Introduction

Intimate partner violence (IPV) against women is a pressing social issue, and some researchers have even described it as epidemic in our society (Briere and Jordan 2004). A majority of international IPV studies focuses exclusively on violence against female victims. The main argument is that women suffer more severe consequences and seek help more frequently than male IPV victims (Archer 2000; Campbell 2004; Dixon and Browne 2003; Frye et al. 2006). A nationwide survey on domestic violence in Norway found that 10% of the women and 2% of the men had experienced severe IPV (Haaland et al. 2005). Recent reviews of IPV have emphasized that there is a paucity of studies on violence escalation and interaction, dynamics of violence, and prediction and coping of IPV incidents as perceived and experienced by the target of violence (Langhinrichsen-Rohling 2005; Schwartz 2005; Arriaga and Capezza 2005; 2004; Cano and Vivian 2001; Heckert and Gondolf 2004; Mears and Visher 2005; Messman-Moore and Long 2003; Winstok 2007). The research we are reporting here was initiated to address four main researches of the experiences of 157 women: (1) their responses to violence escalation and violence perpetrated by the partner, (2) the effect of pregnancy and motherhood on IPV, (3) the possible influence of immigration and ethnicity, and (4) support and treatment. In this article we only address the first research issue, and we do so by reporting findings from the following areas of research: responses and conduct before, during and after the IPV; dynamics of violence; effect of alcohol and drug intoxication; and revictimization.

An Interactional Perspective on IPV

A major consideration in efforts to assess, predict and change IPV pertains to the manner in which aggression is conceptualized. Theoretical approaches to the analysis of IPV have emphasized variables such as conflict e.g., (Straus 1979), and power and control e.g. (DeKeseredy and MacLeod 1997). There is growing evidence that behaviors and perceptions of behaviors in intimate relationships should be investigated as separate entities, and that it may be important to assess how individuals perceive not only their own communication but also their partners’ communication e.g., (Rhoades and Stocker 2006). In a recent article Winstok (2007) argues for an integration of different approaches in an interactional understanding of IPV. Interactional models emphasize the importance of person–situation interactions in efforts to understand both personality and behaviour. The interactional understanding is also used within other branches of research on violence, such as violence committed by persons with major mental disorder e.g., (Bjørkly 1993; Monahan 1988). The main idea is that aggression involves an influential and continuous interaction between individuals and the various situations they encounter. Situations are defined as the actual episodes are perceived, interpreted, and assigned meaning (Magnusson 1981). Consequently, since we investigate different types of IPV in the present study within an interactional perspective it is acknowledged that a woman who has been subject to for instance psychological and physical IPV on separate occasions probably will have different experiences and perceptions of these interactions. Accordingly, for instance the escalations leading to psychological and physical IPV reported by the same woman will be treated as two separate analytical units because the focus is on person–situation interaction rather than two different IPV situations treated as if they were identical because the same victim and perpetrator were involved.

The nature of the IPV that battered women experience is not uniform. Instead, battered women struggle with a variety of combinations of physical violence, psychological violence and sexual violence. The severity of each type of violence or abusive behaviour also varies. It is unclear what causes this variability and how these differences may affect the impact of violence on battered women, and their revictimization (Dutton and Corvo 2006). Although the body of IPV research has focused on patterns, causes and interventions, systematic literature reviews have found that studies on dynamic and interactional aspects are scarce (Bjørkly 2003; Winstok 2007), and that little is known about how women cope in long-term abusive intimate relationships (Lund and Greene 2003; Zink et al. 2006). A review of situational determinants in IPV concluded that IPV research rarely examines domestic violence events as interactional and dynamic phenomena (Wilkinson and Hamerschlag 2005). There is a paucity of studies that investigate situational or event factors and their surrounding contexts in a systematic way. In their review, Wilkinson and Hamerschlag (2005) suggested that, by collecting these kinds of data on heterogeneity of IPV, we can bridge huge gaps in our comprehension of this complex and serious problem. In our study, we investigated women’s perceptions of interpersonal factors within an interactional perspective, with an emphasis on interactions in which IPV behaviours emerge.

It is claimed that recognition of recurrent warning signs, as specific individual precursors of violence, is of great importance in successful treatment and relapse prevention (Bjørkly 2000; Steadman and Silver 2000). There appears to be a paucity of instruments available for the accurate and clinically useful measurement of warning signs of IPV escalation (Bjørkly 2003). Research findings suggest that many women will underestimate the severity of the early stages of IPV escalations (e.g., Campbell 2004). Accordingly, one aim of our research was to collect information on the responses the women had when they observed their partners’ precursors of violence, another was to find out more about whether such observations helped the women to cope with the escalation in time, or if it only made things worse for them.

Impact of Alcohol and Drug Intoxication

The impact of alcohol and drug intoxication on IPV is a controversial issue. Research exploring the link between alcohol use and IPV has repeatedly failed to differentiate the nature of men’s violence to women from men’s violence to men (Galvani 2006). Accordingly, the most prevalent conclusion in feminist research on IPV is that the impact of intoxication is not significant in domestic violence (Galvani 2006). Meta-analytic reviews evaluating the evidence on the relationship between drug use and IPV concludes that increases in drug use and drug-related problems are significantly associated with increases in IPV (Moore et al. 2007; Testa 2004). Another review found that about 20% of men and 10% of women were drinking before the most recent and severe act of violence (Klostermann and Fals-Stewart 2006)

Some investigators have argued that association between alcohol use and IPV may vary considerably as a function of the characteristics of the person drinking and the circumstances under which intoxication occurs. One conclusion was that women with few social roles and responsibilities have a stronger inclination to report problematic alcohol use (Kerr-Corrêa et al. 2007).

