An Estimate of Consequences of Adult-Nonadult Sex
An Estimate of Consequences of Adult-Nonadult Sex
for the Nonadults in the General Population
by Bruce Rind and Robert Bauserman
What are the consequences for children and adolescents who have sexual encounters with adults? This is a question that received very little attention before the 1970s, but has received increasing attention ever since (Okami, 1990). Over the past decade a number of reviews have been published attempting to synthesize the findings of a broad range of studies that have examined these consequences (Beitchman, Zucker, Hood, DaCosta, & Akman, 1991; Beitchman, Zucker, Hood, DaCosta, Akman, & Cassavia, 1992; Browne & Finkelhor, 1986; Constantine, 1981; Conte, 1985; Kendall-Tackett, Myers, & Finkelhor, 1993; Kilpatrick, 1987; Urquiza & Capra, 1990). Most of these reviews have concluded that children and adolescents are typically harmed by these early sexual experiences. For example, both Browne and Finkelhor's (1986) review, which focused exclusively on girls' experiences, and Urquiza and Capra's (1990) review, which focused exclusively on boys' experiences, concluded that these early sexual experiences with adults often lead to: (1) emotional reactions such as anxiety, anger, depression, guilt and shame, and low self-esteem; (2) behavioral problems such as aggression, self-destructive behaviors, homicidal and suicidal thoughts, and substance abuse problems; and (3) sexual problems such as sexual dysfunction, relationship problems as adults, confusion or fear regarding sexual identity, and sexual aversion and inhibited sexual desire.
But not all reviews have reached these predominantly negative conclusions (Constantine, 1981; Kilpatrick, 1987). Constantine's (1981) review concluded that reactions are much more variable and depend on factors such as: (1) the child's or adolescent's own perception of how willingly he or she participated in the sexual interaction; (2) the younger person's knowledge about sex; and (3) his or her acceptance of the moral negatives about sex which pervade modern Western culture. Constantine found that the worst outcomes for the child or adolescent occur when he or she gives passive consent (i.e., participates without being forced to, but actually does not want to participate), is ignorant about sex, and has absorbed and accepted the moral negatives about sex. The outcomes are generally neutral, or even positive, Constantine found, when the younger person perceives his or her participation as being willing, is knowledgeable about sex, and has not absorbed the moral negatives about sex.
These different conclusions among the reviews require explanation. Constantine (1981) reviewed both clinical and legal samples, as well as nonclinical/nonlegal samples. When he contrasted the findings in these two types of samples, he found that outcomes in the clinical and legal samples tend to be negative, but outcomes in the nonclinical/nonlegal samples are much more variable, ranging from negative to positive. In other words, clinical and legal samples are biased toward the inclusion of subjects who have been harmed by these encounters or who have had negative experiences. People who see clinicians because of these early experiences are self-selected; they are already likely to have problems and that is why they see a clinician. Those who do not feel harmed, or are satisfied with their adjustment, are much less likely to come to the attention of a clinician. Experiences that are brought to the attention of the criminal justice system are also likely to be those that are of a more negative nature. Experiences that are viewed as neutral or positive by the younger person are less likely to induce him or her to make a legal complaint. Thus, three conclusions can be made about clinical and legal samples: (1) they are likely to indicate that sexual experiences between nonadults and adults are harmful for the nonadults; (2) they are not representative of the entire population of nonadults who have such experiences; and (3) literature reviews that focus on clinical and legal samples (e.g., Beitchman et al., 1991; Beitchman et al., 1992; Browne & Finkelhor, 1986; Kendall-Tackett et al., 1993; Kilpatrick, 1987; Urquiza & Capra, 1990) are therefore providing an incomplete, and perhaps even distorted, account of the consequences of these early experiences in the population.
In contrast to the dim picture that generally emerges from clinical and legal samples, convenience samples have shown that these early experiences can be positive (Bernard, 1981; Ingram, 1981; Leahy, 1992; Money & Weinrich, 1983; Okami, 1991; Sandfort, 1984; Tindall, 1978). For example, Sandfort (1984) studied 25 Dutch boys between the ages of 10 and 16 who were sexually involved with men in the context of a relationship. Sandfort concluded that "for practically all of the boys their sexual contact with their older partners emerged as a predominantly positive experience" (p. 136). Ingram (1981) reported on 74 English boys from six to 14 years of age who had had sexual contacts with men. He concluded that "I do not think there is any evidence from my study that any of the children were worse off for the activity; many, no doubt, may be better off for a relationship with a loving adult outside the family" (p. 186). Money and Weinrich (1983) investigated longitudinally the consequences for two boys who were involved in long-term sexual relationships with men. They concluded that both boys benefited from the relationships rather than being harmed by them. One of the boys commented that "To me…there was never any harm physically or mentally,…it's probably the best relationship I've ever had with anyone outside my own family–maybe even to go so far as to say with anyone in general because of the openness that it brought out" (p. 47). As is the case with the clinical and legal samples, these convenience samples are also biased. Researchers may deliberately seek out those who feel their experiences were positive (e.g., Okami, 1991), recruit subjects from organizations that have an interest in making a favorable impression (e.g., Sandfort, 1984), or report specifically on positive cases known to them (e.g., Money & Weinrich, 1983). Two conclusions can be drawn from the "positive" convenience samples: (1) they refute the hypothesis that sexual contacts between adults and nonadults are necessarily harmful or will necessarily be experienced negatively; and (2) they are not representative of the population of persons who, as nonadults, have sexual experiences with adults.
