Characteristics associated with repeat homicide offenders relative to single homicide offenders. An explorative study on a 35-year sample of homicide offenders in the Netherlands

(1)

Characteristics associated with repeat

homicide offenders relative to single

homicide offenders

AN EXPLORATIVE STUDY ON A

35-

YEAR SAMPLE OF HOMICIDE

(2)

Characteristics associated with repeat homicide offenders relative to single homicide offenders.

An explorative study on a 35-year sample of homicide offenders in The Netherlands.

Author L.A. van Erk

Student number s1762494

Supervisor Dr. M.C.A. Liem

Second reader Dr. P.G.M. Aarten

Institute University of Leiden, the Netherlands

Faculty of Governance and Global Affairs

Study Master Crisis and Security Management

Course Master thesis Crisis and Security Management 2016-2017

Submission date July 17th_{, 2017}

(3)

Abstract

This explorative study presents the characteristics of a sample of repeat homicide offenders (RHO’s) in the Netherlands between 1980 and 2016. The research involves an analysis of forensic mental health reports, the Dutch homicide monitor, Dutch jurisdiction and newspaper articles for 28 homicide offenders who committed homicide more than once and a sample of homicide offenders who committed homicide in ‘only’ one event (n=640). Descriptive analyses are used to examine and explore characteristics of RHO’s relative to single homicide offenders (SHO’s). This analysis reveals at least partial evidence that the sampled RHO’s and SHO’s differ from each other. Herein, RHO’s seem to be specifically characterised by: unstable lifestyles (being unemployed and single, living alone or temporarily together with someone); a younger age at the time of the homicide offences; instrumental motives when committing the homicides while being limited influenced by alcohol or drugs (or both); victimizing individuals they do not know; having a criminal history; and having psychological problems as a result of a problematic youth. Additionally, the majority of the sampled RHO’s are commonly presented with a below averaged intelligence, limited relationships with family and friends, and associations with criminals. Nevertheless, RHO’s and SHO’s seem to be similarly overrepresented by the male gender and are similarly born in the Netherlands or another country than the Netherlands.

(4)

Preface

This thesis serves as the final examination for the master Crisis and Security Management of the University of Leiden. The study of this thesis aims to describe and explore the little understood phenomenon of repeat homicide offending in the Netherlands between 1980 and 2016. This study was conducted from February until July 2017.

The research question of this study was defined in cooperation with my thesis supervisor, dr. M.C.A. Liem. The process of answering this question was sometimes complex but challenging and always interesting. In fact, the combination of extended (qualitative) literature research and quantitative data analysis eventually ensured a better understanding about RHO’s in the Netherlands.

This work would not have been possible without the provision of the NIFP to collect data from their database. The data of the NIFP provided this work with the necessary information to study the research group and consequently added great value to the quality and depth of the results of this study. Additionally, collecting data at the NIFP also added to my personal learning experience. Therefore, I am especially indebted to MSc Tien and dr. Kempes from the NIFP who have been supportive of my research and who worked actively to provide me with the necessary information to answer the research question of this study. However, the opportunity to collect data at the NIFP would not have been possible without the commitment of my thesis supervisor. I would like to thank her for her effort and guidance.

Liza van Erk July 17th_{, 2017}

(5)

List of figures and tables

Figure 1: Sample selection process of RHO’s ...20

Figure 2: Gender of RHO’s and SHO’s, 1980-2016 (by perpetrator) ...25

Figure 3: Age division of RHO’s and SHO’s, 1980-2016 (by perpetrator per incident) ...26

Figure 4: Birth country of RHO’s and SHO’s, 1980-2016 (by perpetrator) ...27

Figure 5: Professional situation of RHO’s and SHO’s, 1980-2016 (by perpetrator per homicide event) ...28

Figure 6: Housing situation of RHO’s and SHO’s, 1980-2016 (by perpetrator per homicide event) ...28

Figure 7: Civil situation of RHO’s and SHO’s, 1980-2016 (by perpetrator per homicide event) ...29

Figure 8: Patterns of alcohol abuse among RHO’s and SHO’s, 1980-2016 (by perpetrator per homicide event) ..29

Figure 9: Patterns of substance abuse among RHO’s and SHO’s, 1980-2016 (by perpetrator per homicide event) ...30

Figure 10: Psychological mind of RHO’s and SHO’s, 1980-2016 (by perpetrator per incident)...31

Figure 11: Previous convictions of RHO’s and SHO’s, 1980-2016 (by perpetrator) ...32

Figure 12: Type of homicides committed by RHO’s and SHO’s, 1980-2016 (by victim) ...33

Figure 13: Motives of RHO’s and SHO’s, 1980-2016 (by victim) ...33

Figure 14: Relationship between the victims and RHO’s and SHO’s, 1980-2016 (by victim) ...34

Figure 15: Crime scene of homicides committed by RHO’s and SHO’s, 1980-2016 (by victim)...34

Figure 16: Level of planning and intention of RHO’s and SHO’s, 1980-2016 (by perpetrator per homicide event) ...35

Figure 17: Type of sentences imposed to RHO’s and SHO’s, 1980-2016 (by perpetrator per homicide event) ...35

Table 1: Average age of RHO’s and SHO’s, 1980-2016 (by perpetrator per homicide event)...26

Table 2: Age difference between the first and second offence, differentiating between convicted and not convicted RHO’s, 1980-2016 (by perpetrator) ...27

Table 3: Intelligence of RHO’s, 1980-2016 (by perpetrator) ...27

Table 4: Youth history of RHO’s, 1980-2016 (by perpetrator) ...30

Table 5: Current relationship of RHO’s with family, 1980-2016 (by perpetrator) ...31

Table 6: Current relationship of RHO’s with friends, 1980-2016 (by perpetrator) ...31

Table 7: Association with criminals among RHO’s, 1980-2016 (by perpetrator per homicide event) ...32

(8)

List of abbreviations

DHM Dutch Homicide Monitor EHM European Homicide Manual

NIFP Nederlands Instituut voor Forensische Psychiatrie en Psychologie RHO Repeat Homicide Offender

RHO’s Repeat Homicide Offenders

SD Standard Deviation

SHO’s Single Homicide Offenders

(9)

(10)

Chapter 1: Introduction to the research

On the 10th_{of February 2014 former Dutch minister of Healthcare, Els Borst, was murdered at the age}

of 81 by Bart van U. who said to have received a divine command to kill Borst. In December 2015 van U. committed a homicide again, this time he killed his sister. On the 4th_{of February 2016 van U. was}

arrested for the homicide on his sister and only then was recognized as a suspect for the homicide on Els Borst. Media attention regarding this case was big and incomprehension dominated among citizens, especially when public authorities admitted to have failed to assess the danger people as van U. can pose to society (Nu.nl, February 2016).