The most consistent mediator appears to be presence of other factors that are causally implicated in IPV. For example Fals-Stewart’s model of alcohol use, IPV and antisocial personality presumes that men with high levels of antisocial personality characteristics have a tendency to be physically aggressive, even when they are not intoxicated (Klostermann and Fals-Stewart 2006) Still, the study of the link between alcohol and IPV is methodologically complex, and so far, even the relative frequency of physical violence in problem drinking couples, compared with non-problem drinking couples, has not been empirically settled (Kelly and Halford 2006). In line with this, one aim of this study was to investigate the role of alcohol as an interactional predictor of IPV from the perspective of female victims of IPV.

Predictors of IPV and IPV Victimization in More Than One Partnership

Research indicates that some women return to abusive relationships after having been exposed to IPV (Rhatigan et al. 2005). A review of explanatory models for the termination of violent relationships concluded that general approaches, like reasoned action/planned behaviour and investment models, may be a better way to understand such complex and multifaceted decisions than theories of learned helplessness, traumatic bonding and psychological entrapment e.g., (Rhatigan et al. 2006). It was claimed in the review that the general approaches provided a non-pathological understanding of the women’s decisions to leave an IPV partnership. These models suggest that victimized women take into account the same types of information as non-victimized women do, in deciding whether to terminate their relationships. Preventing reassault/revictimization is especially important to survivors’ recovery from interpersonal violence because exposure to multiple traumatic experiences may negatively affect a person’s capacity to recover from subsequent traumatic events (Dutton et al. 2006).

Most of the research on risk factors for IPV focuses on characteristics of the perpetrator, rather than the victim (Norlander and Eckhart 2005; Schumacher, Feldbau-Kohn and Slep 2000a). Information on risk factors pertaining to the victim may contribute to the development of indicated prevention strategies for women. However, if one addresses victim risk factors it is important to acknowledge that risk factors are not necessarily causal variables and do not imply that victims cause the IPV. A review of risk factors for physical IPV concluded that only fewer years of education, unemployment and childhood emotional or verbal victimization are significant mediators of risk (Schumacher et al. 2000a). Yet, even for these associations, only moderate effect sizes were found.

A review of the literature that addressed the risk and protective factors for psychological IPV concluded that psychological IPV may be more difficult to predict than physical IPV (Schumacher et al. 2000b). In general, no empirical evidence exists concerning socio-economic status discrepancies as predictors of psychological IPV. Schumacher et al. (2000a, b) found that certain relationship variables were significantly associated with psychological IPV with moderate to strong effect sizes. However, this association is difficult to interpret, because these relationship variables exhibit a high degree of conceptual and operational overlap with psychological IPV (Schumacher et al. 2000b). Still, it is interesting that studies of clinical samples with the Psychological Maltreatment of Women Inventory (PMWI) have found control behaviours to be risk factors for psychological IPV of medium to large effect sizes (Schumacher et al. 2000b).

A brief review focusing on the risk and protective factors for male-to-female sexual abuse found that several associations were of a moderate effect size, in spite of the fact that partner sexual abuse is a low base rate behaviour (Black et al. 2000). Unemployed women, and women from low-income households, were more likely to report this type of IPV. Prior unwanted sexual experiences and the severity of physical aggression were also associated with IPV (Black et al. 2000). Other reviews and recent research focusing on child sexual abuse demonstrated that women with a history of childhood sexual abuse or childhood exposure to parents IPV were at increased risk of revictimization (Filipas and Ullman 2006; Kogan 2005; Macy 2007; Messman-Moore and Long 2003; Miller 2006). Within an interactional perspective it is of particularly interest to find out more about the impact of adverse childhood experiences and previous IPV victimization on subsequent IPV experiences.

In summary, the scope of this study was to investigate IPV within an interactional perspective. For this purpose, we used a semi-structured interview questionnaire with items that tapped into event and situational factors. We explored women’s responses and conduct before, during, and after the IPV by focusing on (1) the dynamics of violence, (2) the impact of alcohol and drug intoxication, and (3) IPV victimization in more than one partnership. Specific research questions were as follows.

  • Were there significant differences between IPV categories (physical, psychological, and sexual) concerning the women’s responses and conduct before, during and after the IPV?

  • Were the women able to predict the IPV?

  • Did the women have coping strategies, and did they perceive these as self-preserving and effective?

  • Did alcohol and drug intoxication have an impact on the IPV interactions?

  • Were there significant differences between physical, psychological, and sexual IPV in relation to the impact of alcohol and drug intoxication?

  • Were rates of women who had been IPV victimized in more than one partnership high?

  • Were there significant differences between women who had been IPV victimized in more than one partnership and those who had not been IPV victimized in previous partnerships?

Method

The study was approved by the Regional Norwegian Ethics Committee. Written informed consent was obtained from the participants. The interviews were carried out between April 2005 and April 2006 in eight Norwegian counties.

Sampling and Recruitment

One of the authors (SV) contacted shelters, police, and family counseling offices introduced the study, and made contact with institutions that were willing to participate in the study. These shelters, police, and family counseling offices were selected to secure proportional recruitment from the capital of Norway, small towns and the countryside. The shelters, police and family counseling offices that participated asked every woman who had made contact after exposure to IPV if she would volunteer to participate in the study. Inclusion criteria for the women were.

  • a minimum age of 18 years;

  • had contacted a shelter, the police or a family counselling office after being subject to IPV (index IPV exposure); and

  • the index IPV exposure had occurred no more than 6 months before the interview.

Exclusion criteria were:

  • refusal to participate; or

  • ethical, psychological or medical reasons for not participating in the research.