Neither the clinical and legal samples on the one hand, nor the convenience samples on the other, provide us with a valid picture of how children and adolescents in the general population react to these early sexual experiences with adults and what the long-term consequences of these experiences are for these children and adolescents. Such a valid picture can only be obtained from representative samples. Representative samples, however, are almost nonexistent in this area. Nevertheless, one such example does exist (Baker & Duncan, 1985). In Baker and Duncan's (1985) study, face-to-face interviews were conducted with a nationally representative sample of subjects 15 years of age and over in Great Britain. It was found that, of the male respondents who had sexual encounters with adults before they were 16, 4% reported permanent damage, 33% reported being harmed at the time, but with no long-lasting effects, 57% reported no effects at all, and 6% reported that these encounters had improved the quality of their life. For the female respondents in this study, 13% reported permanent damage, 51% reported harm at the time which was not long-lasting, 34% reported no effects at all, and 2% reported improvement in the quality of life. Several conclusions can be drawn from this study: (1) males and females react differently to these early sexual experiences, with only 37% of males showing an initial negative reaction, but 70% of females showing an initial negative reaction; (2) these results do not conform to the clinical and legal profile; (3) these results do not conform to the profile presented by the convenience samples; and (4) these results indicate a profile intermediate to the clinical/legal and convenience profiles.
Although the Baker and Duncan (1985) study represents an advance over studies using non-representative samples, the study contains several important methodological shortcomings that may have biased its results. First, all interviewers were female, which could have affected the responses of male subjects (Barnes & Rosenthal, 1985). Second, the interviews were not anonymous, which could have led to socially desirable responses, given the sexual nature of the inquiry (see Crowne & Marlowe, 1964; Rosenthal, 1977). Third, all respondents were given a card to read, on which it was stated that anyone under 16 is "sexually abused" who is involved in any activity with a sexually mature person which leads to their sexual arousal (p. 458). When respondents indicated that they had had such an experience, they were asked questions such as their relationship with the "abuser" and their age when they were first "abused." It is likely that these value-laden terms influenced respondents to underreport positive experiences because of expectancy effects and demand characteristics – i.e., these negative terms indicated to respondents how they should respond (see Orne, 1962; Rosenthal, 1977). Furthermore, the possibility of negatively biased responses by respondents is suggested by a study conducted by Rind and Bauserman (1993), who demonstrated experimentally that negative sexual terminology (e.g., "sexual abuse") can negatively bias judgments about adult-nonadult sexual contacts.
The biasing influences in the Baker and Duncan (1985) study indicate that a consideration of additional research is warranted in order to answer the question: What are the consequences for children and adolescents in general who have sexual encounters with adults? It was the purpose of the present paper to examine additional research so as to address this question. To this end, research conducted on college students who had early sexual experiences with adults was used because of the several advantages this research offers. First, although not representative of the entire population, a majority of the population – at least in the United States where most of this research has been conducted – has some college experience (Fritz, Stoll, & Wagner, 1981). Second, over the last decade or so a fair number of such college studies have been conducted. Third, these studies have generally obtained data anonymously and have used neutrally-worded questions, thereby avoiding
the potential biases in the Baker and Duncan (1985) study. Fourth, the college studies have used more similar control groups than clinical studies have, allowing for stronger causal inferences to the extent that such inferences can be made in correlational research. To the extent that the college data come from diverse samples (e.g., geographically), are consistent with each other, and are consistent with the Baker and Duncan (1985) data, greater confidence will be reached in the profile of consequences in general which emerges.
Location of Studies
The literature was searched for all available reports of college students who, before adulthood, had sexual contact with adults. This search was achieved by using the PsycLit database for the years 1974-1986 and 1987-1993. The key words "sexual abuse" and "children" or "adolescents" were entered. Also, the Dissertation Abstracts International database for the years 1984-1993 was also searched, using the same key terms. Colleges studies were included in the present review only if they met the following criteria: (1) they included both male and female students; (2) some kind of psychological or behavioral outcome measure was reported; and (3) outcomes were reported separately for males and females. Based on these criteria, nine college studies were located (see Table 1).
