The shortcomings of both the qualitative and quantitative literature reviews point to the need for further investigation of the nature of CSA effects or correlates.

Qualitative reviews present mixed conclusions regarding the commonly assumed CSA properties of causality, pervasiveness, intensity, and gender equivalence and are limited by sampling bias, subjectivity, and imprecision.

The meta-analytic reviews, after correcting for Jumper's (1995) community sample effect sizes, show low intensity of CSA effects or correlates (in terms of effect size).

However, their contributions regarding causality, pervasiveness, and gender equivalence are either absent or wanting because of inadequate reports in the primary studies or the small number of samples included in the analyses.

The purpose of the current review was to address these shortcomings and to achieve a more accurate and precise understanding of CSA in the general population. To do so, we meta-analytically examined the literature on CSA-outcome relations in college samples.

College samples were used for several reasons.

First, this population provides the largest group of studies on nonclinical populations, which are essential for understanding CSA in the general population. The college population is useful for addressing questions regarding the general population because about 50% of U.S. adults have some college exposure ( Fritz, Stoll, & Wagner, 1981 ; U.S. Bureau of the Census, 1995 ).

Second, studies using college samples provide the most extensive data on moderators of CSA-adjustment relations. Many of these studies have examined confounding variables such as family environment, making them useful for examining causality as well as the magnitude of CSA-adjustment relations.

Third, many of these studies have reported a rich variety of other results useful for addressing the issues of pervasiveness of effects and gender equivalence. The CSA literature on college students includes numerous male samples, allowing for a more thorough comparison of the genders than previously reported.

In addition, this literature has never been systematically reviewed before, and many studies based on college samples have never

[Page 27]

been published but should be more widely known to counteract a possible publication bias.

A possible shortcoming of focusing on the college population is that college students may be too young for symptoms to have appeared, or they may be better able to cope with CSA stresses than persons in other populations ( Jumper, 1995 ).

However, younger and older adults did not differ in CSA-adjustment relations in Neumann et al.'s (1996) meta-analysis.

Furthermore, mean effect sizes from college samples, as reported by Jumper, were similar to those from national samples (Rind & Tromovitch, 1997 ), nonclinical samples ( Neumann et al., 1996 ), and community samples ( Jumper, 1995 , after corrections).

Therefore, the argument that college students are better able to cope and thus present fewer adverse reactions than people in other nonclinical populations lacks empirical support.

We addressed the assumed CSA properties of causality, pervasiveness, intensity, and gender equivalence in several ways.

First, we meta-analyzed effect sizes for CSA-symptom relations to estimate the magnitude (i.e., intensity) of the relationship between CSA and adjustment in the college population.

Second, we performed semipartial correlation and contrast analyses on the effect sizes to examine gender differences (i.e., gender equivalence), as well as other moderator variables.

Third, we meta-analyzed results from self-reported reactions to and effects from CSA to examine gender differences further.

Additionally, we analyzed these self-reports to examine the prevalence of negative effects.

Fourth, we meta-analyzed relations between CSA and family environment, as well as between symptoms and family environment, to examine the causal role of CSA in producing symptoms. We addressed the issue of causality more directly by examining the results of statistical control from studies that reported this information.