Most (92.9%, n = 192) of those who were approached to participate in the study volunteered to take part. Of these, 5 (2.6%) were not able to be contacted to make an appointment for the interview, 13 (6.8%) changed their mind before entering the interview, 7 (3.6%) had to cancel because of somatic illness or because they had moved to other parts of the country, and 10 (5.2%) did not show up after the appointment had been made. This resulted in a final sample of 157 participants, corresponding to 73.4% of all the women who were initially approached to participate.

The sample was recruited from 10 shelters (n = 73), 5 police districts (n = 41) and 6 family counselling offices (n = 43). According to official records, shelters, police, and family counseling offices cover about 85% of help-seeking IPV victims in Norway (Norwegian Ministry of Justice 2003, 2005). One hundred and five women were native Caucasian Norwegians and 52 were born in other countries. There was a significant correlation between country of origin in the sample and the distribution of country of origin among immigrant women in Norway in general (Kendall’s tau-b = 0.69, p < 0.001). Sixteen per cent were born in Africa, 50% in Asia, 6% in America/Oceania, 15% in Eastern Europe, 2% in Western Europe, and 11% in other Nordic countries. The women were recruited on the basis of the latest official records to obtain a representative sample concerning proportion of immigrants, sociodemographic characteristics, and annual proportion of women contacting the three recruitment instances (police, shelters and family counselling offices). Until each subgroup concerning immigration and recruitment instances were filled up every woman who had made contact after exposure to IPV was asked if she would volunteer to participate in the study. When one of the subgroups was complete one of the authors (SV) sent a message to every recruitment instances that they could stop recruitment for that particular subgroup. After the recruitment was closed the sample was statistically controlled concerning other demographic variables. The sample was not significantly different from the population it was recruited from.

Subjects

Mean age of the participants (n = 157) was 36 years (SD = 9.51; range = 19–74 years). High school was the median (and modal) education level. Mean income level of the women was identical to that of the general female population of Norway. Sixty-one percent perceived their health as “good” or better. Over 80% had close social relations outside the partnership. Most of the sample (77.1%) grew up with both biological parents until at least the age of 16 years. About half (49%) of the women had lived with their partner for more than 7 years. Over 81% had decided not to return to their partner after the last IPV incident. Close to 75% of the women reported their risk of IPV revictimization to be low, regardless of whether they stayed with the partner or entered a new partnership. Detailed information about socio-demographic variables and relationship with the partner is given in Table 1.

Table 1 Women’s socio-demographic profile and characteristics of relationship with partner (n = 157)
Full size table

Procedure

Data were gathered using a semi-structured face-to-face interview carried out by one of the authors (SV), a female clinical psychologist. Before the interview, respondents were informed that some of the questions were of a very intimate and confidential nature. They were assured that their participation was voluntary, that they were free to withdraw their informed consent at any time during or after the research interview, and that refusal to participate would not affect services that they received at the recruitment facility. Each woman was also informed that information she gave would be coded and stored in a way that made it impossible to disclose its origin. The interviews lasted from one to three hours. In 12 of the interviews a professional interpreter was used. After the interview, participants were asked to give brief feedback on their experience of the interview. More than 76% of the women reported that they felt that they had been taken care of and that the interview was meaningful to them, and about 23% reported the same, but that a few questions were unpleasant. Only one woman reported that she had not been optimally taken care of. Still, she expressed that she found the interview to have been meaningful. Interviews took place at shelters, police stations, and family counselling offices or at the interviewer’s office. There was no financial incentive for participation. Because every woman was already in contact with the actual recruitment facility, none of the participants disclosed the IPV for the first time in the research interview.

Pilot study

We learned from a pilot study of 20 female victims of IPV that a semi-structured interview format with multiple-choice alternatives made it easier for participants to answer intimate questions. We also found that this format yielded reliable and valid information without taking too much time. To make sure that important information was not lost, however, most of the questions also had an open-ended option to elicit further information from the participant.

Interview Questionnaires

Definitions of violence and IPV

Violence is defined as intentional attempts at, threats of, or actual, and intended infliction of psychological or bodily injury or harm on another person (Zillman 1979). IPV is defined in the World Health Report on Violence and Health as: “any behavior within an intimate relationship that causes physical, psychological, or sexual harm to those in the relationship. Such behaviors include acts of physical aggression, psychological abuse, forced intercourse, and other forms of sexual coercion, and various controlling behaviors such as isolating a person from their family and friends, monitoring their movements, and restricting their access to information or assistance” (Winstok 2007). See Tables 2, 3 and 4 for operational examples of the three IPV main categories. In our study, the term IPV includes violence from intimate partners that lived in the same household and those who did not. As noted by Hicks, most studies define IPV as physical and/or sexual abuse only (Hicks 2006). In our study, we also incorporated psychological violence to identify both minor and serious violence. This was partly done because it is our clinical experience that the experiential consequences of psychological IPV may cause serious health problems for the victim. Another reason was that we wanted to analyse possible differences and similarities between physical and sexual IPV, as opposed to psychological IPV. For this particular purpose, psychological IPV was limited only to episodes that had occurred without physical and/or sexual IPV.

Table 2 Most frequent types of physical IPV (%)
Full size table
Table 3 Most frequent types of psychological IPV (%)
Full size table
Table 4 Most frequent types of sexual IPV (%)
Full size table

Structured socio-demographic and health questionnaire

This 93-item questionnaire covers the following themes:

  • Socio-demographic profile: Information such as age, education, employment, economy, housing standard, religious beliefs, substance use, marital status, relation with the partner, and social support was collected. This questionnaire was based on Statistics Norway’s Level of Living survey (Statistisk sentralbyrå, Levekårsundersøkelsen 1995).