Homicide offenders as Bart van U. generally receive a disproportionate amount of attention with important implications for public perceptions of crime, fear of crime, and public policy. In fact, such reactions commonly focus on the perceived failure of the criminal justice system in preventing homicide offenders to commit a subsequent homicide (Cale, Plecas, Cohen and Fortier, 2010). However, despite of this public- and policy relevance of preventing repeat homicides, there is relatively little information on these so called Repeat Homicide Offenders (RHO’s) (Cale, Plecas, Cohen and Fortier, 2010). Additionally, one critical research question that has not yet been explored extensively is whether there are any differences between Single Homicide Offenders (SHO’s) and RHO’s (Cale, Plecas, Cohen and Fortier, 2010). In part, this lack of research reflects the infrequency of repeat homicide, especially compared to other types of violent crime. Nevertheless, because of its relevance to the public, this research empirically explores the characteristics of RHO’s and the extent to which they share characteristics with those who committed homicide ‘only’ once. Specifically, this study aims to answer the following question: “To what extent do repeat homicide offenders in the Netherlands

between 1980 and 2016 share characteristics and to what extent can these characteristics be distinguished from single homicide offenders?” In order to answer this research question, the

following sub questions are formulated:

1. How is repeat homicide offending defined? 2. How is single homicide offending defined?

3. What theoretical characteristics are related to repeat- and single homicide offenders? 4. What was the nature of repeat- and single homicide offenders in the Netherlands between

1980 and 2016?

The purpose of the research question is two-fold; it is descriptive and explorative. First, this question is descriptive as it aims to provide an overview of the characteristics of RHO’s in the Netherlands. Second, this question aims to explore the nature of the phenomenon of RHO’s by comparing their characteristics to the characteristics of SHO’s. Herein, this research does not intend to provide conclusive evidence on characteristics of RHO’s. Nevertheless, it helps to understand the phenomenon of RHO’s better by developing preliminary ideas about it and moving toward refined research questions for more conclusive research (Neuman, 2014).

Answering the research question brings clarification on RHO’s to both the academic world and practice. With regards to academic knowledge, answering the research question fills the gap of knowledge on repeat homicide offending. Herein, this research would be a step towards engaging the academic community into a discussion on whether RHO’s should be viewed as a discrete offender group when compared to SHO’s. In practice, a clarification on RHO’s makes it easier to identify RHO’s earlier. Consequently, this research could be the beginning of a process towards answering criminal

(11)

justice policy issues that are related to repeat homicide offending. These policy issues mainly refer to prevention, treatments and sentencing (Cale, Plecas, Cohen and Fortier, 2010).

The purpose and added value of this research is studied for the Netherlands. The Netherlands is a parliamentary democratic constitutional monarchy, located in North-West Europe. The country has approximately 16,5 million inhabitants, consisting mostly of ethnic Dutch groups (Liem, 2013). Predominant ethnic groups include Turkish (2.41_{%), Surinamese (2.06%), Moroccan (2.25%), and Dutch}

Antillean (0.88%) groups (CBS, 2015). This study focuses on one country since data sources across nations are hardly comparable because of the different definitions of homicide used in each country (Smit, de Jong and Bijleveld, 2012). Within this line of reasoning, the Netherlands is chosen as the geographical scope of this study because of language reasons and possibilities to access homicide related data sources. In fact, since information on homicide offenders is generally very sensitive, organisations that possess such information are careful in sharing it with external parties. Therefore, having admission to access information on homicide offenders is a big consideration in selecting the geographical scope of this study. Herein, as repeat homicide is infrequent (Cale, Plecas, Cohen and Fortier, 2010), this research studies a timeframe of 35 years in order to be able to select a sufficient sample of RHO’s.

In order to answer the research question, this study applies both a qualitative and quantitative approach. As common in descriptive and explorative research, this study first involves a qualitative data collection by means of a literature study. Based on the insights of this literature study, quantitative data is then collected and analysed in order to be able to answer the research question (Creswell, 2009). Herein, the literature study of this research is presented in chapter two. This framework elaborates on the definitions related to repeat homicide offending and addresses several theories that are related to repeat offending and homicide recidivism in order to define several characteristics that can explain repeat homicide offending. Having reviewed the literature, chapter three elaborates on the methodology of this research. This chapter builds on the literature by operationalizing the identified characteristics and defining the methods that are applied to study these characteristics for the identified sample. Thereafter, chapter four presents the outcome of the methodology by describing the results of the data collection. Finally, chapter five answers the research question of this study by linking the results to the theory.

(12)

Chapter 2: Theoretical background

This chapter elaborates on the literature related to repeat homicide offending. Herein, this chapter first discusses the definitions related to repeat homicide offending. Thereafter, several theoretical explanations on repeat offending and homicide recidivism are discussed in order to be able to define characteristics that can be used to study RHO’s.

2.1 Defining repeat homicide offending

Previous studies have paid little attention to the phenomenon of repeat homicide offending in the broadest sense of the term. In fact, Cale, Plecas, Cohen and Fortier (2010) argue that there is no universally accepted definition of repeat homicide offending. As a consequence, previous research on repeat homicide offending have studied differing phenomena. Herein, the big majority of research focuses on homicide recidivism and multiple homicide offending (Björkly and Waage, 2005; Cale, Plecas, Cohen and Fortier, 2010). Homicide recidivism is defined as a form of homicide in which the offender commits a murder, is subsequently institutionalized in a prison or psychiatric facility, and commits a homicide again after release into the community (Björkly and Waage, 2005). With regards to homicide recidivism Cale, Plecas, Cohen and Fortier (2010), for example, studied the characteristics of ‘repeat homicide offenders’ in Canada between 1975 and 2005 and matched them to a sample of homicide offenders who only committed one homicide. Additionally, a key study conducted by Eronen, Hakola, and Tiihonen (1996) studied homicide recidivism in Finland over a 13-year sample in which 36 recidivist were identified from a sample of 1,584 homicide offenders. Multiple homicide, on the other hand, is defined as homicide in which more than one victim is killed (Björkly and Waage, 2005). The literature on multiple homicide generally defines several subgroups, including the killing of more than one victim on a single occasion, sometimes referred to as mass murder; the killing of more than one victim at a single event over an extended period of time, sometimes referred to as spree murder; and the killing of a single victim on three or more occasions with a cooling-off period in between, sometimes referred to as serial murder (Björkly and Waage, 2005; Douglas, Ressler, Burgess and Hartman, 1986). The difference between spree murder and mass murder is the location of the murders: spree murders moves beyond different locations where mass murder takes place at one location (Douglas, Ressler, Burgess and Hartman, 1986). Herein, Delisi and Scherer (2006), for example, studied the linkages between multiple homicide offenders (while referring to repeat homicide offenders) and single homicide offenders in the United States in 2003. Although both the definitions of homicide recidivism and multiple homicide refer to repeat homicide offending as the overarching research subject, they differ significantly from each other. Therefore, in order to come to a definition that covers the full meaning of repeat homicide offending, the following paragraphs will elaborate on the definitions of ‘homicide’ and ‘repeat offending’.

