Intimate partner violence in a police setting. Offender characteristics and typologies found within reported cases from Funen Police

(1)

INTIMATE PARTNER VIOLENCE

IN A POLICE SETTING

- Offender characteristics and typologies found

within reported cases from Funen police

CHRISTINA B. JØRGENSEN

Degree project in Criminology Malmö University

120 credits Health and society

Criminology, Masters’ Programme 205 06 Malmö May 2018

(2)

INTIMATE PARTNER VIOLENCE

IN A POLICE SETTING

- Offender characteristics and typologies found

within reported cases from Funen police

CHRISTINA B. JØRGENSEN

Jørgensen, C. Intimate partner violence in a police setting. Offender

characteristics and typologies found within reported cases from Funen Police.

Degree project in Criminology 30 credits. Malmö University: Faculty of Health

and Society, Department of Criminology, 2018.

Background: Intimate partner violence (IPV) is a serious problem, and the

consequences extend beyond the physical damage that may occur. Police officers are often first responders, why it is important that they possess awareness and skills that facilitate them to identify IPV and the characteristics for the involved parties, in order to reduce further consequences and escalation. Aim: The aim of this thesis was to explore offender characteristics and typologies of IPV as it occurs within police reports from Funen police district, and to discuss how this information can be used in a police setting. Method: A random sample of 400 reported cases of violence were coded, and 60 reported intimate partner violence cases were identified. Descriptive statistics and a two-step clustering were used to explore the 60 cases of IPV. Results: Overall, it was found, that the offenders of the reported IPV cases in Funen's police are in accordance with studies of offenders of IPV in other contexts. A typology was developed from the clusters which resulted in three different offender types; generally violent, alcohol-related violent and family violent. These clusters also resemble groups found in previous studies. Discussion: The results can be used to give an insight into how IPV occurs in the police district, and form a basis for the understanding and

developing of the prioritization of the phenomenon. However, the descriptions and typologies are to be understood as a snapshot and the results are not universal or a general description of IPV, why further and deeper investigations will be necessary.

Keywords: Intimate partner violence, offender characteristics, offender typology,

(3)

INTRODUCTION

Intimate Partner Violence (IPV) is a serious problem worldwide, where 1 in every 3 woman are found to have experienced threats, violence or stalking by a partner (WHO, 2013). In Denmark it is estimated that approximately 46.000 women and men are annually exposed to violence by a current or former partner (Helweg-Larsen, 2012). In the area of violence committed in the public, men constitute the largest group of victims, whereas in the private sphere it is the opposite; women constitute the largest group of victims (Rasmussen, 2014). Violence committed in the home has historically been treated as a private matter and of less serious nature than other acts of violence. Men's abuse of violence against woman has previously been regarded as a natural part of the gendered roles and power relations in a relationship (ibid). Until 1981, textbooks within the police also showed that their most important assignment within domestic dispute was to ‘preserve the marriage’ (Kühlmann, 2008). Later, there has been a greater focus on IPV cases and a significant change in attitude towards this area, although today it is still an area characterized by a large dark number (Pedersen, et al., 2017; Balvig & Kyvsgaard, 2006; Helweg-Larsen, et al., 2010). An increased focus has especially derived from the great number of research that illustrate how the consequences of IPV extend beyond the physical damage that may occur for both men and women. This can be seen in the major financial costs associated with the episodes, for example, medical treatment, lost earnings, contact with social

authorities and the criminal justice system, etc. (Gelles, 2016). Furthermore just as the episodes has far-reaching consequences for the parties involved, it is widely acknowledged that children in violence-based families also can be considered victims of violence, even though the violence is not directed against them (Edleson, 1999; Plauborg, et al., 2012). The consequences for the children are identified by several studies as; an increased risk of becoming victims or offenders later in life, abuse, depression, behavioral difficulties, PTSD, etc. (Felitti, et al., 1998; Stéphanie, et al., 2017; Helweg-Larsen, et al., 2010). These consequences have presented how important it is to prevent IPV or a further escalation of the violence.

BACKGROUND

There is no definitive definition of the concept of intimate partner violence neither nationally or internationally, as both "partner" and "violence" are two concepts that can be interpreted in different ways. Therefore, many terms are also used for the phenomenon, including; “domestic violence”, “violence within close

relationships”, and “Intimate Partner Violence” (IPV) which will be used

throughout this thesis. One of the most widely used definitions is from the World Health Organization, which state that IPV; “Refers to behavior by an intimate

partner or ex-partner that causes physical, sexual or psychological harm, including physical aggression, sexual coercion, psychological abuse and controlling behaviours” (WHO, 2017). In this thesis only the physical violence,

that can be derived from police reports will be included in the definition of IPV. Since 1995, Denmark has conducted nationwide victim surveys, which have provided a gender-based statistical overview of the extent and development of the violence in the Danish society, including IPV (Rasmussen, 2014). In addition,

(5)

since 2004, the National Institute of Public Health has conducted monitoring of violence in close relationships every fourth years. However, the criminological and sociolegal research in the area seems to be less extensive in Denmark than in other Nordic countries (Rasmussen, 2014).

Dynamics and processes of intimate partner violence

The dynamics and processes within IPV is previously studied and related to all types of intimate relationships, as well as a wide range of violence involving both physical, psychological and sexual violence (NCK, 2009). In these studies, the oppression and offenses committed by the offender to the victim is relatively universal and is described as the normalization process or the cycle of violence (Lundgren, et al., 2001). Eva Lundgren's (2001) theory of normalization is particularly widespread in Scandinavia, and provides a processual understanding of the dynamic within IPV (Schacke, 2009). The normalization process is

described by Lundgren et al. (2001) as a process that not only affects the victim and prevents them from leaving the violent relationship, but also affects the offender, which makes it a very difficult process to get out of (Lundgren, et al., 2001). It is described, that the violence becomes a latent threat to everyday life, and both victim and offender develop strategies to control and adapt to the situation (Schacke, 2009). Initially in the process the intimates are experiencing individual events that can be defended as an accident and thus be downplayed and excused. Then, these violent events become harder and harder to explain why more and more events are accepted and the perception of reality is displaced, as well as shame being performed (Schacke, 2009; Lundgren, et al., 2001). The central, according to Lundgren et al. (2001), is that the violence has been controlled and is performed in such a way that it is not perceived by the victim and surroundings (Lundgren, et al., 2001).

