Abstract
Child sexual abuse (CSA) is a well-established risk factor for adult victimization in women, but little is known about the importance of relationship to perpetrator and exposure to other violence types. This study interviewed 2,437 Norwegian women (response rate = 45.0%) about their experiences with violence. Logistic regression analyses were employed to estimate associations of multiple categories of childhood violence with adult victimization. Women exposed to CSA often experienced other childhood violence, and the total burden of violence was associated with adult rape and intimate partner violence (IPV). Researchers and clinicians need to take into account the full spectrum of violence exposure.
Introduction
Childhood violence and abuse have been linked to a wide range of adverse outcomes in adulthood, such as adult mental health problems (Clark, Caldwell, Power, & Stansfeld, 2010; Cohen, Brown, & Smailes, 2001; Kessler et al., 2010), suicide attempts (Dube et al., 2005), somatic problems (Dong et al., 2004; Felitti et al., 1998), and various adverse functioning issues, including intimate relationship problems (Colman & Widom, 2004; Dennerstein, Guthrie, & Alford, 2004), work participation (Strøm et al., 2013), and exposure to new adverse experiences (Widom, Czaja, & Dutton, 2008). Specifically, there is ample evidence that exposure to childhood violence is a risk factor for adult violent victimization (Barnes, Noll, Putnam, & Trickett, 2009; Classen, Palesh, & Aggarwal, 2005; Kimerling, Alvarez, Pavao, Kaminski, & Baumrind, 2007). This phenomenon is known as revictimization, and it is associated with mental health problems in adult life (Jonas et al., 2011; Kimerling et al., 2007). The results from the robust research tradition on revictimization imply that childhood experiences with violence make an individual vulnerable to new experiences of violence and abuse. Thus, it seems that childhood experiences are carried into adulthood, leading to an increased likelihood of re-exposure to violence. It is uncertain, however, which pathways are involved and which aspects of violence are most important for adult victimization.
Traditionally, child sexual abuse (CSA) has been the most studied childhood event, and its association with adult sexual violence has been repeatedly identified (Classen et al., 2005). CSA is quite prevalent in the general population, particularly in girls. Prevalence estimates from different countries suggest that CSA occurs in 7-36% of girls (Finkelhor, 1994). Studies from Norway show comparable results, indicating 9-11% CSA in girls (Mossige & Stefansen, 2007; Steine et al., 2012). Factors that may represent pathways between CSA and adult victimization include risk behavior (Walsh et al., 2013), posttraumatic stress symptoms (Ullman, Najdowski, & Filipas, 2009), and learning processes, such as learned helplessness (see review by Messman-Moore & Long, 1996). Characteristics of the CSA experience may influence the risk of revictimization. For example, betrayal trauma theory states that the impact of trauma can depend not only on fear but also on betrayal. Dependency is crucial to betrayal; thus, the most devastating psychological effects of CSA will occur when a child is abused by a caregiver upon whom she is dependent (Freyd, 1996). Other trauma theorists concur that sexual abuse has particularly damaging effects when perpetrated by parents. For instance, Herman (1992) compared child abuse by parents to political captivity and described children as captives due to their dependency. It may also be that CSA perpetrated by parents is more severe in terms of early onset (Trickett, Noll, & Putnam, 2011). Evidence diverges on whether health consequences are more severe when the perpetrator of CSA is a parent (Bal, De Bourdeaudhuij, Crombez, & Van Oost, 2004; Edwards, Freyd, Dube, Anda, & Felitti, 2012; Ketring & Feinauer, 1999; Lange et al., 1999; Lawyer, Ruggiero, Resnick, Kilpatrick, & Saunders, 2006). There is some empirical support for the suggestion that the experience of parental trauma may result in a compromised capacity to detect social betrayal, possibly increasing the risk of later revictimization (DePrince, 2005). Gobin and Freyd (2009) found that individuals who experienced high-betrayal trauma were more likely to experience a subsequent high-betrayal trauma, such as intimate partner violence (IPV), in adulthood. Thus, there is some indication that the perpetrator relationship in CSA is important for the revictimization risk, though the literature remains scarce. In particular, there is a lack of studies investigating the victim’s relationship to the perpetrator and revictimization in light of exposure to other categories of childhood violence, such as physical or psychological violence, or childhood neglect.
