National Probability Samples
To repeat, our society has come to believe in the last few decades that CSA is "a special destroyer of adult mental health." This implies that, in the typical person, whether male or female, if they have experienced CSA, it will have caused intense harm. The best way to test this assumption would be to examine everyone in the entire population. We cant do this, of course. The next best thing that we can do is to take a representative sample from the population and try to make inferences from it. In various countries, researchers have done this: they have obtained "national probability samples," which are just samples that have been chosen so as to be representative of the population of a given nation. The data from these samples regarding the relation between CSA and adjustment are very important, because they much better represent the typical case than do data from clinical samples.
A few years ago, we gathered together the results from all studies based on national samples that examined CSA-adjustment relations. Our first table (see Table 1) is a listing of these studies, showing some of their attributes. First of all, we can see that four studies were conducted in the U.S., and one each was conducted in Canada, Great Britain, and Spain. Several studies used face-to-face interviews; others were done by telephone; two used a self-administered questionnaire that subjects filled out while the researcher waited nearby; and one was a mail survey. Two studies examined only CSA that subjects felt was unwanted; the other five samples studied both willing and unwanted CSA events. As we can see in the table, sizable numbers of subjects participated in all of these studies. The percent of subjects that had experienced CSA ranged from 6% to 36% for males and from 14% to 51% for females. The percents varied so widely because the definitions of CSA in the studies also varied widely. Excluding two studies that had definitions that seemed overly broad (for example, including willing sexual experiences with siblings as CSA), the percents ranged from 6% to 15% for males with an average of 11% and from 14% to 28% for females with an average of 19%. Thus, at the present time the best available estimates for the prevalence of CSA are 11% for males and 19% for females.
Attributes of Seven Studies Using National Probability Samples to
FTF=face to face interviews; SAQ=self-administered questionaires; Mail=mail survey; Tele=telephone survey
Ages qualifying as "child" given first; in parentheses, ages for other person and any other conditions; C=contact sex, NC=noncontact sexual experience
Includes number of respondents used in data analyses in studies assessing adjustment; otherwise, indicates number of actual participants
Based on actual number of respondents who participated.
These studies reported two types of results that were useful for evaluating the popular assumptions about CSA. One was self-reported effects--that is, how subjects felt the sexual experience had affected them in a negative, neutral, or positive way. The second were objective measures of psychological or sexual adjustment.
Lets talk about the self-reported effects first. Table 2 shows the results of the three studies that made this inquiry. In the Badgley study, subjects were asked to tell about the first unwanted sexual experience they had, if they had one. When asked whether they had been emotionally or psychologically harmed at that time by this experience, only 7% of males with such an experience said yes, compared to 24% of females. Note that this was based on unwanted experiences, and also that this shows a substantial sex difference.
In a second study conducted by Baker and Duncan in Great Britain, subjects were asked about CSA experiences and effects that occurred before the age of 16. The following distributions were found regarding self-perceived effects (see bottom of Table 2): for the males with CSA, 4% said their experience caused permanent damage; 33% said it was harmful at the time, but with no lasting effects; 57% said it had no effect; and 6% said it improved the quality of their life. The distribution for the females with CSA was: 13% reported permanent damage; 51% said it was harmful at the time, but with no lasting effects; 34% said it had no effect; and 2% said it improved the quality of their life. These results strongly contradict popular views that CSA typically scars its victims for life: only 4% of males and 13% of females thought the harm was permanent. As we can see (in the top part of Table 2), 37% of males felt harmed in some way, meaning that 63% did not; the percents were just the opposite for females, with 64% reporting at least some harm. Once again, we see a sex difference. In the last study, Laumann asked subjects about CSA experiences they may have had before puberty. For males, 45% reported some negative effect; 70% of females reported some negative effect. Again we see a sex difference.
percentage of Male and Female Self-Reports of Negative Psychological Effects of Child Sexual Abuse in National Samples
: Data based on first unwanted sex, about two thirds of which occurred prior to age 18
b: Data based on CSA under age 16
c: Data based on sexual touching before puberty with older persons
Together, these three studies show that only a minority of boys perceive some negative effect, but a majority of girls do. Further, permanent harm is rare. These findings cast doubt on the assumptions that harm is generally lasting, that harm is pervasive (especially for boys), and that boys and girls react in an equivalent fashion.