Offender typology

The high recidivate rates in IPV cases illustrates why it is important to identify perpetrators at risk. This is also why there for many years have been made efforts to identify typologies of IPV offenders. One of the most acknowledged and replicated offender typology has been developed by Holtzworth-Munroe and Stuart (1994) (Huss & Ralston, 2008; Cunha & Gonçalves, 2013; Grana, et al., 2014). Holtzworth-Munroe and Stuart (1994) have proposed that IPV offenders could be categorized on three descriptive dimensions: (a) severity and frequency of violence, (b) generality of violence, and (c) psychopathology and personality disorder. With these dimensions, three subtypes of offenders have been identified throughout previous studies: (1) men who carry out ‘family-only’ violence, (2) men who can be categorized as ‘dysphoric/borderlines’ offenders and (3) men who are ‘generally violent/antisocial’ (Holtzworth-Munroe & Stuart, 1994). The ‘family-only’ offenders are characterized by a relatively low level of violence and the violence takes place only within the family. Compared to the other groups this is the least violent. Furthermore, abuse and depression have a limited extent in this group (ibid). The ‘dysphoric/borderlines’ offenders are characterized by committing more serious violence and violence outside the family. Additionally, they are more prominent to substance abuse (ibid). Lastly the ‘generally

violent/antisocial’ offenders are characterized as being more often criminal across all situations and environments and abusers (ibid).

Subtypes of IPV offenders can be of practical importance in order to prevent IPV, as the types are different in their frequency as well as the severity of the

(6)

committed violence (Holtzworth-Munroe & Stuart, 1994). Identifying the offenders to the different subtypes may increase the likelihood of distributing adequate interventions, hence preventing IPV. However, Holtzworth-Munroe and Stuart’s (1994) subtyping has been criticized in relation to some different points in previous literature. For example, several studies have followed the same approach but identified a fourth typology (Holtzworth-Munroe, et al., 2000; Johnson, et al., 2006; Thijssen & de Ruiter, 2011) Further, the typologies have been criticized due to the dimension of psychopathology. It is described that, especially among

practitioners and non-clinicians, it may be difficult to categorize offenders in the different subtypes (Boyle, et al., 2008). In order to meet this criticism some studies have attempted to identify subtypes based on the generality of violence dimension alone, regarding whether the offender is violent only with his wife and family or in other situations as well. This has resulted in two subtypes: ‘family-only’ and the ‘antisocial’ offenders, which are somewhat similar to the original ‘family-only’ and ‘generally violent/antisocial’ offenders proposed by

Holtzworth-Munroe and Stuart (1994) (Boyle, et al., 2008, Goldenstein, et al., 2015). In a police context, this may be highly relevant as it is often only this information that is available. Recent studies made in a Swedish police context have, also with information from the generality of violence dimension, found that ‘antisocial’ offenders are younger than ‘family-only’ offenders, and they have present significantly more risk factors for IPV than the ‘family-only’ offenders (Petersson, et al., 2016). ‘Antisocial’ offenders further recidivate to a greater extent than family-only offenders and the majority of those who did recidivate in both groups did so within the first year from the index crime (Peterson & Strand, 2017). These results indicate that it is important with different intervention strategies depending on the offender's subtype in order to prevent future violence.

Police and intimate partner violence

In the research there is no doubt about the importance of police action in the field of IPV, which is reflected by the large group of research on risk assessments and management by the police (Storey, et al., 2014; Nesset, et al., 2017; Belfrage & Strand, 2012; Belfrage, et al., 2011; Sherman, et al., 1992). Some of the first studies found that arrest of the male offender reduced the risk of future partner violence by 50% and that this was more effective than merely advising (Sherman & Berk, 1984)However, later studies found that there were time-limited effects for the arrests and that they varied in relation to the characteristics of the offender (Sherman, et al., 1992; Garner, et al., 1995). This gives a perspective on the importance for social actors, like the police, to have awareness and ability to identify high risk people in recidivism. One way to obtain this knowledge is by the use of risk assessment tools. Risk assessment is an evidence-based approach to identify and estimate the risk of violence in order to be able to effectively manage the offender in the situation, and implement the necessary measures (Douglas, et al., 1999). The purpose of these tools are to prevent the continuation and

development of IPV. It is therefore a prerequisite that violence has taken place before using the tool. Many countries use a wide range of risk assessment tools as part of the police management and prevention of IPV (Rasmussen, et al., 2016). Studies of risk assessment tools used in Swedish contexts have shown mixed results, however, some studies have shown that the rate of recidivism is lower in high-risk cases (Belfrage & Strand, 2012; Storey, et al., 2014).

As mentioned, criminological and sociolegal research on IPV is limited in

(7)

assessment tools are not a coherent part of the management of IPV, although the National Prevention Center introduced the risk assessment tool SARA (Spousal Assault Risk Assessment) in 2016. The plan is to implement the tool in Funen Police during 2018. Without previous research on how IPV occurs in the Danish police, a theoretical framework of the culture in the police as an organization are included.Lars Holmberg (1999) has developed a theory and understanding of police work and has a broad focus on the estimate the police officers have to make in a given situation. He seeks to reveal attitudes and patterns in the ways the officers perform this estimate and how they enforce legislation (Holmberg, 1999). Holmberg (1999) theorizes that since the officers are free to define what is

relevant to law enforcement or not, there are some dominant understandings for what can be conceived as "good" and "bad" police assignments (ibid). Cases with violence, but especially IPV cases, are an example of a "bad" police assignments, because they often end up in a claim against claim with little or no evidence available. Michael Lipsky (2010), has a similar theoretical framework of the culture based on the police being Street-Level Bureaucrats. Lipsky (2010) also argue, that the police have some degree on discretion on how they enforce the law and make decision on who to arrested and who not to arrest. In order to manage the different and large number of clients the Street-Level Bureaucrats develops routines and simplifications referred to as categorization. Through the

categorization of people, it is possible for the officers to assess and manage the various situations they are facing on a daily basis (Lipsky, 2010). Thus, in the assessment processes, the officer becomes part of a political context in which it is necessary to decide whether the legal context or individual decision making should be used (ibid). The power advantage plays a major role in managing violent situations, and knowledge and understandings about individual cases is of major importance. At the worst, it could mean that the police officers build a wrong kind of stereotype, which is the main argument for the importance of implementing police research (ibid).

Relevance of the study

When police officers go to private homes on suspicion of IPV, a process starts where the police will communicate with the involved parties and, based on the available information, assess whether to, or how to intervene. The nature of the assignment, however, depends not only on what has actually happened in the situation, but also on the collaboration between the involved and the police and furthermore their understanding of the phenomenon of IPV. Therefore, it is important for the police to be able to identify IPV and the characteristics for the involved in order to reduce further consequences and escalation. As the police are often the first public authority that victim and offender meet in cases of IPV, it is important how this meeting go and how the police manage the event. After reviewing the existing literature, and theoretical understandings of the phenomenon and the police involvement, it is clear that characteristics and understandings of the offenders and the situations are essential in order to handle IPV cases in the best way possible. However, since there is a limited knowledge about the patterns and characteristics of IPV, among the Danish professionals who should be able to act and intervene, such mapping is needed. In order to

understand whether the already existing knowledge of IPV is generalizable for a Danish police context more research is needed. Since IPV is covered by a major dark figure, it is further interesting to examine whether the characteristics found within reported cases, looks like characteristics found when studying IPV in different ways. This in order to provide a basis for a discussion of the most

(8)

appropriate management and targeted intervention against IPV in a Danish police context.