Violence and abuse are currently conceptualized in a variety of ways, and concepts may differ between those researchers focusing on children and those focusing on adults, as well as between various academic and clinical fields. The World Health Organization (WHO) defines violence against children as encompassing physical and psychological violence and childhood neglect, as well as CSA (WHO, 2002), thereby employing a comprehensive definition of violence. This definition was used in the current study, and we use the term violence as an overarching concept including physical violence, witnessing parental IPV, psychological violence, sexual abuse, and neglect. The focus on CSA in revictimization literature has recently been expanded, and researchers have investigated revictimization in relation to a broader range of childhood violence (Whitfield, Anda, Dube, & Felitti, 2003; Widom et al., 2008). Several studies have found that other forms of childhood abuse are associated with adult victimization, such as child physical abuse (Fiorillo, Papa, & Follette, 2013; Messman-Moore, Walsh, & DiLillo, 2010), childhood neglect (Villodas et al., 2012), and emotional abuse (Obasaju, Palin, Jacobs, Anderson, & Kaslow, 2009). One prospective study found that although all forms of childhood abuse were associated with adult victimization, individuals exposed solely to childhood neglect had significantly more revictimization than those exposed solely to physical abuse or sexual abuse (Widom et al., 2008). In addition, exposure to various categories of child abuse and neglect tend to overlap (Herrenkohl & Herrenkohl, 2009; Kessler et al., 2010); that is, the experience of one form of childhood abuse increases the likelihood of experiencing another. CSA may be only one part of the violence a child experiences.
Several studies have found an additive effect of multiple forms of abuse on adult health outcomes; for example, the Adverse Childhood Experiences study (ACE study) found associations between number of adverse experiences in childhood (including sexual, physical, and psychological abuse, and parental IPV) and diseases such as depression, alcoholism, ischemic heart disease, cancer, and liver disease in adulthood (Anda et al., 2006; Felitti et al., 1998). This underscores the importance of studying not only various categories of childhood violence but also their co-occurrence when adult health is the focus. Little is known about the way in which the combined burden of various categories of childhood violence relates to adult victimization. However, there is some support for the hypothesis that individuals who experience multiple forms of abuse are at a heightened risk for revictimization (Whitfield et al., 2003; Widom et al., 2008).
Given what we know about the overlap between different forms of childhood adversity, their additive effect, and the potential importance of the relationship with the perpetrator, there is a need for revictimization research that encompasses a comprehensive assessment of childhood experiences of violence. We investigated adult victimization and its association with CSA, relationship to the perpetrator, and other forms of parental childhood violence in a recent cross-sectional general population study of Norwegian women’s experiences with violence. The study thus focuses on the overlap between various childhood and adult victimization, and does not aim to investigate mechanisms by which such overlap occurs. We examined the following research questions:
Method
Study and Response Rate
The current sample comprised 2,437 women between the ages of 18 and 75 (M age 45.2, SD 15.8). This sample is part of a larger study that assessed violence and sexual abuse in a sample of 6,500 Norwegian men, women, and youths. The response rate among those reached by telephone, which is comparable with random digit dialing procedures, was 45.0% for women and 40.8% for men. In a previous publication, we investigated selection bias by analyzing whether our sample differed from the general Norwegian population, and whether responders differed from non-responders, in characteristics such as marital status, education, and income. We found indications of a moderate positive bias in terms of marital status and income compared with the general population. Once we had established contact, women were more likely to be willing to participate than men, and responders were slightly older than non-responders. We also investigated whether our study variables correlated with the number of calls necessary to obtain contact with participants, under the hypothesis that the more calls needed to reach an individual, the more similar that individual would be to non-responders. There were few significant differences in the number of calls necessary to contact those that had been exposed to violence compared with those who had not been exposed, though women who had experienced physical violence in childhood seemed to be slightly more available than women who did not report such experiences (Thoresen, Myhre, Wentzel-Larsen, Aakvaag, & Hjemdal, 2015). Most participants (65.4%) were either married or cohabited with a partner. Only a few participants (3.9%) had a non-Nordic cultural background, defined as having two parents born outside the Nordic countries (of these, most parents were born in Europe). Approximately half (52.6%) had completed higher education after high school (university or university college), and most (90.6%) perceived their financial situation as average or above. Furthermore, 247 women (10.1%) had experienced CSA before the age of 13, 150 (6.2%) women had experienced at least one forcible rape in adulthood, and 224 (9.2%) had experienced IPV (Thoresen et al., 2015).
