Firesetters or Murderers? Arson-Homicide Offenders in the Netherlands

(1)

Master Thesis

Leiden University

Faculty of Governance and Global Affairs

Crisis and Security Management, 2016-2017

Firesetters or Murderers?

Arson-Homicide Offenders in the Netherlands

June 8, 2017

Author:

N. Boer

Student ID:

1216260

Supervisor:

Dr. M.C.A. Liem

Second Reader:

Dr. G.M. van Buuren

(2)

1 Master Thesis Crisis and Security Management

Word count: 19.118 words (including bibliography and appendices) Student e-mail address: n.boer@umail.leidenuniv.nl

(3)

2

Executive Summary

The use of fire in a case of homicide significantly obstructs the research process in a fatal fire incident, regardless of whether the victim died as a direct consequence of the fire or not. The burning of a body can prevent or delay identification and destroy evidence of the primary cause of death or other evidence present at the crime scene (Sapp and Huff, 1996, p.2). Moreover, it raises questions with regard to the nature of the crime: did the offender intend to fatally harm other individuals, or was the offender solely intending to commit arson? And what motivated the offender to commit such a crime?

The primary purpose of this study is to increase knowledge on the offenders involved in arson-related homicides, by identifying event- and offender characteristics. The main question is whether arson-homicide can primarily be explained as homicidal acts, or if they are primarily driven by firesetting behaviours. The study was conducted by statistical analysis of data from the Dutch Homicide Monitor, which contains information on the victim-, offender-, and event characteristics of all known homicides that have taken place in the Netherlands between January 1, 2009 and December 31, 2015. To ensure completeness and quality of the data, arson-homicide cases in the database were updated on the basis of content analysis of news reports and judicial documents on relevant cases. Analysed variables include (but are not limited to) the gender and age of the offenders, the primary method of violence causing death, and the intention of firesetting of the involved offenders.

The results of this study suggest that arson-homicide offenders can primarily be regarded as homicide offenders, and that they are not necessarily driven by firesetting behaviour. This implies that risk management- or prevention strategies for countering the prevalence of arson-homicides should be aimed primarily at homicidal offenders rather than firesetters. The author recommends further research into demographics and backgrounds of the offenders, and into the meaning of fire for different ethnic or cultural populations offenders might come from.

(4)

3

(5)

4

List of Tables

Chapter 2

 Table 2.1. Search Terms for Literature Review 9

 Table 2.2. Victims per Type of Arson-Homicide, Chicago 1965-1995 14  Table 2.3. Categorizations and Characteristics of Dutch Firesetters 19  Table 2.4. Types of Homicide in the Netherlands, 1992-2009 21

Chapter 3

 Table 3.1. Operationalization of Key Concepts 24

Chapter 4

 Table 4.1.1. Distribution of Gender and the Use of Arson 26  Table 4.1.2. Distribution of Age and the Use of Arson 27  Table 4.1.3. Distribution of Age and Gender amongst Arson-Homicide Offenders 28  Table 4.2.1. Method of Violence and the Use of Arson 29

 Table 4.2.2. Method of Violence and Type of Arson 31

 Table 4.2.3. Method of Violence and Age of Arson-Homicide Offenders 32  Table 4.2.4. Method of Violence and Gender of Arson-Homicide Offenders 32

 Table 4.3.1. Type of Homicide and the Use of Arson 34

 Table 4.3.2. Type of Homicide and Age of Arson-Homicide Offenders 35  Table 4.3.3. Type of Homicide and Gender of Arson-Homicide Offenders 35

 Table 4.3.4. Type of Homicide and Type of Arson 36

 Table 4.4.1. Type of Arson and Age of Arson-Homicide Offenders 38  Table 4.4.2. Type of Arson and Gender of Arson-Homicide Offenders 39

 Table 4.5.1. Intention and Type of Arson 40

 Table 4.5.2. Intention and Age of Arson-Homicide Offenders 40  Table 4.5.3. Intention and Gender of Arson-Homicide Offenders 41

 Table 4.6.1. Motive and the Use of Arson 42

List of Case Descriptions

 Case 1: Chalet girl 28

 Case 2: ‘Possessed by Spirits’ 33

 Case 3: Secondary Arson-Homicide, Primary Arson-Suicide 37

 Case 4: Zeister Brandmoord 39

 Case 5: Firesetting with Lethal Consequences 41

(7)

6

1. Introduction

1.1. Arson? Homicide? Arson-Homicide?

In September 2013, a 36-year-old male was sentenced to sixteen years in jail for murdering his girlfriend. In January 2013, the victim was discovered by firefighters who were attempting to extinguish an indoor residential fire in Rotterdam. The pathologists struggled to retrieve the identity of the victim, since the fire caused significant damage to the body. Autopsy later revealed that the victim had suffered blunt force traumas prior to being injured by the fire (Rechtbank Rotterdam, 2013). In the same month, firefighters discovered another body after extinguishing a residential fire in Groningen. In this case, autopsy also reveals that the victim was likely killed before the fire was started (Moorzaken.com, 2013). Five months later at a public parking space in a northern province of the Netherlands, passers-by discovered the tied up and charred body of another homicide victim.

Each of these cases presented a significant challenge to authorities and specialists, since the use of fire caused a lot of damage to the bodies as well as a major loss of traces and evidence. Upon discovery of homicides like these, many questions arise: What was the primary cause of death? Was the victim still alive when the fire was set alight? Did the offender intend to kill someone? What motivates offenders to commit such crimes? The present study aims to shed light on the phenomenon of arson-related homicides and the offenders involved.

1.2. Societal and Academic Relevance

In the Netherlands in 2013, over 36,000 fires were registered by the fire departments. Of the outdoor fires, 85 percent were started deliberately, as were 20 percent of the indoor fires (Dalhuisen, 2016; Statistics Netherlands, 2014). Ninety-two people lost their lives in fires in the Netherlands in 2013 (Statistics Netherlands, 2014), and 147 people were registered as victims in homicide cases (Statistics Netherlands, 2016).

Research on a possible relationship between arson and homicide is scarce: only a handful of scientific articles has until now been devoted to research on arson-related homicides and similar projects are yet to be conducted in the Netherlands1. While a relationship may be difficult to identify, several parties would benefit from information on this topic (such as fire-, crime scene- and homicide investigators) (Sapp & Huff, 1996). Researchers have argued, for example, that it is necessary to increase knowledge on arson offenders, in order to assist

(8)

7

involved agencies in the investigation of unsolved arson cases (Douglas, Burgess, Burgess and Ressler, 2006. P. 236). Moreover, in order to design appropriate risk management- and prevention strategies, one needs to know whether to regard arson-homicides primarily as homicides, or as arson with lethal consequences. A first step in identifying the relationship between homicide and arson could be to compare the motives and intent of offenders in cases belonging to the different categories – non-lethal arson, arson-related homicides, and homicides by other means. Moreover, it could prove be helpful to enhance knowledge on several of the characteristics of the population of offenders.