Aim and research question

Throughout the above it was found particularly interesting to explore IPV as it occurs within the Danish police. A collaboration with the Danish police district of Funen was initiated with the intent of developing a descriptive study of what can be characterized within the reported phenomenon of IPV. The thesis takes into account the following research question:

What characteristics and patterns can be identified in relation to the offender and the criminal act in the intimate partner violence cases reported to the police? and how can this information be used in a police setting?

Three working questions have been developed and are used structurally for the analysis:

1. What characteristics of the offenders are prominent in the cases of IPV? 2. What characteristics of the criminal act are prominent for the IPV cases? 3. What typologies can be identified within the IPV cases?

The last part of the research question, will be further examined in the discussion. Due to the limited extent of the thesis, it has been necessary to make some restrictions in regards to the focus of the thesis. Therefore, focus will be on the offenders and the criminal characteristics and not the victims. This focus has been found as the thesis has been performed within a police context, where it

undeniably is the offender and the criminal characteristics whom has the utmost importance.

MATERIALS AND METHODS

Materials - data and coding procedure

The thesis is based on a random sample of 400 reports of simple and serious violence, from 2017 from the Funen police district1_{. The thesis is based on these} police reports, as well as information from intern police systems. Specific

criminal codes have been chosen for violence which constitutes a restriction of the sample. Criminal codes (gerningskoder) are a categorization used in the Danish police and they are based on sections from the law. With the codes the police divide cases into different groups, as simple and serious violence. The reason for this restriction must be found in the defined capacity of the thesis, as well as a desired focus on cases of violence from Funen police.

The first step in processing the data was the development of a codebook specifying the variables in which the cases should be coded. Since the purpose was to provide characteristics of the reported violence, the quantified information is extremely far-reaching, as it may be challenging to identify in advance what information is interesting. However in order to not completely work in blindness,

(9)

the development of the codebook has partly been based on theoretical interests and focus areas for Funen's police district, as well as partly from qualitative readings of the police reports (Layder, 1998).A pilot coding of a selection of the cases were done by four different people. This were done by reading and coding cases independently of each other and thereby testing and evaluating if the variables and code principles could be used meaningfully on the data. The coding was then compared, discrepancies discussed and the codebook revised. This shift between quantification and qualitative reading has played a main role in ensuring the validity of the study.Out of the 400 cases from the sample a total of 389 unique cases were coded. The remaining were sorted off along the way, including cases where they contained too little information for a valid result. In case of doubt or inconsistency, a collective coding was discussed. Since there are no specific criminal codes for IPV, it has been necessary to read all 400 cases of violence manually to identify whether or not intimate partners were involved. Out of the remaining 389 cases 60 cases were identified as IPV. In the 60 cases there are respectively 54 offenders and 54 victims, thus 6 couples are duplets.

Methods

Through the above, it has been possible to code a number of variables based on the case content. However, the amount of coded data was dense, and in order to specify the final variables, the data was condensed to text. Some descriptive statistics, have been performed in order to select the data used for the final analysis.

Descriptive statistics

The included variables can be categorized in two groups; (1) offender information and (2) crime characteristics (appendix A).

Presentation of offender information variables:  Sex  Age  Alcohol influence  Drug influence  Abuse  Mental illness

 Related to a criminal environment

As it was also highlighted in the introduction and the background, there is a distinctive focus on gender in the field of IPV, drawing a picture that the majority of offenders are men. Thus, it is relevant to gather information about the extent to which this picture is similar in police reports. In the same way age is in earlier literature described as a factor that differs in IPV cases, than among other types of crime (Johnson, et al., 2015). Alcohol, drugs and mental illness are also factors described in criminology to be risk factors in regards to most criminal acts, including IPV cases (Murphy, et al., 2001; Thompson & Kingree, 2006; Stenager & Stevenson, 2013). These variables have been coded “yes” if the offender has been affected at the time of the crime. Abuse has been coded as “yes” if the offender has been found to drink more than the recommended amount of alcohol, if they have previously been reported as drug users, if they had a high

(10)

related to a criminal environment is interesting in relation to a characteristic and whether those involved can be said to be characterized by general risk behaviors. The variable can be coded “yes” if the offender has an attachment to groups or people who regularly commit criminal offenses. The offender has not necessarily been active in the criminal act, but have been part of the environment.

Presentation of crime characteristics variables:  Relationship between victim and offender  Shared housing

 Shared children

 Children witness to the violence  Reported Revictimization of violence  Same offender

 Reported recidivism of violence  Told revictimization of violence

 Victim previously felt threatened by offender  Previous contact with police between the couple

Since IPV cases could not automatically be withdrawn from the police reports, it has been necessary to include knowledge about the relationship between the victim and the offender in order to define the case as an IPV incident. The case has been coded as IPV if it from the reports could be concluded that the offender and victim have had a present or former romantic relationship. IPV is generally described as taking place privately, why it makes sense to examine whether the partners are living together with children or if the reported events are more characterized by distance between them. To the aforementioned understanding of “good” and “bad” police work, it has been found interesting to investigate

previous crime histories for the involved. This crime history, “told victimization”, “previously felt threatened” and the “previous contact with police between the couple”, has been coded “yes” if the offender or victim has been committed or subjected to a similar crime within a two-year period from the index crime. These variables are thus only an expression of their registered and told risk behavior. In regards to the severity of the crime, ”objects used to perform violence” was included.

These variables are chosen as they make sense in regards to the previous mentioned literature, and further this knowledge is available to the police when dealing with the cases, why it is useful in this context. The variables are initially used to map and describe IPV from the sample and thus examining what

characterizes the reported cases. It should be noted however, that the presentation is based on relatively few cases.

Cluster analysis

The variables are coded in Excel, and exported to SPSS (IBM, v. 23), where a statistically generated cluster analysis is performed, with the purpose of examining whether there are particular patterns in the IPV cases. The idea of using a cluster analysis is to form clusters that differ as much as possible from each other, but at the same time internally is as homogeneous as possible. With this method, it is possible to explore and condense the many observations and examine how the cases can be grouped and subsequently typified and

(11)

conceptualized (Hussain & Lauridsen, 2017).Since the dataset contains both continuous and categorical variables, a "two-step clustering" method has been used. This is also the most exploratory approach among the different cluster methods, which is consistent with the purpose of the analysis (Verma, 2013). The first step for a cluster analysis is to identify which variables to include. When performing a cluster analysis, it itself cannot differentiate between relevant and irrelevant variables, why it is important which variables to use. Therefore, only relevant variables, which were supported by a theoretical and practical framework were included. These variables must not be obviously dependent on each other, and they should make sense in relation to the area in which they want to generate knowledge (Verma, 2013). Since focus in this thesis is on the offender and the criminal characteristics the following variables selected for cluster analysis are:

 Age of offender  Alcohol influence  Drug influence

 Couple previous in contact with the police  Reported recidivism of violence

 Used object to perform violence  Related to criminal environment

These variables are chosen as they characterize the offender and the relationship between the offender and victim and the criminal act. The variables further resemble the “generality of violence” dimension suggested by Holtzworth-Monroe and Stuart (1994), and can tell something about whether the violence takes place only in family settings or also in general. Furthermore, as also stated when presenting the variables in the descriptive section, these variables are

practically useful for the police in order to identify subtypes within IPV offenders, as the other dimensions can require more clinical training (Boyle, et al., 2008). It is important to note that since the cluster analysis is a descriptive and explorative analysis, there are not only one method but various possible methods to perform it. No statistical test that can determine which cluster formation is the most valid in relation to the desired output (Hussain & Lauridsen, 2017). However, this does not mean that it is not possible at all to validate the results of the cluster

formations and their usefulness. The validation was done by comparing the analysis with an alternative method for clustering. Both K-means as well as two-step clustering’s has been performed, and they had, as expected, some minor variations, but not some that seemed significant enough to change the choice of clustering method. Furthermore, a ANOVA was performed to illustrate how the clusters differ significantly between each other (see appendix B). This show a satisfying significant level under .05 on all the variables, except “drug influence” and “Related to a criminal environment”. Despite this, these variables were kept, based on criterion validity, in which it is assessed whether the statistically established clusters are understandable in relation to theory and the sense of potential relationships that have been found through the coding process (Verma, 2013).

Ethical considerations

A permission to use the data in the current study has been obtained by Funen police. Further this permission has been approved by the local reginal ethics review board (HS2018 löp nr 65, appendix c). Due to the nature of the study, and the fact that it focusses on individuals and their lives and experiences with

(12)

violence, it is important to make some ethical considerations of how to use and collect the data material. As this thesis uses existing documents as the empirical foundation, it also means that there are no direct contact with the individuals being examined (Duedahl & Jacobsen, 2010). Consequently, no consent has been obtained from the offenders involved in the examination.

In Denmark, §10 of the Data Protection Act (Databeskyttelsesloven) provides the possibility that sensitive personal data can be processed without consent if the processing is conducted solely for the purpose of carrying out statistical or scientific studies of social significance, as long as guidelines for social sciences are respected. Of course, Funen police have agreed upon the processing of the data, but in this context it has not been found appropriate to ask the involved offenders for consent. This is due to the fact that this could have negative effects for the research itself, as those informed may question the authority of the police, which may interfere with the police management. In addition, this study has been explorative, which means that at the start of the study there was no clear picture of what the study would find. Thus it can be extremely difficult to inform about the study. Since the thesis has been done in cooperation with the police, this also is a factor that could cause problems in relation to consent, since contact with the police for many people involves an asymmetric power relation (Holmberg, 2011). Should consent be obtained in this situation, the only possibility would be after the withdrawal of the sample, the police would assist to contact the parties that has been involved. This situation should be carefully considered since this force those involved to relive the situation, without having asked for it as they now have to deal with the researcher's presence. In such a situation, consent may not always be considered voluntary and fully informed simply because of the presence of the police in the project. Furthermore, there are cases that have not been processed and therefore the offenders are still in police custody. Here consent must be regarded as an illusion and the use of the information cannot be considered more ethical correct with a consent, as this cannot be perceived as voluntary. In

Denmark, it is further stated in the Guidelines for Research Ethics in Social Sciences, that the researcher "must avoid unnecessary inconvenience and trouble

(...) and carefully consider any negative effects of its research" (SSF, 2002: 2,3).

There is a real risk that an inconvenient request for consent may help to start for example a confrontation between a suspect and the police, so that the police must resort to more drastic means than necessary. Based on the above, it is considered that, since a person with a consent can end up worse according to the police and the justice system than he/she would have been otherwise, this is a far greater ethical problem than the lack of consent. However, it is important to note that this thesis presents only aggregated data, and thus it is neither possible to identify cases or individuals (De Vaus, 2014). Furthermore, all data has been stored securely and has not been distributed to any external. The data set will be destroyed from the author's possession when the project has been completed.

RESULTS

From the total number of cases of violence (n=389), violence in close

relationships represents 20,8% (n=81). Of these cases, IPV represents the vast majority, corresponding to 72,8% (n=60). In addition to this comes the mentioned fact that IPV is less reported to the police than other violence, why the real figure

(13)

may be expected to be higher (Pedersen, et al., 2017; Balvig & Kyvsgaard, 2006; Helweg-Larsen, et al., 2010).

Prominent characteristics of the offender

Figure 1. Distribution of age in the sample of IPV, 2017, N=60 cases.

As seen in figure 1, there is an overrepresentation of offenders in the 21-25year age group, at 23,3% (n=14). The second largest group of offenders is in the 26-30year age group, which contains 16,6% (n=10). Beside this age difference, a big gender difference is found as well, where the gender equality is dominated by male offenders. In 98,1% (n=59) of the cases, the offender is a man and the victim a woman.

In only 2 cases (3,3%) it could be concluded that the offenders suffer from mental health problems, defined as having either been insane at the time of the crime or experiencing a long-term and severe state of mental illness within the last two years. This is apparently not a major factor in relation to the reported IPV cases. In relation to alcohol impact, in 21,6% (n=13) of the cases, the offenders were under the influence of alcohol at the time of the crime, and it is also seen that the offenders have a more frequent tendency to be under influence than in the case of the victims. In 5 cases it is has not been possible to define if there were alcohol influence or not. The same tendency applies for abuse, in 16,6% (n=10) of the cases the offenders have an alcohol or drug abuse and in 6 cases it is unknown.

Prominent criminal characteristics

There is the same amount of unique offenders (n=54) and victims (n=54) in all of the reported cases. Furthermore, the majority of the cases, 76,7% (n=46), takes place within a private residence, which includes apartments, flats, villas, houses and the like. In 53,3% (n=32) of the cases the victim and the offender share the same housing, and in 45% (n=27) they also have shared children. In the cases where there have been shared children, the children have witnessed the violence in 51,9% (n=14) of the cases. In 10 (16,7%) of the cases weapons or other objects were used to perform the violence.

In 14 of the cases (23,3%), the victim has within the last two years before the reported case, been exposed to violence, where they have reported it to the police. In all except one case, the violence was committed by the same offender. Half of

11-1 5 ye ars 16-2 0 ye ars 21-2 5 ye ars 26-3 0 ye ars 31-3 5 ye ars 36-4 0 ye ars 41-4 5 ye ars 46-5 0 ye ars 51-5 5 ye ars 56-6 0 ye ars 0 5 10 15 20 25 Age of offenders % o f o ff en de rs

(14)

the victims (n=27), furthermore tells that they have previously been victims of violence, but that it has not been reported. In addition, there has in 40% of the cases (n=24), prior to the included report in the sample, been contact with the police and the couple on other criminal codes than violence (e.g. domestic disputes, unrest and disorder, investigation - incident without case, etc.)2_{. In} 26,6% (n=16) of the cases, the offender has previously been charged with violence. Furthermore, in 6 cases the offenders are in some way related to a criminal environment. In 2 cases it is unclear whether the offender has been related.

