Parental reflective functioning in fathers who use intimate partner violence: Findings from a Norwegian clinical sample.

Few studies have examined fathering in an intimate partner violence (IPV) context outside the US. The present study included 36 Norwegian men who were voluntarily participating in therapy after perpetrating acts of IPV. They were interviewed with the revised Parent Development Interview, which is designed to assess parental reflective functioning (parental RF), and screened for alcohol- and substance-use habits and trauma history. At the group level, participants exhibited poor parental RF, high relational trauma scores, and elevated alcohol intake. Parental RF did not correlate with education level, alcohol or substance use, or compound measures of trauma history. There was a moderate negative relationship between having experienced physical abuse in childhood and parental RF.

Intimate partner violence (IPV) is a serious public health problem worldwide, also in Norway. IPV “includes physical violence, sexual violence, stalking and psychological aggression (including coercive tactics) by a current or former intimate partner (i.e. spouse, boyfriend/girlfriend, dating partner, or ongoing sexual partner)” (Breiding et al., 2015). A Norwegian national prevalence study (Thoresen & Hjemdal, 2014) revealed that 16.3% of men and 14.4% of women had been exposed to minor forms of physical violence from their partner, while 1.9% of men and 9.2% of women had experienced severe physical violence from their partner. Another Norwegian study found that the majority of men seeking therapeutic help for their perpetration of IPV were fathers (Askeland & Heir, 2014).

There has been a growing focus on the parenting role of men who are perpetrators of IPV in family violence research. The need for a thorough understanding of the father–child relationship in families with abusive fathers is important for two reasons. First, living with IPV puts children at elevated risk for different forms of child maltreatment, as about half of the men who are physically violent toward their partner also physically abuse their children (Edleson, 1999). Witnessing IPV affects children’s neurological, emotional, and cognitive development, and correlates with a range of negative health outcomes (Carpenter & Stacks, 2009; Hamby et al., 2010). Current IPV in a relationship also affects maternal mental health and maternal parenting quality negatively, which in turn has been associated with behavior problems and aggressiveness in children exposed to the violence (Levendovsky et al., 2006). Second, the majority of children continue to have some form of contact with the perpetrator after episodes of IPV (Rothman, Mandel & Silverman, 2007). This has been a source of concern, as perpetrators may use child contact to exert continued control over their partner after separation (Bancroft, Silverman & Ritchie, 2011). However, studies have also found that children who do not have contact with their fathers following separation after episodes of IPV show higher psychological maladjustment than do children who do have father visitation. It seems that the severity of violence is a better predictor of children’s symptoms than frequency of visitation in the aftermath of IPV (Stover et al., 2003). Notably, there is heterogeneity in IPV, pertaining to severity and type of abuse, duration of IPV relationships, and degree of responsibility the perpetrator takes for the violence (Johnson, 2008). Consequentially, we can expect IPV fathers to differ with respect to their acknowledging and understanding the impact of their violence on the father–child relationship. Indeed, one study found that father–child relationships in IPV families differed in emotional closeness, the child’s dependence on the father for physical needs, and the degree of paternal caregiving responsibility. All these factors were relevant in the mediation of IPV’s impact on children’s psychological reactions (Kiser et al., 2014).

While research in the field of family violence has advanced our knowledge regarding perpetrator characteristics and IPV’s impact on children, IPV fathers’ understanding of how their violence affects the father–child relationship is still an understudied topic. We need to know more about the psychological mechanisms underlying parenting behavior in perpetrators of IPV in order to shape interventions that can enhance their understanding of children’s needs.

Perpetrators of IPV as fathers

In the following, we will present research on fathering in an IPV context. We focus on four themes that repeatedly have been linked to IPV and fathering: aspects of parental reflective functioning (parental RF; Slade, 2005), alcohol and substance use habits, relational trauma history, and the ability to perceive an impact of IPV on children.