Table 1: College Studies of the Consequences for Children and Adolescents after Sexual Contacts with AdultsStudy / Region / Boys
Man / Boys
Woman / Girls
Man / Girls
Woman / Short term
Boys % / Short term
Girls % / Long term
Boys % / Long term
Finkelhor (1979) / Northeastern USA / 84a / 84a / 239 / 239 / pos+neut 62
neg 38 / pos+neut 34
neg 66 / friend 3.2
scale 1-5, neg 5 / all 4.0
Fischer (1991) / Northwestern USA / 15b / 10 / 59 / 0 / - / - / wanted 28%
no problem 21% / wanted 5%
no problem 7%
Fritz, Stoll & Wagner (1981) / Northwestern USA / 8a / 12 / 38 / 4 / - / - / problems 10% / problems 23%
Goldman & Goldman (1988) / Australia / 19b / 21 / 179 / 9 / pos 39
neg 30 / pos 17
neg 68 / evaluation 3.3
scale 1-5, neg 5 / evaluation 4.2
Haggard & Emery (1989) / Middle Atlantic USA / 21a / 21a / 80 / 80 / - / - / very positive now and earlier 33% # / very positive now and earlier 4%
Landis (1956) / Western USA / 181b / 35 / 523 / 8 / pos 6
neg 46 / pos 2
neg 76 / harm:
lasting 0% / harm:
O'Neil (1990) / Puerto Rico / 39b / 10 / 76 / 9 / pos 41
neg 82 / pos 10
neg 82 / pos+neut 56%
neg 44% / pos+neut 31%
Sarbo (1989) / Southeastern USA / 24a / 24a / 62 / 62 / - / - / no difference from control group / no difference from control group
Schultz & Jones (1983) / Middle West USA / 40a from 35 / 40a from 35 / 70a to 9 / 70a to 9 / pos 68
neg 8 / pos 28
neg 52 / pos 61%
neg 9% / pos 25%
a. number of persons; b. number of experiences; # also standardized tests for measured behavior/attitude
Coding of Studies
Each study was coded for: (1) the geographical area in which it was conducted; (2) the number of males and females who had an early sexual experience with an adult; (3) the short-term reactions; and (4) the long-term reactions or outcomes. This information is summarized in Table 1.
The method of assessing reactions and outcome for each study was also coded. "Self-report" studies asked students to evaluate the impact of their experiences themselves (e.g., how they felt about their experience at the time: very positive, positive, neutral, negative, or very negative). "Standardized measure" studies used some form of standardized test to evaluate adjustment. Some studies used both types of methods for assessing outcome. Finally, information was coded that was expected to relate to the reactions and outcomes such as willingness of participation and relatedness of the adult.
Overview of Studies
Of the nine studies based on the college samples, only two were based on samples outside the United States: Goldman and Goldman (1988) from Australia and O'Neil (1990) from Puerto Rico. Of the seven American college studies, two were conducted in the Northwestern part of the country, and one each was conducted in the West, the Midwest, the Northeast, the Mid-Atlantic region, and the Southeast. Thus, the nine college studies were geographically diverse. Seven of the nine studies used only self-report measures, one used only standardized measures, and one used both.
Self-reports: Short-term. Students' reactions at the time to their early sexual experiences with adults ranged from negative to positive. Five studies provided breakdowns of these reactions. Only a minority of males reacted negatively in all these studies: 38% (Finkelhor, 1979); 30% (Goldman & Goldman, 1988); 46% (Landis, 1956); 45% (O'Neil, 1990); and 8% (Schultz & Jones, 1983). These results contrast with negative initial reactions by females which were, for these studies, 66%, 68%, 76%, 82%, and 52%, respectively. The unweighted mean initial reactions for males were: 33% negative and 67% neutral or positive. For females, these mean reactions were: 69% negative and 31% neutral or positive. These findings are remarkably similar to Baker and Duncan's (1985) results: 37% negative vs. 63% non-negative for males, and 64% negative vs. 36% non-negative for females. These results therefore indicate a consistent sex difference: about two-thirds of males react non-negatively, whereas two-thirds of females react negatively.
Positive reactions ranged from 6% to 69% for males and from 2% to 28% for females. The mean positive reactions for males and females, respectively, were 39% and 14%. This different pattern of negative and positive initial reactions by males and females indicates that the genders differ substantially in how they react to sexual encounters with adults, and that inferring from the experiences of females to those of males is invalid. The college data indicate further that more positive reactions occur in the population than were found in Baker and Duncan's (1985) study.
Self-reports: Long-term. Reports of how students felt currently about their experiences followed the pattern obtained with the short-term reactions. Positive current feelings were more common for males than females, and negative current feelings were more common for females than males. In Fischer's (1991) study, 28% of males liked their experience versus only 5% for females. In Haugaard and Emery's (1989) study, 33% of males thought their experiences were very positive then and now versus only 4% for females. In O'Neil's (1990) study, 56% of males felt neutral or positive about their early experiences versus only 31% for females. In Schultz and Jones' (1983) study, 91% of males felt their experiences were positive or neutral versus only 53% for females. Both Finkelhor (1979) and Goldman and Goldman (1988) found that the typical male reaction was neutral, but that the typical female reaction was negative. Fritz et al. (1981) found that only 10% of males reported problems resulting from their early experiences compared with 23% for females. Landis (1956) found that 0% of males reported permanent harm versus 3% for females. As with immediate reactions, current feelings were often neutral or positive for males and negative for females.