  • Mental health: An assessment of mental health was based on the Hopkins Symptoms Check List-25 (HSCL-25) (Derogatis et al. 1974), and parts of the SCL-90-R (Derogatis 1992), which yield information about subjective distress such as anxiety, panic attacks, loneliness, worries, sleep difficulties, and fatigue.

  • Self-efficacy: Measurement of self-efficacy was based on Generalized Self-efficacy (Jerusalem and Schwarzer 1992), which provides information about the individual’s general experience of coping, self-esteem and self-efficacy.

Semi-structured IPV questionnaire

A 229-item multidimensional interview questionnaire was designed with items on perceptions and personal experiences with IPV. The items that measure interactional aspects were developed especially for this research, but items addressing characteristics of IPV were drawn from the Conflict Tactics Scales (CTS-2) (Straus 1979; Straus et al. 1996), and items addressing the last IPV incident the women had been involved in were drawn from the British Crime Survey 1996 (Mirlees-Black 1999). The questionnaire addressed three main categories of IPV: A. Physical IPV (42 items), B. Psychological IPV (40 items), C. Sexual IPV (43 items). The three main categories of IPV were repeated for IPV during pregnancy: D. Physical IPV during pregnancy (34 items), E. Psychological IPV during pregnancy (30 items), and F. Sexual IPV during pregnancy (34 items). Every woman was interviewed about each main section. If she did not report any experience with any of the optional types of IPV (see Tables 2, 3 and 4 for examples) in the actual main section, the interview moved on to the next main section. Around 75% of the items were questions pertaining to IPV interactions with the woman’s last partner. In this context, the term “last partner” signifies the partner that caused the index IPV exposure. Each section in the questionnaire addressed the following topics:

Characteristics of the IPV

This section of the questionnaire was designed to obtain detailed information concerning interactional factors, such as coercive means used by the partner, consequences of IPV, actions taken as a result of injuries, warning signs, coping strategies, feelings of responsibility, guilt and shame, and the severity, duration, and frequency of the three main categories of IPV (physical, psychological and sexual) perpetrated by the partner. This part of the questionnaire was repeated for each of the six IPV sections.

Physical IPV (42 items)

This part covers a wide range of physical IPV (CTS-2)., such as immobilisation, punches, kicks, strangleholds, pushes, hitting with objects, pushing down stairs, stabbing, and shooting. Another item related to the kind of immediate physical consequences the physical IPV resulted in. Optional answers were: none, broken bones, wounds, bruises, dislocation, ocular lesions, and others. A further question addressed the woman’s responses to being physically attacked. Optional answers were: Nothing I can think of; give in and obey him; keep still; escape; try to reason with partner; call for help; cry; shout; defy aggressor; counterattack; and an open “other” category. Another item addresses whether the woman knew any warning signs from her partner that could help her predict physical IPV.

Psychological IPV (40 items)

The items tap information concerning occurrence of, for instance: insults, humiliation (private or public), isolation from family and friends, economic abandonment, pursuit and harassment, not allowed to work, not allowed to leave the house, and an open “other” response category.

Sexual IPV (43 items)

Types of forced sexual activity are: genital intercourse with penetration, anal intercourse, oral sex, objects inserted in vagina or anus, partner masturbation in front of woman, homosexual intercourse, intercourse with animals, prostitution, intercourse while others watched, had to put on certain clothes, use objects or substances, watch pornographic films or photographs, etc. Another item asked whether there had been physical aggression during the sexual abuse. Optional answers were the same as for the Physical IPV main category.

Physical IPV during pregnancy (34 items)

This part covers all the types of physical IPV mentioned above, but also has (5) items specific to this section. For instance, one item addresses the type of consequences the physical IPV had for the fetus/born child. Optional answers were: none, abortion, stillborn, premature birth, harm during delivery, underweight, and an open “other” category. Another example of an item in this section is: was the physical IPV triggered by any specific interaction or occasion? Optional answers were: no, when telling him you were pregnant, when knowing the sex of the fetus, and other. A third item relates to how often the participant has been physically assaulted by her partner during the pregnancy. Optional answers were: only once, occasionally, almost every month, more than twice a month, and more than once a week.

Psychological IPV during pregnancy (30 items)

This part contains all the types of psychological IPV mentioned above, and the items related to pregnancy presented for physical IPV.

Sexual IPV during pregnancy (34 items)

This part covers all the types of sexual IPV mentioned above and the items addressing pregnancy presented for physical IPV.

Lifetime History of Victimization

Childhood victimization (15 items)

Each participant was asked separate questions about the incidence of physical, psychological, and sexual abuse during her childhood. Questions covered the following aspects: If she ever had been subject to (1) physical, and/or (2) psychological, and/or (3) sexual violence by (a) her father or (b) another of her mother’s intimate partners before she was 12 years old. The same questions were also posed concerning (c) her mother or any (d) other of her father’s intimate partners, and for (e) violence by her siblings. Optional answers were: never, occasionally, and frequently.

Victimization in adolescence (15 items)

Questions identical to those concerning childhood victimization were asked one by one, but this time they referred to when the woman was between 12 and 17 years old.

Witnessing inter-parental violence (24 items)

Data were obtained about the incidence of violence between the parents in the woman’s family of origin. Questions covered information about the type of violence, the perpetrator, and the victim.

Violence by previous partners

In cases where a woman had been in more than one partnership, she was asked if she had been a victim of IPV in previous partnerships too.

Partner's use of violence in other relationships

Each woman was asked if she knew about other relationships (friends, relatives etc) in which her partner had been violent. Three questions covered the following types of violence: (1) physical, (2) psychological, and (3) sexual violence in other relationships. Optional answers for each type of violence were: against her children from previous relationships, his children from his previous relationships, common children, relatives, colleagues, friends, previous partner, and others.