2.1.1 Defining ‘homicide’

Homicide is the most serious form of violent crime (Smit, de Jong and Bijleveld, 2012). It is uniquely harmful and affects what is most important to almost everyone – life (Brookman, 2005). However, there is no universally accepted definition of homicide (Smit, de Jong and Bijleveld, 2012). Björkly and Waage (2005) for example refer to homicide as an act in which one person kills another following purposely inflicted injury. Somewhat different, Smit, de Jong and Bijleveld advocate that homicide occurs when there is a dead person and the cause of death can be attributed to another person (Smit, de Jong and Bijleveld, 2012: 5). On the national level, homicide is often defined by a selection of articles from the national criminal code. In the United States, for example, homicide is the illegal wilful

(13)

deprivation of one’s life. Whoever kills a person is guilty of murder in Germany (Smit, de Jong and Bijleveld, 2012). Regardless of some small differences between these definitions of homicide, the term homicide generally seems to cover situations where a person is killed by another individual (Smit, de Jong and Bijleveld, 2012).

On the national level, definitions on homicide also seem to agree with regards to the legal aspects of homicide, namely premeditation and intent (Smit, de Jong and Bijleveld, 2012). ‘Premeditation’ indicates that some degree of planning is present within the homicide where ‘intent’ refers to the intention of the offender to take the life of the victim (Smit, de Jong and Bijleveld, 2012). Premeditation and planning is often judged by assessing the time and the opportunity the offender had to consider what he was doing (de Hullu, 2003). This assessment is often executed by judges which are generally allowed to take into account all factors that may have influenced the degree of planning and intention, such as: the severity of the offence, the circumstances in which the offence is committed and the person of the accused (Koenraadt, Mooij and Mulbergt, 2007). Consequently, judges often have wide discretionary powers when meting out sentences; they can highly influence both the type and severity of criminal punishment (Smit, de Jong and Bijleveld, 2012). This is especially evident in the Dutch criminal law which does not define any mandatory or minimum sentences but only maximum sentences (Koenraadt, Mooij and Mulbergt, 2007). The Dutch criminal law defines sentences based on the level of premeditation and intent. Herein, the Dutch criminal law system distinguishes murder (article 289 of the Dutch criminal code) from manslaughter (article 287 of the Dutch criminal code). Here, murder is defined as someone who intentionally and with premeditation takes the life of someone else. Manslaughter, on the other hand, refers to intentionally taking the life of someone else without premeditation. Additionally, the Dutch criminal law also distinguishes infanticide which is referred to as the murder of an (unborn) child under the age of zero (Koenraadt, Mooij and Mulbergt, 2007). Depending on the level of premeditation and the amount of violations and proof, both perpetrators of murder and manslaughter can be sentenced for life in the Netherlands. However, when there are no other violations or when there is not enough proof or level of premeditation, sentences for manslaughter and infanticide are limited to fifteen years. For murder, sentences are limited to thirty years (article 287 and 289 of the Dutch criminal code). Besides prison sentences, Dutch judges also have the option to impose treatment-based sentencing to offenders that are not, or partly, held accountable for their actions because of their state of mind at the time of the offence. These treatment based sentences can also be imposed in combination with a prison sentence (Ganpat and Liem, 2012). However, defining homicide in practice is not simple. Different levels of motivation, involvement and responsibility of the perpetrator(s) complicates assessing whether the death of a person is caused by premeditation and/or intention or not (Brookman, 2005; Smit, de Jong and Bijleveld, 2012). Non-intentional killings, for example, can extremely complicate assessing the level of premeditation and intention. Think about cases as: abortion, euthanasia, assisted suicide, infanticide, assault leading to death, dangerous driving and justified killings (e.g. killings by police officers) (Smit, de Jong and Bijleveld 2012). When presented with such cases, it is commonly sufficient that the offender is proven to be culpable of the victim’s death in order to hold him or her responsible (Smit, de Jong and Bijleveld, 2012).

Nevertheless, if homicide is identified, it still remains difficult to state an exact number on the scope of this problem. In fact, homicide rates are commonly seen as a ‘dark figure’ as they only represent the tip of the iceberg (Brookman, 2005). There will always be an unknown and uncertain mass of hidden

(14)

homicides as bodies are never found, persons are not registered missing and deaths are not identified as homicide while they were (Brookman, 2005).

In conclusion, homicide in this research is defined as the premeditated and intended action to take the life of someone else. Herein, premeditation indicates that some degree of planning is present. ‘Intent’ refers to the intention of the offender to take the life of the victim (Smit, de Jong and Bijleveld, 2012). However, one should be aware of the practical difficulties of this definition when applying it.

2.1.2 Defining ‘repeat offending’

Literature on repeat offending is often associated with recidivism. Another synonym that is commonly used for repeat offending is re-offending. All words mean the same: the engagement of someone in repeated crime (Fargan and Mazerolle, 2011). Herein, Eck (2003) defined a repeat offender as someone who commits crimes more than once or several times. With regards to this definition of Eck (2003), repeat offenders tent to compose a relatively small subset of the offender population but contribute disproportionately to offending behaviour – especially, serious offending (Fargan and Mazerolle, 2011). Therefore, most legal system punish repeat offenders more severely for the same offense than non-repeat offenders (Emons, 2007). The principle of sanctions based on offense history is also called penalty escalation and is so widely accepted that it is embedded in many penal codes and sentencing guidelines, as in the Netherlands (Koenraadt, Mooij and Mulbergt, 2007).

2.1.3 Coming to a definition of ‘repeat homicide offending’

Combining the definitions of ‘homicide’ and ‘repeat offending’ this study defines repeat homicide offending in the broadest sense of the term as: a premeditated and/or intended action to take the life of someone else (Smit, de Jong and Bijleveld, 2012) more than once or several times (Eck, 2003). Applying this definition to the phenomena that are commonly referred to as repeat homicide offending in the literature, namely recidivism and multiple homicide, several parts of these phenomena do not correspond with the definition of repeat homicide offending applied in this study. First, with regards to homicide recidivism, RHO’s as defined in this study do not necessarily have to be institutionalized between the homicide offenses. Second, when referring to multiple homicide, mass murder and spree murder do not fit the definition of repeat homicide offending of this research. In fact, both types are characterized by one single homicide occasion where repeat homicide offending is defined by more than one homicide event (Eck, 2003). Serial homicide does fit the definition of repeat homicide offending of this research. However, repeat homicide offending in this study does not have to involve a minimum of three homicides with a cooling off period in between. In summary, homicide recidivism and serial homicide do fit the definition of repeat homicide offending as applied in this study, however do not completely cover this definition. To illustrate: a repeat homicide offender can commit two homicides in two occasions while not being institutionalised between the offences. Consequently, such a homicide offender is not defined as a recidivistic homicide offender as he or she is not institutionalised between the homicides. Additionally, he or she is also not defined as a serial homicide offender as he or she did not committed homicide on single victims in three or more occasions (Björkly and Waage, 2005).