1.3. Aim and Research Questions

The aim of this research project is to contribute to a better understanding of arson-homicides and the offender involved in these incidents. Therefore, this research project aims to answer one overarching research question:

Can acts of arson-related homicide be explained primarily as homicidal acts, or are they primarily driven by firesetting behaviour? Or can arson-related homicide be seen as a different phenomenon altogether?

In order to present the results of this research in a structured way, the overarching question is divided into six different sub-questions. Questions 1-3 provide information on the population of offenders with regard to their personal characteristics and characteristics of the events they are involved in. Questions 4-6 aim to shed light on the relation between firesetting, homicidal acts and arson-homicides. The questions are defined as follows:

1. What distinguishes offenders of arson-homicides based on their gender and age?2

2. What methods of violence causing death do offenders resort to when committing arson-related homicide?

3. What are the different types of homicides that offenders of arson-related homicides are involved in?

4. What types of arson do homicide offenders resort to?

5. What has been the primary intention of firesetting of offenders in cases of arson-related homicide?

6. What motivates offenders to commit arson-related homicides, and how does this relate to motivations of other homicide- and arson offenders?

Next to answering these questions, several characteristics of the population of offenders in cases of arson-related homicides in the Netherlands will be discussed and where possible, the results will be compared to the results of prior research from other regions.

(9)

8 1.4. Reading Guide

The following chapter (Chapter 2) will introduce the topics of arson3, homicide, and arson-related homicide by outlining the prevalence of these incidents in the Netherlands and by reviewing prior empirical research on these phenomena. Moreover, it will discuss attempts in the literature to categorize homicide- and arson offenders and incidents on the basis of their characteristics, motivations, or the relationship of offenders with their victims. The chapter will conclude by outlining the knowledge gap to which the present study aims to contribute to. Chapter 3 will explain the research design and methodologies of the present study. In the subsequent chapter, the results of the study will be displayed and compared to the results of prior research, in an attempt to answer the sub-questions discussed above. The general research question will be answered in the concluding chapter, which will also discuss the shortcomings of this study, together with some suggestions for future research.

3_{The terms ‘arson’ and ‘firesetting’ will be used interchangeably throughout this document, referring to the}

same act. The same applies to the terms ‘arson-homicide’ and ‘arson-related homicide’. The precise operationalization of these concepts is discussed in Chapter 3 (Methodology).

(10)

9

2. Literature Review

The present chapter introduces the topics of arson, homicide and fire-related deaths by discussing the prevalence of these incidents within the Netherlands and describing previous research on these topic. Subsequently, the chapter reviews the existing literature on arson-related homicides and discusses the themes and general tendencies found in the literature. The chapter concludes by discussing several pieces of literature on the typologies and categorization of arson- and homicide offenders. The literature discussed was retrieved through Leiden University’s catalogue and via Google Scholar. The table below presents the main search terms per subject. Besides searching for relevant articles online, the snowball method was used in order to retrieve additional literature. The most important sources in applying this method were Dalhuisen (2016) and Brookman (2005).

Table 2.1. Search Terms for Literature Review

Subject Search terms

Arson ‘arson’; ‘firesetting’; ‘arson offenders’; ‘firesetting offenders’; ‘firesetters’;

‘firesetting motives’;

Homicides ‘homicide’; ‘homicide offenders’; ‘motives for homicide’;

Arson-homicides ‘arson-homicide’; arson-related homicide’; ‘homicidal burning’; ‘lethal arson’;

‘arson-related homicide’; ‘burn deaths’.

2.1 Homicides, Arson, and Fire-related Deaths

2.1.1. Homicides

Researchers have increasingly been conducting systematic, empirical research on homicide in the Netherlands since the last decades, but in the past, efforts in doing so have been relatively limited in comparison to systematic research on homicide in other Western countries such as the United Kingdom, the United States and Australia (Ganpat & Liem, 2012, p. 329). Statistical registration of unnatural causes of death and characteristics of homicide cases tended to concentrate on either individual victims, offenders, or event characteristics, hereby failing to provide a complete overview of homicides in the Netherlands (Nieuwbeerta & Leistra, 2003). Uniform registration of Dutch homicides from 1992 onwards, however, has been performed in the Dutch Homicide Monitor (Ganpat & Liem, 2012, p. 331). All known cases of murder4 and manslaughter5 that have taken place from 1992 onwards have been

4_{Article 289 and 291, Dutch Code of Criminal Law} 5_{Article 287, 288 and 290, Dutch Code of Criminal Law}

(11)

10

registered in this database (idem). Monitoring homicides in this database has facilitated research on the prevalence of homicide in the Netherlands and comparison of national trends with international trends. Similar to other European countries, homicide rates in the Netherlands have decreased over the past few decades (Aebi et al., 2010; Ganpat & Liem, 2012). Ganpat and Liem (2012, p. 332) point out at the Dutch population has increased with one million individuals since the beginning of the 1990s, and that the homicide rate decreased from 1.7 to 1.5 per 100.000 inhabitants from the early nineties onwards, remaining steady for some years but further decreasing to a rate of 1.2 per 100.000 inhabitants in 2012. Nieuwbeerta and Leistra (2007) stated that in the Netherlands, 223 individuals per year die as a consequence of homicidal acts (as cited in Ganpat & Liem, 2012, p. 331).

2.1.2. Arson

Until 2014, Statistics Netherlands (Centraal Bureau voor Statistiek) published annual reports on the prevalence of (reported) fires in the Netherlands and information on the nature of these fires, as well as the performances of the fire departments. Between 2000-2013, an average of 44.5 thousand fires were annually reported in the Netherlands (Dalhuisen, 2016, p. 4). An estimated 22.3% of these fires are the result of firesetting behavior (11.2% of the indoor fires, and 27.9% of the outdoor fires). Despite the fact that there the number of reported fires has been declining over the past decades – which has been caused primarily by a remarkable decline in outdoor fires – firesetters still present a major challenge to authorities. In the Dutch Criminal Code6_{, firesetting is referred to as the intentional or neglectful act of firesetting, and} is regarded a crime posing a threat to the general safety of persons and property7_{. Dalhuisen} (2016) describes that

occasionally, despite the severity of this offence, firesetting is seen as subordinate to other crimes like murder. In these instances, firesetting is seen as the means to carry out the murder or to cover up evidence, and firesetting subsequently remains unmentioned in the indictment. (p.4)

From this perspective, it might seem like firesetting is not all that destructive and maybe even less relevant than crimes resulting in more direct harm to individuals. However, whether occurring accidentally or set deliberately, it must be recognized that fires can present lethal danger to human lives. The following section therefore discusses the prevalence of fire-related deaths.

