The aim of this study was to characterize young dating violent offenders (DVO), and to compare them to the general population and to young offenders with violent crimes directed against other victims. We have used data from the Development of Aggressive Antisocial Behaviour Study, in all 262 young men, 18 to 25 years, convicted of violent crimes and imprisoned in the Western Region of the Swedish Prison and Probation Services. We found that young DVO offenders differed from the general population in all investigated areas; however, the group did not differ in comparisons to other young violent offenders. Our results highlight the antisocial aspects of dating violent crime being rooted in aggressive antisocial behaviour, lacking signs of any specific offender type characteristics, thus questioning the validity of crime specific treatment programs in prison for young offenders of dating violence.
El presente estudio ha tenido por objetivo la caracterización de los jóvenes delincuentes violentos en sus relaciones sentimentales (“DVO” en su acrónimo inglés) y su comparación tanto con el conjunto de la población como con otros delincuentes violentos jóvenes que han cometido delitos con otro tipo de víctimas. Hemos empleado datos del estudio sobre desarrollo de conducta antisocial violenta con 262 jóvenes varones de entre 18 y 25 años de edad condenados por delitos violentos y encarcelados en la Región Oeste del Servicio de Prisiones y Libertad Vigilada de Suecia y hallado que los jóvenes delincuentes de tipo DVO se diferenciaban del resto de la población en todas las áreas objeto de estudio. Sin embargo, no mostraban discrepancias con respecto a otros delincuentes violentos jóvenes. Nuestros resul-tados destacan que los aspectos antisociales de los delitos violentos en las relaciones sentimentales tienen su base en una conducta antisocial violenta, sin indicios de vinculación con un tipo específico de delincuentes, lo cual cuestiona la validez de los programas penitenciarios para el tratamiento de delitos específicos orientados a delincuentes jóvenes condenados por violencia en las relaciones sentimentales.
Violence, regrettably, is a matter also concerning the young. World Health Organisation (WHO) states that youth homicide victims (people aged 10-20) make up 43% of the total number of homicide victims globally (WHO, 2015). Being young is also a risk factor for becoming a violent crime offender, while young women between the ages of 16 to 24 are most likely victims of partner violence (Rennison & Welchans, 2000). Partner violence in adolescence and early adulthood, commonly referred to as dating violence, courtship violence, or adolescent relationship abuse, is generally defined as physical, sexual, and/or psychological violence between dating partners in the ages of 15 (or younger) up until 24. It is regularly discussed as a prevalent and serious social and health problem (Centers for Disease Control and Prevention, 2016). Thus, identifying offender characteristics in dating violence is considered an urgent quest.
However, definitions of the concept vary: some include sexual violence (Dardis, Dixon, Edwards, & Turchik, 2015; Silverman, Raj, Mucci, & Hathaway, 2001) and some do not (Lewis & Fremouw, 2001). Investigated age groups vary as well, as some include, or study exclusively, teenagers (Fernández-González, Wekerle, & Goldstein, 2012), adolescents (Giordano, Copp, Longmore, & Manning, 2016) and/or young adults (Gómez, 2010). Overall, the span of investigated ages in studies of dating violence varies from 12 years to 25 years. Dating violence victimization rates found in the United States, where incidences are reported for the last 12 months, have been found ranging from 1.6%, including physical and sexual violence among 12-17 years old, (Wolitzky-Taylor et al., 2008) to 20.9% among girls and 10.4% among boys, in a nationally representative survey in grades 9-12 (Vagi, Olsen, Basile, & Vivolo-Kantor, 2015). Despite any exact estimates of the prevalence of dating violence, contributing also to the severity of the problem is the fact that it afflicts young people. As was suggested in the earliest work in the field (Makepeace, 1981), this fact entails certain consequences: young age means that the process of socialization is ongoing and, in cases where the pattern of abuse is established already in adolescence, the severity of adult intimate partner violence is often greater. This is found to be true for both perpetration and victimization. In a conditional probability analysis by O’Leary et al. (1989) it was found in couples where violence had occurred during the dating period and after 18 months of marriage that the likelihood of new violent acts at 30 months was .72 for women and .59 for men. Smith, White, and Holland (2003) have presented results from a longitudinal study of women in university, showing an increased risk of revictimization. Among the young women who had experienced prior physical abuse in a romantic relationship in adolescence, the relative risk for revictimization was 2.96 during their first year of university studies. The negative health consequences of victimization of partner violent abuse have been investigated in a number of studies (Campbell, 2002; Plichta, 2004; Roberts, Klein, & Fisher, 2003) and associated with numerous conditions, such as substance abuse, diet related behaviours, and overuse of health and social services.