Next, we examined the relation between CSA and psychological or sexual adjustment by examining the data that compared people with CSA to control subjects. As shown in Table 3, five of the studies provided relevant data. The effect sizes are shown in the table separately for males and females. Again, these effect sizes indicate the percent of variability in adjustment among all subjects that CSA accounts for. For males, this ranged from 0.16% to 1.44%. For females, it ranged from 0.25% to 4.00%. The average effect sizes were 0.49% for males and 1.00% for females. These results show several things. First, both males and females with a history of CSA showed poorer adjustment than control subjects. Second, although statistically significant, these differences are small. For example, for males, 99.51% of the variability in their adjustment scores would have to be explained by factors other than CSA. This result, contrary to popular assumptions, does not implicate CSA as a major factor in affecting psychological or sexual well-being in the average person with this experience.
Percent of Adjustment Variance Accounted by CSA in Studies Using National Samples
* indicates a statistically significant result
In summing up this meta-analysis, we can draw these conclusions. First, its findings are considerably more relevant to trying to understand the typical case of CSA in the general population than are clinical findings. The results contradict the assumptions of widespread, lasting harm. Further, these results contradict the common belief that CSA produces intense harm -- the effect sizes were small, but should have been large, or at least medium, to infer intense harm. Additionally, boys reacted much less negatively than girls, which contradicts the assumption that boys and girls react in an equivalently negative fashion.
The final assumption needing of scrutiny is whether the small but statistically significant differences in adjustment found between CSA and control subjects reflects the effect of CSA--that is, did CSA cause this somewhat poorer adjustment? In talking about causality, we should first review some basic methodology. In the U.S., Whites score on average 15 IQ points higher than Blacks. Can you then conclude that race causes IQ differences? If you did, you would be called a racist, and justifiably so. Blacks and Whites differ not only in their race, but in their socioeconomic status, as well as other important factors. It could well be that coming from a poorer environment produces this IQ difference, rather than race. Home environment does have a big impact on intellectual development, so it may play the role of a third variable that completely accounts for the association of the two main variables--race and IQ.
Incidentally, a 15 point IQ difference between the races can be expressed in this way: race accounts for 34% of the variability in IQ scores among Whites and Blacks. In our national samples, CSA accounted for only 1% of the adjustment variation for females and only one half of one percent for males. By comparison, race was 34 to 68 times stronger in accounting for IQ variation. Thus, if we can argue that the race difference in IQ is caused, not by race, but by a poorer home environment, then surely we could consider making this argument for CSA: that the small differences in adjustment that were found may have been attributable to differences in home environment. This is a reasonable possibility. Children in broken homes are less supervised and are more prone, and willing, to engage in counternormative behavior, such as using drugs, skipping school, or engaging in taboo sex (such as sex with adults). In this scheme, the poor home environment not only predisposes them to CSA, but also predisposes them toward becoming less well adjusted. This scenario suggests that the relation that we found between CSA and adjustment could be spurious (that is, false), or, if causal, even weaker than it was.
The researcher Finkelhor was involved in two of the national studies. He and his colleagues used statistical techniques to factor out, or control for, several other variables that might have been responsible for the statistically significant CSA-adjustment relations they found. In both studies, these relations remained statistically significant after this procedure. He and his colleagues argued that this showed that CSA really does cause poorer adjustment. In criticism of Finkelhors approach, however, his group did not control for variables that other researchers have shown can account for the CSA-adjustment relation. Among these variables are physical abuse and emotional neglect, which tend to be confounded with CSA--that is, occur along with CSA experiences. The researcher Wisniewski, for example, examined CSA in 32 samples of college students chosen to be nationally representative of college students in the U.S. When she applied statistical control factoring out nonsexual abuse variables, she found that the CSA-adjustment relations dropped out. She concluded that the "data do not support child sexual abuse as a specific explanation of current emotional distress. The data are best interpreted as supportive of other factors such as family violence...as having the greatest impact on current emotional adjustment." We will return to this issue of causality and statistical control when reviewing the results of our second meta-analysis.