IPV fathers and aspects of parental reflective functioning

Parental RF refers to

the parent’s capacity to understand the nature and function of her own as well as her child’s mental states, thus allowing her to create both a physical and psychological experience of comfort and safety for her child. (Slade et al., 2005, p. 283)

The depth of parents’ mental representations of the parent–child relationship correlates with the emergence and quality of children’s social skills and affect–regulation capacities (Sharp & Fonagy, 2007). Earlier research on parental RF has primarily targeted mothers and found that parental RF associated with sensitive parenting and attachment status in children (Grienenberger, Kelly & Slade, 2005; Schechter et al., 2005; Slade et al., 2005; Suchman et al., 2008). Recently, the first studies examining parental RF in fathers have been published (Esbjørn et al., 2013; Stover & Kiselica, 2014; Stover & Spink, 2012). Two studies were based on the same IPV sample (Stover & Kiselica, 2014; Stover & Spink, 2012), and found that fathers who perpetrated IPV exhibited poor parental RF. Several studies found deficits in IPV fathers pertaining to different aspects of parental RF. These included less empathy for their children and a negative bias in the perception of children’s emotional expressions compared to nonviolent fathers (Fox & Benson, 2004; Francis & Wolfe, 2008). This bias could be present at the child’s birth and become more aggravated during the first year of the child’s life (McGuigan, Vuchinich & Pratt, 2000). Maliken and Katz (2013) found that IPV fathers’ exhibited inadequate emotion regulation of their children in toddlerhood, which in turn predicted the children’s behavioral problems in adolescence. In contrast, one study reported that fathers’ elaborate working models of parenting were associated with an authoritative parenting style and low child abuse potential (Nicholson, Howard & Borkowski, 2008). Parental RF still remains to be studied in a sample of IPV fathers outside the USA, and in men who voluntarily engage in therapy for IPV.

Alcohol and substance use

There is a high co-occurrence between IPV and alcohol and substance use. In Norway, one study found that the majority of IPV episodes resulting in stays at women’s shelters were perpetrated under the influence of alcohol or substances (Lund, 2014). Several studies of fathering in an IPV context included paternal alcohol or substance abuse as a central variable (Eiden, Chavez & Leonard, 1999; Eiden & Leonard, 2000; Eiden et al., 2004; Finger et al., 2010; Stover & Kiselica, 2014, 2015; Stover & Spink, 2012). Paternal alcohol abuse correlated with harsh and insensitive fathering (Eiden & Leonard, 2000; Eiden et al., 1999, 2004). Spousal conflict mediated the association between paternal alcohol abuse and child maladjustment in these families (Finger et al., 2010). A study on fathers with co-occurring IPV and substance-abuse problems demonstrated how parental RF correlated negatively with the severity of substance abuse, positively with the level of education, but not significantly with self-reported parenting behaviors (Stover & Kiselica, 2014). Substance abuse mediated the relationship between hostility and aggressive parenting in substance abusing fathers who had perpetrated IPV (Stover & Kiselica, 2015). Most of the mentioned studies were based on samples reporting clinical levels of alcohol and substance abuse. We still know little about how alcohol and substance use influences the father–child relationship in men who voluntarily engage in therapy for IPV, and who are not in treatment for alcohol or substance abuse.

Relational trauma history

It appears that IPV perpetrators as a group have a high prevalence of relational trauma (Askeland, Evang & Heir, 2011; Dutton, 2007). Askeland (2015) found that the majority of men who sought help after having perpetrated IPV reported physical, emotional, or sexual abuse during their childhood. Unresolved childhood trauma is associated with an insecure or unresolved adult attachment status (Main, 2000), which in turn predicts insensitive and harsh parenting in fathers (McFarland-Piazza et al., 2011; Madigan, Benoit & Boucher, 2011). Only one previous study has examined the relationship between lifetime traumatic experiences of IPV fathers and RF and fathering (Stover & Kiselica, 2014). There was no association between lifetime prevalence of trauma and RF, but a correlation between trauma and hostile-aggressive parenting, confirming findings from nonclinical samples (McFarland-Piazza et al., 2011; Madigan et al., 2011). The impact of early relational trauma on the father–child relationship has so far yielded inconclusive results and should be studied in more detail in violent men.