Regarding the question of aetiology of dating violence, the cycle of violence theory (Widom, 1989) has been in the centre of attention of several studies (Gómez, 2010; Whitfield, Anda, Dube, & Felitti, 2003; Wood & Sommers, 2011). The cycle of violence, or intergenerational violence, is rooted in the social learning theory and proposes that growing up being witness or subjected to violent behaviour teaches the child acceptance for violent behaviour (Bandura, 1977).
In the vicinity of the question of aetiology, you find the search of potential risk factors or correlates. However, although risk factors do not carry any one-directed argument of causality, they bring important knowledge of occurrence of dating violence. In a review of longitudinal studies, Vagi et al. (2013) presented that risk factors for dating violence clustered together in the key dimensions: mental health problems, attitudes, behaviours such as the use of aggressive media, aggressive behaviours towards peers or others, substance abuse, precocious sexual behaviour, and having antisocial peers. The review concluded that effective prevention programs ought to focus on youths who had experienced maltreatment and other adverse childhood events, had particular mental health problems, behaved aggressively and had aggressive attitudes, used substances, and were in hostile or unhealthy relationships. Capaldi and Owen (2001) investigated associations between physical aggression, injury, and fear in an at-risk community sample, and showed that physical aggression among partners was associated to antisocial behaviour. The only study examining a total birth cohort (Magdol, Moffit, Caspi, & Silva, 1998) reported that the most consistent predictor of intimate partner abuse at age 21 was the presence of early problem behaviours, such as aggressive delinquency and substance abuse at the age of 15. Also, concerning the victimization of dating violence, several studies have reported that the victims of dating violence are found among groups of youth exhibiting social and health risk behaviours (Nilses, Blom, Heimer, & Danielsson, 2011; Roberts & Klein, 2003; Silverman et al., 2001), such as non-violent and violent delinquency, increased levels of substance abuse, and sexual risk behaviours. It has been suggested, in accordance with the theory of gender symmetry (Straus, 2006), that the roles of victim and offender oscillate in violent relationships (Muftić, Finn, & Marsh, 2012).
These facts indicate that both offenders and victims of dating violence display social vulnerability, potentially in need of extra societal support, and attention. For example, experience of maltreatment and other adverse childhood events, early onset disruptive behaviours, mental health problems, aggressive behaviour, and substances abuse disorders are all factors presenting an elevated risk of being convicted of crimes when entered into adulthood (Af Klinteberg, Almquist, Beijer, & Rydelius, 2011; Shukla & Wiesner, 2016; Reavis, Looman, Franco, & Rojas, 2013).
However, studies aiming at exploring risk and protective factors of dating violence have almost exclusively used samples derived from high school, college, or community populations. The large amount of general risk behaviours neighbouring dating violence point to the need of inquiry into samples of delinquent adolescents and socially marginalized and/or criminal young adults. The age group of 18-25 years is of particular interest since it is the range where, in most Western cultures, the transition into adulthood is made. We consider an examination of imprisoned dating violent offenders of this age group to be a relevant task partly because it has not been done before, partly because an assumption that the offenders of the most severe violence towards dating partners are the ones convicted to prison. Knowledge regarding imprisoned dating violent offenders could thus add to the knowledge of what characterizes individuals who are at the high end of the spectrum of dating violence.
Based on previous research and theory of aetiology, potentially important areas of investigation could be: occurrence of adverse childhood experiences (e.g., experiences of neglect and abuse in childhood, household social dysfunction, and levels of adaptation to school and work), clinical variables (such as major mental disorders (MMD), personality disorders, substance use disorders, and neurodevelopmental disorders), and overall aggressive antisocial behaviours. A likewise important area to investigate would be the importance of type of offender victim relationship. Offender victim relationship is the denominator for the entire discourse of both dating violence and intimate partner violence (IPV). Research has shown that offender victim relationship, not necessarily being an intimate or dating one, can be a precipitating and aggravating factor of violent crime: the closer the relationship, the more severe the victim injury (Heller, Ehrlich, & Lester, 1983). It has also been found that lethal violence (i.e., homicides) where the offender knew the victim more often is associated with expressive violence, while homicides committed of unknown victims more often are associated with instrumental motives, as such homicides occur in conjunction with other crimes (e.g., robbery, breaking an entry) (Drawdy & Myers, 2004). In spite of several findings of similarities between adult partner violent offenders and violent offenders of other crimes, there have not been any comparisons of offender characteristics among young violent offenders based on victim relationship as the distinguishing factor (e.g., dating violence versus violence against acquaintances).
