Impact of Child Sexual Abuse. A Review of the Research

Abstract

This article reviews studies that have tried to confirm empirically the effects of child sexual abuse cited in the clinical literature. In regard to initial effects, empirical studies have indicated reactions—in at least some portion of the victim population—of fear, anxiety, depression, anger and hostility, aggression, and sexually inappropriate behavior. Frequently reported long-term effects include depression and self-destructive behavior, anxiety, feelings of isolation and stigma, poor self-esteem, difficulty in trusting others, a tendency toward revictimization, substance abuse, and sexual maladjustment. The kinds of abuse that appear to be most damaging, according to the empirical studies, are experiences involving father figures, genital contact, and force. The controversy over the impact of child sexual abuse is discussed, and recommendations for future research efforts are suggested.

Full text

Psychological Bulletin Copyright 1986 by
the American
Psychological Association, Inc.
1986, Vol. 99, No. I, 66-77 0033-2909/86/$00.75
Impact of Child Sexual Abuse: A Review of the Research
Angela Browne and David Finkelhor
Family Violence Research Program and Family Research Laboratory, University of New Hampshire
This article reviews studies that have tried to confirm empirically the effects of child sexual abuse
cited in the clinical literature. In regard to initial effects, empirical studies have indicated reactions--
in at least some portion of the victim population--of fear, anxiety, depression, anger and hostility,
aggression, and sexually inappropriate behavior. Frequently reported long-term effects include depression
and self-destructive behavior, anxiety, feelings of isolation and stigma, poor self-esteem, ditficulty in
trusting others, a tendency toward revictimization, substance abuse, and sexual maladjustment. The
kinds of abuse that appear to be most damaging, according to the empirical studies, are experiences
involving father figures, genital contact, and force. The controversy over the impact of child sexual
abuse is discussed, and recommendations for future research efforts are suggested.
Although clinical literature suggests that sexual abuse during
childhood plays a role in the development of other problems
ranging from anorexia nervosa to prostitution, empirical evidence
about its actual effects is sparse. In this article we review the
expanding empirical literature on the effects of child sexual abuse,
discuss its initial and long-term effects, review studies on the
impact of different kinds of abuse, and conclude with a critique
of the current literature and some suggestions for future research.
Child sexual abuse consists of two overlapping but distin-
guishable types of interaction: (a) forced or coerced sexual be-
havior imposed on a child, and (b) sexual activity between a
child and a much older person, whether or not obvious coercion
is involved (a common definition of"much older" is 5 or more
years). As might be expected, not all studies relevant to our pur-
poses share these parameters. Some have focused on experiences
with older partners only, excluding coerced sexual experiences
with peers. Others have looked only at sexual abuse that was
perpetrated by family members. Such differences in samples
make comparisons among these studies difficult. However, we
include all the studies that looked at some portion of the range
of experiences that are bounded by these two criteria. (See Table
1 for a breakdown of sample composition of the studies reviewed.)
Two areas of the literature are not included in our review. A
small number of studies on the effects of incest (e.g., Farrell,
1982; Nelson, 1981), as well as one review of the effects of child
This research was supported by National Institute of Mental Health
Grant (MH 1516 l), National Center for Child Abuse and Neglect Grant
90CA 0936/01, and the Eden Hall Farm Foundation.
The authors would like to thank the following people for their assistance
and comments during the preparation of this article: Christopher Bagley,
Larry Baron, John Briere, Jean Ellison, William Friedrich, Mary Ellen
Fromuth, Linda Gott, Judith Herman, Karin Meiselman, Diana Russell,
and the members of the Family Violence Research Seminar. This article
is one of a series on child sexual abuse and family violence published by
the Family Violence Research Program at the University of New Hamp-
shire.
Correspondence concerning this article should be addressed to Angela
Browne, Family Research Laboratory, 128 Horton Social Science Center,
University of New Hampshire, Durham, New Hampshire 03824.
sexual experiences (Constantine, 1980), combine data on con-
sensual, peer experiences with data that involve either coercion
or age disparity. Because we were unable to isolate sexual abuse
in these studies, we had to exclude them. Secondly, we decided
to limit our review to female victims. Few clinical, and even
fewer empirical, studies have been done on male victims (for
exceptions, see Finkelhor, 1979; Rogers & Terry, 1984; Sandfort,
1981; Woods & Dean, 1984), and it seems premature to draw
conclusions at this point.~ Under "empirical" studies, we include
any research that attempted to quantify the extent to which a
sequelae to sexual abuse appeared in a specific population. Some
of these studies used objective measures, whereas others were
based primarily on the judgments of clinicians.
Initial Effects
By initial effects, we mean those reactions occurring within 2
years of the termination of abuse. These early reactions are often
called
short-term
effects in the literature. We prefer the term
initial
effects, however, because "short-term" implies that the
reactions do not persist--an assumption that has yet to be sub-
stantiated.
Emotional Reactions and Self-Perceptions
Although several empirical studies have given support to clin-
ical observations of generally negative emotional effects resulting
from childhood sexual abuse, only two used standardized mea-
sures and compared subjects' scores to general population norms.
In an early study of the effects of sexual abuse on children,
DeFrancis (1969) reported that 66% of the victims were emo-
tionally disturbed by the molestation: 52% mildly to moderately
66
The whole literature on sexual abuse poses problems for differentiating
according to gender of victims. As Table l shows, many studies contain
a small number of men included in a larger sample of women. Unfor-
tunately, many of these studies do not specifically mention which effects
apply to men, so it is possible that some of the scqualae described apply
only to the men. However, we believe that most of the sequelae described
relate primarily to women.

IMPACT OF CHILD SEXUAL ABUSE 67
disturbed, and 14% seriously disturbed. Only 24% were judged
to be emotionally stable after the abuse. However, because this
sample was drawn from court cases known to Prevention of Cru-
elty to Children services or to the police, and because the subjects
came primarily from low income and multiple-problem families
who were on public assistance, these findings may have little
generalizability.
In investigating a different type of special population, Ander-
son, Bach, and Griffith (1981) reviewed clinical charts of 155
female adolescent sexual assault victims who had been treated
at the Harborview Medical Center in Washington and reported
psychosocial complications in 63% of them. Reports of "inter-
nalized psychosocial sequalae" (e.g., sleep and eating distur-
bances, fears and phobias, depression, guilt, shame, and anger)
were noted in 67% of female victims when the abuse was intra-
familial and 49% when the offender was not a family member.