6_{Article 157 and 158, Dutch Code of Criminal Law} 7_{Dutch Code of Criminal Law, Section VII.}

(12)

11

2.1.3. Fire related deaths

In the Netherlands in 2009, 57 victims died in fire incidents. This number increased in 2010, when 65 victims, one of which was a fire-fighter on duty, were fatally injured (Statistics Netherlands, 2011). In 2013, over 36,000 fires were registered by the fire departments. Of the outdoor fires, 85 percent were started deliberately, as were 20 percent of the indoor fires (Dalhuisen, 2016; Statistics Netherlands, 2014). Ninety-two people lost their lives in fires in the Netherlands in 2013 (Statistics Netherlands, 2014). In 2016, 41 residential fires with fatally injured victims have been registered by the fire departments. Of these fires, 32 were marked as ‘fatal residential fires’, in which 37 victims died (Instituut Fysieke Veiligheid, 2017). These 32 incidents, however, were not registered to be a consequence of arson, homicidal- or suicidal acts. Six fatal residential fires with 7 victims are presumed to be cases of homicide or suicide.

Prevention of fires resulting in fatal injuries

The literature on fire-related deaths adopts several different perspectives on the phenomenon. Several authors have addressed the issue of fatal fire injuries and prevention methods, often focusing on specific groups of people appearing to be at a higher risk of being fatally injured in a fire. Some authors tend to focus on age groups predicting higher vulnerability, such as young children or the elderly (Elder, Squires & Busuttil, 1996), while others research the relationship between ethnicity and higher risk: Bishai and Lee (2010) also concentrated their analysis on older generations, but looked specifically at the higher risk of fatal injuries of Native Americans and African Americans and concluded that the higher risk was mainly caused by housing conditions. Again, other authors tend to focus on a specific time span rather than groups of people (Mulvaney et al., 2008). However, most of these research projects were performed in order to advise agencies on effective methods of prevention and intervention in cases of accidental fires rather than cases of arson. The possibility exists, however, that a fatal fire was the result of deliberate firesetting or even that fire was used as a modus operandi in a homicide case.

Arson-related homicides

Like homicides in the context of robbery and burglary, homicides that take place in the context of arson attacks are often stranger homicides, which means that the offenders and victims had no identifiable connection prior to the homicide (Brookman, 2005, p. 147). In these cases, it is difficult to establish the motive, and find out whether homicide was the primary intent, or whether fatally injuring the victims was unintended (Brookman, 2005, p.

(13)

12

147). So far, research has shown that the intent of firesetting offenders is rarely to related to killing or injuring others, but that arson was often the main intent (Canter & Fritzon, 1998; McEwan & Freckelton, 2011, p. 320). Either way, the use of fire in a case of homicide significantly obstructs the research process in a fatal fire incident, regardless of whether the victim died as a direct consequence of the fire or not. The burning of a body can prevent or delay identification and destroy evidence of the primary cause of death or other evidence present at the crime scene (Sapp and Huff, 1996, p.2). The paragraph below will discuss the existing empirical research on arson-related homicides.

2.1.4. Regional Differences: A Review of Empirical Studies on Arson-related Homicides

Research Methods and Population size

The majority of studies discussing arson-related deaths or homicides focus on specific geographical regions and make use of autopsy reports of local or regional mortuaries and/or case files of regional law enforcement agencies as their data sources. Studies on arson-homicides in Chicago, US (Drake & Block, 2003) and Australia (Davies & Mouzos, 2007; Ferguson, Doley, Watt, Lyneham & Payne, 2015) made use of national homicide databases. The majority of studies tends to focus on victim characteristics, more specifically their gender, age, and the primary cause of death and timing of burning. In some cases, socio-demographic characteristics of the victims were discussed (Drake & Block, 2003; Kumar et al., 2015; Sapp & Huff, 1996). Offender characteristics were discussed for arson-homicides in Australia and the United States (Davies & Mouzos, 2007; Drake & Block, 2003; Ferguson, et al., 2015; Sapp & Huff, 1996). Significant differences in population size and the periods studied are prevalent between the studies.

In the United States, 183 primary victims died as a consequence of arson-related homicide between 1985 and 1994 (Sapp & Huff, 1996). This study, however, included the primary victim of each individual case in the analysis, thus excluding all other victims from multiple-victim cases. In Chicago, US, between 1965 and 1995, 461 multiple-victims (2% of all homicide victims) died as a consequence of arson-related incidents, distributed over 269 cases (Drake & Block, 2003, p. 227). In Cape Town, South Africa, analysis of autopsy reports of individual morgues showed that 54 homicides involved the use of fire in the first six months of 1986 (Duflou et al., 1988); 35 individuals were victimized in this area in 1991-92 (Lerer, 1994). Similar research methods were adopted for two regions in India. From both these research projects it appears that the incidence of burn deaths (accidental, suicidal and homicidal) in

(14)

13

India is remarkably higher than it is in other Western countries, but it appears that numbers in South Africa are even higher (Ambade & Godbole, 2006; Kumar, Verma, Singh, & Singh, 2015). For the region of Nagpur in the year 1998, 21,6% of all medicolegal deaths were due to burning. Out of a total 384 burn deaths, 10 cases (2,6%) were classified as homicidal burning (Ambade & Godbole, 2006). Regional differences exist within India as well. The rate of homicidal burn deaths was significantly higher for North India compared to the rates in Nagpur (Ambade and Godbole, 2006): out of the total 1369 cases of burns in the period 2008-2012, 536 cases (38.5%) were homicidal deaths (Kumar et al., 2015), whereas this was 2,6% in Nagpur in 1998 (Ambade & Godbole, 2006). Between 1993 and 2003, 12 victims died as a consequence of criminal burning in Lyon, France (Fanton, Jdeed, Tilhet-Coartet & Malicier, 2006). In Australia, 123 arson-related homicide incidents occurred in the period from 1989-2010, out of a total 6,265 homicide cases (Ferguson et al., 2015). These 123 cases involved 170 victims and 152 offenders (Ferguson et al., 2015).

Cause of Death and Type of Arson

While some studies predominantly analysed the cause of death of the victims of fire-related incidents, other studies (Davies & Mouzos, 2007; Drake & Block, 2003) have created typologies of arson-related homicides. From different studies it appears that gunshot wounds, blunt force trauma or sharp object injuries are the most common causes of death in post-mortem arson-homicides. Retrospective analysis of autopsy reports revealed that between 1998 and 2008 in Ankara, Turkey, thirteen individuals were homicide victims burned post-mortem, five of which were killed by a firearm (Tümer et al., 2012). In South Africa, all victims of arson-related homicide showed signs of either blunt object- or sharp object trauma or gunshot wounds (Duflou et al., 1988, p. 292). According to the authors, homicidal burning is often seen in situations of civil unrest in South Africa. These practices are referred to as necklace murders, in which “an automobile tire filled with gasoline is placed around the torso of the victim and then set afire” (Duflou et al., 1988, p. 293). In cases of necklace murder, bodies typically show a large number of other injuries next to the burns, but the primary cause of death in these cases was blunt force trauma (Duflou et al., 1988). Lerer (1994) found that 14% (N=5) of the arson-homicides in Cape Town in 1991-92 were due to necklacing. In their research on arson-homicides in the US, Sapp and Huff (1996) concluded that in almost 60 percent of the cases with multiple victims, the fire-related injuries were received ante-mortem, while this was found in roughly 24 percent of the single-victim cases (1996, p. 27).