We used a computer-assisted telephone interview (CATI), a method that allows for flexibility in the interview. Our manual was designed after a strategy developed by Kilpatrick and colleagues (Kilpatrick et al., 2003; Resnick, Kilpatrick, Dansky, Saunders, & Best, 1993), where endorsement of items asking about experiences leads the respondent to a series of supplementary questions about events. Questions about experiences with violence were, as much as possible, behaviorally specific (Kilpatrick et al., 2003). Although the telephone interview was designed according to the second wave of the National Adolescent Study (Kilpatrick et al., 2003), questions were adapted to fit the Norwegian context. In addition, the interview was expanded with a broad assessment of childhood violence.
Interviewers were instructed to make sure that participants had sufficient privacy when answering questions, by asking whether the participant was alone and could answer the survey without being overheard by others. If the participant did not have sufficient privacy, the interviewers offered to call back at a more suitable time. In addition, questions were designed so that answers were neutral (e.g., “yes” or “no”), ensuring further privacy for the respondents. At the end of the interview, all participants were asked a series of follow-up questions designed to assess their need for assistance. Those who were in need of assistance were offered referrals to mental health services. The study was approved by the Regional Committee for Medical and Health Ethics in Norway.
Measures
CSA was measured using the following question: “We will now ask you a few questions about sexual acts that may take place during childhood. Sometimes children can be tricked, rewarded or threatened into sexual acts that they do not understand or are not able to stop. Before you were age 13, did anyone who was 5 or more years older than you ever have sexual contact with you?” This question was taken from the National Stressful Events Web Survey (Kilpatrick, Resnick, Baber, Guille, & Gros, 2011). All women who answered this item affirmatively were defined as having been exposed to CSA. Those who were exposed to CSA were asked follow-up questions. These questions included relationship to the perpetrator, age when the event happened, and whether it was a single event or an event that occurred multiple times. Relationship with the perpetrator was recorded on a comprehensive list of potential relationships and, for the purpose of this article, categorized into parental relation (biological parents, step-parents, or mother’s or father’s girlfriend or boyfriend), other known perpetrators (other family members or people the respondent knew, such as teachers, leaders of activities, friends, and neighbors), or strangers (both children and adults). In the category “other known perpetrators,” the most common groups were adult relatives (other than parents) and acquaintances. Characteristics of abuse included age of onset, for the purposes of this study dichotomized as before the age of 10 or older (Kliegman, Nelson, & Behrman, 2011); whether it was a single event or multiple incidents; whether abuse involved penetration; whether the respondent feared for her life or feared serious injury during the abuse; and whether she sustained physical injuries. We considered early onset, multiple incidents, penetration, fear for life or serious injury, and sustaining physical injury as indicators of the severity of abuse.
Parental physical violence was defined by the following four items: having been beaten with a fist or hard object, kicked, beaten up, or otherwise physically attacked by a caregiver (Kilpatrick et al., 2003). An affirmative answer on at least one item was defined as having experienced parental physical violence. Psychological violence was measured by one item from the Stressful Life Event Screening Questionnaire (Goodman, Corcoran, Turner, Yuan, & Green, 1998), asking whether a caregiver repeatedly ridiculed, put down, ignored the respondent, or told the respondent that she was no good; this item was scored according to a yes/no format. Emotional neglect was measured by one item asking respondents how often in their childhood they felt loved. Responses were given on a 5-point Likert-type scale ranging from “never” to “very often or always” and were coded as emotional neglect if “never,” “rarely,” or “sometimes” was endorsed. Parental IPV was defined by an affirmative answer on at least one of the following five items: having seen or heard one parent or caregiver slapping the other, beating the other with a fist or hard object, kicking the other, choking the other, or otherwise physically attacking the other (Kilpatrick et al., 2003).