In contradiction to repeat homicide offending, single homicide offending is defined as the premeditated and/or intended action to take the life of someone during one occasion (Douglas, Burgess, Burgess and Ressler, 2006). Here, one occasion refers to a single time and a single location. A single homicide offender is thus an individual who commits homicide in a single occasion (Cale, Plecas, Cohen and Fortier, 2010). Consequently, the main difference between repeat- and single homicide

(15)

refers to the amount of homicide events. Herein, Douglas, Ressler, Burgess and Hartman (1986) advocate that having committed one homicide does not preclude that the individual will not repeat the crime in the future. In fact, they see single-time offenders as caught repeat offenders for which the chance to repeat homicide is deprived (Douglas, Ressler, Burgess and Hartman, 1986). Consequently, one may ask to what extent single- and repeat homicide offenders differ from each other, substantiating the research question of this study.

2.2 Theoretical explanations for repeat homicide offending

Research on general delinquents has found several mechanisms that can explain why individuals repeat their actions (Liem, 2013). These mechanisms are commonly divided into three theoretical categories, namely static theories, dynamic theories (social learning theories, life course theories, developmental theories) and typological theories (Liem, 2013; Blokland and Nieuwbeerta, 2005). In fact, these theories have repeatedly been used to explain homicide recidivism (Liem, 2013). Although homicide recidivism does not fully cover the phenomenon of repeat homicide offending as defined in this study, it is expected that the posed theories are able to provide this research with characteristics that can be used to study RHO’s.This paragraph therefore discusses these theories in order to define features that may characterise RHO’s. Herein, this paragraph also elaborates on the extent to which RHO’s and SHO’s are similar to each other from a theoretical point of view.

2.2.1 Static theories

Static theories ascribe repeat offending behaviour to some latent characteristics that develop before and during early childhood (Liem, 2013). These theories argue that the male gender, an ethnic background, a below averaged intelligence, and a problematic youth are characteristics that commonly predict repeat criminal behaviour at an early age (Bloland and Nieuwbeerta, 2005). Herein, studies have commonly found an overrepresentation of the male gender among homicide recidivists (Vries and Liem, 2011; Liem, 2013; Cale, Plecas, Cohen and Fortier, 2010). Vries and Liem (2011), for example, found an overrepresentation of the male gender within homicide recidivism among juveniles in the Netherlands between 1992 and 2007. Additionally, Cale, Plecas, Cohen and Fortier (2010) found that 83 of the 86 homicide recidivists in Canada between 1975 and 2005 were men. Nevertheless, they also found an overrepresentation of the male gender among SHO’s (Cale, Plecas, Cohen and Fortier, 2010). In fact, homicide offenders generally seem to be overrepresented by the male gender (Brookman, 2005). To illustrate, approximately 90-91% of all homicide perpetrators in the Netherlands between 1992 and 2009 are male (Ganpat and Liem, 2012; Nieuwbeerta and Leistra, 2003). Consequently, one may argue that the male gender is not a specific predictor of repeat homicide offending and therefore does not differ from SHO’s.

Further, in terms of ethnicity, Cale, Plecas, Cohen and Fortier (2010) found that the majority (59%) of the homicide recidivists were White. However, they also found that the majority of the SHO’s were also White (Cale, Plecas, Cohen and Fortier, 2010). Nevertheless, Roberts, Zgoba and Shahidullah (2007), who studied homicide recidivism in the United States between 1990 and 2000, found that the majority (62%) of homicide recidivists were Black. Consequently, one may argue that ethnicity division among homicide recidivists differs per country. Therefore, since the majority (58%) of the homicide offender population in the Netherlands between 1992 and 2009 had another ethnic background (Ganpat and Liem, 2012; Nieuwbeerta and Leistra, 2003), it is expected that both the majority of RHO’s originate from other countries than the Netherlands. Herein, it is expected that RHO’s do not differ from SHO’s (Cale, Plecas, Cohen and Fortier, 2010).

(16)

Additionally, static theories also argue that a below averaged intelligence increases the risk of repeat criminal behaviour (Stattin and Klackenberg-Larsson, 1993). Herein, low intelligence is said to be more a characteristic of recidivists than of one time offenders (Stattin and Klackenberg-Larsson, 1993). Several studies have aimed to explain this intelligence-delinquency relationship. Historically, individuals with a low intelligence are said to be unable to distinguish right from wrong and therefore are more likely to repeatedly commit crime (Goddard, 1914). Another explanation of the intelligence-delinquency relationship is that a low intelligence causes social frustration which motivates individuals to engage in antisocial more than individuals with a higher intelligence (Hirschi and Hindelang, 1977). Further, Lynam, Moffit and Stouthamer-Loeber (1993) explain that low intelligence children are more impulsive and less adept at self-control and therefore have a higher likelihood to repeatedly become involved in criminological actions. Herein, a lack of self-control refers to a lack of inhibitory control to regulate emotions, thoughts and behaviour in the face of temptations and impulses that can result in offending behaviour (Baumeister, Vohs and Tiee, 2007; Gottfredson and Hirschi, 1990). Consequently, it is expected that RHO’s have a lower averaged background than SHO’s.

Finally, static theories argue that experiences of child abuse and neglect are commonly identified with serious mental disorders that, in turn, are linked to severe and repeated offences (Eronen, Hakola and Tiihonen, 1996). Similarly, Cale, Plecas, Cohen and Fortier (2010) concluded that child abuse and neglect were among the strongest predictors of committing a subsequent homicide. In fact, RHO’s are said to experience significantly more child abuse and neglect than SHO’s which (Putkonen, Ryynänen, Eronen and Tiihonen, 2002). Affirmative, Eronen, Hakola and Tiihonen (1996) found that mental disorders in relation to repeat homicide is best explained by childhood abuse. Herein, they found that nearly two thirds of homicide recidivists was clinically diagnosed with a personality disorder (Eronen, Hakola and Tiihonen, 1996). Consequently, it is expected that RHO’s more often experienced a problematic youth and therefore suffer from mental disorders more often than SHO’s.

In summary, the static perspective expects that SHO’s and RHO’s are similar to each other with respect to an overrepresentation of the male gender and a more common origin from other countries than the Netherlands. Nevertheless, static theories also expect that RHO’s differ from SHO’s. Herein, RHO’s are expected to have a lower average intelligence and have suffered a problematic family background and therefore mental problems more often.

2.2.2 Dynamic theories

Dynamic theories differ from static theories in the length of the time in which they render a direct causal influence on crime (Liem, 2013). In fact, where static theories mainly focus on childhood, dynamic theories allow changes in life circumstance during and after adolescence to influence the criminal behaviour in that period (Blokland and Nieuwbeerta, 2005). Dynamic theories are commonly explained in life course theories, social learning theories, and developmental theories (Liem, 2013; Blokland and Nieuwbeerta, 2005).