(15)

14

Using the timing of burning of the victim as a guideline, Drake and Block (2003) have defined four different types of arson-related homicides for their research on homicides in Chicago (US). The first of these is ‘primary arson-homicide’, which is defined as a homicide in which the victim was killed in an arson fire set to a structure or vehicle. When arson is committed after the homicide, this is classified as a ‘secondary arson-homicide’ (Drake & Block, 2003, p. 227). Cases in which the victim was burned to death because he or she was set on fire, but was not in an arson, are classified as ‘person burned/set on fire’. The last type of arson-related homicide is the type ‘body burned’, which is used to describe cases in which the victim’s body is burned after a homicide (Drake & Block, 2003, p. 227). To distinguish between these types, it is necessary to determine whether the victim died as a direct consequence of the fire, or whether the burns were produced after death. Determination of when the fire-related injuries were received, is thus a critical part of the investigation (Polson, 1956; Sapp & Huff, 1996, p. 3). The findings of Drake and Block (2003) are summarized in the table below8:

Table 2.2. Victims per type of arson-homicide, Chicago 1965-1995.

Davies and Mouzos (2007) conducted research for arson-related homicides in Australia, using categorizations of arson-homicides similar to those of Drake and Block (2003). Results showed that 68% of the arson-related homicides occurring between 1990-2005 involved the use of fire as a primary modus operandi, and 29% involved secondary arson (Davies & Mouzos, 2007). Follow-up research on Australian arson-homicides by Ferguson and colleages (2015) did not apply the categorizations used by Davies and Mouzos (2007). Fanton and colleagues (2006, p. 88) found that in Lyon, 75% of the victims (N=9) were subject to post-mortem burning, and that the predominant causes of death were either firearm- or blunt object injuries. Three victims (25%) were burned alive (Fanton et al., 2006).

8_{Adapted from Drake & Block (2003, p. 231).}

Type of Arson-Homicide Number of victims % of arson-homicides

Primary arson 293 64

Person burned (ante-mortem) 43 9

Secondary arson 78 17

Body burned (post-mortem) 47 10

(16)

15

Victim Characteristics

Over one-third of the victims from the United States were aged between 18-29 years, the mean age of victims being 36.3 years. Overall, female victims were slightly younger than male victims (Sapp & Huff, 1996). In Chicago, children until 14 years of age and adults over sixty were overrepresented as victims of arson-homicides compared to victimization in non-arson homicides (Drake & Block, 2003, p. 231). In Lyon, the mean age victims was 38 years (Fanton et al., 2006). Turkish victims were older: their mean age was 43.5 years (Tümer et al., 2012). Individuals aged between 25-34 years were most likely to be victimized in Cape Town (Lerer, 2994, p. 345). The gender distribution of victims also differs between regions. In the United States, 54% of the victims of arson-homicides in the period 1985-1994 were female, 46% of the victims were male. In Chicago, women victimization was also quite high: Drake and Block stated that “43% of primary arson victims and 50% secondary arson victims were women” (2003, p. 233). Of the victims whose bodies were burned ante-mortem, 30% were women, as were 39% of the victims whose bodies were burned post-mortem (Drake & Block, 2003, p. 233). Female victim preponderance was significant in Nagpur, India, as 70% of the victims of homicidal burning in 1998 were female. This female preponderance might partially be explained by the phenomenon of ‘bride burning’, also called the Dowry system, in which women are burned because their dowries are deemed insufficient by their husbands or in-laws (Ambade & Godbole, 2006; Kumar et al., 2015). This often leads to injuries, but can be fatal in some cases, thus adding to the number homicidal burnings. Kumar and colleagues (2015) also concluded that the incidence of burn deaths was higher for females in North India for the years 2008-2012. In Australia between 1989-2010, 57% (N=96) of arson-homicide victims were male, and 43% (N=74) of the victims were female (Ferguson et al., 2015). French arson-homicide victims were equally distributed (Fanton et al., 2006). According to Lerer (1994) victims in Cape Town in 1991-92 were predominantly males (85%, N=29). A similar gender distribution for victims in Ankara between 1998 and 2008 (Tümer et al., 2012).

Offender Characteristics

Difficulties in analysing offender characteristics are caused by the high number of unsolved cases. Sapp and Huff (1996) found that for the 183 victims who deceased between 1983-1994, only 31 cases were cleared by arrest (1996; p. 5). Although some regard arson as a ‘male’ offense, Drake and Block (2003) found that there were no significant differences between (primary) arson-homicides and non-arson homicides with regard to the distribution of gender amongst the offenders. Between 1965 and 1995, 15% of the primary arson-homicides were

(17)

16

committed by female offenders, as were 12% of the non-arson homicides. This distribution differed in Australia for the period between 1989-2010, where 74% (N=113) of arson-homicide offenders were male, and 26% (N=39) of the offenders were female (Ferguson et al., 2015). With regard to the motives of offenders in arson-related homicides, however, only Davies and Mouzos (2007) have attempted to present an oversight. The authors found that in arson-related homicides in Australia occurring between 1990 and 2005 showed that most offenders were motivated by a desire for revenge or provoked by domestic arguments or arguments occurring in other situations. The consumption of alcohol also motivated offenders in a small number of cases (Davies & Mouzos, 2007, p. 5).

However, what stands out from the literature on arson-related homicides is that most authors focus on the victims rather than the perpetrators. The majority of the articles discussed above retrieved data from autopsy reports, and analyzed the characteristics of victims like their gender, age, socio-demographic characteristics and the primary causes of their deaths. However, perpetrators are often left out of the research projects. This might partially be explained by the fact that many perpetrators are never caught and these homicides end up on the shelves between other cold cases. Moreover, there is a large dark number in homicide research (Brookman, 2005), and it is likely that the numbers of known homicides are skewed in relation to the real number of homicides occurring. While this is true also for arson- and homicide cases, a wide range of literature exists that attempts to create typologies of, or categorize offenders involved in such incidents. This literature is discussed in the following paragraphs.

