in the commission of the offence.

Body disposal factors

Across the 26 cases, 15 (57.7%) victims had their body moved after the murder. Twelve (46.2%) offenders made attempts to conceal the victim’s corpse. Eleven cases involved the body being both moved and concealed (42.3%).

An attempt to destroy the body only occurred in one (3.8%) case; this was performed by setting the body alight. In one (3.8%) case, it was unclear if attempts were made to destroy the body. No witnesses (excluding accomplices) to the crime or body disposal were reported in any case.

When body transportation took place, 66.7% (n = 10) of victims were dragged to their respective disposal site, while 33.3% (n = 5) of victims were taken by car. Ten victims (38.5%) were removed from the property where the killing occurred, while five were repositioned on the same property.

Sexual behaviour associations with body disposal

The associations between sexual homicide behaviours and body disposal are displayed in Table 4. Notable associations included asphyxiation occurring more frequently when body movement took place than when it did not (60.0% vs. 9.1%). Conversely, humiliation occurred less frequently where body movement took place (13.3% vs. 36.4%). When the body was moved, victims were also targeted less often (26.7% vs. 63.6%), and it was less frequently a spontaneous offence (26.7% vs. 45.5%). Premeditation for murder was also less common (13.3% vs. 36.4%). When victims were moved, it was more often a sexually motivated crime (60.0%) compared to those committed for other reasons (26.7%). Regarding body concealment, asphyxiation appeared to take place more frequently in cases where concealment occurred (50.0% vs. 26.6%), as did the use of bondage (50.0% vs. 28.6%). No other clear associations stood out.

Table 4. Associations between sexual homicide behaviours/characteristics and body disposal factors^a.

		Overall		Body Moved				Body Hidden
		(n = 26)		N (n = 11)		Y (n = 15)		N (n = 14)		Y (n = 12)
Behaviours/Characteristics		n	%	n	%	n	%	n	%	n	%
Asphyxiation	No	14	53.8%	10	90.9%	4	26.7%	9	64.3%	5	41.7%
	Yes	10	38.5%	1	9.1%	9	60.0%	4	26.6%	6	50.0%
	Unknown	2	7.7%	0	0.0%	2	13.3%	1	7.1%	1	8.3%
Sexual Behaviours
Humiliation	No	18	69.2%	6	54.5%	12	80.0%	8	57.1%	10	83.3%
	Yes	6	23.1%	4	36.4%	2	13.3%	4	28.6%	2	16.7%
	Unknown	2	7.7%	1	9.1%	1	6.7%	2	14.3%	0	0.0%
Bondage	No	15	57.7%	7	63.6%	8	53.3%	9	64.3%	6	50.0%
	Yes	10	38.5%	4	36.4%	6	40.0%	4	28.6%	6	50.0%
	Unknown	1	3.8%	0	0.0%	1	6.7%	1	7.1%	0	0.0%
Torture	No	20	76.9%	8	72.7%	12	80.0%	10	71.4%	10	83.3%
	Yes	5	19.2%	3	27.3%	2	13.3%	3	21.4%	2	16.7%
	Unknown	1	3.8%	0	0.0%	1	6.7%	1	7.1%	0	0.0%
Non-insertional Genital Tortureb	No	22	84.6%	9	81.8%	13	86.7%	11	78.6%	11	91.7%
	Yes	1	3.8%	0	0.0%	1	6.7%	0	0.0%	1	8.3%
	Unknown	3	11.5%	1	9.1%	1	6.7%	3	21.4%	0	0.0%
Rape/Sodomy/Foreign object insertion	No	6	23.1%	2	18.2%	4	26.7%	2	14.3%	4	33.3%
	Yes	18	69.2%	7	63.6%	11	73.3%	10	71.4%	8	66.7%
	Unknown	2	7.7%	2	18.2%	0	0.0%	2	14.3%	0	0.0%
Necrophilia	No	16	61.5%	8	72.7%	8	53.3%	8	57.1%	8	66.7%
	Yes	5	19.2%	1	9.1%	4	26.7%	2	14.3%	3	25.0%
	Unknown	5	19.2%	2	18.2%	3	20.0%	4	28.6%	1	8.3%
Motivation/Premeditation
Premeditation	Spontaneous	9	34.6%	5	45.5%	4	26.7%	5	35.7%	4	33.3%
	Unstructured Premeditation	0	0.0%	0	0.0%	0	0.0%	0	0.0%	0	0.0%
	Premeditation for Violence	4	15.4%	1	9.1%	3	20.0%	1	7.1%	3	25.0%
	Premeditation for Murder	6	23.1%	4	36.4%	2	13.3%	4	28.6%	2	16.7%
	Unknown	7	26.9%	1	9.1%	6	40.0%	4	28.6%	3	25.0%
Targeted Victim	No	12	46.2%	4	36.4%	8	53.3%	7	50.0%	5	41.7%
	Yes	11	42.3%	7	63.6%	4	26.7%	7	50.0%	4	33.3%
	Unknown	3	11.5%	0	0.0%	3	20.0%	0	0.0%	3	25.0%
Sexual Motivation	Yes	14	53.8%	5	45.5%	9	60.0%	7	50.0%	7	58.3%
	Other	10	38.5%	6	54.5%	4	26.7%	6	42.9%	4	33.3%
	Unknown	2	7.7%	0	0.0%	2	13.3%	1	7.1%	1	8.3%