Limited understanding of the impact of IPV on children

Two large quantitative studies found that a substantial proportion of IPV fathers were not concerned about negative effects of their perpetration of IPV on their children (Rothman et al., 2007; Salisbury, Henning & Holdford, 2009). Several qualitative studies have described how IPV fathers struggle with fully integrating the consequences of their use of IPV on the father–child relationship. In general, IPV fathers focused on the close and nurturing aspects of the father–child relationship (Perel & Peled, 2008; Veteläinen, Grönholm & Holma, 2013). They were able to acknowledge the negative impact of IPV on their children, but distanced themselves from the role they had played as perpetrators. They could only partially grasp the impact of their past violence on their ongoing relationship with the child. They rejected violence against children in general, at the same time considering their use of parental violence against children to be necessary under certain circumstances (Perel & Peled, 2008; Veteläinen, Grönholm & Holma, 2013). In addition, fathers would often hold their children responsible for episodes when they lost their temper (Harne, 2005).

The present study explored parental RF, IPV, and fathering in a Scandinavian context. We expected to find poor parental RF in our sample. Based on the literature review, we further hypothesized a negative correlation between alcohol/substance use and parental RF. We expected to find a high prevalence of trauma, and a relationship between childhood trauma and level of parental RF. We expected fathers to acknowledge difficulties in the father role, as our study consisted of men who voluntarily engaged in therapy. To our knowledge, this is the first European study to assess parental RF in the fathering representations of violent men, and one of few studies assessing parental RF in fathers.

Method

Sample and procedures

We recruited 36 participants from four therapeutic facilities for perpetrators of IPV in Norway between March 2012 and December 2014. Inclusion criteria were Norwegian ethnicity, enrollment in a therapeutic process following IPV, and at least one visitation per week. In order to avoid confounding variables, exclusion criteria were current treatment at a psychiatric hospital, and severe chronic mental or physical illness or disability of the child. Those who consented to participate were contacted by phone and scheduled to attend an independent appointment at the facility where they usually received therapy. Of thirty-six participants, all but one fulfilled the screening procedure, and all fulfilled the parental RF assessment.

Measures

Demographic variables

Age of the fathers and age of the target child were assessed continuously and reported as mean age. We reported education level as a mean of total years of education, primary school included. Current residence status was assessed using a check list giving the following options: (a) I live with my partner, (b) I live on my own, (c) my partner is currently staying at a shelter (d) My partner is currently living with others, (e) I am living with others. Marital status was assessed by asking whether participants were (a) married or cohabiting, (b) having a non-live-in girlfriend, or (c) single. Participants’ marital and residence status was combined and reported in three categories: (a) living with the child’s mother, (b) not currently living in an intimate relationship, and (c) living with a new partner. Number of children did include non-biological children participants had a parenting relationship to.

Clinical variables

Parental RF

We used the Parent Development Interview-Revised (PDI-R2; Slade et al., 2003) to assess parental RF. The questions in the PDI-R2 give the interviewee the opportunity to reflect upon his own as well as the child’s feelings, thoughts, and intentions, and how they might influence both the child’s and the parent’s behaviors and mental processes. Interviews were transcribed from audio files, and transcripts were compared to audio files for accuracy by the first author. Protocols were scored for parental RF on an 11-point scale from –1 to 9, with a score of 5 and above indicating an adequate level of RF (Slade et al., 2003). The lack of qualified scorers of RF on the Norwegian version of the PDI-R2 made it necessary for the first author to score all PDI-R2 protocols, some of them from interviews that he had administered himself. We aimed to control for the risk of bias arising from this by letting external researchers who were blind to the study score ten of these protocols. When there was disagreement between raters, the conclusion of the external coder was used in the analyses.

IPV

We assessed IPV based on the intake interview, the PDI-R2, and on the referral to the therapeutic facility. Fathers were categorized as being physically violent when they confirmed episodes of recurrent mild physical violence, or single or recurrent episodes of severe physical violence. Mild physical violence referred to behaviors such as pushing, holding, slapping, and shoving, while severe violence referred to behaviors such as choking, punching, and inflicting bruises or broken bones. We defined mild psychological violence as behaviors such as screaming and shouting during arguments, while severe psychological violence referred to threats to use physical violence, control of the partner’s privacy and social network, verbal denigration, and destruction of inventory in the presence of partner or children.