2.2 Typologies and Categorization of Offenders

2.2.1. Arson Offenders

A growing body of research devotes attention to firesetting offenders, attempting to categorize the arsonists on the basis of their motives or their mental health (Dalhuisen et al., 2015, p. 440). The purpose of such classification systems is to allow for better understanding of a complex whole by reducing it to smaller, more manageable parts (Doley, 2003, p. 346). The earliest attempts of categorizing firesetters focused primarily on the motivations of the offender, following the attempt made by Lewis and Yarnell (1951). Icove and Estepp (1987) have created motive-based offender profiles of arsonists and offenders involved in other fire-related crimes. The authors defined motive as the “inner drive or impulse that is the cause, reason or incentive that induces or prompts specific behaviour” (Icove & Estepp, 1987, p. 17).

(18)

17

Various authors have acknowledged, however, that identifying clear motivations for the act of firesetting is a difficult task, and that offenders may have more than one reason to undertake their acts (Dalhuisen, 2016; Doley, 2003; Fineman, 1995; Prins, Tennent & Trick, 1985; Rider, 1980; Wood, 2000). Moreover, limited disclosure by offenders leaves researchers guessing for their motives (Dalhuisen, 2016). Despite these difficulties, several authors have argued that researching motivations is important to gain more knowledge on firesetters. Dividing firesetters into different motivational categories can be useful to identify different types of arson. For example, Douglas, Burgess, Burgess and Ressler (2006) argue that offenders motivated to set fire in order to conceal other criminal activities may be attempting to hide a homicide or suicide, but stress that their firesetting act could also be about burglary or car theft. Different indicators found at the crime scene can aid in determining the type of scenario in which the incident has occurred (Douglas et al., 2006, p. 274). It is perhaps not surprising that numerous articles have been published in which motivations were the primary factor on the basis of which categorizations of arson offenders were made9.

Inciardi (1970) analysed the motives of convicted firesetters in the US for the period 1961-1966, and concluded that over half of the firesetters were motivated by revenge, and that these offenders often targeting known victims after confrontations with them. Other frequently occurring motives were jealousy or hatred. The second largest group (almost one in five) were excitement firesetters (Dalhuisen, 2016; Inciardi, 1970). Prins, Tennent and Trick (1985), conducting their study in the United Kingdom in the period of 1980-1984, also concluded that revenge was the most dominant motive amongst convicted firesetting offenders (Dalhuisen, 2016). A similar study on convicted firesetters in the United Kingdom for 1983-1993 confirmed revenge to be the most important motivational factor (Rix, 1994). In his research, the author made a distinction between motives often seen in male and female offenders, differing in this regard from previous authors providing categorizations (Dalhuisen, 2016). Revenge was the most common motive in both groups. Excitement firesetting (including cases of sexual arousal) was more commonly observed in male firesetters, while using firesetting as a cry for help or to achieve rehousing were more commonly seen in women. Young boys often set fire as a means of vandalism, while this motive was not applicable for women (Dalhuisen, 2016; Rix, 1994).

9_{An elaborate overview of studies basing categorizations solely on the motivations of firesetting offenders can}

(19)

18

Since it is often difficult to find one distinct motive, partly due to non-disclosure by firesetters and retrospective assessment of the cases, categorizing firesetters on the basis of their motivation is generally a difficult task (Dalhuisen et al., 2015, p. 440). Canter and Fritzon (1998) combined event- and offender-related characteristics to create a typology of firesetters. The authors distinguished four different themes in the behaviour of firesetters: their objectives could be expressive (internal) or instrumental (external), and their target could be a person (internal) or an object (external) (Canter & Fritzon, 1998; Dalhuisen et al., 2015). Dalhuisen, Koenraadt and Liem (2015) stated that motives of crimes involving firesetting are often hard to identify, which is why the mental sanity of the offender is often questioned. Subsequently, suspected arsonists are often subjected to a forensic mental health examination before the trial starts (Dalhuisen et al, 2015, p. 440; Prins, 1994; Van Kordelaar, 2002). Mental health disorders are frequently diagnosed in arson offenders (Dalhuisen, 2016). Ducat, McEwan and Ogloff (2013) attempted to compare the background and characteristics of firesetters to the characteristics of non-firesetting offenders. In this study, they concluded that overall, there is a lack of differentiation between these two groups (Ducat et al., 2013, p. 561). The authors suggested that this could be due to the prevalence of criminal versatility in firesetters, and that these offenders are prone to offending behaviour in general, since firesetting was also commonly seen in combination with other types of violent behaviour (Ducat et al., 2013, p. 561-562). Examples of other types of violent behaviour could range from sexual offending to homicide crimes.

Dalhuisen (2016) researched the characteristics of Dutch firesetters in order to individualise and categorize firesetting offenders, studying populations of firesetters from three different sources: firesetters having undergone inpatient pre-trial forensic mental health assessments at the Pieter Baan Centre in the period 1950-2010, analysing assessments reports; convicted firesetters in the period May 2008-September 2012, analysing police files; and firesetters registered in the forensic registration and information system (FRIS database). The researched variables were not limited to motives, but also included mental health of the offenders, their demographic characteristics and backgrounds. Based on the results of this study, she proposed six subtypes of firesetters appropriate to categorize Dutch firesetters. An overview of these subtypes and their characteristics is displayed in table 2.3 on the following page.

(20)

19

Table 2.3. Categorization and Characteristics of Dutch Firesetters10_.

Subtypes Characteristics of offense Motivations Personal characteristics

Vandalism firesetters

Targeting freely accessible objects, without causing harm to other individuals.

Boredom; thrill-seeking; sensationalism; vandalism.

Young (school-attending), native boys with below-average intelligence. Usually not intoxicated at the time of the offense.

Disordered firesetters

Impulsive acts, targeting persons (often themselves or their own homes).

Influenced by psychosis or suicidal tendencies.

Often male (but sometimes female) in the age group 30-39, single, with average intelligence. Psychiatric diagnoses and prior convictions are common.

Disturbed-relationship firesetters

Targeting a known victims, often (ex-)partners, resulting in danger to these persons.

Instrumental motives; revenge; retribution; relationship

problems.

Firesetters with (below-) average intelligence, aged in their late twenties, functional in their social lives.

Opportunistic firesetters

Often work with accomplices, targeting houses or vehicles in residential areas. Financial or other instrumental motives; Crime concealment.

Dutch, single males with below-average intelligence, in their mid-twenties, often with a daily occupation. Usually not intoxicated at the time of the offense.

Desiring firesetters

Targeting randomly chosen objects; Fires impulsively, often only endanger property.

Fascination for fires and their aftermaths; fire as a means to alleviate tension.

Young, Dutch, single males with low intelligence. Often exhibit behavioural problems in their youth.

Multi-problem firesetters

Relatively high chance of bodily harm.

Revenge; retribution; cry for help.

Highest proportion of female

firesetters (16%). Prior convictions and recidivism are common. Limited to poor social networks. High levels of hostility.