Prominent typologies within the partner violence - cluster analysis

The statistical cluster analysis reduces the number of cases, and the analysis includes 54 cases of the total 60. A case can only be included in the cluster formation if it contains a real value on each of the 7 included variables. When removing the 6 cases with missing data, a complete case analysis has been performed.

As expected, three subtypes of offenders were identified; generally violent, alcohol-related violent and family violent. The results as a whole resemble the typology proposed by Holtzworth-Munroe and Stuart (1994).

Table 1. Cluster analysis performed on seven variables

Cluster 1 (Generally violent) 16 cases (29,6%) Cluster 2 (Alcohol-related) 16 cases (29,6%) Cluster 3 (Family violent) 22 cases (40,7%) Age at crime (average) 28,38 39,06 34,18

Alcohol influence No (100%) Yes (81,2%) No (100%)

Drug influence No (100%) No (93,8%) No (100%)

Couple previously in

contact with police Yes (75%) No (56,2%) No (100%)

Reported recidivism of violence Yes (62,5%) No (81,2%) No (100%) Used object to perform violence No (87,5%) No (50%) No (100%) Related to a criminal environment No (87,5%) No (93,8%) No (95,5%)

(15)

Cluster 1 - Generally violent

The first cluster consists of 16 cases (29,6%). This group are those who participate in most violence and have the most extensive history of related

criminal behavior and illegal involvement. Alcohol and drug impact at the time of the crime are not present. A majority of the few cases were the offender are related to a criminal environment are also found within this group. The cluster furthermore, consists of the youngest offenders within the three clusters, with an average age of 28,38 years.

Cluster 2 - Alcohol related

The next cluster that emerged from the exploratory cluster analysis are also based on 16 cases (29,6%). Contrary to the previous group, there are no overwhelming previous crime history of similar violence in this group. However, it is only a small overweight that has not been in previous contact with the police in a couple context. In addition, the most dysphoric factors are found within this group, which are reflected in the group's alcohol influence at the time of the crime, where the majority of the group has been affected. In this group the oldest offenders are found, with an average age of 39,06 years. Furthermore, within this group most cases where a subject has been used to perform the violence are found.

Cluster 3 - Family violent

The last and largest group, is based on 22 cases (40,7%). In this cluster, the violence is solely limited to the intimate partners. There are found no one with a former history of similar violence, and there are no alcohol or drug influence at present. The age-average in this group are 34,18 years.

DISCUSSION

The aim of this thesis was to explore and describe the offender and criminal characteristics within IPV in reported cases in the police of Funen. The main findings are that males were almost exclusively offenders and the majority are in their twenties. Mental illness, alcohol and/or drug use did not seem to be a dominating factor in the cases. In over a third of the cases there have been previous contact with the police and the reported violence occurs most often in private residence. Further three subtypes of offenders were identified on the basis of the empirical data, which supports previous findings of heterogeneity among offenders in IPV (Holtzworth-Munroe & Stuart, 1994; Cunha & Gonçalves, 2013; Grana, et al., 2014).

Discussion of the results

If the prominent gender difference is compared to previous research in the field, both national and international studies confirm that women report IPV more often than men, as well as they constitute the majority of the victims (Pedersen, et al., 2016; Stéphanie, et al., 2017). However, other studies have shown that women, to a greater extent than assumed, also commit IPV (Straus, 2011). When female offenders are not found among the reported cases, it may indicate that there is a difference in the tendency to report between the genders. In research within genders this can be explained by the fact that men reporting IPV can be exposed to social stigma and other deviations from their masculinity (Lupri & Grandin, 2004; Arnocky, 2014; Rambøll, 2013). Furthermore, IPV committed by women

(16)

against men is generally less recognized by society than IPV against women, which may act as a further hindrance for men reporting their situation (Lupri & Grandin, 2004; Arnocky, 2014).

The relationship between age and crime is one of the most studied and robust understandings in criminology. This relationship shows that crime rises in the early adulthood, peak in the early to mid-20s and then desist again (Rocque, et al., 2016) This understanding has been termed the ‘age-crime curve’. Although most offenders desist after adulthood, criminal career research has shown that not everyone is following this, and this is especially true for certain types of crimes. Gambling, drunk driving and family violence are more common among those over 30 years (ibid). This can also be seen in the results, where the age of the

offenders’ peak later and desist later than in the general age-crime curve. This may be due to the fact that it takes time to establish an intimate relationship, why it also takes longer before the violence can take place (Johnson, et al., 2015; Volsing, et al., 2018).The age tendency can also be understood in regards to the normalization process, where the violence at the beginning of the relationship often will be downplayed or rationalized by the victim (Lundgren, et al., 2001). Enforcement of these crimes should therefore note their seriousness in later life and prevention efforts should focus on older people as well as adolescents (Rocque, et al., 2016).

Contrary to the results, earlier results from Danish studies showed considerably higher cases of both alcohol and drug influence, as well as abuse and mental illness in IPV cases (Balvig & Kyvsgaard, 2006; Helweg-Larsen, 2012; Pedersen, et al., 2016; Pedersen, et al., 2017). Several other studies have further found a correlation between alcohol impact and abuse as well as aggressive behavior, including violence against partners and the risk of being exposed to violence (Murphy, et al., 2001; Thompson & Kingree, 2006; Stenager & Stevenson, 2013). Because the police reports do not consistently state whether abuse or alcohol influence is present, the results shown here might be an under representation of the actual numbers. Further, the offender could be arrested several hours or days after the violent event, why it can be difficult for the police to assess whether there has been alcohol or drugs involved. The deviating results could, however, indicate that attention to a changed registration practice regarding abuse and influences may give a more realistic picture of the situations. This may contribute further to the understanding of abuse, and whether treatment thereof could result in a reduced risk of committing IPV.

It is not surprising that IPV is a relatively private matter, which unfolds behind the four walls of the home. The home is additionally characterized by the absence of witnesses as capable guardians, which makes the violations difficult to investigate and clarify (Cohen & Felson, 2014). This is in coherence with Holmberg’s (1999) understanding of the prioritization of this kind of assignments as “bad” within the police. To this comes that weapons are rarely used, which also is demonstrated in other studies, where on average weapons are used in approximately 20% of the cases of IPV (Pedersen, et al., 2017).