Alcohol use

The alcohol use disorders identification test (AUDIT; Babor et al., 2001) is a 10-item self-report questionnaire about a person’s alcohol use. It assesses the frequency and quantity of alcohol consumption, the degree of alcohol abuse, and dependence symptoms.

Drug use

The drug use disorders identification test (DUDIT; Berman et al., 2007) is an 11-item self-report questionnaire about a person’s use of substances other than alcohol and prescribed medications. It assesses the frequency and quantity of substance use, the degree of substance abuse, and dependence symptoms.

Trauma

The traumatic events checklist (TEC; Nijenhuis, Van Der Hart & Kruger, 2002) is a 29-item self-report questionnaire that examines the presence, duration, and subjective impact of different potentially traumatic life events. It assesses single experiences, such as having been exposed to an accident, as well as clusters of recurring traumatic events in childhood, namely emotional neglect, emotional abuse, physical violence, sexual harassment, and sexual abuse.

Parental self-evaluation

The short form of the Parental Stress Inventory (PSI-SF; Abidin, 1990) is a 36-item self-report questionnaire that examines a subject’s experience of parenting a child 12 years of age or younger. We used a single item from the PSI-SF to assess how fathers evaluated themselves as parents: I feel that I am (a) not a very good parent, (b) someone who has problems being a parent, (c) an average parent, (d) a better than average parent, (e) a very good parent.

Data analysis

We examined all variables for outliers and skewness. Several variables were statistically skewed. We checked single cases and found that skewness did not occur due to measurement error. We decided to accept non-normal distribution of the data. We computed means, standard deviations and range on all measures. We scored the presence of two specific traumatic experiences (growing up with parental alcohol abuse and witnessing domestic violence) as categorical variables and reported them in percentages. We conducted Spearman’s rank-order correlations applying two-tailed test of significance and accepting a significance threshold of .05 for all statistical tests. Correlations were calculated between our study variables, and between the different types of relational childhood trauma in the TEC and the other study variables. We calculated Inter-rater reliability on the PDI-R2 as Intraclass Correlation Coefficient, using both one-way random single measures and one-way random average measures.

Results

Sample characteristics

Fathers’ age, level of education, marital status, as well as number of children and age of the target child are presented in Table 1. Gender of target children was equally distributed.

Table 1.

Descriptive statistics.

	Mean (SD)	N (%)
Father’s age (years)	36.2 (7.6)
Father’s length of education (years)	13.9 (2.5)
Child’s age (years)	5.7 (2.5)
Living with the child’s mother		18 (50%)
Not in an intimate relationship		13 (36%)
Living together with a new partner		5 (14%)
One child		7 (19.5%)
Two children		16 (44%)
Three or more children		13 (36%)

Type of violence

Distribution of participants who reported mild versus severe physical or psychological violence toward partner and children is presented in Table 2. Among men who admitted physical violence toward their partner, 29% also reported use of physical violence toward a child.

Table 2.

Type of violence perpetrated by the participants.

	N	%
Physical violence toward partner	24	66
• Physical violence toward partner (severe)	14	39
• Physical violence toward partner (mild)	10	28
Physical violence toward children (total)	11	31
• Physical violence toward children (severe)	6	17
• Physical violence toward children (mild)	5	14
Physical violence toward both partner and child	7	19
Psychological abuse of mother	33	92
Psychological abuse of child	26	72

Study variables

Scores for parental RF, alcohol and substance use, and trauma history are presented in Table 3. All 36 participants were interviewed with the PDI-R2. Almost nine out of ten had an RF score of 4 or lower, indicating a non-adequate level of parental RF; the most common RF score was 3 (55%, N = 18). The interrater consistency on the PDI-R2 reached satisfactory levels with an intraclass correlation of 0.77 on single measures and 0.87 on average measures. Thirty-five participants completed the AUDIT, DUDIT, and TEC screening. AUDIT scores revealed that 57% of the participants had a low risk of developing an alcohol-abuse disorder, 38% had an elevated risk, and 5% had a high risk. DUDIT scores revealed that 80% of the participants had not used substances other than alcohol during the past year.

Table 3.

Descriptive statistics for study variables.