2.2.2.1 Life course theories

According to life course theories, several types of resources related to life course reduce the likelihood of recidivism (Liem, 2013; Sampson and Laub, 2005). Herein, Putkonen, Ryynänen, Eronen and Tiihonen (2002) argue that stable personal relationships, community and family support, positive employment prospects, stable housing situation and the absence of alcohol and drug abuse decrease the likelihood of homicide recidivism (Blokland and Nieuwbeerta, 2005). In fact, the stronger the ties to family and work, the less criminal behaviour (Sampson and Laub, 2005). When, however, areas of

(17)

social control (through employment or marital and parental attachment) are reduced, offenders are said to have smaller incentives to abstain from reoffending (Cale, Plecas, Cohen and Fortier, 2010). Herein, Cale, Plecas, Cohen and Fortier (2010) provided evidence that unstable lifestyles are important in explaining some of the differences between SHO’s and RHO’s. However, Cale, Plecas, Cohen and Fortier (2010), did not found confirming evidence that substance abuse increases the likelihood of homicide recidivism. In fact, they found the opposite in that RHO’s are less likely than SHO’s to exhibit patterns of substance abuse (Cale, Plecas, Cohen and Fortier, 2010).

Although life course theories do not (yet) seem to agree on the impact of all key lifestyles on the likelihood of homicide recidivism, this study expects that key lifestyle indicators are important in explaining differences between RHO’s and SHO’s. Herein, it is expected that SHO’ are more likely to represent stable lifestyles than RHO’s since stable lifestyle factors are said to decrease the likelihood of homicide recidivism (Blokland and Nieuwbeerta, 2005). In fact, it is expected RHO’s are more likely to be presented with the following lifestyle characteristics: job instability, unstable marital attachment, unstable contact with friends and relatives, unstable housing situation and more severe patterns of substance abuse.

2.2.2.2 Social learning theories

According to social learning theories, offending behaviour is learned through observation and imitation of the behaviour of other (criminal) people (Espelage, Boswordt and Simon, 2000). These theories argue that differential association with criminals make individuals more likely to develop norms that favour crime, and, therefore make them more likely to engage in recidivism (Liem, 2013: 20). Herein, criminal behaviour can be learned inside- and outside prison. Outside prison, individuals can, for example, become involved in gangs or the criminal circuit. Within these environments, criminal behaviour is often praised or even rewarded encouraging individuals to commit a subsequent crime (Espelage, Boswordt and Simon, 2000). With regards to association inside prison, Baaij, Liem and Nieuwbeerta (2012) who studied recidivism in The Netherlands between 1996 and 2004, found that contact with negative social bonds in prison leads to acquiring skills that facility future crime. Social learning theories thus expect that RHO’s are more commonly associated with people who favour criminal behaviour than SHO’s, whether inside- or outside prison.

2.2.2.3 Developmental theories

Rather than defining characteristics that increase the likelihood of recidivism, developmental theories point to reasons for abstaining from criminal behaviour (Liem, 2013). These theories argue that delinquent behaviour reduces as offenders age: offenders who are older at the time of their first homicide, are less likely to reoffend (Gove, 1985). In this line of reasoning, Cale, Plecas, Cohen and Fortier (2010) found that RHO’s exhibit a significant earlier age at the time of the homicide event compared to SHO’s. This declining chance of reoffending can be explained by biological and psychological factors. Herein, Gove (1985) explains that the older someone becomes, the less physical- strength and energy and psychological drive he or she has to commit a homicide. In fact, this is also evident for the general population of homicide offenders in the Netherlands between 1992 and 2009 in which roughly 1% of the homicide offenders was past retirement age when committing homicide (Ganpat and Liem, 2012; Nieuwbeerta and Leistra, 2003). Developmental theories thus expect that SHO’s are generally older at the time of the homicide event than RHO’s (at the time of their first homicide event).

(18)

2.2.3 Typological theories

Finally, typological theories assume that delinquents who engaged in antisocial behaviour at multiple stages of life are relatively immune to changes in life circumstances and therefore are more likely to repeatedly commit crime (Liem, 2013). This theory is supported by Wolfgang (1958) who noted that homicide offenders in general are unlikely to be a first time offender when the homicide is committed. Herein, Putkonen, Ryynänen, Eronen and Tiihonen (2002) identified a strong link between the likelihood of repeat violent offending and individuals leading a criminal lifestyle. Confirmatively, Cale, Plecas, Cohen and Fortier (2010) found that RHO’s were more likely to be convicted prior to the first homicide than SHO’s. Consequently, it is expected that RHO’s are more likely than SHO’s to have engaged in criminal activities prior to the homicide.

Rather than differentiating homicide offenders according to whether they have been involved in criminal behaviour before, several studies also differentiate homicide offender groups by the type of homicide they committed (Liem, 2013). Herein, types of homicides are commonly related to different types of motives (Smit, Bijleveld and van der Zee, 2001). These types generally consider: criminality, robbery, family sphere, (non-familial) arguments, revenge, jealousy, separation, triviality, hate crime, threat to offender, mental illness, altruism, and rape or sex (Smit, Bijleveld and van der Zee, 2001; Granath et al., 2011). These motives are commonly divided into two main categories, namely: instrumental and reactive-expressive (Wolfgang and Strohm, 1956; Thijssen and de Ruiter, 2011). Herein, instrumental homicide occurs when the offender aims to achieve a goal (e.g. money, personal belongings, sex, territory), and commits homicide as a means to this end (Thijssen and de Ruiter, 2011). Reactive-expressive homicides, on the other hand, occur in the context of an emotional response to frustration or ego threats, such as insult or personal failure. Within these homicides, influence of alcohol and drugs are often important factors increasing the risk of homicide (Wolfgang and Strohm, 1956). In other words, in certain contexts, such as bars and clubs, aggression can escalate to homicide in part because of the effects of alcohol or drugs (or both). Similarly, domestic assaults can more easily escalate when both parties involved are under the influence of alcohol or drugs (Bennett, 1995). With regards to these two motive categories, approximately 64% of the homicides committed in the Netherlands between 1992 and 2009 are committed with an reactive-expressive motive. Herein, most victims are killed during an argument (23-25%), closely followed by individuals that are killed by their partner (19-23%) (Ganpat and Liem, 2012; Nieuwbeerta and Leistra, 2003). With regards to the instrumental perspective, criminal homicides accounted for 13-14%, robbery homicides for 8-9%, and sexual homicide for 3-4% (Ganpat and Liem, 2012; Nieuwbeerta and Leistra, 2003).

Considering these two main motives when characterising RHO’s and SHO’s, research seems to agree that repeat homicides are generally cold-blooded, purposeful, and instrumental (Thijssen and de Ruiter, 2011). Single homicides, on the other hand, are more commonly situational or reactive. Herein, single homicides are often emotionally charged situations involving family members, intimates, or acquaintances rather than strangers (Kraemer, Lord and Heilbrun, 2004). Additionally, SHO’s are also expected to commit more homicides in bars and domestic environments since situational homicides are commonly related to the influence of alcohol or drugs (or both) (Bennett, 1995). Nevertheless, Cale, Plecas, Cohen and Fortier (2010) did not find distinguishing features between SHO’s and RHO’s with regards to the crime scene.