Studies have shown that the risk of recidivism and revictimization in relation to IPV offenses are high (Walby & Allen, 2004; Hendricks, et al., 2006). Research indicates that when an offender has begun to commit violence towards their partner, the behavior is difficult to change (ibid). When the results show that the

(17)

told revictimization and recidivism is higher than the reported, it reflects that IPV is affected by a large dark number (Pedersen, et al., 2017; Balvig & Kyvsgaard, 2006; Helweg-Larsen, et al., 2010). The personal relationship between the offender and the victim may also be considered to affect the tendency of filing a report. However, when it is seen that the police in 40% of the cases have been in contact with the couples prior to the violence, it may be considered whether it have been possible for the police to predict the violence. Warning signs are often present before an IPV incident and they often go unnoticed (Hendricks, et al., 2006). The high rate of previous contact and told revictimization could indicate that police officers estimate the contacts too easygoing and thus does not initiate the necessary interventions in order to avoid future violence between the couple. It is a fundamental principle of law that the police cannot make an arrest without sufficient proof that something illegal has been committed. However, in the Police Act, paragraph §1 it is stated, that the police must work for reassurance, security, peace and order in society. The police must promote this purpose through

preventive, helping and enforcement activities (Politiloven, 2015:956, §1). Risk

assessment tools could be used in a preventive perspective, and be used even with other events than violence. In practice, however, as it is explained with

Holmberg’s (1999) theory, there are no tradition for weighing the responsibility of prevention as much as the enforcement obligation (Rasmussen, et al., 2016). Furthermore, with a perception of IPV as a ‘bad’ police assignment, signs of an ongoing normalization process can easily be overlooked. The use of a risk assessment tool could maybe help the police officers to systematically assess the events and help them to not overlook warning signs in high risk cases.

The first cluster identified were closest to the understanding proposed by

Holtzworth-Munroe and Stuart (1994) of a ‘generally violent/antisocial’ offender, as being more often criminal across all situations and environments. However, alcohol and drug impact at the time of the crime is not seen as a factor, which may be a bit surprising in this group as this type in similar typologies is described as being prone to having alcohol and drugs influences (Holtzworth-Munroe & Stuart, 1994). Again the registration practice of the police should be noticed, which has made it difficult to conclude conclusively whether alcohol and drugs have or have not been present at the time of the crime. Holtzworth-Munroe and Stuart (1994) estimate, according to previous studies, that this group contains of approximately 25% of the offender sample, which is consistent with these results (ibid).

In the second cluster, violence is almost only seen between the two parties involved, and it is only in a few cases that violence occurs outside of the reported case. However, this is also the group where most of the serious incidents appears, where the violence has been performed with a subject. This is the only cluster where alcohol influence is seen, and further the oldest offenders are found in this group, which indicates that age plays a role in the performance of the violence in alcohol-influenced state. Even without a large amount of violence outside of the family, this group can be compared to the ‘dysphoric/borderlines’- group by Holtzworth-Munroe and Stuart (1994), where estimates from previous studies indicate that this group, should contain about 25% of the offender samples (ibid). In the third cluster, no violence has been committed prior to the reported case and the couples has no previous contact with the police. This could indicate that the IPV is committed relatively privately and controlled or for the first time. Within the offenders in this group, it could maybe be of importance to be aware of signs from the normalization process, and whether the violence has been controlled and

(18)

is performed in such a way that it is not perceived by the victim and surroundings (Lundgren, et al., 2001). Further, with reference to Holmberg's (1999)

understanding of police work, this group is likely to be the most difficult group for the police to identify and handle, as there are no immediate visible risk factors. This group is similar to Holtzworth-Munroe and Stuart’s (1994), ‘family-only’- group where they suggest that this group may contain up to 50% of the offender samples.

In addition to the original typologies from Holtzworth-Munroe and Stuart (1994), the subgroups found can also be compared to newer adjusted typologies.

Especially the ‘family-only’ and ‘generally-violent’ groups have similarities with ‘antisocial’ and ‘family-only’ offenders found in other contexts. Previous studies showed that the ‘antisocial’ offenders were significantly younger than the ‘family-only’ offenders, which is similar to the results found in this thesis (Petersson, et al., 2016). Further the ‘antisocial’ offender was found to be more likely to recidivate than the ‘family only’ offender, which also can be seen in the reported cases (Petersson & Strand, 2017). The results found indicate that in spite of another set of measures and another context of the sample, a similar pattern and characteristics of the offenders could be found in the reported cases from Funen police. It can thus be argued that it is not only sufficient to prevent within a limited understanding of a homogeneous offender, but that a need for variation in efforts should be taken into account if the full width of the IPV is to be reached. When there is a need to understand that offenders of IPV are a heterogeneous group, it is also because there are different ways in which the groups should be managed and treated (Holtzworth-Munroe & Stuart, 1994; Cunha & Gonçalves, 2013; Grana, et al., 2014). For example, it is in other studies found that violence within the ‘family-only’ group often can be seen as low-level violence (Petersson, et al., 2016), why the victims may not report the incident. This can be contrary to the ‘generally violent’ group, where victims may be more likely to report both threats and violent, as they are familiar with the offender’s capability to commit violence (ibid). In addition, it is also previously illustrated how the majority of those offenders that recidivate do it within the first year, why interventions and protective measures should be initiated as soon as possible (Petersson & Strand, 2017). In the end however, the typologies of the offenders will prove relatively useless if they remain descriptive; their potential usefulness comes from the ability to use them to better understand the causes and functions of the various types of offenders (Holtzworth-Munroe & Stuart, 1994).

Practical implications

First, the results can contribute to an understanding of the reported IPV and thus provide broader insight. When discussing practical implications for the police, risk assessment tools are often highlighted as useful for identifying different risk scenarios; high, low and medium risk groups (Storey, et al., 2014; Douglas, et al., 1999). It is therefore necessary to distinguish between typologies of offenders and risk assessments. The purpose of a risk assessment is to help the police or other authorities to “predict” and assess the likelihood that the violence will continue or escalate (Messing & Thaller, 2014). Typology identification can be used to develop appropriate risk assessment processes, as they can give insight into the nature of future behavior as well as disposition to engage in different types of interventions (Boxall, et al., 2015). Typologies and offender descriptions can thus lead to more accurate identifications, risk assessments and interventions.

(19)

However, it is difficult to navigate between the different types, why detailed information about the offender is necessary. In particular, information about mental health of the offender and information regarding personality disorder has been shown to be of major importance for the most accurate subtyping (Boxall, et al., 2015; Holtzworth-Munroe, et al., 2000) Nevertheless, this kind of information is generally not available to the police, which is also seen in the results. Even if this information was available to the police, it would be unrealistic that the officers would be able to use it as they are rarely clinically trained to identify specific features of psychopathology or specific personality disorders (Boxall, et al., 2015). However, the typologies in this thesis are based on the information available to the police, and the results may therefore be useful as guidance for applications and prioritizing of interventions.