Discussion

Location-specific research is particularly important in the study of uncommon social phenomena such as sexual homicide, for testing the homogeneity of international findings and assessing the utility of profiling within such a population (Eichinger & Darjee, 2021). This descriptive study aimed to provide this assessment of generalisability to the New Zealand context, with findings relatively consistent with many of the patterns found in the existing international sexual homicide literature. This research provides a realistic account of sexual homicide offences in New Zealand, highlighting offender and victim features as well as key behaviours observed in these offences; this is particularly valuable for the detailing of rare behaviours which cannot be subject to causal or statistical modelling (Gerring, 2012).

Findings regarding age and gender of both offenders and victims reflected those found in previous research. As in Karakasi et al.’s (2017) review and Eichinger and Darjee’s (2021) research in Australasia, this study found offenders were typically male with an average age around 30 years old. While a lower percentage than that found in previous studies (Chan & Heide, 2016), Caucasian males made up nearly 50% of the offenders in this study’s sample. In addition, Caucasian perpetrators exhibited a preference for targeting individuals of the same ethnicity, whereas minority groups such as Māori demonstrated a more even distribution of victims across different ethnicities. This finding is consistent with the international sexual homicide literature (Chan et al., 2010; Safarik et al., 2002).

In the current study, victims were all female, and in more than 60% of cases they were Caucasian. Victims tended to have more distant relationships with perpetrators, with over half of incidents involving strangers. Only two incidents involved a close relationship between the perpetrator and victim, and in both cases the same perpetrator was involved. This replicates patterns of relationships between perpetrator and victim seen in previous sexual homicide research (Carter et al., 2017; Darjee & Baron, 2013). Selecting victims with whom the perpetrator has minimal or no pre-existing relationship may serve as a psychological defence mechanism, allowing them to dissociate themselves from the consequences of their violent actions and avoid feelings of guilt or remorse. Alternatively, targeting individuals outside their social circles may reduce a perpetrator’s chances of being recognised or linked to the victim, in turn making it harder for law enforcement to identify them as the perpetrator (Chan, 2024).

Consistent with some prior research, asphyxia was found to be the most common cause of death (Abrahams et al., 2017; Häkkänen-Nyholm et al., 2009). The current research found that lethal asphyxiation was typically achieved through melee or an airway-restrictive device. Conversely, firearms were rarely used by perpetrators, with close contact weapons being more likely to be used than more distant means of killing. This is consistent with previous research (Chan et al., 2019; Chan & Heide, 2008). It has been suggested these methods may provide offenders with a greater sense of intimacy and control during the incident, fulfilling some kind of psychological need that may partially motivate the offence. Stefanska et al. (2019) found at least one-third of their sexual homicide perpetrator sample could be classified as sexual sadists; expressing anger, power, or a combination of the two through the use of close contact weapons, providing offenders with experiences of sexual gratification or euphoria (Chan et al., 2011). As noted earlier, firearm regulation may also play some role in the rarity of firearm use in our sample. Firearm availability in New Zealand is tightly regulated, particularly for handguns, and this may limit the range of weapons available to those who offend. Examining the types of firearms used in these offences could be a valuable avenue of exploration in future research.

Engagement in multiple forms of sexual behaviour during the offence was also uncommon in the current sample. Just over half of analysed cases involved only one sexual behaviour, indicating that not all sexual homicide offenders involve intricate offences. Interestingly, at a descriptive level it appears the presence of multiple sexual behaviours is more closely associated with non-sexual motivations than sexual, suggesting that sexual homicides may become more grotesque when they arise from emotional responses such as anger or jealousy, or in the process of other criminal offending. Homicide literature has consistently connected excessive violence, referred to as “overkill” with emotional, impulsive, or anger-related motives (Chopin & Beauregard, 2021; Kopacz et al., 2022); similar associations may exist between these motivations and sexual behaviours as well. The prevalence of non-consensual vaginal penetration was higher than that found in previous research (Beauregard & Martineau, 2013; Darjee & Baron, 2013), occurring in just under two-thirds of cases in our sample; this is still below what might be expected for an offence of this nature. This provides further evidence against vaginal penetration being a defining feature of sexual homicide.