2.2.2. Homicide Offenders

While classifications regarding firesetting focus mainly on the offenders, classification of homicides are often made on the level of individual incidents of homicide, taking into account multiple characteristics related to the event, the offender, or the victim. Examples include the gender and age of the individuals involved, the motivation or intention of the offender, and the relationship between the victim and offender. Despite a wide range of possible characteristics, tendencies in the literature exist to create dichotomized typologies. Psychological literature on violence and homicide often display a classic distinction between ‘expressive’ and ‘instrumental’ types of violence (Megargee, 1972, 1982; Block & Block, 1991, 1992; cf. Brookman, 2003, p. 83). Violent incidents in which the offender seeks some kind of gain or improve his or her position are classified ‘instrumental’ types of violence, while ‘expressive’ types of violence are more likely to be motivated by personal confrontation

(21)

20

(Brookman, 2003; Luckenbill, 1977-78) or a desire for retaliation or revenge (Decker, 1996; Felson, 1978). In this expressive/instrumental distinction, motivations for homicides are to some extent dichotomized. The same is often done when creating typologies of homicides based on the relationship between the victim and the offender. Some researchers describe these relationships as stranger/nonstranger relationships (Messner & Tardiff, 1985; Sampson, 1987) while other authors have defined victim-offender relationships as primary or secondary (Parker & Smith, 1979; Smith and Parker, 1980; cf. Decker, 1996). In these classifications, primary- or nonstranger relationships occur between intimate partners, family members or close friends, and secondary- or stranger relationships include acquaintances or strangers (individuals with little to no connection prior to the incident). Cross-classification between these typologies is also often seen in the literature. Generally, offenders of primary or non-stranger homicides – connected to their victims more intimately – are presumed to have expressive motives (Loftin, 1986; Maxfield, 1989). Offenders involved in secondary or stranger homicides, on the other hand, tend to have instrumental motives for committing homicide (Cook, 1987; Daly & Wilson, 1982; Block, 1981; Riedel, 1981, 1987; Riedel & Przybylski, 1993; Rojek & Williams, 1993; cf. Decker, 1996, p. 428).

As in research on firesetting, offender motivations are regularly used in homicide research to distinguish between different types of homicide (Granath et al., 2011; Miethe & Regoeczi, 2004). While the categorizations described in the paragraph above are dichotomized, other authors have divided homicidal events over more than two categories. Ganpat and Liem (2012) identified four main areas of research on homicide in the Netherlands: the epidemiology of homicides, offenders and victims; homicide subtypes based on the victim-offender relationship; homicide subtypes based on motive; and the sentencing of homicide offenders. In the third research area, however, motives may be predominant but are not the sole factors taken into account when creating subtypes. Next to offender motivations, the relationship between the victim and offender, and the context in which the homicide takes place are also taken into account when categorizing homicidal incidents (Ganpat & Liem, 2012, Granath et al., 2011). When researching homicides in this research area, authors often focus on one specific type of homicide, such as sexual homicides (Van Beek, 1999), honour killings (Nauta & Werdmölder, 2002; Van Eck, 2001; Yecilgöz, 1995) and criminal liquidations (Van de Port, 2001 cf. Ganpat & Liem, 2012). Ganpat and Liem (2012) provided an overview of the types of homicides occurring in the Netherlands between 1992 and 2009. These types are primarily based on the victim-offender relationships and the contexts of the

(22)

21

homicide. In total, nine different categories can be distinguished: four different types of domestic homicide (intimate partner-, child-, and parental homicides, and other family homicides), homicides that take place in the context of a fight or argument, robbery homicides, and criminal homicides (Ganpat & Liem, 2012, p. 333). The distribution of homicides (1992 to 2009) across these categories found by Ganpat and Liem (2012) is presented in the table below.

Table 2.4. Types of Homicide in the Netherlands, 1992-2009.

2.2.3. Knowledge gap: Offenders of Arson-Homicides

From the literature review, it can be concluded that a wide range of literature has been devoted to the characteristics of offenders, victims and incidents of arson or homicide, and the subsequent creation of typologies or categorizations of these offenders based on their characteristics. However, although some authors discuss offender characteristics in their research on arson-homicides, these discussions are often quite superficial and predominantly descriptive. There is a lack of literature attempting to categorize these offenders, which justifies the effort to gain more knowledge on this specific population in facilitate the creation of typologies or categorizations. The present study attempts to take a first step in this direction, and the methods and results will be discussed in the subsequent chapters.

Type of Homicide % of Homicides

Domestic Intimate partner Child killing Parental killing Other family 30 17 4 2 7 Argument 21 Criminal 12 Robbery 7 Sexual 3 Other 18

(23)

22

3. Methodology

3.1. Research Design

The primary aim of this research project is to improve knowledge on the scarcely researched topic of arson-related homicides in the Netherlands. The central research question and the sub-questions are based on observations from a small amount of studies, rather than a pre-existing theory on the subject. This implies that the research design is inductive rather than deductive. One of the aims of this project is to make a comparison of offender characteristics between arson-related homicides and homicide cases in which no arson was used. To answer the main research question and sub-questions, it is necessary to measure the extent of variation between and within groups of offenders, for which quantitative research methods are most appropriate (Kumar, 2014). The majority of data necessary to conduct this research was already available through the Dutch Homicide Monitor, but content analysis of additional sources was used to expand an correct the pre-existing data. The focus of this expansion was on motives and intent of offenders in of arson-related homicides and the surrounding incident characteristics. Thus, the research design is not purely quantitative, but rather one of mixed methods. Data is processed and analysed with IBM SPSS Statistics (23). The coding of additional data is based on the Coding Manual of the Dutch- and European Homicide Monitor (provided by Granath et al., 2011).

3.2. Data Collection / Triangulation of Methods

Both primary and secondary data sources were used to collect data for the research project. In first instance, the Dutch Homicide Monitor was used to get an overview of all known cases of homicide in the Netherlands that have occurred in the Netherlands in the years 2009-2016. This database is regularly updated and contains a wide variety of variables on all cases. To ensure completeness of data, additional information was retrieved from newspaper articles and judicial documents. The analysis of these newspaper articles and judicial documents was then used as a basis to update the data in the Dutch Homicide Monitor. The additional information was registered in the database according to the Coding Manual of the Dutch- and European Homicide Monitor (provided by Granath et al., 2011). Three new coding schemes were established to register the fact whether arson was used or not, the type of arson used in each individual case and the offenders’ intention of the use of arson11_.

(24)

23

3.3. Case selection

Cases of arson-related homicide in the database were found by examination of case descriptions and the registered modus operandi of all cases. First, the cases for which the use of fire or smoke were registered as modus operandi were selected and coded for the use of arson in a new variable (USE_ARSON; if an offender used arson, code = 1). Subsequently, the database was checked for any remaining cases in which fire or smoke were used. IBM SPSS Statistics (23) offers the possibility to find relevant cases by searching for words or numbers within a selected variable. This function was used to find all remaining cases (select variable ‘DESCRIPTION’ → Ctrl+F → search brand, vuur (fire) and rook (smoke)). These remaining cases were then also coded for the use of arson (USE_ARSON=1). The total number of cases that found to fit the definition of ‘arson-related homicide’ was 50, and all of these cases were selected for further analysis (SPSS → Data → Select Cases → Select Cases If USE_ARSON=1) .