In-depth analyses and descriptions of IPV within a police area can help to form a theoretical understanding of how to work with the phenomenon. This can help to challenge existing thoughts and myths. A disadvantage of not having theoretical knowledge about the dynamics of IPV is that all knowledge comes from practical experience and prevailing attitudes in the working environment. This means that as an officer you may risk working in an environment that is characterized by limited knowledge and ability to deal with IPV in the best way possible. As described, much of the police work depends on estimates and assumptions from the officers which gives them the opportunity to assess what the individual cases requires. In comparison to Lypsky's (2010) theory of Street-Level Bureaucrats where the officers already divide people into categories and stereotypes,

empirically found detailed understandings and description may help to upgrade these estimates and assumptions. Too much dependency on theoretical knowledge can however conversely, also pose some challenges as it can be difficult to

evaluate experiences of IPV that do not immediately fit into the descriptions and typologies. Thus, typologies can amplify myths or stereotypes, which can lead to misidentification and poor management of the cases (Wangmann, 2011). The presented typologies and descriptions can therefore only be used as a benchmark in a police setting, and it should be used solely informatively to support the

experiences, risk assessments and estimates from the police, so it does not become a tool of discrimination. To develop the findings, future research could examine how the different types of offenders respond to different management and interventions in the light of risk factors and recidivism.

Methodological discussion

Police reports as empirical data

In this thesis police reports have been used as the empirical fundament. This way of using information is based on a documentary methodology, which derives knowledge of a phenomena from existing documents (Duedahl & Jacobsen, 2010). The filing police officer’s perception and thus also his understanding of the situation are therefore in the coded material further interpreted. However, the development of the codebook has sought to limit one’s own attitude and influence of the coding.

It is furthermore important to note that the report documents only contain what actually took place and what is relevant in regards to prosecution, hence the quality and quantity of information varies. Furthermore, the officers do not add information on the absence of variable and actions. Since the reports have been

(20)

developed with a different aim than the present research question, a critical approach to the material has been necessary to ensure the quality of the assessment (ibid). Further it has been necessary to question whether it can be concluded that a variable (e.g. alcohol influence) has not been present if it is not described in the report. The scientific principles where absence of evidence is not evidence of absence have been followed, and an estimate has been made in each case. In the reports and registers there could be indications of the opposite. In cases of divergent or contradictory testimonies, the victim's statement has been the one taken into account. This is because it is assumed that the victim have the least incentive to make false statements (Heinskou & Liebst, 2017). In relation to the discussed, the use of police reports can not only be considered as a reading of information but also to some extent an analytical interpretation. A standardized way to conduct information by the officers would qualify the police to more systematically compare and assess cases. Further this could be advantageously when performing risk assessments, where some exact information is needed.

Reliability and validity

The development of a very detailed codebook has largely contributed to ensuring the validity and reliability of the coded variables. However, it is obvious that among the more objective variables (gender, age, etc.) there is a greater reliability than among the more subjective variables (motives, previously felt threatened, etc.). In addition, there is the mentioned challenge that the police reports are written with a legal intention. The reports give an insight into the actual event of the violence, but they only provide limited insight into social backgrounds for the involved offender and victim. Furthermore, when the reports are quantified, it is likely that the transformation into numbers will involve some loss in nuances. However, it should be noted that in order to determine and describe the actual incident, the police obtain as many information’s and sources as possible. Therefore, images of injuries, witnesses, hospital reports and likewise are taken into account when examining what has happened. In this thesis, all of such materials, both subjective and objective information from the reports are evaluated, as to address all aspects of the case.

Although the 400-sample represents a statistically representative sample of all reported cases of violence in 2017, it does not constitute a representative sample of all IPV episodes committed. Again, it is a well-known fact, that not all cases of violence, and especially not those in close relationships, come to the knowledge of the police (Pedersen, et al., 2017; Balvig & Kyvsgaard, 2006; Helweg-Larsen, et al., 2010). However, as shown several key findings reflect trends found in other studies, which can be considered a mitigating circumstance in relation to the representativeness of the thesis.

Limitations

Because of the time and word limit for the thesis, the study was conducted only on a small size of the reported cases. Furthermore, the sample has only been

conducted for the year of 2017. Therefore, to generalize the results the study should have used more cases and throughout more years. The thesis is solely able to give a snapshot of how the situation looks like. In addition, although the sample has been treated on the basis of gender neutrality, there have been mostly male offenders and female victims. This also means that Holtzworth-Munroe and Stuart's (1994) typology development has been appropriate to use, as it is based on these measures. If the sample had shown a larger proportion of female offenders,

(21)

a different typology comparison would have been necessary. In subsequent research, the proposed typology and descriptions needs to be validated. It should further be noted, that the results are based on cases and not unique offenders and victims. As stated there were 54 unique offenders and victims in the sample, which means that 6 couples are represented twice. When this approach is chosen, it is because the aim of the thesis was to describe the reported cases and the characteristics that could be found within them. Therefore, it is relevant to include all offenders from the cases, and even though there are duplicates, the

characteristics found, express what the police show up to when they have a case of IPV. Furthermore, some of the characteristics for the offenders differ in the different cases. For example, age has changed for some of the duplicates, given that they have had a birthday in between their IPV offences. If only unique offenders were used, then it would be necessary to cherry pick which case to use, which are not found to be good research practice. However, it is important to notice when this approach is used, then the frequency of the variables could also be skewed.

Police registration practices are complicated and as it is found intimate partner events is registered on various criminal codes. The restriction of which criminal codes to include in this thesis has therefore presented some limitations. In future studies an inclusion of more offences like i.e. sexual offences within intimate partners, would give a more precisely picture of the offender and the criminal characteristic.

The used methods of cluster analysis also have some limitations. In relation to its use of age parameters. The mean age is calculated for the clusters, which are based on the frequency of the measurements. But when the quartiles are 25 and 40 (as they are for cluster 2 e.g. - see appendix B) then the average age for the cluster become 30. This is obviously not representative, but statistically correct. Further, when using a complete case analysis, the results can only be found valid, when the cases with complete data have the same characteristics as the cases with missing data (Verma, 2013). In order to ensure the validity and capacity of

typologies found, it is therefore necessary to determine whether the typologies can cover the nuances of the overall picture from the cases. By examining the dropout there was a pattern that the majority of the drop-off could be explained by missing information of alcohol influence. However, it was found, that the remaining characteristics for the group did not differ from the found typologies, and

therefore the group of missing cases did not represent a type that could not fit into the developed typologies.

Finally, it is important to clarify that the offenders in this study are only suspects rather than sentenced since not all of the cases has been finalized.

CONCLUSION

The present thesis is based on 60 reported IPV cases identified through a sample of 389 reported violent cases within the police of Funen. The sample were

collected in 2017, and the findings are thus merely a snapshot of how the situation looked during this period. Similarly, the thesis is based solely on reported cases in the police, why it is only the committed IPV as it appears to the police that are

(22)

examined. Overall, it can be concluded that the offenders of the reported partner violent cases in Funen's police are in accordance with studies of offenders of IPV in other contexts. Through the description of the most prominent characteristics, it is shown that the most common offenders are men who commit violence against women. The offenders are a bit older than the general offender, which can be explained in the light of the type of violence. In regards to the use of alcohol and drugs, there is a lower appearance of these characteristics, than it is seen in other studies. The criminal act is also characterized by a high level of recidivism, and furthermore, in more than a third of the cases there has been contact between the police and the couple prior to the actual case. These findings indicate that an increased focus on events between partners could be beneficial, rather than focusing exclusively on violence between partners.