Adult rape: Respondents were asked questions about four forms of rape: “Has anyone ever forced you into (a) intercourse, (b) oral sex, (c) anal sex, or (d) put fingers or objects in your vagina or anus by use of physical force or by threatening to hurt you or someone close to you?” If a respondent had experienced at least one of these items when she was 18 years or older, the event was defined as adult rape. Adult IPV: Respondents were asked six questions about violent acts they might have experienced: having been beaten with a fist or object, kicked, choked, beaten up, threatened with a weapon, or otherwise physically assaulted after they had turned 18. All items had yes/no response categories. Relationship to the perpetrator was asked in supplementary questions, and respondents who identified a partner or ex-partner as the perpetrator were categorized as endorsing adult IPV. Measures of adult rape and adult IPV were adapted from the National Adolescent Study (Kilpatrick et al., 2003).
Adjustment variables were age, ethnicity (having a non-Nordic background, that is, having two parents born outside of Norway and the Nordic countries), parental mental health problems (as measured by Felitti et al., 1998), and education (high school completion).
Statistical Analyses
In Tables 1 and 2, groups of CSA perpetrator relationships were compared. Some respondents experienced CSA both from parents and from other people they knew or from known and unknown perpetrators. To ensure that each respondent was only represented in one category, we represented the relationship with the perpetrator in a hierarchical variable in which the closeness of the relationship determined where a respondent was placed in cases of overlap. Parental relationships were defined as the closest, whereas other known perpetrators were defined as less close than parents but closer than strangers. Thus, a respondent with both a parental perpetrator and another known perpetrator was placed in the parental perpetrator category, whereas a person with both another known perpetrator and an unknown perpetrator was placed in the known perpetrator category. Overall, 17 women reported such an overlap (eight had an overlap between parents and other known perpetrators, and nine had an overlap between other known perpetrators and strangers). In the multivariable models (Table 4), relationship with the perpetrator was not defined according to this hierarchy. Instead, CSA perpetrated by someone who was not a parent was included as a separate dichotomous variable, whereas CSA from a parent was included in the parental violence variable. Thus, a respondent with both parental and other CSA was scored as exposed on both variables. One person did not report her relationship with the perpetrator and was excluded from the analyses.
Characteristics of CSA According to Relationship to the Perpetrator of CSA, n = 247 (total N = 2,437).
Note. CSA = child sexual abuse; IPV = intimate partner violence.
Any woman who reported that CSA was committed by a parent.
Women who reported that CSA was committed by someone they knew that was not a parent; if two categories were answered affirmatively and one was a parent, the respondent was categorized in the parental perpetrator category.
Only stranger(s) as perpetrator(s); if two categories were answered affirmatively and one was a parent or other known perpetrator, the respondent was categorized in the parental perpetrator category (if any CSA by parent) or other known perpetrator (if no CSA from parent but any from other known).
χ2 analyses between the three groups of perpetrator relationships.
Number of Other Categories of Parental Violence, by Perpetrator of CSA.
Note. CSA = child sexual abuse.
χ2 p value for difference between the three perpetrator groups.
We adjusted for sociodemographic variables (age and ethnicity), and for variables that may indicate social disadvantage in childhood (parental mental health and high school completion).
Chi-square tests were employed to test differences in event characteristics between different groups of perpetrator relationships (Tables 1 and 2). Where small cells occurred, exact tests were employed, using a Monte Carlo procedure with 100,000 replications if necessary. Logistic regression analyses were employed to test associations between various forms of childhood violence and perpetrator relationships with two dichotomous outcome variables: adult rape and adult IPV (Tables 3 and 4). Because the amount of missing information was very low in this sample (out of 2,437 respondents, 13 did not answer questions about adult rape, and three did not answer questions about adult IPV), complete case analyses were implemented. All regressions within the same table were run on the same selection of individuals.
Logistic Regression Analysis Displaying Associations Between Different Forms of Childhood Victimization and Adult Rape and IPV (n = 2,323).