The aim of the present study was 1) to provide a characterization regarding psychosocial background factors of imprisoned young Swedish male offenders convicted of dating violent crimes towards a partner in comparison to the general Swedish population, and 2) to compare them to imprisoned offender groups convicted of violent crimes with other victim-offender relationships, that is, offenders with a known victim, and offenders with an unknown victim, regarding i) clinical features and personality traits and ii) aggressive antisocial behaviours and psychopathic traits.
MethodThe Development of Aggressive Antisocial Behaviour StudyThe Development of Aggressive Antisocial Behaviour Study (DAABS) was led and coordinated by the universities of Lund and Gothenburg. Its primary aim was to study the prevalence of developmental and clinical disorders in a nationally representative cohort of young adult male violent offenders in Swedish prisons. The study began in March 2010 with the recruitment of inmates from nine prisons in the Western Region of the Swedish Prison and Probation Services (SPPS), representing one fifth of the whole SPPS. In all, 270 male inmates aged 18 to 25 and convicted of violent crimes, including hands-on sexual crimes, serving time in these prisons were recruited for the study. Other inclusion criteria were language skills, speaking and understanding Swedish sufficiently well to participate in clinical assessments, and duration of stay, that is, at least four weeks in order for the assessments to be planned and executed. There were no differences between those who consented and those who declined participation regarding median age, previous convictions, or country of origin.
Present Study PopulationFrom the DAABS cohort, 262 offenders were retrieved into the present study. In five cases information concerning relation to victim was missing or not known. There were also three perpetrators of sexual violent crimes towards victims younger than 12 years of age that were excluded from the present study because of the potential paedophilic element present, leaving a total of N=262 accessible for this study. Moreover, complete data collection was not possible for all participants, leading to some variations in the number of subjects for each measure (see Wallinius et al., 2016, for a more detailed description of the procedures and measures used in the DAABS).
The study population was divided into three groups according to the offenders’ relation to the victim of their violent crime. This division is pragmatic and in line with a common praxis of classification of offenders in Swedish prisons. For example, when deciding appropriate program for the offender in question, committing a violent crime towards a partner is the necessary prerequisite to take part in programs especially designed to understand and change such behaviours. The dating violent offender group (DVO) consisted of 42 offenders convicted of a violent crime towards a partner (always a female), including seven who had committed violent sexual offences towards their partner. In this group, some offenders had also committed other minor offences, such as violations of the legislation against carrying arms/knives in public places, drug offences, and drunk driving. The two other groups were constituted depending on whether the offenders had committed a violent crime towards (an) unknown victim(s) (unknown victim offender, UVO) or if they had prior knowledge of their victim(s), (known victim offender, KVO). The UVO group consisted of 135 offenders, of which 12 had female victims, 97 had male victims, and in 26 cases there were both male and female victims. The KVO group consisted of 85 offenders, of which 62 had male victims, 17 had female victims, and 6 had both male and female victims. In the UVO group, three offenders (2%) were convicted of sexual crimes, and in the KVO group the same number was 17 (20%). The same criteria were applicable for these groups as for the DVO group: they might have been convicted of lesser criminal offences as well, such as the above mentioned; however, the main offence was a violent crime directed towards at least one victim, known or unknown.
Comparisons to the general Swedish population regarding psychosocial background factors were made through results from a study based on the total Swedish population born in 1987 to 1992 by the Centre for Health Equity Studies (CHESS). CHESS was initiated by the Swedish government and is a network consisting of researchers from Karolinska Institute, Stockholm University and the Swedish Council for Social Research (SFR) for research with health inequalities as a key factor.
Regarding academic achievements, results were compared to data from The Swedish National Agency for Education (2010). The study population of dating violent offenders was born between 1985 and 1992, reaching the age of graduation for secondary school between the years 2004 and 2011. The year of comparison between the study population and the general Swedish population was chosen to be the year in the middle of this time span, that is, the graduation of secondary school for the year of 2008.
Data for levels of unemployment were found in the site http://www.ekonomifakta.se/. This is a site that presents economic information collected from public Swedish and international statistic sources, primarily from Statistics Sweden, OECD, and Eurostat.
MeasuresA broad battery of instruments, both expert-rated by licensed psychologists and self-rated, was used in the DAABS project, of which the following domains were covered:
Psychosocial background factorsThe DAABS structured interview covered a broad range of psychosocial background factors, ranging from social circumstances during childhood to previous criminality. Among these variables the following risk-related background factors were included in the present study: if offenders had witnessed repeated violence against a parent; if offenders had completed compulsory and/or secondary school, if they had experienced bullying during school years.
Susceptibility factors, such as having been in need of special support or care, were also included, as well as if offenders had been taken into care at a youth institution; if they had received psychiatric care during childhood or adolescence; and, finally, whether or not they were previously convicted (any criminality), and whether or not they had had a full-time job prior to being incarcerated for the current crime.
