METHOD

Location of Studies
Criteria for Inclusion
Table 1
Coding of Studies

Location of Studies

Studies including males who had sexual contact with adults in childhood or adolescence were located using a variety of standard literature search techniques.

	First, reference lists of both clinical and nonclinical studies of adult-juvenile sexual contacts already known to one or both authors were reviewed.
	Second, reference lists and summary tables of the comprehensive reviews available to the authors were scanned for studies that included both male and female subjects or that included males only

(Beitchman et al., 1991, 1992; Bolton, 1989; Browne and Finkelhor, 1986; Constantine, 1981; Conte, 1985; Kendall-Thckett et al., 1993; Kilpatrick, 1987; Mendel, 1995; Urquiza and Capra, 1990; Watkins and Bentovim, 1992).

Third, computer database searches were used. The PsycLit database for the years 1974 to 1993 was searched utilizing the key words "sexual abuse" and "children" or "adolescents." Abstracts were read on-screen to determine if a particular article was a study of outcomes or effects and if the research population was nonclinical. The Dissertation Abstracts Intemational database was also searched, utilizing the same key terms, for the years 1984 to 1993. This procedure located a number of unpublished dissertations that seemed potentially includable. Book or journal references to research with nonclinical samples were then located on shelf or via

[Page 109]

interlibrary loan and examined in detail; unpublished dissertations were located on UMI microfilms.

Criteria for Inclusion

Because the focus of this review was effects on male adolescents and children of sexual contacts with adults, and the intent was to assess both the range of responses and variables that might account for that range, several criteria were established in addition to the obvious criterion that the studies had to include male subjects.

	First, when a study included both males and females, outcomes had to be reported separately on at least one measure. Reporting of results that did not distinguish males from females clearly precluded examination of outcomes for males alone.
	Second, a study had to include at least two males so that some comparison of a range of outcomes was at least possible. This criterion led to the exclusion of single-case studies, although such reports were typically from clinical sources anyway.
	Third, the study had to report some kind of psychological or behavioral outcome measure. Measures could include self-report or self-evaluation of the experience, scores on various standardized measures of personality or adjustment, or reports of clinical impressions or diagnoses. This criterion led to the exclusion of reports that only indicated the incidence of sexual contacts, but did not offer information on any outcome measures.

Based on the criteria outlined above, we located 35 nonclinical studies that could be included in the review. These studies included reports appearing as journal articles, book chapters or books, and unpublished dissertations. The included studies were published between 1956 and 1994 and are summarized in Table I.

Coding of Studies

All studies included in this review were coded for date of publication, country in which the study was conducted, sample type, sample size, age of subjects at time or the sexual experience or relationship, current age of subjects, the method(s) of assessing outcomes, and the range of outcomes. Each author coded a portion of the studies.

For sample type, samples were coded as "college" if the subjects were college students; as «convenience" if the researcher(s) used subjects found from several sources or from unusual sources such as print advertisements; and "general population" if the subjects were a random population sample from a specific area.

[ Pages 110 & 11: Table I; Page 112]

Sample size is coded in Table I as the number of male and female subjects reporting childhood or adolescent sexual contact with an adult. Thus, fur college and general population studies the total sample is larger than the figure listed because fewer than 100% of the subjects report such experiences. In some cases, reported results were based on the total number of experiences rather than the number of subjects reporting any experience(s).

Method of assessing outcome was coded as one of three possibilities.

	Studies were coded as "self-report" if subjects were asked to evaluate the impact of their experiences themselves. Self-report measures typically involved the subject rating the experience on a positive-to-negative scale, although self-report also included perceptions of being victimized or not victimized, or the presence or absence of perceived problems or harm.
	In cases in which the researchers used some form of standardized test to evaluate personal or social adjustment or mental health (e.g., the Hopkins Symptom Checklist), the study was coded as "standardized measure."
	Finally, in cases where reports of outcome consisted of the clinical diagnoses or impressions of the researcher(s), the study was coded as "clinical impression."

Studies could receive more than one code if more than one type of outcome measure was used.

The coding of outcomes was performed so that the most commonly noted outcome was put first, followed by the range of reported outcomes.

Positive outcomes were claimed in a study only if at least one of the following criteria was met:

	(i) some of the subjects evaluated their sexual experience with an adult as positive,
	(ii) the author of the report claimed that some subjects benefitted, or
	(iii) standardized measures indicated better adjustment in some fashion compared to controls.

The first two of these criteria are similar to those used by Constantine (1981).

In nearly all cases, claims of positive outcomes were from the subjects themselves. Because it is possible that claims of positive outcomes might be misinterpreted as justifying sexual exploitation or abuse, or as inappropriate for other reasons (cf. Okami, 1992), it should be noted that most such claims stem from the individuals who would be labeled "victims."

In effect, positive outcomes were defined as either personal perception that the experience was beneficial or as better psychological adjustment compared to controls without early sexual contact with adults.

Neutral outcomes were claimed if at least one of the following criteria was met:

	(i) some of the subjects evaluated their experiences as neutral,
	(ii) authors reported that some subjects were asymptomatic, or
	(iii) standardized tests revealed no significant differences between "abused" and control groups.

In many studies "mixed" outcomes (in which experiences were claimed by subjects or researchers to have both positive and negative effects) could not be disentangled from the idea of truly "neutral" outcomes (in which subjects' adjustment did not significantly differ from that of controls in either a positive or neutral direction). In

[Page 113]

practice, both type of experiences were included under the neutral category.

Finally, negative outcomes were coded based on

	(i) self-evaluations by subjects,
	(ii) clinical diagnoses or judgments, and
	(iii) poorer adjustment on standardized tests compared to controls.

Very negative outcomes were coded to indicate variations in the number and severity of symptoms among subjects. If a study provided such information, then it was coded as including both negative and very negative outcomes.