"Externalized sequelae" (including school problems and running
away) were noted in 66% of intrafamilial victims and 21% of
extrafamilial victims. However, no standardized outcome mea-
sures were used, so the judgments of these effects may be sub-
jective.
In what is probably the best study to date, researchers affiliated
with the Division of Child Psychiatry at the Tufts New England
Medical Center gathered data on families involved in a treatment

68 ANGELA BROWNE AND DAVID FINKELHOR
program restricted to those children who had been victimized
or revealed their victimization in the prior 6 months. Standard-
ized self-report measures--the Louisville Behavior Checklist
(LBC), the Piers-Harris Self-Concept Scale, the Purdue Self-
Concept Scale, and the Gottschalk Glesser Content Analysis
Scales (GGCA)--with published norms and test validation data
were used, so that characteristics of sexually abused children
could be contrasted with norms for general and psychiatric pop-
ulations. Subjects ranged in age from infancy to 18 years and
were divided into preschool, latency, and adolescence age groups.
Data were gathered on four areas: overt behavior, somaticized
reactions, internalized emotional states, and self-esteem.
In evaluating the initial psychological effects of child sexual
abuse, Tufts (1984) researchers found differences in the amount
of pathology reported for different age groups. Seventeen percent
of 4- to 6-year-olds in the study met the criteria for "clinically
significant pathology," demonstrating more overall disturbance
than a normal population but less than the norms for other chil-
dren their age who were in psychiatric care. The highest incidence
of psychopathology was found in the 7- to 13-year-old age group,
with 40% scoring in the seriously disturbed range. Interestingly,
few of the adolescent victims exhibited severe psychopathology,
except on a measure of neuroticism.
Friedrich, Urquiza, and Beilke (in press) also used a stan-
dardized measure in their study of 61 sexually abused girls.
Subjects were referred by a local sexual assault center for eval-
uation or by the outpatient department of a local hospital. Chil-
dren in this sample had been abused within a 24-month period
prior to the study. Using the Child Behavior Check List (CBCL;
see Achenbach & Edelbrock, 1983, for a description of this mea-
sure), Friedrich et al. reported that 46% of their subjects had
significantly elevated scores on its Internalizing scale (including
fearful, inhibited, depressed, and overcontrolled behaviors) and
39% had elevated scores on its Externalizing scale (aggressive,
antisocial, and undercontroiled behaviors). This was compared
with only 2% of the normative sample who would be expected
to score in this range. Younger children (up to age 5) demonstrated
a tendency to score high on the Internalizing scale, whereas older
children (ages 6-12) were more likely to have elevated scores on
the Externalizing scale.
Breaking down emotional impact into specific reactions, we
find that the most common initial effect noted in empirical stud-
ies, similar to reports in the clinical literature, is that of fear.
However, exact proportions vary from a high of 83% reported
by DeFrancis (1969) to 40% reported by Anderson et al. (1981).
Because of its use of standardized measures, we would give the
most credence to the Tufts (1984) study, which found that 45%
of the 7- to 13-year-olds manifested severe fears as measured by
the LBCs, compared with 13% of the 4- to 6-year-olds. On the
adolescent version of the LBC, 36% of the 14- to 18-year-olds
had elevated scores on "ambivalent hostility," or the fear of being
harmed.
Another initial effect in children is reactions of anger and hos-
tility. Tufts (1984) researchers found that 45% to 50% of the 7-
to 13-year-olds showed hostility levels that were substantially el-
evated on measures of aggression and antisocial behavior (LBC),
as did 35% on the measure of hostility directed outward (GGCA).
Thirteen percent to 17% of 4- to 6-year-olds scored above the
norms on aggression and antisocial behavior (LBC), whereas 25%
of 4- to 6-year-oids and 23% of the adolescents had elevated
scores on hostility directed outward (GGCA). In his study of
court cases, DeFrancis (1969) noted that 55% of the children
showed behavioral disturbances such as active defiance, disruptive
behavior within the family, and quarreling or fighting with siblings
or classmates. DeFrancis' sample might have been thought to
overselect for hostile reactions; however, these findings are not
very different from findings of the Tufts study for school-age
children.
Guilt and shame are other frequently observed reactions to
child sexual abuse, but few studies give clear percentages. De-
Francis (1969) observed that 64% of his sample expressed guilt,
although this was more about the problems created by disclosure
than about the molestation itself. Anderson et al. (1981) reported
guilt reactions in 25% of the victims. Similarly, depression is
frequently reported in the clinical literature, but here too, specific
figures are rarely given. Anderson et al. (1981) found that 25%
of female sexual assault victims were depressed after the abuse.
Sexual abuse is also cited as having an effect on self-esteem,
but this effect has not yet been established by empirical studies.
Fifty-eight percent of the victims in the DeFrancis (1969) study
expressed feelings of inferiority or lack of worth as a result of
having been victimized. However in a surprising finding, Tufts
(1984) researchers, using the Purdue Self-Concept Scale, found
no evidence that sexually abused children in any of the age groups
had consistently lower self-esteem than a normal population of
children.
Physical Consequences and Somatic Complaints
Physical symptoms indicative of anxiety and distress are noted
in the empirical literature as well as in clinical reports. In their
chart review of female adolescent victims, Anderson et al. ( 1981)
found that 17% had experienced sleep disturbances and 5%-7%
showed changes in eating habits after the victimization. J. Peters
(1976), in a study of child victims of intrafamilial sexual abuse,
reported that 31% had difficulty sleeping and 20% experienced
eating disturbances. However, without a comparison group, it is
hard to know if this is seriously pathological for any group of
children, or for clinical populations in particular. Adolescent
pregnancy is another physical consequence sometimes mentioned
in empirical literature. DeFrancis (1969) reported that 11% of
the child victims in his study became pregnant as a result of the
sexual offense; however, this figure seems far too high for a con-
temporary sample. Meiselman (1978), in analyzing records from
a Los Angeles psychiatric clinic, found only 1 out of 47 incest
cases in which a victim was impregnated by her father.
Effects of Sexuality
Reactions of inappropriate sexual behavior in child victims
have been confirmed by two studies using standardized measures
(Friedrich et al., in press; Tufts, 1984). In the Tufts (1984) study,
27% of 4- to 6-year-old children scored significantly above clinical
and general population norms on a sexual behavior scale that
included having had sexual relations (possibly a confounding
variable in these findings), open masturbation, excessive sexual
curiosity, and frequent exposure of the genitals. Thirty-six percent
of the 7- to 13-year-olds also demonstrated high levels of distur-

IMPACT OF CHILD SEXUAL ABUSE
69
bance on the sexual behavior measure when contrasted to norms
for either general or clinical school-age populations. Similarly,
Friedrich et al. (in press), using the CBCL to evaluate 3- to 12-
year-olds, found that 70% of the boys and 44% of the girls scored
at least one standard deviation above a normal population of
that age group on the scale measuring sexual problems. Inter-
estingly, sexual problems were most common among the younger
girls and the older boys.