3.4. Method for Retrieving Additional Data

Preliminary analysis of the cases in the database resulted in 50 individual cases involving the use of arson in one form or another. Additional information on these cases was retrieved from primary sources such as newspaper articles available in the Lexis-Nexis database, and freely accessible judicial documents, retrieved from www.moordzaken.com, www.jure.nl and www.rechtspraak.nl. Generally, articles and documents could be found by using event descriptions (brand, moord, brandmoord) in combination with the names, descent, age or gender of victims and/or offenders (21-jarige Pool; 22-jarige prostituee brand; Tessa van Bokhorst; Dennie J. moord), the area in which an event took place (i.e. Cuijck; brandmoord Rotterdam; Tjeukemeermoord) and important dates as search terms. In LexisNexis the following settings were applied: Power Search; Search Type = Natural Language; Specify Date = Date is After ‘Date of Homicide’; Select Source = All Dutch News.

Using multiple sources for the same data increases the level of reliability of data exploitation. Moreover, it should be possible to use the same data source for answering the central research question for different European countries. However, using both the Dutch Homicide Monitor and newspaper articles and judicial documents on arson-related homicides might diminish the possibility to carry out this research project in other parts of Europe, since these data sources may be less accessible in other countries.

(25)

24

3.5. Operationalization of Concepts

The operationalization of concepts and their definitions are consistent with those used in the Dutch Homicide Monitor, in order to perform the research as systematically and consistently as possible. In the table below, definitions and indicators for the key concepts of this study are presented.

Table 3.1. Operationalization of Key Concepts

Concept Definition Indicators Data Sources

Homicide “An intentional criminal act of violence by one

or more human beings resulting in the death of one or more other human beings” (European Homicide Monitor, 2011).

Cases of homicide as registered in the data source. Dutch Homicide Monitor (hereafter: DHM).

Arson Intentional firesetting to a structure or individual

Prevalence of arson in homicide cases

DHM.

Motive The underlying reason of an offender to

commit a violent crime.

Motive as declared by offender, as registered in the data source. DHM; Newspaper Articles; judicial documents.

Intention The intention of the offender in setting a fire. For instance, the offender might have had the intention to solely set a fire, or the offender might have had the intention to kill another person and used arson as a modus operandi.

Intention of fire-setting as declared by offender, as

registered in the data source DHM; Newspaper Articles; judicial documents. Arson-related homicide Primary arson Person burned Secondary arson Body burned Evidence burned

A case of homicide with one or multiple lethal victims, in which arson was used either as a

modus operandi to kill the victim(s), or in

which a fire is set after a homicide is committed. Five types are distinguished12,13_:

∙ Setting fire to a structure or vehicle in which the victim is present (ante-mortem);

∙ Setting fire to the body of the victim (ante-mortem);

∙ Setting fire to a structure or vehicle in which the victim’s body is present (post-mortem); ∙ Setting fire to the body of the victim (post-mortem);

∙ Setting fire aimed at pieces of evidence (flight vehicles, for example), leaving the victim unaffected by firesetting (post-mortem).

Cases of Homicide that involve the use of fire in any way, as registered in the data source.

DHM; News Articles.

12_{Adapted from Drake and Block (2003).}

(26)

25

3.6. Unit of analysis

The units of analysis for this research project are the offender characteristics of the analysed cases of arson-related homicides. Drake and Block (2003) have argued that it is necessary to analyse data on arson-related homicides at the incident level of individual incidents, stating that using offender-level analysis, cases will be analysed multiple times if they have multiple offenders and that this might cause skewed results. However, the focus of these authors was not specifically on the offenders, as opposed to this research project. Thus, since cases can have multiple offenders, and the goal of this research project is to compare offender characteristics, data has been analysed on the level of individual offenders rather than cases. This also implies that not only primary offenders have been included in the analyses, but that all offenders have been taken into account. The unit of observation is thus offender-level case data on arson-related homicides. This approach also in line with data registration in the Dutch Homicide Monitor: all individuals involved in a specific case (both victims and offenders) are placed on a new row in the dataset, allowing for separate registration of an individual’s characteristics.

3.7. Internal and External Validity

In designing the research, it was attempted to establish as much internal and external validity as possible. However, the validity depends on appropriateness and completeness of the measurements (Babb, 2012, p. 77). To increase internal validity, definitions and operationalization of the concepts has been inspired by other literature on firesetting, firesetters, and arson-related homicides (for instance, Dalhuisen, 2016; Drake & Block, 2003; Sapp & Huff, 1995). It is likely that the external validity and the possibility for generalization are higher for countries in which homicide- and firesetting statistics are comparable to those in the Netherlands, as opposed to countries in which these statistics differ significantly. Moreover, the use of the pre-established Dutch Homicide Monitor database, which is coded in line with the European Homicide Monitor database and has been used as a data source for prior research, increases external validity.

(27)

26

4. Results

In total, 1377 offenders are registered in the Dutch Homicide Monitor Database, involved in homicide cases in the Netherlands between January 1, 2009 and December 31, 2016. Of these offenders, 70 individuals (5%) were involved in 50 individual cases of arson-related homicides between 2009 and 2016. In 26 of these cases, the perpetrator has to date remained unknown; in 4 cases, the offender passed away. Of the 20 cases in which the offenders were known, there are 12 single-offender cases, and 8 multiple-offender cases.14

The sections below each discuss a separate part of the research findings, based on the sub-questions posed in the introductory chapter. Adjacent to the results, each section provides a short narrative on a relevant case.

4.1. What distinguishes offenders of arson-related homicides based on their gender and age?15

Table 4.1.1. Distribution of Gender and the Use of Arson

Gender of the Individual Use of Arson

No Yes Total n % of group % of total n % of group % of total (%) Male 1040 89 85 42 91 3 1082 (89) Female 132 11 11 4 9 0 136 (11) Total 1172 100 96 46 100 4 1218 (100) Missing 159

Of the complete homicide offender population of which the gender is known (n=1218), 89% (n=1082) is male and 11% (n=136) is female. The gender is known for 46 out of 70 offenders of arson-related homicides. Of these perpetrators, 91% was male, and 9% was female. This divide is approximately the same compared to the complete population of offenders, as in this group, 89% of the offenders is male, and 11% is female. Ferguson and colleagues found that in 1989-2010 in Australia, the percentage of female offenders involved in arson-homicides was relatively high: 26% (n=39) of the offenders in arson-homicides were female, whereas 13% (n=923) of the complete population of homicide offenders (including arson-homicides) were female ((Ferguson et al., 2015). Dalhuisen (2016, p. 193) found the mean age of firesetters (evaluated at the Pier Baan Centre, n=389) was 29.1 years, and that the population of firesetters consisted for 91% of male offenders and 9% female offenders.

