There is also a considerable overlap between physical, emotional and sexual abuse, and children who are subject to one form of abuse are significantly more likely to suffer other forms of abuse (Briere and Runtz 1990; Bifulco et al. 1991; Mullen et al. 1996; Fergusson et al. 1997; Fleming et al. 1997). Mullen and colleagues (1996) found women with histories of child sexual abuse had over five times the rate of physical abuse, and were three times as likely to also report emotional deprivation.
It could be that family circumstances conducive to child sexual abuse are also productive of other forms of abuse. This hypothesis is supported by the clear overlap between the risk factors for all three types of abuse. The second possibility is that the apparent comorbidity could reflect a data collection artefact created by individuals who are prepared to disclose one type of abuse being prepared to disclose other forms of abuse (Fergusson and Mullen in press).
The possibility has been raised that characteristics such as physical attractiveness, temperament or physical maturity might increase the risks of children being sexually abused (Finkelhor and Baron 1986). Child molesters are reported to selectively target pretty and trusting children (Elliot et al. 1995). A recent study suggested early sexual maturation in girls may be associated with increased vulnerability to abuse (Fergusson et al. in press). Fleming et al. (1997) reported girls who were socially isolated with few friends of their own age were almost twice as likely to report having been sexually abused.
Interpreting correlation studies
The tendency for child sexual abuse to co-vary with disturbed family backgrounds, other forms of abuse and possibly even victim characteristics, creates profound difficulties when it comes to interpreting correlational studies. This is particularly the case when examining long-term deleterious effects that could theoretically result from child sexual abuse itself, or from those other childhood traumas and disadvantages with which it is so often associated.
In some cases, the adverse outcomes attributed to child sexual abuse may be related as much to the disrupted childhood backgrounds, in the context of which the abuse arose, as to the child sexual abuse itself. There are reports that poor family functioning may account for many of the apparent associations between a history of child sexual abuse and adult psychopathology (Fromuth 1986; Conte and Schueman 1987; Friedrich et al. 1987; Wyatt and Mickey 1987; Harter et al. 1988).
Mullen et al. (1993) in a study on New Zealand women found positive correlations between a history of child sexual abuse and mental health problems in adult life. However, the overlap between the possible effects of child sexual abuse and the effects of the matrix of disadvantage from which abuse so often emerges was so considerable as to raise doubts about how often, in practice, child sexual abuse could operate as an independent causal element.
When examining all subjects with histories of child sexual abuse, it was found that the risks of women victims, who came from stable and satisfactory home backgrounds, developing significant adult psychopathology were no higher than for non-abused controls from similar backgrounds. This did not, however, hold for those who gave histories of the most physically intrusive forms of abuse involving actual penetration. This group, which contained a significant proportion of women subjected to chronic penetrative abuse in an incestuous context, did have significant increases in psychopathology, even when account was taken of the confounding influence of disrupted and disorganised family and social backgrounds.
Fleming et al. (in press), in a study of Australian women, found mental health problems to be associated with a history of child sexual abuse. However, when a multivariate analysis taking into account social and family background variables was employed, it was again only in those whose abuse had involved penetration that the association remained significant.
These findings go some way to reconciling the observations of clinicians who discern clear and dramatic relationships in their patients between prior child sexual abuse and current symptoms of specific mental disorders, and epidemiologists who extract from their data less specific correlations that barely survive confrontation with confounding variables.
The clinician sees, almost exclusively, the most severely abused whereas the epidemiologist studies the full range of reported child sexual abuse in a community. The clinician extrapolates from the individual case where dramatic personal experiences like child sexual abuse inevitably seem to explain the occurrence of disorder (particularly when patient and therapist start from the assumption that child sexual abuse deserves primacy), whereas the epidemiologist studying differences in incidence of disorders in a population is drawn to broad sociocultural and environmental influences that explain the bulk of the variation in populations.
Both perspectives have their place, and with that place comes limitation. Clinicians who, on the basis of experiences with individual cases, seek to describe the role of the full range of child sexual abuse in generating disorder and disease in our community are likely to fall into error, just as epidemiologists fall into error when they attempt to deny any reality, or therapeutic benefits, to the meaningful connections constructed between child sexual abuse and current difficulties in a treatment process.
Long-term Effects of Child Sexual Abuse
by Paul E. Mullen and Jillian Fleming