Effects on Social Functioning
Other aftereffects of child sexual abuse mentioned in the lit-
erature include dit~culties at school, truancy, running away from
home, and early marriages by adolescent victims. Herman ( 1981)
interviewed 40 patients in therapy who had been victims of fa-
ther--daughter incest, and compared their reports with those from
a group of 20 therapy clients with seductive, but not incestuous,
fathers. Of the incest victims, 33% attempted to run away as
adolescents, compared with 5% of the comparison group. Sim-
ilarly, Meiselman (1978) found that 50% of the incest victims in
her sample had left home before the age of 18, compared with
20% of women in a comparison group of nonvictimized female
patients. Younger children often went to a relative, whereas older
daughters ran away or eloped, sometimes making early marriages
in order to escape the abuse. Two studies, neither with comparison
groups, mentioned school problems and truancy. Ten percent of
the child victims in J. Peters's (1976) study quit school, although
all of his subjects were under the age of 12 at the time. Anderson
et al. found that 20% of the girls in their sample experienced
problems at school, including trauncy or dropping out.
A connection between sexual abuse, running away, and delin-
quency is also suggested by several studies of children in special
treatment or delinquency programs. Reich and Gutierres (1979)
reported that 55% of the children in Maricopa County, Arizona
who were charged with running away, truancy, or listed as missing
persons were incest victims. In addition, in a study of female
juvenile offenders in Wisconsin (1982), researchers found that
32% had been sexually abused by a relative or other person close
to them.
Summary of Initial Effects of Child Sexual Abuse
The empirical literature on child sexual abuse, then, does sug-
gest the presence--in some portion of the victim population--
of many of the initial effects reported in the clinical literature,
especially reactions of fear, anxiety, depression, anger and hostility,
and inappropriate sexual behavior. However, because many of
the studies lacked standardized outcome measures and adequate
comparison groups, it is not clear that these findings reflect the
experience of all child victims of sexual abuse or are even rep-
resentative of those children currently being seen in clinical set-
tings. At this point, the empirical literature on the initial effects
of child sexual abuse would have to be considered sketchy.
Long-Term Effects
Emotional Reactions and Self-Perceptions
In the clinical literature, depression is the symptom most
commonly reported among adults molested as children, and em-
pirical findings seem to confirm this. Two excellent community
studies are indicative of this. Bagley and Ramsay (1985), in a
community mental health study in Calgary utilizing a random
sample of 387 women, found that subjects with a history of child
sexual abuse scored more depressed on the Centre for Environ-
mental Studies Depression Scale (CES-D) than did nonabused
women (17% vs. 9% with clinical symptoms of depression in the
last week), as well as on the Middlesex Hospital Questionnaire's
measure of depression (15% vs. 7%). S. Peters (1984), in a com-
munity study in Los Angeles also based on a random sample,
interviewed 119 women and found that sexual abuse in which
there was physical contact was associated with a higher incidence
of depression and a greater number of depressive episodes over
time, and that women who had been sexually abused were more
likely to have been hospitalized for depression than nonvictims.
In a multiple regression that included both sexual abuse and
family background factors (e.g., a poor relationship with the
mother), the variable of child sexual abuse made an independent
contribution to depression.
The link between child sexual abuse and depression has been
confirmed in other nonclinical samples as well. Sedney and
Brooks (1984), in a study of 301 college women, found a greater
likelihood for subjects with childhood sexual experiences to re-
port symptoms of depression (65% vs. 43% of the control group)
and to have been hospitalized for it (18% of those depressed in
the childhood experience group vs. 4% of women in the control
group). These positive findings are surprising, in that the re-
searchers used an overly inclusive definition of sexual experiences
that may not have screened out some consensual experiences
with peers. Their results are consistent, however, with those from
a carefully controlled survey of 278 undergraduate women by
Briere and Runtz (1985) using 72 items of the Hopkins Symptom
Checklist, which indicated that sexual abuse victims reported
that they experienced more depressive symptoms during the 12
months prior to the study than did nonabused subjects.
Studies based on clinical samples (Herman, 1981; Meiselman,
1978) have not shown such clear differences in depression be-
tween victims and nonvictims. For example, although Herman
(1981) noted major depressive symptoms in 60% of the incest
victims in her study, 55% of the comparison group also reported
depression. Meiselman (1978) reported depressive symptoms in
35% of the incest victims whose psychiatric records she reviewed,
compared with 23% of the comparison group; again, this differ-
ence was not significant.
Both clinical and nonclinical samples have shown victims of
child sexual abuse to be more self-destructive, however. In an
extensive study of 153 "walk-ins" to a community health coun-
seling center, Briere (1984) reported that 51% of the sexual abuse
victims, versus 34% of nonabused clients, had a history of suicide
attempts. Thirty-one percent of victims, compared with 19% of
nonabused clients, exhibited a desire to hurt themselves. A high
incidence of suicide attempts among victims of child sexual abuse
has been found by other clinical researchers as well (e.g., Har-
rison, Lumry, & Claypatch, 1984; Herman, 1981). Bagley and
Ramsay (1985), in their community study, noted an association
between childhood sexual abuse and suicide ideation or deliberate
attempts at self-harm. And Sedney and Brooks (1984) found
that 39% of their college student sample with child sexual ex-
periences reported having thoughts of hurting themselves, corn-

70 ANGELA BROWNE AND DAVID FINKELHOR
pared with 16% of the control group. Sixteen percent of these
respondents had made at least one suicide attempt (vs. 6% of
their peers).
Another reaction observed in adults who were sexually vic-
timized as children is symptoms of anxiety or tension. Briere
(1984) reported that 54% of the sexual abuse victims in his clin-
ical sample experienced anxiety attacks (compared with 28% of
the nonvictims), 54% reported nightmares (vs. 23%), and 72%
had difficulty sleeping (compared with 55% of the nonvictims).
In their college sample, Sedney and Brooks (1984) found 59%
with symptoms indicating nervousness and anxiety (compared
with 41% of the controls); 41% indicated extreme tension (vs.