In order to measure experience of abuse and neglect during childhood, the Childhood Trauma Questionnaire-Short Form (CTQ-SF, Bernstein et al., 2003) was used. The CTQ-SF is a self-report questionnaire comprising 28 questions, measuring five clinical subscales: Emotional Abuse (verbal abuse or verbal violations directed towards a child, in order to violate or humiliate its self-esteem), Physical Abuse (physical attacks from adults towards the child, aiming to hurt), Sexual Abuse (sexual contact or behaviour between adult and child), Emotional Neglect (omission to provide to basic emotional and psychological needs of a child), and Physical Neglect (omission to provide to basic physiological needs of a child, such as providing food, clothes, etc.). Each subscale consists of five questions, rated on a 5-point Likert scale ranging from never true to very often true. There are also three additional questions forming a subscale of Minimization/Denial, measuring the possible denial and underreporting of maltreatment. In addition, there are threshold scores for each subscale for exposure: “none to low”, “low to moderate”, “moderate to severe”, and “severe to extreme”, respectively. In interpreting our findings, we used the threshold values identifying “moderate to severe” exposure. Raw scores and the corresponding categorization of severity of the scales based on the reference group “male university students” were chosen (Gerdner & Allgulander, 2009). Research has shown good support for the reliability and validity of the CTQ-SF (Bernstein et al., 2003).
Clinical assessmentsStructured Clinical Interviews for the DSM-IV disorders (SCID), based on the DSM-IV diagnostic criteria for axis I and II disorders, were used for the diagnostic assessments (First, Gibbon, Spitzer, Williams, & Benjamin, 1997; First, Spitzer, Gibbon, & Williams, 1996). The SCID was used but combined with a DSM-IV checklist (i.e., the diagnostic criteria from the manual) for those categories not covered by the SCID-I. Participants’ answers were collected in connection with the SCID interview, and diagnoses were established in consensus between the investigating psychologist and one of the responsible researchers. Axis II disorders are presented in clusters: A (consisting of paranoid, schizoid, or schizotypal personality disorder); B (antisocial, narcissistic, histrionic, or borderline personality disorder); and C (avoidant, dependent, or obsessive-compulsive personality disorder). Assessments for diagnosing neurodevelopmental disorders, such as Attention Deficit Hyperactivity Disorder (ADHD), were made by the use of complementary interviews based on DSM-IV criteria, and of autism spectrum disorder (ASD), by the use of the Asperger Syndrome/high-functioning autism Diagnostic Interview/ASDI (Gillberg, Gillberg, Råstam, & Wentz, 2001).
Aggressive antisocial behaviours and psychopathic personality traitsThe Aggression Questionnaire, originally created by Buss and Perry (1992), was adapted to Swedish (the Aggression Questionnaire–Revised Swedish Version, AQ-RSV) by Prochazka and Ågren (2001), and found to be comparable to the original US version showing considerable internal consistency. AQ-RSV is a 29-item questionnaire in which the individual assesses his expression of aggressiveness as Physical Aggression (nine items), Verbal Aggression (five items), Anger (seven items), and Hostility (eight items). Each item is answered on a 4-point Likert scale ranging from 1 (least characteristic) to 4 (most characteristic).
Psychopathic personality traits were studied by the Psychopathy Checklist-Revised (PCL-R) two factor, four facet model (Hare, 2003). PCL-R is a 20-item rating scale, where each item is scored on a three-point scale from 0 (does not apply) to 1 (may apply or in some respects applies) to 2 (does apply). Items are summed up into a total PCL-R score (range 0 to 40), and four facet scores. The two-factor model identifies the underlying psychological construct, consisting of an affective/interpersonal factor (Interpersonal facet and Affective facet) and an impulsive/antisocial lifestyle factor (Lifestyle facet and Antisocial facet). PCL-R is an internationally frequently used instrument for measuring psychopathic traits that it is well validated against different prison populations (Folino, 2015; Salekin, Rogers, & Sewell, 1996; Salvador, Arce, Rodríguez-Díaz, & Seijo 2017).
EthicsThis study was approved by the Research Ethics Committee at Lund University as a multicentre study, thereby involving several sites or authorities. As a project within the correctional system it also had to follow the research guidelines of the Swedish Prison and Probation Services. All participants received both oral and written information about the study, stressing that participation was voluntarily and that it in no way would affect their penalty. Everyone that accepted participation in the study had to give a written informed consent before participating in the actual assessments. Participation included receiving a small monetary reward of 200 SEK (equal to approximately 25 USD). These procedures were fully in line with the Swedish legal requirements for this kind of research.