Exposure of the women’s children to IPV

Data were obtained about incidence of each woman’s child or children witnessing IPV against her from her last partner. Optional answers were: never, occasionally, and frequently. Other separate questions addressed the following aspects: age of each child the first time he/she witnessed/was exposed to (1) physical, (2) psychological and (3) sexual IPV. Another question addressed each child’s behavioral responses when he/she was exposed to each of the three main categories of IPV. Optional answers were: cried, shouted, tried to defend you by attacking the aggressor, tried to defend you by getting between you and your partner, defended you verbally, called for help, and other kind of behavior.

Child victimization

Women were asked about their partners’ physical, psychological, and sexual victimization of children living in the family. The questions were restricted to children as victims. Apart from that they were identical to the ones used for “Partner’s use of violence in other relationships” described above.

Statistical Analysis

Since 90% of the participants had been victims of combinations of physical, psychological, and sexual IPV, statistical tests for related samples were used. The McNemar test was used as a non-parametric alternative for testing differences between IPV categories measured in two related samples for binary data. For continuous data in two related samples, the Wilcoxon signed-rank test was chosen when the scores did not meet the assumptions for parametric tests. We used the Friedman test for more than two related samples. Mann–Whitney U tests were estimated to control for possible independent group differences for variables with non-parametric distributions. The Chi Square Test was used for nominal data and unrelated groups. As a measure of non-parametric correlation for ranked variables Kendall’s tau-b was used. Finally, univariate and multivariate logistic regression analyses were conducted with IPV victimization in more than one partnership as the dependent variable. Only significant variables from the univariate logistic regression analyses were used in the multivariate logistic regression analyses. All statistical analyses were performed using the statistical program package SPSS, version 15.0.

Results

Characteristics of IPV

Detailed information about physical, psychological and sexual IPV is given in Tables 2, 3 and 4. Including the three categories of violence during pregnancy, the women (n = 157) were exposed to 16 different combinations of physical, psychological, and sexual violence. Percentage distributions of combinations of the three main categories of IPV are presented in Fig. 1.

Fig. 1
figure 1

Combinations of physical, psychological and sexual IPV the women (n = 157) had been subjected to

Full size image

Not every woman who had been exposed to physical violence had also been exposed to separate incidents of psychological violence beyond that caused by the physical abuse. Ninety percent had been targets of physical violence, 96.8% of psychological violence, and 36.3% of sexual violence.

Physical violence

The most frequent violent acts are presented in Table 2. Bruises (86.5%) were the most frequent injury after physical violence, but more than 40% of the women had injuries such as swellings, wounds, bumps, and bleeding, 22.7% had fractured/broken bones, and 13.5% had dislocations. A high proportion of the women (47.5%) had to see a doctor, and 19.9% had to see a psychologist or psychiatrist as a consequence of physical violence. Almost 43% had made a formal complaint to the police, whereas 22.9% had contacted the police at least once, but without making any formal complaints. Over 32% received an emergency alarm from the police and around 31% of the partners had visiting restrictions imposed by the police.

Psychological violence

Verbal attacks, forced control and power measures, and homicidal threats were the most frequent violent acts (see Table 3). The proportion of the women who had sought professional help was high: 32.2% had seen a psychologist or psychiatrist, 27.6% had seen a medical doctor, and 22.1% had seen police.

Sexual violence

About 70% of the sexually abused women were also physically attacked at the same time. More than 50% were physically injured as a consequence of sexual violence and 18.2% needed medical treatment. Only one woman made an appointment with a psychologist or psychiatrist as a consequence of sexual violence, and none of the women had contacted the police after sexual IPV. About 30% had wounds, internal haemorrhages, and bruises. Over 10% attempted suicide as a consequence of the sexual violence. For 61 of the 157 (38.9%) women, unwanted sexual relationships with the partner occurred at least once a week.

An Interactional Perspective on IPV

About 25% of the sample (n = 157) had been exposed to IPV for more than 10 years. The severity (Friedman (54) = 28.353, p < 0.001) and duration (Friedman (54) = 19.810, p < 0.001) were significantly different in the three IPV categories. Thus, the women had been psychologically abused for a longer time period than they had been exposed to physical (W (137) = 3.716, p < 0.001), or sexual violence (W (54) = 3.728, p < 0.001). About 45% of the women who had been subject to psychological violence had been so on a daily basis. Psychological abuse was also more frequent than physical (W (137) = 8.865, p < 0.001), and sexual IPV (W (54) = 4.882, p < 0.001), and occurred with a higher regularity than physical violence (W (137) = 4.146, p < 0.001). More than 50% were able to predict IPV, and there were no differences between IPV main categories in this respect. The experience of being in mortal danger was highest for physical IPV (79.4%: McN (137) = 9.031, p < 0.005), and lowest for sexual IPV (39.3%: McN (54) = −4.491, p < 0.001). Less than 50% perceived that they were able to prevent the IPV incident. Women able to predict physical violence reported that this enhanced their ability to prevent it (χ 2 (141) = 18.530, p < 0.001). The perception of being able to stop sexual IPV was less frequent than the perception of being able to stop psychological IPV (W (54) = 2.590, p < 0.01), or physical IPV (W (54) = 3.094, p < 0.005). About 60% used coping strategies if IPV was predicted and if they felt at the time that they were able to prevent the IPV escalating further. On this point, there were no differences between IPV categories. Fifty-five percent (physical IPV), 57.9% (psychological IPV) and 75% (sexual IPV) reported that nothing they had tried had any effect as a coping strategy. The most effective coping strategy reported by the women was to give in and obey him (physical IPV = 13.5%, psychological IPV = 9.9%, sexual IPV = 14.3%), and less than 10% found escape, call for help, and counterattack to be effective. More than 95% of the women reported that any IPV was unacceptable.