Variable	Mean (SD)	Actual range	Possible range	Prevalence (%)
RF (PDI-R2)	3.4 (0.9)	2–6	–1–9	–
Alcohol use (AUDIT; at risk of abuse, %)	7.6 (5.0)	0–25	0–40	43
Drug use (DUDIT; at risk of abuse, %)	1.5 (4.1)	0–19	0–44	11
Lifetime traumatic experiences (N)	7.9 (3.9)	0–15	0–29	–
Childhood emotional neglect	5.0 (4.9)	0–12	0–12	50
Childhood emotional abuse	5.5 (4.6)	0–12	0–12	53
Childhood physical abuse	4.3 (5.3)	0–21	0–21	43
Childhood physical sexual abuse	–	–	–	14
Growing up with parental alcoholism				33 (N = 12)
Witnessed DV				14 (N = 5)

Correlations between study variables

Correlations between parental RF and alcohol use, substance use, education, number of lifetime traumatic experiences, and childhood trauma are presented in Table 4.

Table 4.

Correlations between study variables.

Variable	RF	Alcohol use	Drug use	Education	Lifetime trauma (N)	Childhood trauma
RF	–	−.33	−.03	.23	−.16	−.21
Alcohol use		–	−.33	−.43*	.35*	.10
Drug use			–	−.27	.50**	.43**
Education				–	−.43**	−.30
Lifetime trauma					-	.71**
Childhood trauma						–

p < .05.

p < .01.

Correlations between parental RF and different types of childhood trauma are presented in Table 5.

Table 5.

Correlations between aspects of childhood trauma and study variables.

	RF	Alcohol use	Drug use	Education	Lifetime trauma (N)
Childhood trauma	–.21	.10	.43**	–.30	.71**
Physical abuse	–.34*	.28	.39*	–.43**	.71**
Emotional abuse	–.09	–.08	.16	–.37**	.49**
Emotional neglect	–.20	.19	.50**	–.14	.66**
Sexual abuse	–.05	.02	.44**	.16	.37*

p < .05

p < .01.

Self-evaluation as parents

In our sample, 35 participants provided PSI-SF scores. Here, 40% rated themselves as average parents, 49% as better than average, and 11% stated that they had difficulties being a parent.

Discussion

We assessed a Norwegian sample of domestically violent fathers in terms of parental RF, alcohol- and substance-use habits, trauma history, and perception of own parenting. The sample consisted of men voluntarily participating in therapy after perpetrating acts of IPV. Our main findings were that parental RF was poor despite low prevalence of substance abuse and fairly high levels of education. Contrary to our expectations, neither education nor alcohol and substance use correlated significantly with parental RF. RF was not correlated with a compound measure of childhood trauma. However, it was negatively correlated with having experienced physical abuse in childhood. Alcohol use was subclinical, yet elevated at group level. The majority reported no illicit drug use at all. As expected, there was a high prevalence of relational trauma. Contrary to what we expected, almost half of the sample rated themselves to be better than average as parents.

First, parental RF was generally low in this study. More specific, we found that reflections on the child’s mental states were shallow, and any descriptions of mental processes were predominantly made in relation to the father’s own experiences. This reflects findings from a previous study using parental RF as an index of fathers’ mental representation of the parent–child relationship (Stover & Kiselica, 2014). A specific feature of our sample was that parental RF did not correlate with education and alcohol and substance use. The reason for this may be a fairly high level of education and absence of major alcohol- and substance-abuse problems in the sample. This finding suggests that low education and substance abuse do not necessarily explain poor mentalization in violent fathers. In addition, there was a moderate negative relationship between having experienced physical abuse in childhood and parental RF. The basis for mentalization is a secure attachment relationship, and physical violence toward a child undermines secure attachment (Fonagy & Target, 1997). Physical abuse can force children to refrain from mentalizing their early attachment relationships, in an attempt to regulate overwhelming emotional states (Fonagy, 2003; Green & Goldwyn, 2002). In addition, families where child abuse occurs may have little focus on fostering reflective capacities in children. As a result, shallow mentalization may develop as a stable organizing principle of close relationships, manifesting itself in poor emotion regulation, poor empathic understanding of others, and fragmented ability to experience own and others’ emotional states as meaningful (Fonagy & Target, 1997). However, scores on the RF scales do not tell anything about the underlying psychological mechanisms linking physical violence experiences in childhood to poor parental RF in adulthood. A qualitative analysis of PDI transcripts may be a method to gain more insight into this matter.