Statistical AnalysisThe number of offenders varies due to missing data in some of the studied variables. In results, valid percentages are used.
The level of significance was set to p ≤ .05. In order to investigate group differences of categorical variables, Fischer's exact test was used throughout all analyses, and the effect size was calculated with Cramer's V. Standardized residual equal to or above±1.96 was used as a measure to decide which cell contributed to the effect when the Fischer's test showed a significant association. Due to the large differences in group size, statistics were checked in order to determine the distribution of normality with regard to continuous data. Thus, in cases where Levene's test showed a result less than the significance level p ≤ .05, further assessments of normality were conducted. Those included checking for skewness and kurtosis, and possible outliers were recognized by conducting box plots. Group differences regarding continuous data were investigated by the use of one way analysis of variance (ANOVA), followed by a post hoc Bonferroni test to establish significant group differences. In addition, strength of association was provided by calculations of η2.
In addition, one sample non-parametric test (i.e., chi-square) was used to test if the frequencies regarding some of the psychosocial background factors found in our dating violent offender group differed from frequencies found in the general population.
Cramer's V was interpreted according to the following recommendation: .10=small effect, .30=medium, and .50=large effect, and η2 was interpreted in accordance with the subsequent recommendation: .01=small effect, .06=medium effect, and .14=large effect (Cohen, 1988).
ResultsCharacterization of Young Male Offenders of Dating ViolenceStatistics on parents’ abuse and frequency of institutional care of the offender during childhood for the DVO group are presented and compared to the general population in Table 1. Testing the distributions found in the DVO group against the values of the general population showed that the former were significantly different from the latter.
Comparison of Frequencies of Parents’ Alcohol and Substance Use, and of Institutional Care of the Offender before the Age of 13 and during the Age of 13 to 19.
| DVO, born 1985-1992 (n=41) | General population, born 1987-1992a (n=655,722) | |
|---|---|---|
| Parents’ abuse | ||
| Alcohol (%) | 46.3* | 2.5 |
| Substances (%) | 17.1* | 1.5 |
| Institutionalization during childhood | (n=24) | |
| Before the age of 13 (%) | 33.3* | 1.5 |
| Between the age of 13 to 19 (%) | 66.7* | 1.9 |
Note. DVO=Dating Violent Offenders. a Data from CHESS.
Comparisons between the DVO group and the Swedish general population showed that the two groups were quite similar regarding primary school achievement (at 16 years of age). As for graduation from secondary school and establishment in the job-market, the differences between the groups grew remarkably (Table 2). The distribution found in the DVO group did, however, significantly differ from the values of the general population for all studied variables.
Dating Violent Offenders in Comparison to the General Population Regarding Academic Achievements and Establishment in the Job Market.
| DVO (n=42) | General population (collected from different sources) | |
|---|---|---|
| Completed primary school at expected age (%) | 76.2* | 88.0a |
| Completed secondary school at expected age (%) | 21.4* (n=40) | 68.9b |
| Reduced course of studies (%) | 40.0* | 1.0c |
| Unemployed prior to current incarceration versus rate of youth unemployment for 2010 (%) | 47.6* | 24.8d |
Note. DVO=Dating Violent Offenders;
In comparison to cut-offs established for male university students (n=81) (Gerdner & Allgulander, 2009), the DVO group reached, or exceeded, the cut-off values for the moderate to severe range of all CTQ-SF scales except for Sexual Abuse (Table 3).
Distribution of Mean (SD) for the Childhood Trauma Questionnaire-Short Form for Dating Violent Offenders Compared to the Reference Group of Young Male University Students.
| DVO | Moderate to severe cut-off | |
|---|---|---|
| Emotional Abuse M (SD) | 10.8 (± 6) (n=31) | > 10-11 |
| Physical Abuse M (SD) | 10.2 (± 5.7) (n=33) | > 7 |
| Sexual Abuse M (SD) | 5.7 (± 3.5) (n=33) | > 7 |
| Emotional Neglect M (SD) | 16.7 (± 5.5) (n=32) | > 13-15 |
| Physical Neglect M (SD) | 11.9 (± 2.6) (n=29) | > 9-10 |
Note. DVO=Dating Violent Offenders
Comparisons of clinical variables/mental illness revealed overall a high prevalence of mental disorders, but only few statistically significant differences between the groups. The differences were found between the DVO and KVO groups on one side, and the UVO group on the other (Table 4).