(19)

In summary, typological theories expect that RHO’s are more commonly convicted in different stages of life prior to the homicide event. Additionally, typological perspectives assume that RHO’s more often commit homicides with an instrumental perspective where SHO’s are expected to commit more homicides with an expressive-reactive intent. Subsequently, it is expected that RHO’s are less often under the influence of alcohol or drugs and less often commit homicide in a bar or at home. Finally, it is assumed that SHO’s are more often convicted for manslaughter as they react to situations more and therefore less often premeditate to commit homicide.

In conclusion of these theories explaining repeat offending and homicide recidivism, the following hypotheses can be answered:

 Null hypothesis (H0): RHO’s have different characteristics than SHO’s.

 Alternative hypothesis (H1): RHO’S share similar characteristics as SHO’s

From a theoretical point of view, it is expected that RHO’s and SHO’s mainly differ from each other. In fact, the theories expect that SHO’s only compare to RHO’s with regards to the overrepresentation of the male gender and the majority coming from another country than the Netherlands. Consequently, it is expected that RHO’s: are younger at the time of the first offence; have a lower intelligence; have more severe problematic family backgrounds and as a result more severe mental disorders; more often have unstable marital attachment; more often have unstable relationships with family and friends; more often live in unstable housing situations; are more often unemployed; are more severe substance abusers; are more commonly associated with criminals; are more commonly convicted prior to the homicide event; and more often commit homicides from an instrumental perspective.

(20)

Chapter 3: Methodology

This chapter elaborates on the methods applied to answer the research question of this study. Herein, this chapter first sets the basis of the methodology by indicating the type of research that is applied in this study. Based on this direction, this chapter then operationalizes the theoretical explanations of repeat offending and homicide recidivism into variables that can be empirically studied to describe repeat homicide offending. Thereafter, the population and sample of this research are defined. Further, this chapter elaborates on the sources used- and the methods applied in order to collect the necessary data. Finally, this chapter discusses how the collected data is assessed and analysed.

3.1 Type of research

This research aims to describe and explore the phenomenon of RHO’s in the Netherlands between 1980 and 2016. Herein, RHO’s are described by stating general facts regarding the perpetrator and the homicide events. Additionally, these descriptive outcomes are explored by comparing them to general facts on the characteristics of SHO’s. This research does not provide conclusive evidence on characteristics of the total population of RHO’s and the causal relation between these characteristics. Nevertheless, this study aims to set the basis for future research on RHO’s by describing the general characteristics of the sample of this study. Within this explorative descriptive approach, both RHO’s and SHO’s are mainly studied by applying a quantitative approach. A quantitative approach is applied in order to being able to systematically collect and analyse data (Neuman, 2014).

3.2 Operationalization of the variables

Although specific theoretical explanations on RHO’s in the broad sense of the term seemed to be lacking, theoretical explanations on repeat offending and homicide recidivism, as defined in the previous chapter, are used to define the variables of this study. These theory based variables mainly refer to the perpetrator and the homicide event. Additionally, this study also adds several variables on sentencing because of its importance to the policy issues related to the phenomenon of RHO’s (Cale, Plecas, Cohen and Fortier, 2010). The majority of these variables are operationalized into statistic labels by accessing the European Homicide Monitor (EHM). The EHM is an extensive list of 85 homicide related variables which are chosen, defined and operationalized based on extensive discussions between researchers from the Netherlands, Sweden and Finland (Liem, 2013; Granath et al., 2011). The extensiveness and detailed character of this manual makes it a reliable tool to study the identified theory based variables of this research. However, some of the identified variables are not represented in the EHM, namely: the type of motive (instrumental or expressive); the intelligence; the age categories; the specific psychological disorder; the youth history; the current relationships with family and friends; the level of association with criminals; the extent to which the individual is held accountable and the stages of life in which the individual offended. These variables are operationalized into labels by accessing related literature and the Dutch institute for statistics who standardized several variables based on their experience in studying these subjects (CBS, 1998). The operationalization of these variables is processed in the codebook which is attached in appendix 1. This codebook defines the variable name and its abbreviation for statistical use; states the purpose of the variable (whether it is descriptive or attached to a theory); describes the labels attached to each variable and clarifies each variable by stating definitions and instructions for scoring the variable. Taken together, this research relies on 78 variables. These variables are studied by means of a cross-case approach in which variables are compared across numerous cases (Neuman, 2014).

(21)

3.3 Population and sample selection

3.3.1 Population

The population of this research comprises those individuals that were found guilty for one or more homicide(s) in the Netherlands between 1980 and 2016. However, it is difficult to state an exact number of this population of homicide perpetrators as data on homicide perpetrators is commonly less easier available than data on victims because

of privacy issues (Brookman, 2005). Additionally, short term overviews on homicide offenders are more easily available than detailed long term information on homicide offenders (Nieuwbeerta and Deerenberg, 2004). Consequently, as such long term overviews on homicide offenders in the Netherlands between 1980 and 2016 seem to lack, this study is not able to define the exact number of homicide perpetrators of this study.

3.3.2 Sample selection

In order to answer the research question of this study, this research has to sample both RHO’s and SHO’s. Herein, RHO’s are selected by purposive sampling. Purposive sampling is a non-probability sampling technique in which a wide range of methods are used to locate all possible cases of a highly specific and difficult-to-reach group (Neuman, 2014: 273). Although applying this sampling method rarely selects a sample that represents the entire population, it is appropriate to select cases with an informative purpose (Neuman, 2014) which suits the main goal of this study. Additionally, this study necessities purposive sampling as there is no database available to this research that conveys an overview of all RHO’s between 1980 and 2016. Consequently, when sampling RHO’s, this study needs to rely on the databases that are available to this study, namely: the database of the Netherlands Institute Forensic Psychology and Psychiatry (NIFP1_{), the dataset of}

Dutch Homicide Monitor (DHM), the database of public Dutch jurisdiction and the ‘homicide list’ (Korterink, 2014). Herein, the NIFP database comprises all forensic mental health reports on suspects of homicide that are questioned on their criminal responsibility between 1980 and 2016. The DHM, on

1_{When judges doubt about the mental state of a suspect, they ask the (independent) NIFP to investigate the}

individual and define a diagnosis with regards to criminal responsibility of this individual (Koenraadt, Mooij and Mulbergt, 2007). This institute advises the public prosecutors on whether to impose treatment based sentences or not by investigating the individuals on a psychiatric, psychological and environmental level through observations and conversations1_{(NIFP, 2014).}