29% of the controls), and 51% had trouble sleeping (compared
with 29% of the controls). These findings are supported by results
from community samples, with Bagley and Ramsay (1985) noting
that 19% of their subjects who had experienced child sexual abuse
reported symptoms indicating somatic anxiety on the Middlesex
Hospital Questionnaire, compared with 9% of the nonabused
subjects.
The idea that sexual abuse victims continue to feel isolated
and stigmatized as adults also has some support in the empirical
literature, although these findings come only from the clinical
populations. Sixty-four percent of the victimized women in
Briere's (1984) study reported feelings of isolation, compared
with 49% of the controls. With incest victims, the figures are
even higher: Herman (1981) reported that all of the women who
had experienced father-daughter incest in her clinical sample
had a sense of being branded, marked, or stigmatized by the
victimization. Even in a community sample of incest victims,
Courtois (1979) found that 73% reported they still suffered from
moderate to severe feelings of isolation and alienation.
Although a negative self-concept was not confirmed as an ini-
tial effect, evidence for it as a long-term effect is much stronger.
Bagley and Ramsay (1985) found that 19% of the child sexual
abuse victims in their random sample scored in the "very poor"
category on the Coopersmith self-esteem inventory (vs. 5% of
the control group), whereas only 9% of the victims demonstrated
"very good" levels of self-esteem (compared with 20% of the
controls). Women with very poor self-esteem were nearly four
times as likely to report a history of child sexual abuse as were
the other subjects. As might be expected, self-esteem problems
among clinical samples of incest victims tended to be much
greater: Eighty-seven percent of Courtois's (1979) community
sample reported that their sense of self had been moderately to
severely affected by the experience of sexual abuse from a family
member. Similarly, Herman ( 1981) found that 60% of the incest
victims in her clinical sample were reported to have a "predom-
inantly negative self-image," as compared with 10% of the com-
parison group with seductive but not incestuous fathers.
Impact on Interpersonal Relating
Women who have been sexually victimized as children report
problems in relating both to women and men, continuing prob-
lems with their parents, and difficulty in parenting and responding
to their own children. In DeYoung's (1982) sample, 79% of the
incest victims had predominantly hostile feelings toward their
mothers, whereas 52% were hostile toward the abuser. Meiselman
(1978) found that 60% of the incest victims in her psychotherapy
sample disliked their mothers and 40% continued to experience
strong negative feelings toward their fathers. Herman ( 1981) also
noted that the rage of incest victims in her sample was often
directed toward the mother and observed that they seemed to
regard all women, including themselves, with contempt.
In addition, victims reported difficulty trusting others that
included reactions of fear, hostility, and a sense of betrayal. Briere
(1984) noted fear of men in 48% of his clinical subjects (vs. 15%
of the nonvictims), and fear of women in 12% (vs. 4% of those
who had not been sexually victimized). Incest victims seem
especially likely to experience difficulty in close relationships:
Sixty-four percent of the victims in Meiselman's (1978) clinical
study, compared with 40% of the control group, complained of
conflict with or fear of their husbands or sex partners, and 39%
of the sample had never married. These results are supported
by findings from Courtois's (1979) sample, in which 79% of the
incest victims experienced moderate or severe problems in re-
lating to men, and 40% had never married.
There is at least one empirical study that lends support to the
idea that childhood sexual abuse also affects later parenting.
Goodwin, McCarthy, and Divasto (1981) found that 24% of
mothers in the child abusing families they studied reported incest
experiences in their childhoods, compared with 3% ofa nonabu-
sive control group. They suggested that dit~culty in parenting
results when closeness and affection is endowed with a sexual
meaning, and observed that these mothers maintained an emo-
tional and physical distance from their children, thus potentially
setting the stage for abuse.
Another effect on which the empirical literature agrees is the
apparent vulnerability of women who have been sexually abused
as children to be revictimized later in life. Russell (in press), in
her probability sample of 930 women, found that between 33%
and 68% of the sexual abuse victims (depending on the serious-
ness of the abuse they suffered) were raped later on, compared
with 17% of women who were not childhood victims. Fromuth
(1983), in surveying 482 female college students, found evidence
that women who had been sexually abused before the age of 13
were especially likely to later become victims of nonconsensual
sexual experiences. Further evidence of a tendency toward revic-
timization comes from a study conducted at the University of
New Mexico School of Medicine on 341 sexual assault admit-
tances (Miller et al., 1978). In comparing women who had been
raped on more than one occasion with those who were reporting
a first-time rape, researchers found that 18% of the repeat victims
had incest histories, compared with only 4% of first-time victims.
In addition to rape, victims of child sexual abuse also seem
more likely to be abused later by husbands or other adult partners.
Russell (in press) found that between 38% and 48% of the child
sexual abuse victims in her community sample had physically
violent husbands, compared with 17% of women who were not
victims; in addition, between 40% and 62% of the abused women
had later been sexually assaulted by their husbands, compared
with 21% of nonvictims. Similarly, Briere (1984) noted that 49%
of his clinical sexual abuse sample reported being battered in
adult relationships, compared with 18% of the nonvictim group.
Effects on Sexuality
One of the areas receiving the most attention in the empirical
literature on long-term effects concerns the impact of early sexual
abuse on later sexual functioning. Almost all clinically based

IMPACT OF CHILD SEXUAL ABUSE 71
studies show later sexual problems among child sexual abuse
victims, particularly among the victims of incest. However, there
have not yet been community-based studies on the sexual func-
tioning of adults molested as children, as there have been of
other mental health areas such as depression.
Of the clinical studies, Meiselman (1978) found the highest
percentage of incest victims reporting problems with sexual ad-
justment. Eighty-seven percent of her sample were classified as
having had a serious problem with sexual adjustment at some
time since the molestation, compared with 20% of the compar-
ison group (women who had been in therapy at the same clinic,
but had not been sexually victimized as children). Results from
Herman's (1981) study are somewhat less extreme: Fifty-five
percent of the incest victims reported later sexual problems, al-
though they were not significantly different from women with
seductive fathers on this measure. Langmade (1983) compared
a group of women in therapy who had been incest victims with
a matched control group of nonvictimized women and found
that the incest victims were more sexually anxious, experienced
more sexual guilt, and reported greater dissatisfaction with their
sexual relationships than the controls. In his study of a walk-in
sample to a community health clinic, Briere (1984) found that
45% of women who had been sexually abused as children reported
difficulties with sexual adjustment as adults, compared with 15%
of the control group. Briere also noted a decreased sex drive in
42% of the victims studied, versus 29% of the nonvictims.