Three violent offender groups and DSM-IV diagnoses of mental disorders
| DVO (n=42) | UVO (n=135) | KVO (n=85) | p-value | Cramer's V | |
|---|---|---|---|---|---|
| Axis 1 Clinical disorders | |||||
| Psychotic syndromes n (%) | 0 (0%) | 1 (0.7%) (n=134) | 0 (0%) | 1.0 | .06 |
| Mood disorders n (%) | 27 (64.3%) | 73 (55.0%) (n=133) | 40 (47.1%) | .18 | .12 |
| Anxiety disorders n (%) | 25 (59.5%) | 69 (52.0%) (n=133) | 37 (43.5%) | .20 | .11 |
| Impulse control disorders n (%) | 9 (21.4%) | 25 (19.0%) (n=132) | 20 (23.5%) | .71 | .05 |
| Substance use disorders | |||||
| Alcohol n (%) | 26 (62.0%) | 62 (46%) | 38 (44.7%) | .15 | .12 |
| Cannabis n (%) | 28 (68.3%) (n=41) | 112 (83.0%)a | 61 (72.0%) | .05 | .15 |
| Sedatives n (%) | 16 (40.0%) (n=40) | 77 (57.0%)a | 35 (41.2%) | .03 | .16 |
| Stimulants n (%) | 16(40%) (n=40) | 80 (59.3%)a | 30(35.3%) | .00 | .23 |
| Neurodevelopmental disorders | |||||
| ADHD during childhood n (%) | 25 (61.0%) (n=41) | 88 (66.2%) (n=133) | 50 (58.8%) | .45 | .07 |
| ADHD as adult n (%) | 16 (39.0%) (n=41) | 61(46.0%) (n=133) | 33 (38.8%) | .54 | .07 |
| ASD n (%) | 0 (0%)a | 17 (12.6%) (n=134) | 8 (9.3%) | .03 | .15 |
| Personality disorders | |||||
| Cluster A n (%) | 5 (12.0%) | 12 (9.0%) (n=133) | 8 (9.4%) | .81 | .03 |
| Cluster B n (%) | 24 (57.0%) | 91 (68.4%) (n=133) | 50 (58.8%) | .22 | .11 |
| Cluster C n (%) | 2 (4.8%) | 5 (3.8%) (n=133) | 1 (1.2%) | .33 | .08 |
Note: † group significantly different identified by standardized residual equal to or larger than -1.96 or 1.96. DVO=Dating Violent Offender, UVO= Unknown Victim Offender, KVO= Known Victim Offender
Regarding measures of aggressive antisocial behaviours and psychopathic traits (i.e., AQ-RSV, and PCL-R) there were no significant differences between the three groups (Table 5).
Psychopathic traits and aggressive antisocial behaviors; Distributions, Mean, SD, f-value, p-value and Effect Size for the Three Groups, DVO, UVO, KVO.
| DVO | UVO | KVO | f-value | p-value | η2 | |
|---|---|---|---|---|---|---|
| AQ-RSV | ||||||
| Anger M(±SD) | 21.8 (7.4) (n=34) | 23.0 (6.1) (n=101) | 21.0 (7.0) (n=69) | 1.6 | .20 | .01 |
| Physical Aggression M(±SD) | 34(9.1) (n=32) | 34.5 (7.5) (n=101) | 32.7(8.3) (n=70) | .9 | .37 | .01 |
| Hostility M(±SD) | 24.5 (8.2) (n=32) | 22.8 (7.0) (n=102) | 23.2 (6.9) (n=69) | .7 | .48 | .00 |
| Verbal aggression M(±SD) | 17.1 (3.4) (n=31) | 16.6 (3.7) (n=100) | 16.1 (3.2) (n=69) | .9 | .38 | .01 |
| Total score M(±SD) | 97.2 (21.2) (n=31) | 97.0 (19.9) (n=99) | 93.0 (20.8) (n=69) | .8 | .41 | .00 |
| PCL-R | (n=40) | (n=130) | (n=85) | |||
| Interpersonal M(±SD) | 1.3 (1.7) | 0.8 (1.2) | 1.0(1.4) | 1.8 | .15 | .01 |
| Affective M(±SD) | 3.3 (2.2) | 3.1 (2.2) | 3.2 (2.3) | .0 | .92 | .00 |
| Lifestyle M(±SD) | 5.7 (2.5) | 6.7 (2.4) | 6.5(2.7) | 1.9 | .14 | .01 |
| Antisocial M(±SD) | 5.4(2.7) | 6.6 (2.8) (n=129) | 6.2 (2.9) | 2.4 | .09 | .02 |
| Total score M(±SD) | 16.7 (6.9) | 17.9 (6.7) | 17.9 (7.1) | .4 | .61 | .00 |
Note: DVO=Dating Violent Offender, UVO= Unknown Victim Offender, KVO= Known Victim Offender
The aim of the study was to provide a characterization of the young offender convicted of dating violence, and to compare him to young offenders with violent crimes directed against other victims and with the general population. We found that the young DVO offender differed from the general population in all investigated areas, however, the group did not differ in any comparison to other young violent offenders, regardless of their relationship to the victim.
