I found 34 studies which attempted to account for long-range differential outcomes of sexual experiences which occurred during childhood. These 34 studies are listed in Table 1 with relevant information about each one. Each study was reviewed, considering the concerns previously described, to determine if the data presented were relevant to three hypotheses.

Nineteen studies reported negative outcomes as the more frequently reported effect, with 5 of these reporting very negative outcomes. Each of these 19 studies will be briefly summarized and analyzed in the order that they appear in Table 1.

Sloane and Karpinsky (1942)

The primary strength of this descriptive study of 5 adolescent incest cases is that the authors drew their conclusions from only adolescent cases. Hence differences could be compared with findings on effects of incest on pre-adolescents. There are, however, the sampling problems of no control group, a small number of cases, and the use of a clinical, lower-class population. 
The conclusion that "indulgence in incest in the post-adolescent period leads to serious repercussions in the girl" (p. 673) is, therefore, questionable. The findings cannot be generalized and can, at best, serve as a working hypothesis for other studies. Other variables, such as background characteristics, should also be considered.

Weinberg (1955)

Weinberg was one of the few researchers who adhered to the dictionary definition of incest. He studied 203 lower-class families in the Chicago area in which incest had occurred, primarily between father-daughter. These cases had all been reported to the courts. Although Weinberg used an offender population with no control group, he did study a large number of cases by structured interviews in the home. We can accept his evidence that childhood incest experiences in lower-class families where the perpetrator has been prosecuted lead to long-term harmful effects. Yet we must question how many of these harmful effects were due to the lower-class  background or to the court procedure and incarceration of the family member. 

Greenland (1958) 

Greenland analyzed seven letters written to an advice column in Great Britain regarding the writer's incestuous experiences during childhood and resulting present problems. The very small number of cases studied, lack of a control group, and the self-selected and non-representative sample who wrote because of problems they were having lead us to reject the evidence. His findings of harmful effects may be used as a working hypothesis for more definitive studies.

Kubo (1959) 

This Japanese study of 36 cases of incest from clinic or agency records and interviews defines terms specifically but contains the sampling problems of lack of a control group, a rather small   

number of cases, use of a clinical population, and the combining of age groups. Kubo was quite cautious and conservative when reporting effects, however. Although he saw strong negative behaviors such as crimes, misconduct, and mental disorders, he stated that "it cannot be concluded these were directly attributable to incest" (p. 154). We can acknowledge the definite trend toward harmful effects but must realize these effects may have been caused by background variables.

Vestergaard (1960) 

This author studied 13 cases of parental incest in Copenhagen in which the fathers had been sentenced to prison. Vestergaard conducted interviews with women whose incestuous experiences took place more than 10 years previously. All the women felt the experience was much worse than the court records showed. Although the definitions of terms were clear-cut, the problems of few cases, no control group. offender population, and combining age groups were present. Again, we must be cautious in interpreting the trend which is indicated by these data.

Weiner (1962) 

Weiner's study of 5 paternal incest cases in New York through psychotherapy with the father has the serious flaw of the perpetrator father being the only source of information on long-term effects on the daughter. If daughter had also been studied directly, this would have been a contribution to the literature because few studies have been done with both sexual partners. The other problems of extremely few cases, no control group, and offender population render these findings useless for purposes of generalization about long-range effects. 

Chaneles (1967) *

* Upon close examination of the literature, I found that the Chaneles (1967) and De Francis (1969) studies were both funded under the Child Research Grant R-222, U.S. Children's Bureau. It is not known if both these authors reported on some of the same cases. Therefore, both are included in this review.

Chaneles studied 159 child victim cases as reported by public agencies in a 3-year project of the American Humane Association. Due to the clinical population, lack of control groups, and the  preliminary conclusion by the author that "at present, we may only conjecture long-range effects" (p. 55), this evidence of a strong negative trend must be viewed with caution.

Medlicott (1967) 

This New Zealand study of 27 psychotherapy cases of reported parental incest compares the 17 actual incest cases with the 10 falsely alleged cases. In this sense there is a control group, but the nature of the evidence which permitted the judgment that the alleged cases were actually alleged is not described. No significance 

testing was reported, but the author did find rather large differences in some areas of sexual adjustment of the two groups. The small number of cases and clinical population do not provide conclusive support for the hypothesis, but a trend is indicated.

De Francis (1969) 

This study of 263 sex-offense cases reported by child protection agencies in New York and consisting primarily of lower-class families had no control group, but the sample was large and both objective and subjective measures of consequences were obtained. His findings can not be generalized to nonclinical, middle- or lower-class populations, but the data suggest that child victims of reported sex crimes in lower-class families in large urban areas are likely to experience harmful effects.