Two nonclinical studies show effects on sexual functioning as
well. Courtois noted that 80% of the former incest victims in
her sample reported an inability to relax and enjoy sexual activity,
avoidance of or abstention from sex, or, conversely, a compulsive
desire for sex. Finkelhor (1979), studying college students, de-
veloped a measure of sexual self-esteem and found that child
sexual abuse victims reported significantly lower levels of sexual
self-esteem than their nonabused classmates. However, Fromuth
(1983), in a similar study also with a college student sample,
found no correlation between sexual abuse and sexual self-es-
teem, desire for intercourse, or students' self-ratings of their sexual
adjustment. Virtually all (96%) of Fromuth's respondents were
unmarried and their average age was 19, so it is possible that
some of the long-term sexual adjustment problems reported by
women in the clinical and community samples were not yet in
evidence in this younger population. Still, this does not explain
the discrepancy from the Finkelhor findings.
In another study, Tsai, Feldman-Summers, and Edgar (1979)
compared three groups of women on sexual adjustment mea-
sures: sexual abuse victims seeking therapy, sexual abuse victims
who considered themselves well-adjusted and had not sought
therapy, and a nonvictimized matched control group. Results
indicated that the "well-adjusted" victims were not significantly
different from the control group on measures of overall and sexual
adjustment, but the victims seeking therapy did show a difference.
They experienced orgasm less often, reported themselves to be
less sexually responsive, obtained less satisfaction from their sex-
ual relationships, were less satisfied with the quality of their close
relationships with men, and reported a greater number of sexual
partners. It is hard to know how to interpret findings from a
group of victims solicited on the basis of feeling "well-adjusted."
This seems far different from a comparison group of victims
who were not in therapy, and thus these results are questionable.
A long-term effect of child sexual abuse that has also received
a great deal of attention in the literature is an increased level of
sexual behavior among victims, usually called promiscuity (e.g.,
Courtois, 1979; DeYoung, 1982; Herman, 1981; Meiselman,
1978). Herman noted that 35% of the incest victims in her sample
reported promiscuity and observed that some victims seemed
to have a "repertoire of sexually stylized behavior" that they
used as a way of getting affection and attention (p. 40). DeYoung
(1982) reported that 28% of the victims in her sample had en-
gaged in activities that could be considered promiscuous; Mei-
selman (1978) found 25%. However, in her study of 482 female
college students, Fromuth (1983) found no differences in this
variable and observed that having experienced child sexual abuse
only predicted whether subjects would describe themselves as
promiscuous, not their actual number of partners. This poten-
tially very important finding suggests that the "promiscuity" of
sexual abuse victims may be more a function of their negative
self-attributions, already well documented in the empirical lit-
erature, than their actual sexual behavior; thus researchers should
be careful to combine objective behavioral measures with this
type of self-report.
Another question that has received comment but little em-
pirical confirmation concerns the possibility that sexual abuse
may be associated with later homosexuality in victims. Although
one study of lesbians found molestation in their backgrounds
(Gundlach, 1977), Bell and Weinberg (1981), in a large-scale,
sophisticated study of the origin of sexual preference, found no
such association. Studies from the sexual abuse literature have
also found little connection (Finkelhor, 1984; Fromuth, 1983;
Meiselman, 1978).
Effects on Social Functioning
Several studies of special populations suggest a connection
between child sexual abuse and later prostitution. James and
Meyerding (1977) interviewed 136 prostitutes and found that
55% had been sexually abused as children by someone 10 or
more years older, prior to their first intercourse. Among adoles-
cents in the sample, 65% had been forced into sexual activity
before they were 16-years old. Similarly, Silbert and Pines (1981 )
found that 60% of the prostitutes they interviewed had been
sexually abused before the age of 16 by an average of two people
for an average of 20 months. (The mean age of these children at
the time of their first victimization was 10.) They concluded
that, "The evidence linking juvenile sexual abuse to prostitution
is overwhelming" (p. 410). However, Fields (1981) noted that,
although 45% of the prostitutes in her sample had been sexually
abused as children, this did not differentiate them from a com-
parison group of nonprostitutes matched on age, race, and ed-
ucation, of which 37% had been abused. Although there was no
difference in prevalence between the two groups, Fields did find
that the prostitutes were sexually abused at a younger age--14.5
versus 16.5--and were more apt to have been physically forced.
An association between child sexual abuse and later substance
abuse has also received empirical support. S. Peters (1984), in a
carefully controlled community study, found that 17% of the
victimized women had symptoms of alcohol abuse (vs. 4% of
nonvictimized women), and 27% abused at least one type of
drug (compared with 12% of nonvictimized women). Herman

72 ANGELA BROWNE AND DAVID FINKELHOR
(1981) noted that 35% of the women in her clinical sample with
incestuous fathers abused drugs and alcohol (vs. 5% of the women
with seductive fathers). Similarly, Briere (1984), in his walk-in
sample from a community health center, found that 27% of the
childhood sexual abuse victims had a history of alcoholism
(compared with 11% of nonvictims), and 21% had a history of
drug addiction (vs. 2% of the nonvictims). College student sam-
ples appear more homogeneous: Sedney and Brooks (1984) found
a surprisingly low reported incidence of substance abuse, and
no significant differences between groups.
Summary of Long-Term Effects
Empirical studies with adults confirm many of the long-term
effects of sexual abuse mentioned in the clinical literature. Adult
women victimized as children are more likely to manifest
depression, self-destructive behavior, anxiety, feelings of isolation
and stigma, poor self-esteem, a tendency toward revictimization,
and substance abuse. Difficulty in trusting others and sexual
maladjustment in such areas as sexual dysphoria, sexual dys-
function, impaired sexual self-esteem, and avoidance of or ab-
stention from sexual activity have also been reported by empirical
researchers, although agreement between studies is less consistent
for the variables on sexual functioning.
Impact of Sexual Abuse
In light of the studies just reviewed, it is appropriate to evaluate
the persistent controversy over the impact of sexual abuse on
victims. It has been the continuing view of some that sexual
abuse is not traumatic or that its traumatic impact has been
greatly overstated (Constantine, 1977; Henderson, 1983; Ramey,
1979). Proponents of this view contend that the evidence for
trauma is meager and based on inadequate samples and unwar-
ranted inferences. Because of the general lack of research in this
field, clinicians have only recently been able to substantiate their
impressions that sexual abuse is traumatic with evidence from
strong scientific studies. However, as evidence now accumulates,
it conveys a clear suggestion that sexual abuse is a serious mental
health problem, consistently associated with very disturbing
subsequent problems in some important portion of its victims.