Compared to the Swedish general population, the group showed a troublesome psychosocial background. It has previously been shown that having a parent with (substance) abuse problems entails several negative outcomes for the individual, for example, increased risk of never completing secondary school (Hjern, Berg, Rostila, & Vinnerljung, 2013). Even though youth unemployment rates have decreased since 2012 (Mångs, 2016), being without a secondary school exam is a well-defined risk factor (SOU, 2013) of unemployment in young adulthood. Among those who start secondary school, but never complete their studies, 20% are still unemployed at their mid-twenties. In the project “Children of addicts-long term outcomes of health, cognitive competence, and social adjustment” (Hjern et al., 2013) the authors compared psychosocial background factors such as having a parent with a short secondary education, having parents that had separated, and having lived in a family with economical support from social services, together with having a parent with substance abuse or mental illness. Their results showed that as much as 85% of the increased risk for children of parents with substance abuse and 75% of the increased risk for children of parents with mental illness to not achieve secondary school competence were associated with the here above presented three psychosocial background factors. The prognosis for our offenders of dating violence is far from promising when results such as these are taken into consideration, since, for example, a majority of them have not completed secondary school and about half of them have parents with alcohol abuse problems. They really appeared as significantly exposed in comparison to what was seen for the general population.
Moreover, our results show that the levels of experienced neglect and abuse in childhood were increased in the group of dating violent offenders, compared to a general population of young male university students. This is in accordance with previous results that have found an association between this kind of childhood experiences and dating violence (Widom, Czaja, & Dutton, 2014) and taps into the argument of the impact of intergenerational violence and abuse.
The DVO group was further characterized by early onset of psychosocial problems, neurodevelopmental disorders, and alcohol/substance abuse. The prevalence of substance use disorders was high: about two thirds of the group had substance use disorders, foremost alcohol (n=26, 61.9%) and cannabis (n=28, 68.3%), but also to a large extent stimulants and sedatives. The use of cannabis in the DVO group was about ten times higher than in the Swedish general population (European Monitoring Centre for Drugs and Drug Addiction, 2015). Describing this group clinically, the main problems were found in the area of mood and anxiety disorders, areas in which more than half of the group members were affected, and in the form of personality disorders characterized by antisocial, borderline, and narcissistic traits. Many of the DVO offenders also fulfilled diagnostic criteria for ADHD, both during childhood and as adults. All in all, a plausible interpretation of the extensive level of mental health problems is that they would lead to severe problems for the general ability of the DVO offenders to cope with the problems of daily life.
In comparison to the other two offender groups, there were not any differences in the investigated areas. It thus seems as a majority of young male violent offenders, regardless of relation to their victim, are more alike than different. The DVO group clearly does not deviate from what is known of other groups of young men institutionalized due to behaviour problems and delinquency (Pratt, Cullen, Blevins, Daigle, & Unnever, 2002; Ståhlberg, Anckarsäter, & Nilsson, 2010). Our findings suggest that dating violence perpetrated by young male offenders is part of male violent criminal behaviour at large, and thus (theoretically) is related to aggressive antisocial behaviour in general. This might carry potential meaning for both theory and treatment.
Regarding theoretical developments of adult intimate partner violence (IPV), it has been found that men that perpetrate IPV form a heterogenic group: the predominantly used typology of Holtzworth-Munroe and Stuart (1994) recognizes three subtypes of offenders subsumed according to the dimensions generality of violence, severity of violence, and psychopathology. The Holtzworth-Munroe and Stuart classification rendered three groups: family only, generally violent-antisocial, and dysphoric-borderline. There are also other typologies (Gottman, Jacobson, Rusche, & Shortt, 1995; Johnson, 1995); however, regardless of which typology is used, there is a subgroup of IPV offenders that comprises an antisocial, overall violent offender group. According to the empirical validation studies of the Holtzworth Munroe and Stuart typology, the general-violent antisocial subset is considered to perpetrate the most severe forms of violence towards the partner (see also Holtzworth-Munroe, 2000). In this study, we found that the young dating violence offenders in prison resemble this group, converging in the same propensity of criminal activity, yet separated by the age parameter. Due to the potentially more severe partner violence perpetrated by this “type”, we believe this further underlines the importance to develop knowledge of the young dating violence offenders in prison.