Katan (1973) 

Katan psychoanalyzed six middle- to upper-class women who had experienced oral, anal, or genital rape when they were 1 to 3 years of age. Again we have a small clinical study with no control group, which makes generalizing to other populations impossible.

Benward and Denson-Gerber (1975) 

These authors studied 52 women in Odyssey House drug treatment centers who had experienced incest as children. Of the 93 different incestuous partners reported by the women, intercourse took place with only 34 (37%) of these. The other behaviors ranged from fondling to attempted seduction. Incestuous partners included step-relations, in-laws, and "quasi-family." 
These definitional problems in a clinical setting cloud the study, which is admittedly an exploratory study of incest as a causative factor in antisocial behavior. The author did use comparison groups which support their claim that "incest was a significant factor necessitating further study" (p. 339). This exploratory study should be viewed as such.

Molnar and Cameron (1975) 

Eighteen cases of parental incest in Canada were studied by these authors in psychotherapy interviews in a general hospital psychiatric unit. Behaviors described as "incest" were "a wider range of comportments which mayor may not include intercourse" (p. 373). We also have the few cases, no controls, and clinical population configuration that is typical of many of the studies in this analysis. The author, of course, reported harmful effects because all the patients were there because they were having problems.

James and Meyerding (1977) 

Two studies of a total of 228 prostitutes through questionnaires and interviews constituted this report. Through comparisons of their findings with findings of other studies of "normal" females, some interesting trends emerge which support the authors' conclusion that early sexual experiences and prostitution 

are related. The familiar configuration of a biased population and no control group is present. 

Meiselman (1978) 

Meiselman studied 58 cases of incest which were seen at a Los Angeles psychiatric clinic. She used a control group which consisted of a random sample of 100 patient charts from the previous 5 years. Her definition of incest was clear and specific in regard to both behavior and partner. The primary problem with this study is the use of a clinical population for both the incest group and the control group. We do not know how these cases differ from a nonclinical population. We can, however, use the evidence from her finding regarding a clinical population that the occurrence of incest does predispose the individual to certain kinds of problems, such as difficult relationships with men or sexual maladjustment.

Justice and Justice (1979) 

These authors studied 112 incestuous families in Texas through a survey and therapy with selected cases. No control group was used for the clinical population studied. Conclusions regarding long-term consequences of incest were drawn from their review of the work of others regarding incest as much as from their own study. Therefore, their measures were more a summary of the literature than specific analyses of their own study. It is not possible to extract their own findings from the larger body of literature they discussed.

Gross (1979) 

Again we have a small clinical study (4 cases of incest) , without a control group. A unique feature of this study was that it studied hysterical seizures in adolescent girls and found incest in their backgrounds. As with all small descriptive studies, the finding of the study provides questions for further research. 

Tsai, Feldman-Summers, and Edgar (1979) 

These authors compared : a clinical group of 30 women who had been molested as children and were seeking therapy with a nonclinical group of 30 women who had been molested and also with a control group of 30 women who had not been molested. All of the women were secured through media advertisements, producing a potentially biased sample. Specific behaviors and partners were defined. Although no causal inferences could be made, these findings are certainly worthy of serious consideration. Noteworthy is that not all sexually molested children necessarily experience adult maladjustment and that a later age at cessation, stronger negative feeling, higher frequency and longer duration of molestation are key variables in explaining adult maladjustment.

Sedney and Brooks (1984) 

This study was of 301 middle- to upper-class college women. Those having sexual experiences as children involving other people were compared to women with no such experiences. Definitions were clear, no clinical or offender populations were sampled, age groups were separated for purposes of some analyses, a primarily middle- to upper-class population was used, and there were specific measures of consequences. 
These features, together with the sizable number of women studied, make the findings that  reports of childhood sexual experiences are frequently associated with symptoms of distress later in life acceptable. It must be remembered that this is a retrospective study of primarily middle-class women. 

Cleveland (1986) 

Three women who had experienced incest with fathers were personally selected by the author from three different populations in order to study varied outcomes from a developmental perspective. One was from an offender population, one from a clinical population, and one was referred by a friend. Definitions were clear but the small number studied and the sampling method based on outcome with no control group make the findings interesting but not generalizable.

Of these 19 studies, only 5 met enough of the scientific criteria to be given serious consideration in accepting or rejecting hypothesis I (see Table 2):

Table 2

Studies Supporting Hypothesis 1 by Scientific Criteria