Journal Issue: Sexual Abuse of Children Volume 4 Number 2 Summer/Fall 1994
Prevalence: Estimating the Number of People Who Suffer Sexual Abuse at Some Point During Childhood
Because so much sexual abuse remains undisclosed, many researchers have concluded that the best picture of the scope of the problem is obtained by asking adults about their childhood experiences. (See Table 2 for a summary of 19 such adult retrospective surveys.7-25 The chart contains most of the adult retrospective surveys completed in the United States and Canada since 1980 using community samples and random sampling techniques. Surveys of college students were not included.)
Prevalence studies vary greatly in their definition of abuse, methodological approach, and quality. One problem is the lack of a common definition of abuse: for example, the use of different ages (16 or 18) to define the end of childhood, and the inclusion or exclusion of noncontact experiences or abuse by peers.
The percentage of adults disclosing histories of sexual abuse in these studies ranges from 2% to 62% for females and from 3% to 16% for males. Of these, five were national random samples. The Los Angeles Times survey reported sexual abuse of 27% of the women and 16% of the men of all ages in the United States.13,26 A national survey of the correlates of women's problem drinking estimated a history of sexual abuse in the backgrounds of 19% to 23% of all women over age 21, depending on which definition of sexual abuse was used.23 The National Survey of Children found a history of rape or forced sex in 8% of women and 1% of men in a national sample of 18- to 22-year-olds.17
The National Women's Study reported forcible rape before age 18 to 9% of American women.20,27,28 The Badgley Commission7 revealed sexual abuse to 18% of women and 8% of men in Canada.29
In the past decade, at least 20 adult retrospective studies have been conducted in countries outside North America, including Australia, Austria, Great Britain, Greece, New Zealand, Spain, and the Scandinavian countries. These studies reveal a distribution of findings similar to the North American studies, with a range of 7% to 36% for women and 3% to 29% for men.30
In a review of the findings of 19 surveys, including some student samples, Peters, Wyatt, and Finkelhor31 concluded that the most dramatic variations were not primarily explained by the definitions used, the sampling techniques, the response rates, the socioeconomic status of respondents, or whether subjects were interviewed by phone, in person, or with self-administered questionnaires. Most important was the number of specific questions that were asked to ascertain a possible history of abuse. Five of six studies asking women only a single question had rates under 13%. Seven of eight studies asking two or more questions had rates over 19%. The review concluded that multiple questions were more effective in gaining disclosures because they gave respondents more cues regarding the various kinds of experiences that the study was asking about and because they gave the respondents a longer time and more opportunities to overcome embarrassment and hesitation about making a disclosure.
These prevalence studies have led most reviewers to conclude that at least one in five adult women in North America experienced sexual abuse (either contact or noncontact) during childhood.31,32 This conclusion is based on the fact that the more methodologically sophisticated studies using multiple screen questions and random samples have had findings this high or higher.
The most commonly cited specific figures for females are 27% from the Los Angeles Times study because of its national scope and 34% (contact abuse only) from the Russell study because of its careful methodology. These findings are not without limitations. The Los Angeles Times survey included questions that were vague with regard to the exact types of experiences being asked about, making it far from ideal. (For example, one preliminary screener asked about "anyone trying or succeeding in having any kind of sexual intercourse with you or anything like that" with no follow-up questions about the details of the activities to see what "anything like that" might have meant to the respondent.) The often-cited Russell study, which was among the most meticulous in its methodology (it employed explicit definitions, a good questionnaire design, and extensive interviewer training), was limited to a sample from San Francisco.33,34 A national study using the Russell definition of sexual abuse (although somewhat different screening questions) put the national prevalence at 19%.23 Enough credible figures cluster around or exceed 20% to suggest that the number of female victims has been at least this high.35
The number of male victims is more problematic because it has been the subject of fewer quality studies. The 16% prevalence estimates from the Los Angeles Times survey (often cited as one in six males) is among the highest in the literature based on community surveys13 and is subject to the limitations mentioned earlier. The range of other community studies about males tends to be between 3% and 11%,16,18,24,36-38 but many of these studies used the inferior format of a single question. In light of the limitations of these other studies, use of the 16% Los Angeles Times figure is defensible as the only truly national estimate, but it has less corroboration from other studies than the estimate for women. A more conservative estimate for men of 5% to 10% would have support from a variety of studies.
When interpreting prevalence findings, most researchers have warned that all percentages based on adult retrospective reports are probably underestimates, although not so far off as the official incidence studies cited earlier. For comparison, see Box 3. It has generally been presumed that a certain percentage of sexual abuse victims would fail to disclose their victimization in retrospective studies—no matter how methodologically sophisticated the approach—because of embarrassment, privacy concerns, or simply because they did not remember. The literature on victimization surveys suggests that it is difficult to remember more than a year previously, not to mention the 20- and 30-year time spans required to recall childhood events. Indeed, Williams,39 in a follow-up of 100 girls who were seen in a hospital emergency room with diagnoses of child sexual abuse prior to age 13, found that 38% failed to disclose this episode in a study 17 years later, even in response to a very detailed history-taking questionnaire. Although Williams's abuse victims were younger on average, more seriously abused, and more socioeconomically disadvantaged than typical abuse victims, her finding that 17% of abused women were misclassified as nonabused suggests that prevalence surveys do underestimate abuse and especially the extent of more serious abuse at younger ages.
On the other hand, the question should also be asked whether adult prevalence estimates could be subject to inflation, as well. In other words, do some respondents possibly fabricate or enlarge trivial episodes in a way that exaggerates prevalence estimates? Although this may occur occasionally, no evidence suggests that fabrication is a major threat to the validity of victimization surveys. The weight of evidence from victimization surveys and from studies of sexual behavior and other sensitive subjects is that the withholding of disclosures is a much larger problem than the fabrication of disclosures.40
It needs to be remembered that, even if in present-day America there may appear to be some "secondary gain" to considering oneself a "survivor of sexual abuse," many of the prevalence surveys were completed in an earlier epoch when the climate was quite different. Kinsey's finding from the World War II era that 24% of women had preadolescent sexual contact with older males is consistent with the findings of more recent studies.41
In conclusion, there is considerable accumulated evidence that at least 20% of American women and 5% to 10% of American men experienced some form of sexual abuse as children.