Treatment ImplicationsIn reference to batterer intervention programs, BIPs, in prison, our understanding of young violent offenders imprisoned due to dating violence as acts related to aggressive antisocial behaviour, in general they have potential impact since much concern has been expressed for the poor outcome of BIPs (Akoensi, Koehler, Lösel, & Humphreys, 2012; Haggård, Freij, Danielsson, Wenander, & Långström, 2017). The failed success of BIPs has been investigated in a recent study in which Ruiz-Hernández, García-Jiménez, Llor-Esteban, and Godoy-Fernández (2015) compared adult offenders of IPV and common offenders in prison. Through the comparison, it was found that regarding sociodemographic variables, criminal characteristics, and attitudes favouring IPV there were no differences between the groups. The difference was that the common offender group had a higher prevalence of antisocial personality disorder and the IPV group had a higher prevalence of borderline personality disorder. Referring to the principles of risk, need, and responsivity (as presented in the works of Andrews & Bonta, 2010), Ruiz-Hernández et al. (2015) proposed interventions of a general nature, focused on drug dependence, problems within the field of education and labour, impulse control deficits, interpersonal skills, empathy, conflict resolution techniques, and cognitive distortions about women and the use of violence. The conclusion was thus made that in order to reduce levels of reoffending, an intervention needs to target the criminogenic needs of the individual, (i.e., the antisocial personality pattern, pro-criminal attitudes, substance abuse, close relationships, school and work performance, and involvement with social leisure activities)
The aforementioned findings touch upon the debate of the meaningfulness of dividing offenders into “types” understood as a proclivity of engaging in specific crime, a matter lively debated. It has been shown that most offenders very seldom specialize in crime types – as a rule they are versatile in their criminal activities (Gottfredson & Hirschi, 1990; McGloin, Sullivan, & Piquero, 2009; Piquero, 2000). In a recent study of IPV over the life course, IPV and generally violent offenders showed a significant overlap regarding convictions of violent crimes and involvement in IPV (Piquero, Theobald, & Farrington, 2014). Also, the higher the rate of offences, the more the risk of criminal violence and of IPV – being a high-risk offender also increased the risk of IPV perpetration. Similar results were shown by Richards, Jennings, Tomsich, and Gover (2013) in a study of men convicted of IPV and, to judge from our data, the same pattern seems to be applicable to young male offenders of dating violence.
LimitationsLimitations are an integral part of research and this study is no exception from this rule, why there are several that need to be taken into consideration. One main limitation concerns the fact that our data do not contain any information regarding the seriousness or level of commitment of the relationship between the offender and the victimized partner. However, the classification of index crime victim being a “partner” came about after interviewing the offender, controlling it with information drawn from the written court sentences. Another main limitation consists of the fact that data about psychosocial background factors were collected retrospectively. This may have led to insufficient presentations of childhood adversities and psychopathologies due to recall bias. On the other hand, research has shown that the ability to recall this kind of experiences through clinical interviews is quite good (Hill & Nathan, 2008). Related to the matter of using historical data, there is the matter that the division into mutually exclusive groups is done with only the current conviction as dividing factor, simply because we did not have any information about victim relation for those participants that had been convicted of previous violent criminality. However, if such data had shown a historical overlap between the groups regarding victim relations, this would further support our finding that the perpetration of dating violence is an expression of a general tendency to act with aggressive antisocial behaviours. Another limitation consists of some missing data for some of the used measures, and since it has not been possible to conduct a non-response analysis we cannot determine to what extent this is systematically biasing the results. However, the study group is representative of young Swedish men in prison. Overall, the statistically significant results of differences between groups were of small to moderate effect sizes. In addition, type 1 error is probable because of the great differences in size between the three offender groups in some measures. Thus, we might have been unable to detect true differences between the groups due to small group size in several of our comparisons. Generalizations from these data should therefore be carried out with cautiousness and restricted to similar samples.
ConclusionYoung male offenders of dating violence are overall more similar than different from other young male violent offenders, especially with regard to mental health, aggressive antisocial behaviours, and psychopathic traits. They are, on the other hand, different from the general population with regard to psychosocial circumstances during the childhood and adolescent years as well as regarding educational achievements and establishment in the job-market. Thus, young violent offenders in prison are vastly alike regarding psychosocial adversities, mental health status, and potential risk factors, independent from type of violent crime and victim relation. This fact raises questions of the validity of interventions specialized according to victim relationship for young violent offenders in prison.
Financial SupportWe gratefully acknowledge The Swedish Crime Victim Compensation and Support Authority for a grant to Thomas Nilsson, and to The National Board of Forensic Medicine for research time for Anna-Kari Sjödin. The authors also wish to thank Viveca Spong, Ola Ståhlberg, and Therese Olsson for their help with data collection.
Conflict of InterestThe authors of this article declare no conflict of interest.
