Critical Appraisal of the 1998 Meta-Analytic Review of Child Sexual Abuse Outcomes Reported by Rind, Tromovitch, and Bauserman

• NCJ Number: 194445
• Journal: Journal of Child Sexual Abuse Volume: 9 Issue: 3/4 Dated: 2001 Pages: 135-155
• Author(s): John A. Whittenburg; Pamela Paradis Tice; Gail L. Baker; Dorothy E. Lemmey
• Date Published: 2001
• Length: 21 pages
• Publication Format: Article
• Publication Type: Literature Review
• Language: English
• Country: United States of America
• Note: See articles NCJ-194444 and NCJ-194446 for discussions of the Rind, Tromovitch, and Bauserman findings.

Annotation

This article is a methodological critique of the research contained in 1998's "A Meta-analytic Examination of Assumed Properties of Child Sexual Abuse Using College Samples" by Rind, Tromovich, and Bauserman.

Abstract

This article presents a critique of the methodology use in Rind, Tromovitch, and Bauserman's (Rind et al.) 1998 work "A Meta-analytic Examination of Assumed Properties of Child Sexual Abuse Using College Samples." The authors of this article identified seven key concerns regarding Rind et al.’s research methods. Those concerns are: 1) the study population selected by Rind et al; 2) the research strategy; 3) the meta-analytic coding used; 4) the statistical techniques applied; 5) the psychological adjustment symptoms analyzed; 6) the underrepresentation of males sampled and the confounding of gender and age; and 7) the limitations of the cross-sectional approach. The population selected by Rind et al. consisted entirely of college students. The authors of this article, Whittenburg et al., compared the results of that population to existing research from a national sample and found statistically significant differences in findings across four of six types of child sexual abuse analyzed by Rind et al. and by these authors. The effects of the removal of clinical and legal cases from the Rind et al. research sample, the unavailability of independent judging for the selection of the meta-analytical interrater, the use of Pearson coefficient r to determine relationships, and the use of a nonparametric statistic to test the homogeneity of variance among the samples on the resulting statistical analysis was also presented. Whittenburg et al. identify their most important methodological question as the use of a single information source, the victim’s self-report of effects, in a cross-sectional analytical approach. The authors conclude that Rind et al. utilized flawed methodology and failed to take into account relevant scientific literature in formulating the meta-analytic review. 2 tables, 4 notes, 35 references