(22)

the other hand, is an ongoing data collection effort based on the variables of the EHM and includes information on homicide incidents, victims and perpetrators in the Netherlands between 2008 and 2016 (Liem and Nieuwbeerta, 2010). Further, the public database of Dutch jurisdiction is accessible through the internet and comprises a big majority of (anonymized) jurisdiction of all sorts of offences, including homicide offending (Rechtspraak.nl, 2012). Finally, the ‘homicide list’ is a private initiative which registered the basic information (i.e. age, gender, location, and context) of all homicides in the Netherlands between 1960 and 2014 based on newspaper articles and relevant websites. Accessing these databases, only those RHO’s are selected that are actually convicted for the homicides they committed. This verification is executed by accessing jurisdiction on each identified RHO case. In fact, such a confirmation is necessary in order to avoid selecting criminal offenders that did not committed homicide more than once or several times (Eck, 2003). Following these lines, figure 1 shows the sampling process of the RHO’s and reveals that this study managed to sample a total of 28 RHO’s. Additionally, SHO’s are sampled by applying theoretical sampling. Theoretical sampling is a non-random sample in which specific cases are selected to develop or evaluate theoretical ideas (Neuman, 2014: 276). This sampling method is applied to SHO’s as it is theoretically interesting to compare SHO’s to RHO’s (Cale, Plecas, Cohen and Fortier, 2010). In order to select SHO cases, theoretical sampling is applied to the DHM. Herein, the DHM is chosen to sample the SHO’s of this study because of practical reasons. In fact, data on the cases in the DHM is already collected so SHO’s can be analysed immediately, saving this study a lot of time. When selecting SHO cases from this database, only those SHO’s are sampled that are actually convicted for homicide in order to prevent selecting cases that are not responsible for homicide. Herein, all convicted SHO’s are sampled in order to ensure the highest representativeness possible. Following this selection process, 640 SHO’s are identified in this study.

3.4 Data collection

When collecting data on extreme criminals as RHO’s, researchers must rely on extant criminal records which are not available to everyone (DeLisi and Scherer, 2006). The databases that are exclusively available to collect data for the purpose of this research are the databases of the NIFP and the DHM. However, due to ethical constraints, these databases can only be accessed at fixed locations and therefore cannot complement each other. In fact, access to these databases needs to be constraint in order to limit chances that sensitive data on an individual level becomes available to the public. Herein, individual data should remain confidential because of privacy concerns. Further, in addition to these exclusive databases, this study also collects information from sources that are publically available, namely Dutch jurisdiction and news articles.

By accessing the aforementioned sources, data is collected by means of coding. Data coding refers to systematically reorganizing raw data into a format that is easy to statistically analyse (Neuman, 2014). Herein, the codebook (appendix 1) serves as a set of rules that indicates when to assign certain numbers to variable attributes (Neuman, 2014). In order to assure validity of the results, the coded data is registered by indicating a comment in the source document each time a variable is coded. This way, accuracy and contingency of the coded data can easily be monitored (Neuman, 2014). Further, in addition to data coding, this study also applies open coding in order to identify any characteristics related to RHO’s that are not defined in the literature (Neuman, 2014). In fact, open coding is necessary since the literature did not provide one extensive list of characteristics that are related to RHO’s in the broadest sense of the term. Consequently, this may indicate that there are variables related to RHO’s which are not yet known to the literature. In fact, as a result of applying open coding, this study

(23)

identified a subsequent variable, namely the risk to engage in recidivism. The risk is commonly assessed in the forensic mental health reports and included in jurisdiction documents. The operationalization of this variable is processed in the codebook (appendix 1).

All coded data is registered in one document in the Statistical Package for the Social Sciences (SPSS) (IBM SPSS Version 22.0). SPSS is a statistical program that provides facilities to report and analyse data in a clear overview of all the variables and its options from the codebook. Within this program, each row represents one person (Granath et al., 2011). This indicates, for example, that one homicide event involving one perpetrator and two victims will have three rows in the dataset. Additionally, each variable is represented in its own column.

3.4.1 Collecting data on RHO’s

The data on RHO’s is mainly collected by accessing the forensic mental health reports which are available through the database of the NIFP. In fact, this database is accessed to study RHO’s as it provides information on the full timeframe of this study and since it presents the necessary information which is not available through public sources. Herein, the forensic mental health reports within this database always report on: whether the individual suffers from a mental disorder; to which extend this disorder influenced the individuals’ actions and choices at the time of the offence; the risks on recidivism; and sentences that should be imposed in order to decrease the chance of recidivism (NIFP, 2014). This way, forensic mental health reports present a reliable multidimensional picture of the personality, the motives and abilities, and the particular vulnerabilities and inabilities of the defendant (Koenraadt, Mooij and van Mulbergt, 2007). However, not all forensic mental health reports are similar. In fact, individuals can be investigated by: an environmental assessor, an observational assessor, a psychologist or a psychiatrist. Herein, the psychologist is concerned with answering the court’s question regarding whether the suspect was psychologically limited at the time of the homicide and how dangerous he or she is. When examining the individual, psychologists apply observation and perception, take interviews and psychological tests and access criminal case files in order to define diagnostic considerations (Koenraadt, Mooij and van Mulbergt, 2007). Second, the psychiatrist assesses whether any mental disorders can be identified at the defendant, and if so which of these restricted the individuals’ free will at the time of the offense. Herein, the psychiatrist examines general impressions (how the person makes contact, his awareness and understanding of his illness); cognitive functions (e.g. intelligence, perception, delusions); affective functions (checking for depressive or manic disorders and anxiety); and psychomotor function, urges and will. Additionally, the psychiatrist also assesses the defendant’s life story (Koenraadt, Mooij and van Mulbergt, 2007). Third, the environmental assessor describes the detainee’s life in interaction with his surroundings (Koenraadt, Mooij and van Mulbergt, 2007). Finally, the observational assessor reports on the detainees’ behavioural capabilities and handicaps (Koenraadt, Mooij and van Mulbergt, 2007). The composition of these assessments depend on the situation in which this individual is investigated. Individuals can be investigated within an observational clinic or within their own living environments. When being investigated in an observation clinic, all experts standardly assess the individual (NIFP, 2014). Additionally, when investigated in their own environment, individuals are investigated either by: a psychologist or a psychiatrist alone; both a psychiatrist and a psychologist; a psychiatrist or psychologist in collaboration with an environmental assessor; or a psychiatrist, a psychologist and an environmental assessor (NIFP, 2014).

(24)

Although the forensic mental health reports presents this study with exceptional and detailed information on RHO’s, they do not dispose of enough information to present data on all variables of this study. In fact, information on sentencing and conviction generally lacks in these reports. This lack of information on sentencing within forensic mental health reports is easily explained as these reports usually serve as a source of information for judges to decide on the length and type of sentencing. However, in order to investigate RHO’s as detailed as possible, data of forensic mental health reports is complemented by accessing jurisdiction as well as newspaper articles. Both jurisdiction and newspaper articles are accessed through the internet. Herein, data is first supplemented by accessing Dutch jurisdiction and finally, when information on variables still lacks, by accessing newspaper articles. This specific order is necessary because of reliability issues. In fact, where forensic mental health reports are defined by experts who have access to information from first hand and jurisdiction is mainly based on factual information, news articles are commonly based on unreliable sources and are commonly asked to make some kind of judgment (Neuman, 2014).