Findings of long-term impact are especially persuasive. Eight
nonclinical studies of adults (Bagley & Ramsay, 1985; Briere &
Runtz, 1985; Finkelhor, 1979; Fromuth, 1983; S. Peters, 1984;
Russell, in press; Sedney & Brooks, 1984; Seidner & Calhoun,
1984), including three random sample community surveys, found
that child sexual abuse victims in the "normal" population had
identifiable degrees of impairment when compared with non-
victims. Although impairments in these nonclinical victims are
not necessarily severe, all the studies that have looked for long-
term impairment have found it, with the exception of one (Tsai
et al., 1979).
These findings are particularly noteworthy in that the studies
were identifying differences associated with an event that occurred
from 5 to 25 years previously. Moreover, all these studies used
fairly broad definitions of sexual abuse that included single ep-
isodes, experiences in which no actual physical contact occurred,
and experiences with individuals who were not related to or
emotionally close to the subjects. In all four studies that used
multivariate analyses (Bagley & Ramsay 1985; Finkelhor, 1984;
Fromuth, 1983; S. Peters, 1984), differences in the victimized
group remained after a variety of background and other factors
had been controlled. The implication of these studies is that a
history of childhood sexual abuse is associated with greater risk
for mental health and adjustment problems in adulthood.
Unfortunately, although the studies indicate higher risk, they
are not so informative about the actual extent of impairment.
In terms of simple self-assessments, 53% of intrafamilial sexual
abuse victims in Russell's (in press) community survey reported
that the experience resulted in "some" or "great" long-term ef-
fects on their lives. Assessments with standardized clinical mea-
sures show a more modest incidence of impairment: In Bagley
& Ramsay's (1985) community survey, 17% of sexual abuse vic-
tims were clinically depressed as measured by the CES-D, and
18% were seriously psychoneurotic. Thus, most sexual abuse
victims in the community, when evaluated in surveys, show up
as slightly impaired or normal. It is possible, however, that some
of the impairment associated with childhood molestation is not
tapped by these survey evaluations.
Summarizing, then, from studies of clinical and nonclinical
populations, the findings concerning the trauma of child sexual
abuse appear to be as follows: In the immediate aftermath of
sexual abuse, from one-fifth to two-fifths of abused children seen
by clinicians manifest pathological disturbance (Tufts, 1984),
When studied as adults, victims as a group demonstrate im-
pairment when compared with their nonvictimized counterparts,
but under one-fifth evidence serious psychopathology. These
findings give reassurance to victims that extreme long-term effects
are not inevitable. Nonetheless, they also suggest that the risk
of initial and long-term mental health impairment for victims
of child sexual abuse should be taken very seriously.
Effects by Type of Abuse
Although the foregoing sections have been concerned with the
various effects of abuse, there are also important research ques-
tions concerning the effects of various kinds of abuse. These have
usually appeared in the form of speculation about what types
of abuse have the most serious impact on victims. Groth (1978),
for example, on the basis of his clinical experience, contended
that the greatest trauma occurs in sexual abuse that (a) continues
for a longer period of time, (b) occurs with a more closely related
person, (c) involves penetration, and (d) is accompanied by
aggression. To that list, MacFarlane (1978) added experiences in
which (e) the child participates to some degree, (f) the parents
have an unsupportive reaction to disclosure of the abuse, and
(g) the child is older and thus cognizant of the cultural taboos
that have been violated. Such speculations offer fruitful directions
for research. Unfortunately, however, only a few studies on the
effects of sexual abuse have had enough cases and been sophis-
ticated enough methodologically to look at these questions em-
pirically. Furthermore, the studies addressing these issues have
reached little consensus in their findings.
Duration and Frequency of Abuse
Although many clinicians take for granted that the longer an
experience goes on, the more traumatic it is, this conclusion is
not clearly supported by the available studies. Of nine studies,

IMPACT OF CHILD SEXUAL ABUSE 73
only four found duration associated with greater trauma. (We
are treating duration and frequency synonymously here because
they tend to be so highly correlated.) Three found no relation,
and two even found some evidence that longer duration is as-
sociated with less trauma.
Russell's (in press) study reported the clearest association: In
her survey of adult women, 73% of sexual abuse that lasted for
more than 5 years was self-rated as extremely or considerably
traumatic by the victims, compared with 62% of abuse lasting
l week to 5 years and 46% of abuse occurring only once. Tsai
et al. (1979) found duration and frequency associated with greater
negative effects, when measured with the Minnesota Multiphasic
Personality Inventory and a problems checklist, at least in their
group of adult sexual abuse victims who sought counseling. Bag-
ley and Ramsay (1985) found that the general mental health
status of adult victims--measured by a composite of indicators
concerning depression, psychoneurosis, suicidal ideation, psy-
chiatric consultation, and self-concept--was worse for longer
lasting experiences. Finally, Friedrich, Urguiza, and Beilke (in
press), studying children, found that both duration and frequency
predicted disturbances measured by the CBCL, even in multi-
variate analysis.
However, other studies have not found such relations. Finkeihor
(-1979), in a retrospective survey of college students, used a self-
rating of how negative the experience was in retrospect and found
no association with duration. Langmade (1983) reported that
adult women seeking treatment who had had long or short du-
ration experiences did not differ on measures of sexual anxiety,
sexual guilt, or sexual dissatisfaction. In addition, the Tufts (1984)
study, looking at child victims with more comprehensive mea-
sures than Friedrich et al. (in press), could find no association
between duration of abuse and measures of distress, using the
Louisville Behavior Checklist and the Purdue Self-Concept Scale,
as well as other measures.
Finally, some studies indicated a completely reversed relation.
Courtois (1979), surprisingly, found that adult victims with the
longest lasting experiences reported the least trauma. In addition,
in their college student sample, Seidner and Calhoun (1984) re-
ported that a high frequency of abuse was associated with higher
self-acceptance (but lower social maturity) scores on the Cali-
fornia Psychological Inventory.
In summary, then, the available studies reach quite contra-
dictory conclusions about the relation between duration and
trauma. However, duration is closely related to other aspects of
the abuse experience--e.g., age at onset, a family relationship
between victim and offender, and the nature of the sexual activity.
Some of the contradictions may be cleared up when we have
better studies with well-defined multivariate analyses that can
accurately assess the independent effect of duration.