Contrary to previous research, spontaneous killings were not associated with more body disposal attempts than premeditated killings in the current sample. This discrepancy may result from the highly limited sample sizes used in previous studies, which tend to report that premeditation occurs more frequently in sexual homicides where the body is moved. For instance, Beauregard and Proulx (2002) conducted their analysis on only ten sexual homicide offenders. However, it must be acknowledged that premeditation could not be determined in over one-third of the cases in our current study, and so the ability to identify any effect was greatly limited.

Regarding the association between sexual behaviours and body disposal, few notable associations were identified, with only asphyxia being markedly more frequent when body movement occurred than when it did not. This finding is in contrast to Chai et al. (2021), who reported a trend (although not statistically significant) that those who died by asphyxiation were less likely to have their body moved. This may be due to their larger sample (250 solved sexual homicides) or the extended time period over which their data was sourced from, between 1948 and 2010. The otherwise lack of clear relationships regarding sexual behaviour and body disposal in the current sample was not surprising; the heterogeneity observed throughout previous research studies was repeated here, reinforcing the premise that understanding offence behaviour in sexual homicide cases based on perpetrator and crime characteristics is a daunting task, given the diverse, unconventional, and often unusual methods employed in these crimes. Moreover, the involvement of unique paraphilias and unsystematic decision-making further complicates this process.

Limitations and future research

Although our sample size was arguably moderate compared to many studies in this field, it remains a significant limitation. This is a trend amongst research in this area due to the rare nature of sexual homicide. Nonetheless, the small sample has implications for statistical power and prevented planned bivariate analyses from being conducted.

Additionally, comparing body disposal factors against sexual behaviours has again demonstrated the heterogeneity observed throughout the sexual homicide literature. Future research might endeavour to perform clustering on a larger sample to identify natural variable groupings between sexual behaviours and other crime-scene factors. These clusters might then show stronger associations in regard to body disposal factors than individual variables alone. For example, profiling in this area may benefit from distinguishing incidents by motivation to identify how sexual and non-sexual behaviours cluster in relation to this.

A rarely-noted limitation in this field is the potential difficulty in identifying sexual homicide in male victims. Sexual behaviour on male victims may be more difficult to detect during investigation and consequently go unrecorded as a homicide of this nature. Additionally, it may that be a potential bias exists where investigators and coroners assess female victims more extensively for sexual abuse compared to males. Thus, although the sample in the current study comprised female victims only, this is not necessarily reflective of the true gender identities of all sexual homicide victims. This consideration is important as perpetrators have been shown to utilise different strategies during the crime-commission process when targeting male victims (Chopin & Beauregard, 2023)

Conclusion

The current study was able to provide support for the existence of numerous international trends in New Zealand sexual homicides. In line with global findings, the victims’ gender and age mirrored familiar patterns. Asphyxia emerged as the most common cause of death, and various sexual behaviours and motivations were observed in these offences. Body disposal did not appear strongly associated with the presence of sexual behaviours as initially assumed.

Some important implications arise out of these findings. As necessitated by the rarity of sexual homicide and relatedly emerging nature of sexual homicide research (Gerring, 2012), this study presents a more complete, precise, and valid description of reality than previously seen in New Zealand. The detection of replicable patterns across international samples indicates potential opportunities for law enforcement to focus upon during the investigation of these crimes. However, it is essential to acknowledge that this study highlights the need for more research to be conducted to better understand the nature and heterogeneity of these offences, and the offenders and victims involved. Encouragingly, recent years have witnessed an intensification of research in this domain, although the field’s progress is somewhat constrained by limited sample sizes. Nevertheless, there is a growing body of research uncovering potentially meaningful patterns.

Our findings regarding the prevalence of sexual behaviours during these events also indicate sensationalised media coverage of sexual homicides can perpetuate harmful stereotypes and misperceptions. According to our study, sexual homicide cannot be characterised by overt physical brutality or be primarily attributed to sexual desire. These research findings hold the potential to facilitate responsible and considerate reporting of such cases by educating journalists and media professionals on accurate portrayals.

Finally, the descriptive nature of this research provides a building block for future research; studies which may further validate meaningful patterns in offences to help in the investigation, prosecution, and treatment of those involved in such a rare and impactful crime. Profiling literature in particular benefits from examination of sexual homicide in local jurisdictions, where it is vital that investigators have access to research that supports the application of best-practice profiling approaches that are relevant to the unique local context.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 If exclusive to genitals, coded “No” to distinguish from subsequent category (Non-insertional genital torture).