Methodological Issues in Inferences from Meta-Analysis About the Effects of Child Sexual Abuse

Abstract

Tromovitch and Rind have responded to my methodological criticisms of their meta-analyses of research on the effects of child sexual abuse (CSA). Here I provide theoretical frameworks from the domain of developmental psychopathology to guide interpretation of the empirical findings. I then expand on my concerns about conclusions based on college-student samples and about overly broad definitions of CSA used in primary research. Finally, I review two recent studies of the effects of CSA that used notably better methods. Both studies lead to the conclusion that CSA does have measurable effects on some victims, for outcomes such as borderline personality disorder, depression, and generalized anxiety disorder, even when family environment is controlled.

Key Words:

• Child sexual abuse
• meta-analysis