The results do not yield a simple answer to the question of whether childhood victimization has differential effects by gender on psychiatric outcome. Instead, our findings suggest interesting gender differences in the effects of childhood victimization on psychiatric outcome depending on the age of participants at the time of assessment (i.e. youth versus adult). In general, adult samples show either that victimization is associated with greater risk in females than males (n=4) or that there are no differences by gender (n=7). Only three studies found increased risk in male adults, and one of these studies also showed increased risk in female adults depending on measured outcomes (i.e. victimization was associated with greater risk for social phobia and some substance problems in males, and dysthymia, anxiety disorders, PTSD, and some substance problems in females). By contrast, studies among youth indicate either that there are no gender differences (n=7) or that males tend to have worse outcomes than females (n=12). Although five youth studies demonstrate worse outcomes for females compared to males, four of these also reported worse outcomes in males depending on the type of victimization or outcomes assessed.
With respect to outcome, when gender differences were reported, they were distributed across both internalizing and externalizing categories for both genders. However, we again find evidence for interesting time-of-sampling differences. Within the adult studies that assessed internalizing outcomes, there was a trend for females to fare worse than males, whereas males tended to have worse outcomes than females within youth studies that assessed externalizing outcomes.
It is not clear why the findings are different for youth samples relative to adult samples. In studies with youth samples, the temporal proximity of the victimization experience to measured outcomes may play a role, whereas in studies with adult samples, there may be ‘forgetting’ of past victimization experiences. In the latter case, increased temporal distance from the event would be expected to attenuate the associations observed in youth. As has been noted previously in the literature, reliance on retrospective designs in studies of the psychological impact of childhood victimization poses serious concerns; namely, studies with adults have shown that as many as 38% of individuals with substantiated CSA histories do not recall the abuse accurately (e.g. Williams, 1994
; Widom & Morris, 1997
). This explanation, however, would not account for the gender disparity between youth and adult studies unless there are also gender differences in the effect of time on memory for victimization experiences. There is evidence to suggest that women exhibit better biographical memory for emotional events from childhood (both positive and negative) than do men (Davis, 1999
). However, to our knowledge, there is no evidence to suggest that such a gender difference is specifically observed in memory for victimization in childhood.
Another possible explanation is related to course of illness. When psychiatric outcomes are measured in closer temporal proximity to childhood victimization, rates of psychiatric symptoms and disorders may be higher, and in some cases may dissipate over time. Thus, among adult samples, it is only those symptoms and disorders that persist into adulthood (or have adult onset) that are being measured. The effects of childhood victimization on psychiatric outcome may therefore be different in symptoms and disorders that are short term relative to those that are long term or have delayed onset. This explanation would not account for the gender disparity between youth and adult samples unless there are also gender differences in the course of illness of common psychiatric conditions, with females having a disproportionate number of longer illness episodes, or greater frequency of delayed onset. To our knowledge, there is no evidence to suggest this.
Several limitations should be noted. First, there may be a gender disparity in the general severity of childhood victimization experiences, such that females may experience more serious forms of victimization than males (e.g. more prolonged victimization, more likely to know the perpetrator than males). However, studies that test for statistical interactions between gender and victimization afford the opportunity to examine gender differences in the effects of victimization at any given level of measured victimization. It may be that there are unmeasured gender differences in levels of victimization that are more nuanced. These gender differences would be missed and could represent a confounding variable. It could appear that psychiatric outcomes are worse in females, for example, at a particular level of victimization, but in reality the levels are not discriminating enough to capture differences in severity by gender. Of interest, Pimlott-Kubiak & Cortina (2003)
investigated whether psychiatric outcomes varied by gender across a range of victimization experiences in a sample of 16 000 adults. The researchers standardized levels of victimization across gender, and no differential effects by gender on psychiatric outcomes were found.
Another potential limitation of the present review concerns differential rates of childhood victimization by gender. Specifically, if the rates of childhood victimization are disproportionately higher in females compared to males, then gender differences in psychiatric outcomes could be attributable to these differential rates by gender. Studies that test for gender differences in the effects of childhood victimization on psychiatric outcomes using a gender by victimization interaction term afford the opportunity to examine gender differences in the effect of victimization (by comparing slopes of the regression lines) despite gender differences in rates of different types of victimization.