Women's Feelings of Responsibility, Guilt and Shame

About 30% felt responsible for the IPV in which they had been involved. Perception of being responsible was higher for physical IPV than for psychological IPV (W (137) = 2.047, p < 0.05), and sexual IPV (W (54) = 2.828, p < 0.005). There was a trend that the participants felt more responsibility for psychological than sexual IPV (W (54) = 1.865, p = 0.062). The experience of being guilty was highest for physical IPV, compared with psychological (W (137) = 2.193, p < 0.05) and sexual IPV (W (54) = 2.874, p < 0.005). The perception of shame was lower for psychological IPV than for physical (W (137) = 3.316, p < 0.001) and sexual IPV (W (54) = 1.956, p < 0.05). Women who had been sexually abused were less likely to tell friends or family about their IPV experience than women in the two other IPV categories (McN (54) = 16.690, p < 0.001), whereas the highest proportion of women who informed members of their close private network was found in the psychological IPV group (McN (137) = 3.781, p < 0.05).

The Impact of Alcohol and Drug Intoxication

Forty-five percent of participants reported that they did not use alcohol. Specific information about drinking habits indicated that 55% used alcohol less than once a month. There is a significant association between participants’ general perception of their alcohol consumption (“I am a non-user”, “I am a sometimes user”, etc.) and the specific information they gave of their drinking habits in terms of how often they had consumed alcohol per week during the last 12 months (Kendall's tau-b = 0.631, p < 0.001). More than 70% said that, to their knowledge, neither they nor their partner had consumed any alcohol before the last IPV episode. For detailed information about alcohol and drug intoxication see Table 5. The association between alcohol and IPV seems to vary across the three IPV categories (Friedman (54) = 9.846, p = 0.007) for the partners, and for the women (Friedman (54) = 6.500, p = 0.039).

Table 5 Participants’ and partners’ alcohol and drug intoxication before index IPV episode across main categories of IPV
Full size table

There seems to be a strong association between women’s and partners’ consumption of alcohol. If the woman had consumed alcohol there was a 67-fold increased in the likelihood that the partner had also consumed alcohol, and vice versa (OR = 66.667, CI = 8.422–527.700, p < 0.001).

IPV in More Than One Partnership

Only 22.9% of participants had experienced IPV by previous partners, and 66.9% had no experience of IPV in previous partnerships. About 10% did not have any previous partner experience.

Comparison of Women With or Without IPV in Previous Partnerships

Information about IPV victimization in more than one partnership is given in Table 6. Women who never had a prior partnership were eliminated from these analyses. The most striking finding was that women who had been subjected to sexual abuse by their father, mother, or a parent’s intimate partner were at almost 25 times increased risk of having been an IPV victim in more than one partnership. Women who had been subjected to childhood physical violence or exposed to their parents’ (physical) IPV ran a 2.6 and 2.9 times increased risk, respectively, for IPV in more than one partnership. Exposure to parents’ physical IPV increased the risk of having been victimized in more than one partnership more than having been subjected to childhood physical victimization. There also seems to be a hierarchal association among sexual, physical and psychological childhood victimization and IPV victimization in more than one partnership. Results indicate that if the women had been victims of only psychological violence there is a trend for increased probability of IPV victimization in more than one partnership (OR = 2.483, CI = 0.968–6.370, p = 0.059). If the woman had been the victim of psychological, and physical childhood violence, only physical violence increased the probability of IPV victimization in more than one partnership (OR = 2.564, CI = 1.016–6.471, p = 0.046). If the woman had been the victim of sexual, physical and psychological childhood violence, only sexual violence increased the probability for IPV victimization in more than one partnership (OR = 24.990, CI = 2.048–304.997, p = 0.004).

Table 6 Multivariate logistic regression analyses of the association between IPV victimization in more than one partnership and violence victimization and exposure to parents IPV during childhood, demographic factors, and current alcohol use
Full size table

Victimization during childhood and adolescence

About 30% of the women had been psychologically, physically, and/or sexually victimized by fathers and/or mothers. Before the age of 12 years, fathers were more often physical perpetrators than mothers, but after that there was no significant difference between mother and father (W (140) = −1.817, p = 0.069). This was caused by a decrease in physical abuse caused by fathers (W (140) = 1.985, p < 0.05). There were no such differences in psychological childhood victimization. Whereas, 12.5% of the sample (n = 157) had been sexually abused by their father, only one woman had been sexually abused by her mother.

Witness to/exposed to violence between parents

Less than 30% of the women had been exposed to physical, psychological and/or sexual violence between their parents. In most cases, physical violence was perpetrated by the father against the mother (W (140) = 3.934, p < 0.001). The participants reported that they had more often witnessed physical violence by their father against their mother before the age of 12 than later in life (W (140) = 2.952, p < 0.005). For psychological violence, the difference between mother and father as perpetrator was smaller, but we still found a significant difference for witnessing more violence from the father against the mother than vice versa (W (152) = 2.841, p < 0.005). Only one woman had witnessed sexual violence between parents.

Partners’ use of IPV in his previous relationship

According to the women, 35.8% of the partners had also used physical IPV against previous partners. The corresponding figures for psychological and sexual IPV were 26.5% and 20.4%, respectively.