Relationship to the Offender
Popular and clinical wisdom holds that sexual abuse by a close
relative is more traumatic than abuse by someone outside the
family. Empirical findings suggest that this may be the case, at
least for some types of family abuse. Three studies have found
more trauma resulting from abuse by relatives than by nonrel-
atives: Landis (1956), in an early study asking students about
how they had recovered; Anderson et al. (1981), in a chart review
of adolescents in a hospital treatment setting; and Friedrich et
al. (in press), in their evaluation of young victims. However, other
researchers (Finkelhor, 1979; Russell, in press; Seidner & Cal-
houn, 1984; Tufts, 1984) found no difference in the impact of
abuse by family members versus abuse by others.
It must be kept in mind that how closely related a victim is
to the offender does not necessarily reflect how much betrayal
is involved in the abuse. Abuse by a trusted neighbor may be
more devastating than abuse by a distant uncle or grandfather.
Also, whereas abuse by a trusted person involves betrayal, abuse
by a stranger or more distant person may involve more fear, and
thus be rated more negatively. These factors may help explain
why the relative-nonrelative distinction is not necessarily a con-
sistent predictor of trauma.
What has been more consistently reported is greater trauma
from experiences involving fathers or father figures compared
with all other types of perpetrators, when these have been sep-
arated out. Russell (in press) and Finkelhor (1979) both found
that abuse by a father or stepfather was significantly more trau-
matic for victims than other abuse occurring either inside or
outside the family. The Tufts (1984) study also reported that
children abused by stepfathers showed more distress, but for
some reason it did not find the same elevated level of distress
among victims abused by natural fathers. Bagley and Ramsay
(1985) found a small but nonsignificantly greater amount of im-
pairment in women molested by fathers and stepfathers.
Type of Sexual Act
Results of empirical studies generally suggest, with a couple
of important exceptions, that the type of sexual activity is related
to the degree of trauma in victims. Russell's findings on long-
term effects in adult women are the most clear-cut: Fifty-nine
percent of those reporting completed or attempted intercourse,
fellatio, cunnilingus, analingus, or anal intercourse said they were
extremely traumatized, compared with only 36% of those who
experienced manual touching of unclothed breasts or genitals
and 22% of those who reported unwanted kissing or touching of
clothed parts of the body. The community study by Bagley and
Ramsay (1985) confirms this, in a multivariate analysis that found
penetration to be the single most powerful variable explaining
severity of mental health impairment, using a composite of stan-
dardized instruments.
Moreover, four other studies confirm the relation between type
of sexual contact and subsequent effects by demonstrating that
the least serious forms of sexual contact are associated with less
trauma (Landis, 1956; S. Peters, 1984; Seidner & Calhoun, 1984;
Tufts, 1984). However, some of these studies did not find the
clear differentiation that Russell and Bagley and Ramsay did
between intercourse and genital touching. The Tufts (1984) study,
for example, using measures of children's anxiety, found children
who had been fondled without penetration to be more anxious
than those who actually suffered penetration. Moreover, there
are three additional studies (Anderson et al., 1981; Finkelhor,
1979; Fromuth, 1983) that do not show any consistent relation
between type of sexual activity and effect. Thus, a number of
studies concur that molestation involving more intimate contact
is more traumatic than less intimate contact. However, there is

74 ANGELA BROWNE AND DAVID FINKELHOR
some disagreement about whether intercourse and penetration
are demonstrably more serious than simple manual contact.
Force and Aggression
Five studies, three of which had difficulty finding expected
associations between trauma and many other variables, did find
an association between trauma and the presence of force. With
Finkelhor's (1979) student samples, use of force by an abuser
explained more of a victim's negative reactions than any other
variable, and this finding held up in multivariate analysis. Fro-
muth (1983), in a replication of the Finkelhor study, found similar
results. In Russell's (in press) study 71% of the victims of force
rated themselves as extremely or considerably traumatized,
compared with 47% of the other victims.
The Tufts (1984) study found force to be one of the few vari-
ables associated with children's initial reactions: Children sub-
jected to coercive experiences showed greater hostility and were
more fearful of aggressive behavior in others. Tufts researchers
reported that physical injury (i.e., the consequence of force) was
the aspect of sexual abuse that was most consistently related to
the degree of behavioral disturbances manifested in the child, as
indicated by the LBC and other measures. Similarly, Friedrich
et al. (in press) found the use of physical force to be strongly
correlated with both internalizing and externalizing symptoms
on the CBCL.
Three other studies present dissenting findings, however. An-
derson et al. (1981), in studying initial effects, concluded that,
"the degree of force or coercion used did not appear to be related
to presence or absence of psychosocial sequelae" in the adoles-
cents they evaluated (p. 7). Seidner and Calhoun (1984), in an
ambiguous finding, noted that force was associated with lower
social maturity but higher self-acceptance. In addition, Bagley
and Ramsay (1985) found that force was associated with greater
impairment, but this association diminished to just below the
significance level in multivariate analysis. Despite these findings,
we are inclined to give credence to the studies showing force to
be a major traumagenic influence, especially given the strong
relation found by Finkelhor, Friedrich et al., Fromuth, Russell,
and the Tufts study. Although some have argued that victims of
forced abuse should suffer less long-term trauma because they
could more easily attribute blame for abuse to the abuser
(MacFarlane, 1978), empirical studies do not seem to provide
support for this supposition.
Age at Onset
There has been a continuing controversy in the literature about
how a child's age might affect his or her reactions to a sexually
abusive experience. Some have contended that younger children
are more vulnerable to trauma because of their impressionability.
Others have felt that their naivet6 may protect them from some
negative effects, especially if they are ignorant of the social stigma
surrounding the kind of victimization they have suffered. Un-
fortunately, findings from the available studies do not resolve
this dispute.
Two studies of long-term effects do suggest that younger chil-
dren are somewhat more vulnerable to trauma. Meiselman
(1978), in her chart review of adults in treatment, found that
37% of those who experienced incest prior to puberty were se-
riously disturbed, compared with only 17% of those who were
victimized after puberty. Similarly, Courtois (1979), in her com-
munity sample, assessed the impact of child sexual abuse on
long-term relationships with men and the women's sense of self,
and also found more effects from prepubertal experiences.