3.4.3 Collecting data on SHO’s

Data on SHO’s is collected by accessing the DHM. The information presented in the DHM is based on various sources, among which: homicide related articles generated by the Dutch National News Agency; annual summaries of homicides in the Netherlands published by Elsevier, a weekly magazine; and files from the National Bureau of Investigation, the Public Prosecution Office, the Judicial Information Service, the Ministry of Justice, and the Criminal Justice Knowledge Centre (Liem and Nieuwbeerta, 2010). Despite the fact that the DHM does not represent the full timeframe of this study and lacks to provide data on all the variables defined in this research, time limitations restricts this study to collect information on this group itself. Nevertheless, as the variables of the DHM are based on the EHM, collecting data on SHO’s through this monitor provides enough information to compare the majority of the variables to the data of RHO’s.

3.5 Data analysis

In order to be able to answer the research question, data on both RHO’s and SHO’s is descriptively analysed and compared. Descriptive analysis refers to describing the basic patterns of the data (Neuman, 2014). Herein, this study applies frequency distribution, measures of central tendency (mean, median, mode) and measures of variation (standard deviation). Within these analyses, variables are mainly studied by univariate statistics, indicating that only one single variable is described at the time. Although bivariate analysis may be much more valuable as it describes relationship between variables, this study is not designed to conclude on such relationships as the research sample of RHO’s is too small (Neuman, 2014). The descriptive univariate analyses are executed in SPSS (see appendix 2 for the calculations made in this programme).

In order to assure that the descriptive analysis measures the variable as intended, data on both RHO’s and SHO’s is analysed based on either the victim or the perpetrator. Herein, data that is analysed based on the victim refers to the variables that changes with the victim, such as the motive and the relation to the perpetrator. Additionally, data that is analysed based on the perpetrator relates to the variables that are fixed on the perpetrator, such as gender and birth country of the perpetrator. However, when referring to perpetrators of repeat homicide, several variables may also change per homicide event, such as the professional situation of the individual at the time of the event. Therefore, when referring to perpetrators of repeat homicide, several variables are also analysed based on the perpetrator per homicide incident. With regards to these foundations when analysing the variables, a total of 640

(25)

convicted perpetrators of single homicides are identified who victimized a total of 1153 individuals. Further, with regards to RHO’s, a total of 28 individual perpetrators are identified who are responsible for the homicide on 65 victims. These 28 RHO’s are responsible for 58 homicide events in which 22 individuals are responsible for 2 homicide events and 5 is the highest amount of homicide events per perpetrator. Additionally, 4 of the 28 RHO’s collaborated in pares when committing the homicides. Consequently, an amount of 63 perpetrators of repeat homicides are identified per incident.

Having descriptively analysed the variables based on either the victim or perpetrator, the data is represented in tables and graphs. Graphs are applied to be able to relatively compare the data between RHO’s and SHO’s. Herein, this study applies bar charts as these graphics clearly show similarities and differences between groups (Neuman, 2014). Additionally, tables are applied to those variables for which no data is available on SHO’s. Within these tables and graphs, percentages regarding SHO’s are rounded by one decimal and percentages regarding RHO’s by whole numbers. This difference is made since the sample of RHO’s is so small that presenting decimals does not add any valuable information to the results of this study. With regards to the sampled SHO’s, however, decimals could present this study with valuable information because of the sample size of this group.

3.6 Reliability and validity

Reliability and validity are important concepts in research as they enhance the accuracy of the assessment and evaluation of research work. Herein, reliability refers to the consistency, stability and repeatability of results. Results are considered reliable if similar results are obtained in identical situations but in different circumstances (Neuman, 2014). The reliability of this study is ought to be maximized by strictly following the codebook (appendix 1). Following the codebook ensures that similar results will be found when the research is repeated. Nevertheless, as interpretation is unsurmountable regarding several variables (mainly regarding relationships with family and friends) and assessors of the perpetrators sometimes contradict each other, this study cannot completely assure reliability.

Validity, on the other hand, refers to the extent in which the measuring instruments measure what it intends to measure (Neuman, 2014). The validity of this study is also mainly insured by strictly following the codebook of this research (appendix 1). In fact, the majority of this codebook is compiled through extensive research by known researchers from the field of study. Using the codebook thus scientifically ensures that the indicator measures the construct (face validity) (Neuman, 2014). Validity of the remaining variables is ensured by accessing scientifically based definitions and labels of the variable in order to ensure that the full content of the definitions is measured (content validity) (Neuman, 2014). In addition to face- and content validity, research also commonly speaks of external and internal validity. Here, internal validity is ensured when the independent variable, and nothing else, influences the dependent variable (Neuman, 2014: 298). External validity, on the other hand, refers to the ability of a study to generalize findings beyond that specific study (Neuman, 2014). Internal validity does not apply to this study since this research does not aim to conclude on any causality between variables. External validity, on the other hand, does apply to this study. It is questionable, though, whether this study is capable of generalizing its results to the population since the population of RHO’s is currently unknown and the sample studied is rather small. Nevertheless, this study does not intent to generalize findings. Though, this study aims to bring clarification on the phenomenon of RHO’s as the basis for more conclusive research.

(26)

Chapter 4: Results

This chapter systematically and objectively reports on the results of the data collection. These results present the nature of repeat- and single homicide offending in the Netherlands between 1980 and 2016. Specifically, the characteristics of RHO’s and its similarities and differences with the characteristics of SHO’s are reported (where possible). Herein, this chapter distinguishes individual characteristics from control variables (such as professional situation, civil situation and mental illness) (Cale, Plecas, Cohen and Fortier, 2010) and further elaborates on the characteristics of the homicide event and sentencing regarding RHO’s and SHO’s.

This research studies a total of 28 RHO’s and 640 SHO’s. The sampled RHO’s victimized 58% (38 in total) men and 42% (27 in total) women, where SHO’s victimized 67% (769 in total) men and 33% (379 in total) women. Herein, both the victims of RHO’s and SHO’s had an average age of 39 at the time of the homicide event. Further, with regards to the RHO’s, 14 of the 28 perpetrators recidivated during- or after their sentence. Consequently, 14 perpetrators were not convicted for their first homicide as they were not yet identified as the perpetrator of that homicide.

4.1 Individual characteristics of RHO’s and SHO’s in the Netherlands, 1980-2016

This paragraph presents the results on the individual characteristics of the sampled RHO’s and SHO’s in the Netherlands between 1980 and 2016. The variables are based on the individuals itself and refer to gender, age, birth country, and intelligence. Here, intelligence is not studied for SHO’s.

4.1.1 Gender and age

The results in figure 2 reveal that both repeat- and single homicides in the Netherlands between 1980 and 2016 are predominantly committed by men. Herein, RHO’s are slightly more overrepresented by the male gender than SHO’s.

Figure 2: Gender of RHO’s and SHO’s, 1980-2016 (by perpetrator)

93% (=26) 7% (=2) 90,0% (=576) 10,0% (=64) 0% 20% 40% 60% 80% 100% Men Women

RHO's (n=28) SHO's (n=640) Total: RHO’s = 28 ; SHO’s = 640

Characteristics associated with repeat homicide offenders relative to single homicide offenders. An explorative study on a 35-year sample of homicide offenders in the Netherlands