Second, participants reported elevated but subclinical alcohol use levels. The proportion with AUDIT scores indicative of an elevated risk for developing an alcohol use disorder was twice as high as in the general population (Andreassen, 2011). While the majority did not use illicit substances at all, the proportion of substance use was three to ten times higher than in the general Norwegian population, applying estimates for the two most common drugs, cannabis (Statens Institutt for Rusmiddelforskning, 2014) and amphetamine (European Monitoring Centre for Drugs and Drug Addiction, 2015). However, our sample still seems to differ from previous studies, where alcohol and substance use was higher at the group level (Stover & Kiselica, 2014; Stover & Spink, 2012). Possible reasons for this may be differences in research populations, who often stem from treatment populations, as there are differences in the way treatment for IPV is organized in the US compared to Scandinavia. Notably, US studies have found that alcohol and substance abuse are more prevalent in perpetrators of IPV who were court-mandated into treatment, as opposed to self-referred, and who were violent outside the family, as opposed to men who perpetrated violence only within the family (Illinois Department of Human Services, 2005). In addition, the correlation between alcohol abuse and IPV was stronger in clinical samples from alcohol treatment facilities, and among subjects having alcohol-abuse problems meeting criteria for alcohol dependence (Foran & O’Leary, 2008). In Scandinavia, treatment options for domestically violent men are voluntary. Therefore, severe alcohol- and substance-abuse problems might be less prevalent in Scandinavian IPV treatment populations compared to clinical samples from North-American IPV treatment facilities.

Third, our findings regarding relational trauma are similar to those in studies involving larger clinical samples of male perpetrators of IPV (Askeland, 2015), and can be considered to be representative of a Norwegian population voluntarily seeking help for IPV. The proportion of our participants reporting severe physical abuse in their childhood was eight times higher than in the general Norwegian population (Thoresen & Hjemdal, 2014). Similarly, the proportion reporting emotional abuse and contact sexual abuse was five times higher than in the general population (Steine et al., 2012; Thoresen & Hjemdal, 2014). While there was only a minor difference between our sample and the general population when it comes to witnessing IPV in childhood (Thoresen & Hjemdal, 2014), more than twice as many men in our sample reported parental alcohol abuse (Rossow, Moan & Natvig, 2009). Stover and Kiselica (2014) pointed out that it can be hard to test the association between parental RF and childhood trauma, since the bulk of participants scored either 3 or 4 on the 11-point RF scales, making differentiation within the sample difficult. However, the high prevalence of relational childhood trauma in our sample makes it likely that a substantial proportion of IPV fathers have insecure or unresolved attachment representations in adulthood. Partner abusive men have previously been found to be more often classified with an insecure attachment style than non-violent husbands are (Babcock et al., 2000; Dutton, 2007). Studies testing the association between fathers’ attachment representations and their parenting found that fathers classified as insecure or unresolved showed less sensitive and more hostile parenting (McFarland-Piazza et al., 2011; Madigan et al., 2011). As fathers’ adult attachment status seems to be linked to their children’s socio-emotional development (Steele & Steele, 2005), a focus on fathers’ relational trauma histories and how well they have been integrated in the individual should be a part of any clinical intervention with IPV men. We also found a relatively high percentage of sexually abused men in our sample, confirming findings from a larger Norwegian sample of male perpetrators of IPV (Askeland, 2015). The possible influence of having experienced sexual abuse on the fathering of men with IPV problems has to our knowledge not yet been studied. Our findings suggest that this is a possibly under-communicated topic in IPV therapy.