Discussion

Main Findings

The purpose of this article is to present results from interviews with 157 women about their experiences with three aspects of IPV: (1) responses and conduct before, during and after the IPV, (2) impact of alcohol and drug intoxication, and (3) revictimization. The main findings can be summarized as: (a) The nature and characteristics of the IPV interactions were complex and heterogeneous. There were significant interactional differences between the IPV categories concerning the women’s responses and conduct before, during and after the IPV. (b) More than half of the women reported that they were able to predict IPV, and there were no differences between IPV main categories in this respect, (c) Less than 50% perceived that they were able to prevent the IPV incident, and these participants reported that this enhanced their ability to prevent it. About 60% used coping strategies if IPV was predicted and if they felt at the time that they were able to prevent further IPV escalation. (d) The impact of alcohol and drug intoxication was relatively small on the occurrence of IPV. About 75% reported that alcohol was not consumed before the index IPV incident, neither by the perpetrator nor by herself. (e) The association between alcohol and IPV seems to vary across the three IPV categories. (f) Only 23% of the women had experienced IPV by previous partners. (g) Women who had been subject to childhood sexual abuse in family of origin were at almost 25 times increased risk of IPV victimization in more than one partnership. Exposure to a parent's physical IPV increased the risk of victimization in more than one partnership significantly more than if the woman had been subject to childhood physical victimization.

An Interactional Perspective on IPV

Results from this study suggest that from an interactional perspective, the nature, and characteristics of IPV interactions were not uniform, even when we limited our investigation to a sample of help-seeking women. Other research has also come to a similar conclusion (Winstok 2007; Dutton and Corvo 2006). Our interactional research approach yielded some interesting results in terms of patterns of dynamic factors linked to the three IPV main categories. The finding that a woman who had been subject to different IPV categories on separate occasions had different experiences and perceptions of the escalation, interaction, and consequences concurs with the growing evidence that behaviors and perceptions of behaviors in intimate relationships should be investigated as separate entities (Rhoades and Stocker 2006). One example is the finding that physical IPV was the main category that resulted in highest levels of feeling responsible and guilty for the incident. Compared with the other two main categories, physical IPV was reported to have shorter duration, and lower frequency and regularity. Our study did not confirm earlier findings that physical IPV may be easier to predict than psychological IPV (Schumacher et al. 2000b). However, Campbell’s (2004) finding that 50% of female survivors of IPV were able to predict IPV incidents was supported by our results. The direct association between IPV and life stressors, conflicts and problem solving seems to be stronger in studies where IPV perpetrators and/or couples are interviewed (Cano and Vivian 2001; Richardson and Hammock 2007; Wilkinson and Hamerschlag 2005). Studies using this research design have disclosed consistent differences between men’s and women’s perceptions of motives for IPV (Winstok 2007). Other research has suggested that female victims of IPV are the ones best able to predict their risk of re-assault. Our findings indicated that if the women were able to predict the physical violence they reported, their ability to prevent it was enhanced (Campbell 2004). This result suggests that the women acknowledged the existence of a link between recognizing warning signs in their partners and coping with the threat of IPV. The integration of this link as a therapeutic focus in the treatment of couples with IPV problems may be worthwhile.

More than half of the women victimized by physical IPV experienced that nothing had any effect as a coping strategy, and to “give in and obey him” was reported as the most efficient coping strategy. Three out of four female survivors of sexual IPV reported that nothing they could think of had any effect as a preventive measure. More than half of the women who felt that they could stop sexual IPV reported that in a way they could handle the IPV by “giving in and obey him”. This would turn the interaction into a more “bearable” kind of rape, because it protected against being subject to the physical IPV that would follow any resistance. Even if the women reported that they had a limited ability to predict and cope, the negative long-term prediction “Did not believe that the IPV would stop” was the most frequent reason for having decided to terminate the partnership. In sum, the findings of our investigation support arguments for an integration of different approaches in an interactional understanding of IPV based on the main idea that aggression involves an indispensable and continuous interaction between individuals and the various situations they encounter e.g., (Winstok 2007).

The Impact of Alcohol and Drug Intoxication

Most of the participants in our study reported that they did not come from problem drinking relationships or partnerships with drug-related problems. This, of course, limits the generalizability of our findings concerning the impact of alcohol and drug intoxication. The low prevalence of substance abuse in our study may reflect underreporting or actual cultural differences between Norway and other countries. Findings from a recent meta-analysis suggested that increases in drug use and drug-related problems are significantly associated with IPV (Moore et al. 2007). However, with the exception of a medium-strength link between cocaine use and some of the IPV categories, the effect size of the association was relatively small. Even if our results indicate a relatively small impact of alcohol and drug intoxication, there is some evidence that supports other investigations that have found a stronger relationship between men’s substance use (alcohol and/or other drug use) and physical IPV, than for the other IPV main categories (Testa 2004). Our results also concur with previous findings of a possible relationship between men’s substance use and sexual aggression, and that women’s alcohol use increases the risk for physical victimization (Testa 2004). In our sample of help-seeking women, we did not find evidence of a relationship between women’s substance use and being subjected to sexual abuse. Still, we found that alcohol use increased the probability of IPV victimization in more than one partnership for the two other main categories of IPV (physical and psychological). Our sample was characterized by a high proportion of mothers, good social support, and average level of education and employment. This fact may offer support to the hypothesis that the lower number of social roles and the fewer responsibilities a woman has, the higher the inclination to report problematic drinking (Kerr-Corrêa et al. 2007). Even if substance use may be a risk factor for IPV, the fact that 75% of the IPV was not related to substance use indicates that other interactional and interpersonal risk factors should be explored. Fals-Stewart’s model of alcohol use, IPV and antisocial personality among IPV perpetrators may be a relevant approach to this (Klostermann and Fals-Stewart 2006). Another possibility is to explore whether differences emerge about warning signs and vulnerability situations when comparing IPV interactions with and without substance abuse involved.