Note. IPV = intimate partner violence; OR = odds ratio; CI = confidence interval.
Additionally adjusted for age, ethnicity, parental mental health, education, and each other.
Logistic Regression Analysis Displaying Associations Between Number of Categories of Parental Violence (CSA, Physical Violence, Psychological Violence, Emotional Neglect, and Parental IPV), Extra-Parental CSA, and Adult Rape and IPV (n = 2,345).
Note. CSA = child sexual abuse; IPV = intimate partner violence; OR = odds ratio; CI = confidence interval.
Additionally adjusted for age, ethnicity, parental mental health, education, and each other.
Reference category: No parental violence.
All analyses in the tables were performed using SPSS Statistics 20 for Windows.
Results
Characteristics of Abuse
Among the women with CSA experiences, most had experienced CSA from a non-parental known perpetrator. When CSA was committed by a parental perpetrator, it was more often severe in some characteristics of the event (more than one incident and injury sustained) than if it was committed by another known or unknown perpetrator. However, CSA was not more severe in terms of other characteristics (early onset, fear for life, or severe injury and penetration).
Relationship to the Perpetrator and Other Parental Violence
Women who had experienced CSA had been victims of other forms of childhood violence more often than women without such experiences (all χ2 p values <.001). As shown in Table 1, women who were sexually abused by their parents experienced all of these forms of parental violence to a greater extent than those who were sexually abused by other perpetrators. Table 2 presents the occurrence of non-sexual parental violence in the three perpetrator groups. Those who experienced CSA from a parental perpetrator experienced a high number of other categories of parental violence, with 85.7% experiencing at least one other category of parental violence and 34.3% experiencing three or more other categories. Children who were sexually abused by perpetrators other than parents also reported high levels of exposure to parental violence: 47.6% of those sexually abused by other known perpetrators and 57.5% of those abused only by strangers experienced at least one category of parental violence. Thus, all women who were exposed to CSA were highly burdened by other forms of parental violence, but none as much as the respondents who were sexually abused by their parents.
Adult Victimization
CSA was significantly associated with adult rape and IPV, which occurred 2-3 times more often in exposed respondents than in non-exposed respondents (adult rape: 18.4% in those exposed to CSA, 4.8% in those not exposed to CSA; adult IPV: 18.3% in those exposed to CSA, 8.2% in those not exposed; both χ2 p values <.001). The increased occurrence of adult rape and IPV was observed for all CSA perpetrator groups. There were no significant differences between the different groups of perpetrators in the occurrence of adult rape and IPV (χ2 p values .829 and .285, respectively).
Associations Between Childhood Violence and Adult Victimization
To compare different forms of childhood violence, we examined the association between CSA, non-sexual parental violence, and adult victimization (Table 3). CSA by different perpetrators was collapsed into “any CSA.” Before adjusting for each other, all measured forms of childhood violence were associated with both outcomes. CSA was associated with adult rape, as expected. Parental psychological violence and witnessing parental IPV were also significantly associated with adult rape after adjusting for the other categories of violence and age. CSA was also associated with adult IPV; however, after adjusting for the other categories of childhood violence and adjustment variables, the association was no longer significant. Parental psychological violence, parental emotional neglect, and witnessing parental IPV remained significantly associated with adult IPV.
The Total Burden of Childhood Violence and Adult Victimization
Table 4 presents the associations of the number of categories of parental violence and extra-parental CSA with adult victimization. Our results show that having experienced one category of parental violence in childhood, as opposed to no categories, was significantly associated with rape and IPV in adulthood. Furthermore, our findings suggest a graded relationship between the number of categories of childhood parental violence and both adult rape and IPV, where the odds of adult victimization increase with the number of childhood violence categories. Thus, in our data, the more categories of childhood abuse a woman experienced, the more likely she was to have been a victim of sexual or physical violence in adulthood. After adjusting for parental violence, extra-parental sexual abuse was significantly and uniquely associated with adult rape, though no longer significantly associated with adult IPV. Our findings are consistent with a graded relationship, although not all contrasts were significant.