14_{All 70 offenders were included in the analysis, since many incident-related characteristics are known and}

registered on the row of the offender in the database.

(28)

27 Table 4.1.2. Distribution of Age and the Use of Arson

Age of the Individual

Use of Arson Total

No Yes n % of group % of total n % of group % of total N (%) 17 or below 34 3 3 1 2 0 35 (3) 18-29 years 429 39 37 11 24 1 440 (38) 30-39 years 266 24 23 18 40 2 284 (25) 40-49 years 236 21 21 13 29 2 249 (22) 50-59 years 95 9 8 2 4 0 97 (8) 60 or older 44 4 4 - - - 44 (4) Total 110 4 100 96 45 100 4 1149 (100) Missing 228

Table 4.1.2 shows the distribution of age amongst the complete population of offenders in the database (n=1,149). The largest group consists of offenders between 18-29 years of age (n=440) and comprises 38% of the complete offender population. Offenders between 30-39 years (n=284) and 40-49 years (n=249) are the form the second and third largest groups, covering 25% and 22%, respectively. Similar to the distribution of gender, the distribution of age amongst the non-arson offenders is roughly equal to the distribution of the complete population: the largest group of offenders is between 18-29 years of age (39%, n=429). The second and third largest groups are those offenders aged between 30-39 years (24%, n=266) and 40-49 years (21%, n=236). The distribution of age amongst arson offenders is different: 40% of the offenders in this category are between 30-39 years of age (n=18). The age group of 40-49 years is the second largest (29%, n=13), and the third largest group consists of offenders aged 18-29 years (24%; n=11). There are no arson-homicide offenders over 60 years of age. An analysis of variance showed no significant differences between the two groups of homicide offenders regarding their age distributions (F (5,1143) = 2,145, p=.058). Drake and Block (2003) found that the likelihood of juvenile offending (defined as offending by individuals aged 16 years or younger) in arson-homicides did not differ significantly from the likelihood of juvenile offending in non-arson associated homicides in the years 1965-1995. However, the percentage of juvenile offenders in Chicago homicides was higher than it appears for Dutch homicides: in Chicago, 9% of the arson-related homicide cases involved at least one juvenile offender (Drake and Block, 2003, p. 233), while 2% of the researched Dutch population was aged under 17. It must be noted here that Drake and Block (2003) analysed data on the level of unique incidents, while the present study takes into account all offenders involved in arson-homicides, which makes it somewhat more difficult to compare

(29)

28

the results of these studies. In Australia, offenders in arson-homicides were found to be slightly younger than offenders in all recorded homicide: on average, the arson-homicide offender was 35.6 years of age, while the average homicide offender was aged 38.5 years (Ferguson et al., p. 5).

Table 4.1.3. Distribution of Age and Gender amongst Arson-Homicide Offenders

Age of the Individual

Gender of the Individual

Male Female Total

n % of group % of total n % of group % of total n (%) 17 or below 1 2 2 - - - 1 (2) 18-29 years 11 27 24 - - - 11 (24) 30-39 years 15 37 33 3 75 7 18 (40) 40-49 years 12 29 27 1 25 2 13 (29) 50-59 years 2 5 4 - - - 2 (4) Total 41 100 91 4 100 9 45 (100) Missing 25

The largest group of male arson offenders (n=41) is aged between 30-39 years (37%, n=15). The majority of the female offenders belong to the same age group: three out of four female arson offenders are aged between 30-39 years. One female offender is aged between 40-49 years.

Case 1: ‘Chalet girl’

While working to extinguish fire in a chalet, firefighters discover the body of an 18-year-old girl. Judging on the characteristics of the fire and the place where she was found, her death immediately raises questions, since it appears that she should have been able to escape the chalet before getting injured by the fire. Autopsy later revealed that prior to the fire, the girl had repeatedly been hit on the head with a brick. After this, she had likely been smothered, which caused the girl to lose consciousness – leaving her unable to escape the fire. The suspect in this case is her neighbour, a 36-year-old male. Although convincing evidence is found and the suspect is found guilty, non-disclosure leaves authorities guessing for a motive.16 The offender in this case belongs to the largest age group (30-39 years). The case provides a solid example of the difficulties in coding. Since the fire was set to the structure in which the victim was discovered, and the victim was still alive when the fire was set, this case was coded ‘primary arson’. However, crime scene investigation revealed that the intention of the offender was most likely to silence the victim, and to conceal evidence of the crime. For the variable ‘intention’, this case was eventually coded ‘crime concealment’.

(30)

29 4.2. What methods of violence causing death do offenders resort to when committing

arson-related homicide?

Table 4.2.1. Method of Violence and the Use of Arson

Method of Violence Causing Death Use of Arson No Yes Total n % of group % of total n % of group % of total n (%) Poisoning 4 0 0 - - - 4 (0) Hanging/Strangulation/ Suffocation 127 10 9 6 9 0 133 (10) Drowning 2 0 0 - - - 2 (0) Firearm 449 34 33 27 39 2 476 (35) Bomb or Explosive 2 0 0 - - - 2 (2) Smoke or Fire - - 0 10 14 1 10 (1) Knife/Sharp Object 362 28 26 2 3 0 364 (26) Blunt Object 49 4 4 7 10 0 56 (4) Axe 8 1 1 - - - 8 (1) Push or Shove 9 1 1 - - - 9 (1) Motor Vehicle 31 2 2 - - - 31 (2)

Hitting, kicking or other similar type of physical violence without weapon

120 9 9 4 6 0 124 (9) Other 10 1 1 - - - 10 (1) Unknown 134 10 10 14 20 1 148 (11) Total 1307 100 95 70 100 5 1377 (100) Missing 0

Overall, over one third of all offenders were involved in cases in which the victim was killed by a firearm (35%, n=476). This distribution was roughly similar between the offenders of cases involving arson (39%, n=27), and those not involving arson (34%, n=449). The use of a knife or sharp object is more common for offenders in the non-arson group (28%, n=362) than for offenders using arson (3%, n=2)17. For 10 of the 70 offenders involved in arson-related homicides, smoke or fire were the primary modus operandi (14% of the group). For 20% of the offenders involved in arson-related homicides (n=14), the primary cause of death of their victims was unknown or was not published, which is likely related to the destruction of evidence.

Most research projects analyzed the cause of death on the level of individual victims, rather than using the level of the individual offenders. Involvement of multiple victims or multiple offenders might cause some skewness when comparing the results of studies using different

17_{A chi-square test was completed for exploratory purposes, but showed no valid results due to low expected}