IPV Victimization in More Than One Partnership

As found in other research, our results indicate that a subgroup of women return to or get involved in new violent relationships (Rhatigan et al. 2005). However, frequency of women victimized in more than one partnership was less than 25%. The same trend was found in an investigation of women from Norwegian shelters (Jonassen 2007). Other studies, however, report higher frequencies (Rhatigan et al. 2006). Our findings about victimization in more than one partnership support previous findings that women with a history of childhood sexual abuse, childhood physical victimization, or witnessing physical IPV run a higher risk of revictimization than others (Griffing et al. 2006; Messman-Moore and Long 2003). Our results suggest that being a witness or exposed to parents’ physical IPV is a stronger predictor of adult IPV victimization in more than one partnership than being a childhood victim of physical violence. This predictor of revictimization has also been reported in other research (Filipas and Ullman 2006; Kogan 2005; Macy 2007; Messman-Moore and Long 2003).

According to our investigation, supported main income is one of few demographic factors that increase the probability of IPV victimization in more than one partnership (Schumacher et al. 2000a). It is speculated here that this may indicate that being economically independent of the partner may provide an easier way out of a relationship for the woman when the early warning signs of imminent IPV problems comes to the surface. However, our findings concurs with previous findings that regular drinking increases the probability of revictimization (Messman-Moore and Long 2003). Research indicates that having resided with an abusive partner may place women in greater danger when attempting to leave the relationship (Wilkinson and Hamerschlag 2005). Termination of partnerships also increases the risk of mortal danger (Wilkinson and Hamerschlag 2005). In our study, the women’s explanations of the pros and cons of relationship termination indicate that non-pathological approaches like reasoned action or planned behaviour, and investment models, may be better for understanding this complex and multifaceted decision than theories of learned helplessness, traumatic bonding, and psychological entrapment (Rhatigan et al. 2006).

Clinical, Legal and Policy Implications

Using interview questionnaires with an interactional and interpersonal perspective on IPV may also be useful in the clinical context. This approach may give help-seeking women the opportunity to describe and analyse their IPV interactions in a multifaceted way that is imperative to understand and cope with such traumatic experiences. Systematic and detailed assessment, with an interactional approach to IPV, may guard against some long-lived clinical myths, such as the one that links IPV solely to problem drinking in male partners or in both partners. Childhood sexual abuse, and childhood violence victimization, and exposure to IPV appear to be potent predictors of IPV victimization in more than one partnership. Accordingly, it is our opinion that questions about these subjects should routinely be asked of help-seeking women prior to selection of adequate treatment for IPV sequela.

If an IPV victimized mother has a high risk of IPV victimization in more than one partnership, this should also be taken into consideration in assessments of risk of future child neglect. If further research concurs with our finding that being a witness or exposed to parents’ physical IPV as a child may pose a higher risk of re-entering IPV relationships, this may be taken into consideration in custody and court case procedures.

Methodological Limitations

Findings from our sample of help-seeking women do not necessarily generalize to IPV victims outside Norway due to cultural and social differences. Still, since help seeking has been associated with IPV severity, the opportunity to apply our findings, at least to samples of similar IPV severity, appears relevant (Haaland et al. 2005; Pape and Stefansen 2004; Sartin et al. 2006). We also relied on women’s retrospective self-report of IPV experiences, which may have been subject to recall bias. In other studies, this has been associated with underreporting of IPV (Schwartz 2005). There are also methodological limitations connected to only interviewing the victims and not others involved, such as the partner and, in some cases, the children. Self-report surveys that interview both partners tend to reveal higher rates of IPV and disclose consistent differences between men and women's perception of IPV (Hicks 2006; Archer 2000; Moore et al. 2007). For ethical and safety reasons, only the women were interviewed in this study. In our opinion, the inclusion of perpetrators would have increased the risk of obtaining a selected and non-representative sample of help-seeking women. The fact that one researcher did all the interviews may have increased the risk of systematic measurement error. However, analyses of the score variances yielded no indications of systematic measurement error. Reliability of the study is also enhanced by the use of only one interviewer. Finally, the cross-sectional design of our study has limitations concerning the measurement of the causality between variables. For instance, a prospective longitudinal study would work much better to find out whether IPV revictimization causes drug problems and low income, or vice versa.

Further Research

Large-scale studies with an interactional and interpersonal perspective on IPV are needed. To study the problem of IPV, the field must move beyond a reliance on small-scale convenience samples. Studies with an adequately large and diverse sample, preferably of randomly selected participants, are needed. The interactional approach used in our investigation invites researchers to confront the inherently difficult methodological issues that follow the event perspective on IPV. In particular, we recommend that researchers address the issue of gathering data from both partners. It is of special clinical relevance to explore the two partners’ perceptions of early and immediate warning signs, the role of vulnerability situations, and the predictability of IPV. It would also be useful to conduct comparative studies of different subgroups concerning topics such as: IPV victimization in more than one partnership, marital status, IPV duration, ethnicity and immigration, the impact of pregnancy and motherhood, and help-seeking patterns. To investigate childhood victimization and exposure to IPV as predictors of adult IPV victimization is instrumental to a better understanding of the underpinnings of the cycle of violence. Research focusing on the consequences of unwanted sexual relationships with partner in IPV partnerships is also needed.

The evolving nature of IPV requires longitudinal studies on violent events. Longitudinal data is fundamental to examining the interaction of individual, situational, and contextual variables. Although this design will be costly and time consuming, it has the potential to bring substantial improvement in theory development, measurement, and new suggestions for a policy to develop more efficient prevention and support strategies.