Education may be on a causal pathway between childhood violence and adult victimization; for example, mental health problems and substance abuse resulting from childhood violence may make it more difficult for an individual to complete high school. Therefore, adjusting for education may represent overadjustment. We performed supplementary analyses without adjusting for education. These analyses yielded results that were almost identical to the full models, with highly overlapping confidence intervals.
Discussion
Revictimization is one of the main concerns facing women who have experienced violence. In the present study, we found that not only sexual abuse but also other types of violence in childhood were associated with adult victimization. The strongest association with revictimization was found for those who experienced multiple types of childhood violence.
We found that CSA from parents was associated with some, but not all, indicators of abuse severity. Thus, our findings were inconclusive regarding whether parental CSA is more severe than CSA perpetrated by other known or unknown persons. However, when we considered the co-occurrence of other categories of violence experienced in childhood, clear differences emerged between those abused by parents and those abused by others. It is important to note that in comparison with non-exposed women, all groups of CSA-exposed women, regardless of their relationship to the perpetrator, had an increased occurrence of additional childhood violence. However, women who had experienced parental CSA were particularly prone to report other types of parental violence, namely, emotional neglect, physical and psychological violence, and witnessing parental IPV. In fact, parental CSA rarely occurred alone. Rather, parental CSA seems to fit into a pattern of violence from parents. These results emphasize the particular vulnerability to other types of violence exposure in girls exposed to parental CSA.
Contrary to our hypothesis, revictimization in adulthood was not significantly more common among individuals who were sexually abused by parents. Betrayal trauma theory states that traumas high in betrayal, such as parental sexual abuse, might result in a reduced capacity to detect betrayal in interpersonal relationships, leading to revictimization in adulthood (DePrince, 2005; Freyd, 1996). However, children might experience a high degree of betrayal even when the perpetrator is not a parent. Perpetrators of CSA are typically persons the child trusts, depends upon, or cares for, such as other relatives or acquaintances.
Importantly, we found relatively high levels of exposure to other categories of parental violence among all CSA survivors. Perhaps non-sexual violence from parents is just as likely to create a sense of betrayal as parental CSA.
Our findings imply that both sexual and non-sexual violence in childhood are associated with adult rape and adult IPV. Childhood violence entails an increase in adult victimization that appears largely unspecific; for example, witnessing parental IPV in childhood is associated with adult rape, and childhood psychological violence is associated with adult IPV. However, not all categories of childhood violence were significantly associated with adult victimization in the adjusted model. The more categories of childhood violence a respondent had experienced, the more likely she was to have also experienced adult physical or sexual victimization. It seems that not only are categories of violence other than CSA comparably associated with adult victimization, but that the combination of various types of parental violence is particularly potent when adult victimization is the outcome. Thus, the additive effect of multiple categories of childhood adversity and violence that has been found on mental and somatic health outcomes (Anda et al., 2006; Felitti et al., 1998) seems to apply to various categories of victimization in adulthood as well.
Our findings underscore the need to assess childhood violence in a broad, comprehensive fashion, in line with the recommendations from Finkelhor, Ormrod, and Turner (2007). To better understand the impact of the violence children experience, a range of violent acts should be taken into account. In our study, the total burden of childhood violence was the most important factor for adult victimization. Thus, the adverse effect of multiple categories of childhood violence seems to be present in the general population as well as in more severely exposed populations, as shown by other authors (Widom et al., 2008).
A potentially causal relationship between childhood and adult violence is likely not simple and direct (Pratchett & Yehuda, 2011); many factors influence an individual’s vulnerability. The strong association between childhood violence and adult victimization, and their combined effect on health, nevertheless points to an opportunity for intervention. Clinicians working with children who have experienced one type of violence, such as CSA, can benefit from assessing experiences of parental violence in a comprehensive manner. Our findings imply that such assessment will be of particular importance when CSA was committed by a parent, although it is still recommended with non-parental perpetrators. Screening for violent experiences is not always done in child mental health clinics, and clinicians may experience ambivalence toward asking about such experiences (Hultmann, Möller, Ormhaug, & Broberg, 2014). The systematic use of a screening tool may help clinicians to assess these experiences in help-seeking children.