Furthermore, there is some evidence to suggest that differential rates of exposure by gender to sexual victimization do not account for the elevated rates of internalizing disorders observed in females compared to males. To determine whether gender differences in depression and anxiety symptoms could be attributed to a history of CSA in a sample of youth (n
=1053, ages 16–18 or 18–21 years), Fergusson et al. (2002)
controlled for history of CSA in analyses, hypothesizing that this would significantly reduce the association between gender and depression and anxiety. They found that a history of CSA accounted for a small proportion of the gender difference in rates of depression and anxiety. Specifically, the OR for depression was reduced from 2.5 [confidence interval (CI) 1.9–3.1] to 1.9 (CI 1.4–2.3), p
=0.03, and the OR for anxiety from 2.3 (CI 1.7–3.1) to 1.8 (CI 1.3–2.4), p
A general methodological limitation of the studies we reviewed concerns variability in definitions of childhood victimization. The child abuse literature is plagued by inconsistent definitions of victimization, in some cases too variable to allow comparison across studies. For example, individuals who are identified as having experienced CSA may have also experienced varying degrees of CPA, psychological abuse, or neglect that are either not measured or not reported. Similarly, victimization experiences may vary widely in severity, frequency, or duration, or in the relationship of the perpetrator to the victim. Studies often fail to measure or report characteristics of victimization experiences. Variability in definitions of childhood victimization, however, would only alter the results of this review if inconsistencies differed by gender. We are not aware of evidence for this.
Another general methodological limitation of the literature we reviewed is a paucity of longitudinal studies. We identified only five prospective studies. When examined separately, the results from prospective studies are similar to those found across all studies reviewed. Specifically, four studies report significant gender by victimization interactions: two demonstrate greater risk in males (Schwartz et al. 1998
; Briggs-Gowan et al. 2003
), one demonstrates greater risk in females (Khatri et al. 2000
), and the remaining study shows greater risk in both genders, depending on the type of victimization assessed (Boney-McCoy & Finkelhor, 1996
). As all four studies were with youth, time-of-sampling differences cannot be examined. With regard to psychiatric outcomes, these were distributed across both internalizing and externalizing categories for both genders.
Studies examining psychiatric outcomes of trauma have been excluded from the present review because trauma is a term that encompasses a range of severely stressful experiences in addition to childhood victimization (e.g. military combat; mugging; serious car accident; fire; flood; life-threatening illness; witnessing acts of violence, death or serious injury; discovering a corpse). Unless the role of childhood victimization is explicitly examined by the researchers, the range of events assessed by the term trauma makes it difficult to tease apart whether the gender differences observed in psychiatric outcomes of trauma are attributable to exposure to victimization in childhood or to other stressful events measured.
It is worth noting, however, that a number of community-based studies provide evidence that females exposed to trauma exhibit a higher risk of developing PTSD than trauma-exposed males (e.g. Norris, 1992
; Breslau et al. 1997
), even after controlling for type of trauma or ‘most upsetting’ trauma (e.g. Norris, 1992
; Kessler et al. 1995
; Breslau et al. 1997
). There is evidence to suggest that these gender differences in PTSD may be largely due to the effects of ‘assaultive’ forms of trauma (e.g. threat with weapon, rape; Breslau et al. 1999
) or to violent crime, for which exposed females exhibited twice the rate of PTSD of exposed males (Norris, 1992
). Tolin & Foa (2006)
, in their recent meta-analysis of 40 studies using nonclinical samples, examined the possibility that higher rates of PTSD among females might be attributable to higher lifetime rates of sexual victimization among females compared to males, but found no significant gender differences in risk for PTSD among individuals with a history sexual victimization.
In summary, our findings challenge the longstanding supposition that victimization experiences are associated with differentially poorer psychiatric outcomes in females relative to males. Taken together, the findings from the 30 studies considered in this review indicate that the evidence for gender differences in the effects of childhood victimization is neither simple nor compelling. Many studies find no evidence for gender differences, and when these differences are found, they vary by the age of participants at the time of assessment (i.e. youth versus adult). Thus, gender differences in prevalence rates of internalizing disorders, such as depression, do not appear to be attributable to differential effects of childhood victimization. Our findings underscore the need for elucidating more precise gender- and age-specific paths for psychiatric outcomes. We suspect that descriptions of gender differences that account for differences in context (e.g. frequency, duration, severity, relationship of perpetrator to victim) and significance of victimization experiences may reveal more meaningful differences related to psychiatric outcomes. Ultimately, including gender-related issues in primary and secondary prevention efforts will require more attention to exactly how and why child victimization leads to psychiatric outcomes in some individuals but not in others. Proper attention to these subtle yet important concerns is necessary to link primary and secondary prevention efforts related to downstream psychiatric sequelae of childhood victimization.