Fourth, similar to previous studies (Fox & Benson, 2004; Veteläinen, Grönholm & Holma, 2013), we found that fathers did not see their perpetration of IPV as relevant in their evaluation of the father–child relationship. The majority of our sample rated themselves to be average or better than average parents. One possible explanation for this finding is that the fathers compared their parenting with their own childhood experiences, which in many cases were characterized by violence and neglect. Since most participants talked about memories of abuse and neglect when they were completing the PDI-R2, and stated that they wanted to be better parents than their parents had been, they might have activated a comparison bias when later evaluating themselves as caregivers on the PSI-SF. In addition, several fathers perceived the child as difficult, confirming findings from previous studies (Harne, 2005; Veteläinen, Grönholm & Holma, 2013). Children who are exposed to IPV often show externalizing symptoms, associated with the severity of the violence (Stover et al., 2003), which is one possible reason for perceiving the child as difficult. Another possibility, as suggested by Francis and Wolfe (2008), is that the fathers in our sample misperceived their children’s normal emotional expressions as extremely negative, and consequentially attributed the blame for father–child conflicts to the child.

Finally, the minority of our sample admitted physical violence toward children, applying a definition that counts single acts of spanking or slapping as violence. Less than one-third of men who reported physical IPV also reported physical violence toward their children. While this is a high proportion, it is below the reported prevalence of 40–60% for the concordance between partner and child physical abuse (Edleson, 1999). One possible reason for this may be the criminalization of spanking and the cultural stigma associated with physical aggression toward children in Scandinavian countries (Modig, 2009). In the US, spanking is often culturally accepted and common (Lee, Guterman & Lee, 2008) and may evolve into more severe violence (Straus & Stewart, 1999).

Strengths and limitations

As there are societal, cultural, and judicial differences between the US and Scandinavia pertaining to IPV and fathering, we think that our study can contribute with insights from a Norwegian IPV sample. Since we studied a homogenous Norwegian sample of men voluntarily engaging in therapy for IPV, our sample is more representative of Scandinavian treatment models for IPV than are US studies. However, the small sample size and the homogeneity of the population also make generalizations of our findings difficult. In addition, men who voluntarily participate in treatment for IPV may differ from violent men who do not seek treatment, who may be more prevalent in substance abuse treatment or the criminal justice systems. As we did not use a validated research tool to assess IPV, we relied partly on perpetrators’ self-reports, which may yield biased presentations of both type and severity of IPV. We still tried to describe patterns of more or less severe psychological and physical violence.

Implications for research

We suggest that parental RF represents too general an index of mental processes linking IPV and parenting. To better understand the mechanisms underlying parenting by violent fathers, we suggest that the core constructs of parental RF – specifically, empathy toward the child, affect recognition and affect regulation, and understanding of family dynamics – need to be examined further.

Clinical implications

The findings from this study can inform practice on the following points. First, parenting should be addressed early on with IPV fathers, as our results suggest that they have problems with understanding and regulating children’s emotional signals and needs. Second, a thorough assessment of lifetime traumatic experiences, specifically relational trauma is called for. Third, our findings suggest the presence of subclinical, yet elevated alcohol use. Consequentially, alcohol habits should be assessed as a rule. In therapy, both trauma history and alcohol use should be linked to IPV and potential challenges with mentalizing the father–child relationship. Stover’s (2013, 2015) model for treatment of IPV fathers with co-occurring substance use problems, Fathers for Change, incorporates all of these points, and pilot studies have shown promising results. In addition, Child–Parent psychotherapy (CPP; Lieberman, Ghosh Ippen & Van Horn, 2006) is an approach that has proven good results after traumatic raptures of the parent–child relationship. These treatment models seem to be good therapeutic interventions for male perpetrators of IPV also outside alcohol and substance treatment, as long as safety precautions for all involved family members are in place. Both models address how own childhood trauma affects parenting via underdeveloped or distorted mental models of the parent–child relationship.

Conclusion

Our findings confirm the following tendencies identified in previous research on violent men and fathering: IPV fathers show poor parental RF, elevated alcohol intake, a high level of relational trauma experiences, and a positive bias when evaluating own parenting skills. These tendencies have been confirmed in an ethnically homogeneous sample of Scandinavian men who exhibited a relatively high degree of social integration. In addition, and contrary to previous studies, we found no association between education or substance use and parental RF, but a moderate negative association between having experienced physical abuse in childhood and parental RF.

Disclosure statement

No potential conflict of interest was reported by the authors.