Understanding that childhood violence entails an increased risk of adult violence provides clinicians and others who work with exposed children with an opportunity to prevent subsequent violence and abuse. Our results emphasize children’s need for protection from further violence after experiencing a variety of violent events. In particular, children who experience multiple forms of violence are in need of intervention to prevent revictimization.
When working with adult victims of rape and IPV, clinicians could also benefit from a comprehensive assessment of experiences of childhood violence, so that they can select the appropriate interventions. In addition, being aware of the full range of childhood violence experienced by their adult patients may help therapists to better understand their patients’ current problems. Our findings imply that childhood experiences with violence should be a part of the screening of violence-exposed adults.
Revictimization in adulthood constitutes one of the many negative outcomes in the study of the consequences of childhood violence and abuse. In our opinion, studies of treatment approaches to trauma-related problems in children could benefit from including subsequent violence as an outcome, in addition to health.
This study focuses on the association between childhood and adult experiences with violence. Future prospective studies should identify mediators that may lie on the path between first exposure to violence or abuse and later victimization with a focus on individual coping ability, risk and protective factors in close relationships, and community factors and social or educational deprivation. Identifying these mechanisms will help target interventions to prevent negative long-term development in high-risk children.
This study has several important limitations. Because it is a cross-sectional study, we cannot imply causality. Individuals with experiences of violence in adulthood may recall their experiences of violence in childhood more easily, possibly affecting our estimates of association. The response rate of the study was such that more than half of those we reached by telephone declined to participate, which may have introduced selection bias to our sample. Unfortunately, lower response rates in telephone surveys seem to be a trend (Atrostic, Bates, Burt, & Silberstein, 2001). In studies of violence and abuse, it is hard to evaluate the validity of self-report, as there is no gold standard with which to compare. Nevertheless, there is no accepted alternative to self-report in these studies. The respondents’ lack of willingness to disclose highly sensitive information is perceived by some authors as a greater challenge than false positive reports (Fergusson, Horwood, & Woodward, 2000). We used behaviorally specific questions in this study, and previous studies have demonstrated that this strategy greatly increases participants’ disclosure (Fisher, Cullen, & Turner, 2000). Some studies have investigated test–retest reliability on self-reports of experiences with violence. The results from these studies indicate that people are just as likely to be inconsistent when answering questions about violence and abuse as when they are answering questions about subjects such as lifetime drug use or age of first alcohol use (see Thoresen & Øverlien, 2009, for a discussion). Nevertheless, retrospective report may be biased, as memories of past events may be influenced by current emotional states.
The hierarchical variable we used for Tables 1 and 2 to perform chi-square analyses might introduce a bias by shifting more serious violence (e.g., with multiple perpetrators) in the direction of parental perpetrators or other relatives or known perpetrators. When we performed the analyses for Tables 1 and 2 with the individuals who had experienced CSA with overlapping categories of perpetrators excluded, the results remained largely the same; thus, it is unlikely that our results can be attributed to the hierarchical variable. We lacked information about non-parental violence other than CSA, such as community violence or bullying. Our analyses show that the overlap between childhood and adult violence withstood adjustment for age, ethnicity, education, and parental mental health problems during childhood (Tables 3 and 4). Other indicators of childhood social disadvantage that we were not able to control for may also have influenced revictimization risk (e.g., parental income, parental education, and financial situation in childhood). Current social disadvantage, such as low income or unemployment, could not be used for adjustment, as they may be an outcome of violence exposure, rather than a confounding variable. Participants in this study are Norwegian women, and our results are not necessarily transferable to women from other countries and cultures.
The strengths of this study include the thorough assessment of childhood violence with questions about a variety of events and detailed information about experiences of violence, including the perpetrator relationship.
Footnotes
Authors’ Note
The Norwegian Ministry of Justice and Public Security had no involvement in the study design; in the collection, analysis, or interpretation of the data; or in the decision to submit the article for publication.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Data collection was funded by the Norwegian Ministry of Justice and Public Security. The work of the first author is funded by the Norwegian Women’s Public Health Association.