An offender typology was developed using a two-step clustering method which resulted in three different offender types; generally violent, alcohol-related violent and family violent. The first type is described as being the most criminal group. The second group is described as being alcohol influenced when committing the crime, and with only a few cases of similarly violence outside the family. The third and largest group is characterized by not having a history of similarly crime outside the relationship. The results can be used to give an insight into how IPV occurs now and here, thus forming a basis to understand and develop a

prioritization of the phenomenon. However, it is likely that descriptions and typologies may change over time, especially when there are changes in the management practices internally within the police, why further and deeper investigations will be necessary.

(23)

REFERENCES

Arnocky, S., 2014. Sex Differences in Response to Victimization by an Intimate Partner: More Stigmatization and Less Help-Seeking among Males. Journal of

Aggression, Maltreatment & Trauma, Vol 23(7), pp. 705-724.

Balvig, F. & Kyvsgaard, B., 2006. Vold og overgreb mod kvinder., Copenhagen: Justitsministeriets Forskningsenhed.

Belfrage, H. & Strand, S., 2012. Measuring the outcome of structured spousal violence risk assessments using the B-SAFER: Risk in relation to recidivism and intervention. Behavioral Sciences & the Law, Vol 30(4) pp. 420-430

Belfrage, H. et al., 2011. Assessment and management of risk for intimate partner violence by police officers using the spousal assault risk assessment guide. Law

and human behavior, Vol 1(36), pp. 60-67.

Boxall, H., Rosevear, L. & Payne, J., 2015. Domestic violence typologies: What value to practice?. Trend and Issues in Crime and Criminal Justice. Vol 1(494) pp. 1-9

Boyle, D. J., O’Leary, K. D., Rosenbaum, A., & Hasset-Walker, C., 2008. Differentiating between generally and partner-only violent subgroups: Lifetime antisocial behavior, family of origin violence, and impulsivity. Journal of Family

violence, Vol 23(1), pp. 47-55

Cunha, O. & Gonçalves, R. A., 2013. Intimate partner violence offenders:

Generating a data-based typology of batterers and implications for treatment. The

European Journal of Psychology Applied to Legal Context, Vol 5(2) pp. 131-139.

De Vaus, D., 2014. Surveys in Social Research. London: Routledge.

Douglas, K. S., Cox, D. N. & Webster, C. D., 1999. Violence risk assessment: Science and practice. Legal and Criminological Psychology, Vol 4(1), pp. 149-184.

Duedahl, P. & Jacobsen, M. H., 2010. Introduktion til dokumentanalyse -

Metodeserie for social- og sundhedsvidenskaberne. Odense: Syddansk

Universitetsforlag.

Edleson, J. L., 1999. Children’s witnessing of adult domestic violence. Journal of Interpersonal Violence. Vol. 14(8), pp. 839-870.

Felitti, V. J. et al., 1998. Relationship of childhood abuse and household dysfunction to many oft he leading causes of death in adults. The Adverse Childhood Experiences (ACE) Study. American journal of Preventive Medicine. Vol 14(4), pp. 245-258.

(24)

Garner, J., Fagan, J. & Maxwell, C., 1995. Published findings from the spouse assault replication program: A critical review. Journal of Quantitative

Criminology, Vol 11(1), pp. 3-28.

Gelles, R. J., 2016. Intimate violence and abuse in families: Oxford University Press.

Goldenstein, D. A., Cantos, A. L., Brenner, L., Brenner, L. H., Verborg, R. J., & Kosson, D. S., 2015. Perpetrator type moderates the relationship between severity of intimate partner violence and recidivism. Criminal Justice and Behavior, Vol 43(7), pp. 879-898.

Grana, J. L., Redondo, N., Munoz-Rivas, M. J. & Cantos, A. L., 2014. Subtypes of Batterers in Treatment: Empirical Support for a Distinction between Type I, Type II and Type III. PLOS ONE, Vol 9(10), pp. 1-8.

Heinskou, M. B. & Liebst, L. S., 2017. Gadevold: en sociologisk kortlægning af

vold i byen. Jurist- og Økonomforbundet.

Helweg-Larsen, K., 2012. Vold i nære relationer - omfang, karakter, udvikling og

indsats i Danmark, Copenhagen: Statens Institut for Folkesundhed.

Helweg-Larsen, K., Kruse, M., Sørensen, J. & Brønnum-Hansen, H., 2010.

Voldens pris - samfundsmæssige omkostninger ved vold mod kvinder.,

Copenhagen: Syddansk Universitet & Rockwool Fondens Forskningsenhed.

Hendricks, B., Werner, T., Shipway, L. & Turinetti, G. J., 2006. Recidivism among spousal abusers: Predictions and program evaluation. Journal of

Interpersonal Violence, Vol 21(6), pp. 703-716.

Holmberg, L., 2011. ”Må jeg overvære Deres anholdelse?” - Om politiforskning,

deltagerobservation og informeret samtykke. Copenhagen: Det juridiske Fakultet

ved Københavns Universitet.

Holmberg, L., 1999. Politiets skøn i retssociologisk belysning, Copenhagen: Det juridiske Fakultet ved Københavns Universitet.

Holtzworth-Munroe, A. et al., 2000. Testing the Holtzworth-Munroe and Stuart (1994) batterer typology. Journal of consulting and clinical psychology, Vol 68(6), pp. 1000-1019.

Holtzworth-Munroe , A. & Stuart, G. L., 1994. Typologies of Male Batterers: Three Subtypes and the Differences Among Them. Psychological Bulletin, Vol. 116(3), pp. 476-497.

Hussain, M. A. & Lauridsen, J. T., 2017. Videregående kvantitative metoder, Metoder i Samfundsvidenskab og Humaniora. Samfundslitteratur.

Huss, M. T. & Ralston, A., 2008. DO BATTERER SUBTYPES ACTUALLY MATTER? Treatment Completion, Treatment Response, and Recidivism Across a Batterer Typology. Criminal Justice and Behavior, Vol 35(6), pp. 710-724.

Intimate partner violence in a police setting. Offender characteristics and typologies found within reported cases from Funen Police

INTIMATE PARTNER VIOLENCE

IN A POLICE SETTING

- Offender characteristics and typologies found

within reported cases from Funen police

CHRISTINA B. JØRGENSEN

INTIMATE PARTNER VIOLENCE

IN A POLICE SETTING

- Offender characteristics and typologies found

within reported cases from Funen police

CHRISTINA B. JØRGENSEN

TABLE OF CONTENTS

INTRODUCTION

BACKGROUND

MATERIALS AND METHODS

RESULTS

DISCUSSION

CONCLUSION

REFERENCES