However, four other studies found no significant relation be-
tween age at onset and impact. Finkelhor (1979), in a multivariate
analysis, found a small but nonsignificant tendency for younger
age to be associated with trauma. Russell (in press) also found
a small but nonsignificant trend for experiences before age 9 to
be associated with more long-term trauma. Langmade (1983)
could find no difference in sexual anxiety, sexual guilt, or sexual
dissatisfaction in adults related to the age at which they were
abused. Bagley and Ramsay 0985) found an association between
younger age and trauma, but that association dropped out in
multivariate analysis, especially when controlling for acts in-
volving penetration.
The Tufts (1984) study gave particular attention to children's
reactions to abuse at different ages. Tufts researchers concluded
that age at onset bore no systematic relation to the degree of
disturbance. They did note that latency-age children were the
most disturbed, but this finding appeared more related to the
age at which the children were evaluated than the age at which.
they were first abused. They concluded that the age at which
abuse begins may be less important than the stages of develop-
ment through which the abuse persists.
In summary, studies tend to show little clear relation between
age of onset and trauma, especially when they control for other
factors. If there is a trend, it is for abuse at younger ages to be
more traumatic. Both of the initial hypotheses about age of onset
may have some validity, however: Some younger children may
be protected by naivet6, whereas others are more seriously trau-
matized by impressionability. However, age interacts with other
factors like relationship to offender, and until more sophisticated
analytical studies are done, we cannot say whether these current
findings of a weak relation mean that age has little independent
effect or is simply still masked in complexity.
Sex of Offender
Perhaps because there are so few female offenders (Finkelhor
& Russell, 1984), very few studies have looked at impact ac-
cording to the sex of the offender. Two studies that did (Finkelhor,
1984; Russell, in press) both found that adults rated experiences
with male perpetrators as being much more traumatic than those
with female perpetrators. A third study (Seidner & Calhoun,
! 984) found male perpetrators linked with lower self-acceptance,
but higher social maturity, in college-age victims.
Adolescent and Adult Perpetrators
There are also very few studies that have looked at the question
of whether age of the perpetrator makes any difference in the
impact of sexual abuse on victims. However, two studies using
college student samples (Finkelhor, 1979; Fromuth, 1983) found
that victims felt significantly more traumatized when abused by
older perpetrators. In Finkelhor's multivariate analysis (which
controlled for other factors such as force, sex of perpetrator, type

IMPACT OF CHILD SEXUAL ABUSE 75
of sex act, and age of the offender), age of the offender was the
second most important factor predicting trauma. Fromuth (1983)
replicated these findings. Russell (in press), with a community
sample, reported consistent, but qualifying, results: In her survey,
lower levels of trauma were reported for abuse with perpetrators
who were younger than 26 or older than 50. The conclusion that
experiences with adolescent perpetrators are less traumatic seems
supported by all three studies.
Telling or Not Telling
There is a general clinical assumption that children who feel
compelled to keep the abuse a secret in the aftermath suffer
greater psychic distress as a result. However, studies have not
confirmed this theory. Bagley and Ramsay (1985) did find a sim-
ple zero-order relation between not telling and a composite mea-
sure of impairment based on depression, suicidal ideas, psychi-
atric consultation, and self-esteem. However, the association be-
came nonsignificant when controlled for other factors. Finkelhor
(1979), in a multivariate analysis, also found that telling or not
telling was essentially unrelated to a self-rated sense of trauma.
Further, the Tufts (1984) researchers, evaluating child subjects,
reported that the children who had taken a long time to disclose
the abuse had the least anxiety and the least hostility. Undoubt-
edly, the decision to disclose is related to many factors about the
experience, which prevents a clear assessment of its effects alone.
For example, although silence may cause suffering for a child,
social reactions to disclosure may be less intense if the event is
long past. Moreover, the conditions for disclosure may be sub-
stantially different for the current generation than they were for
past generations. Thus, any good empirical evaluation of the
effects of disclosure versus secrecy needs to take into account
the possibility of many interrelationships.
Parental Reaction
Only two studies have looked at children's trauma as a function
of parental reaction, even though this is often hypothesized to
be related to trauma. The Tufts (1984) study found that when
mothers reacted to disclosure with anger and punishment, chil-
dren manifested more behavioral disturbances. However, the
same study did not find that positive responses by mothers were
systematically related to better adjustment. Negative responses
seemed to aggravate, but positive responses did not ameliorate,
the trauma. Anderson et al. (1981) found similar results: They
noted
21/2
times the number of symptoms in the children who
had encountered negative reactions from their parents. Thus,
although only based on two studies of initial effects, the available
evidence indicates that negative parental reactions aggravate
trauma in sexually abused children.
Institutional Response
There is a great deal of interest in how institutional response
may affect children's reactions to abuse, but little research has
been done. Tufts (1984) researchers found that children removed
from their homes following sexual abuse exhibited more overall
behavior problems, particularly aggression, than children who
remained with their families. However, the children who were
removed in the Tufts study were also children who had experi-
enced negative reactions from their mothers, so this result may
be confounded with other factors related to the home environ-
ment.
Summary of Contributing Factors
From this review of empirical studies, it would appear that
there is no contributing factor that all studies agree on as being
consistently associated with a worse prognosis. However, there
are trends in the findings. The preponderance of studies indicate
that abuse by fathers or stepfathers has a more negative impact
than abuse by other perpetrators. Experiences involving genital
contact seem to be more serious. Presence of force seems to
result in more trauma for the victim. In addition, when the per-
petrators are men rather than women, and adults rather than
teenagers, the effects of sexual abuse appear to be more disturbing.
These findings should be considered tentative, however, being
based on only two studies apiece. When families are unsupportive
of the victims, and/or victims are removed from their homes,
the prognosis has also been shown to be worse; again, these find-
ings are based on only two studies.
Concerning the age of onset, the more sophisticated studies
found no significant relation, especially when controlling for other
factors; however, the relation between age and trauma is especially
complex and has not yet been carefully studied. In regard to the
impact of revealing the abuse, as opposed to the child keeping
it a secret, current studies also suggest no simple relation. Of all
these areas, there is the least consensus on the effect of duration
of abuse on impact.
Discussion
Conclusions from the foregoing review must be tempered by
the fact that they are based on a body of research that is still in
its infancy. Most of the available studies have sample, design,
and measurement problems that could invalidate their findings.
The study of the sexual abuse of children would greatly benefit
from some basic methodological improvements.
Samples.
Many of the available studies are based on samples
of either adult women seeking treatment or children whose mo-
lestation has been reported. These subjects may be very self-
selected. Especially if sexual abuse is so stigmatizing that only
the most serious cases are discovered and only the most seriously
affected victims seek help, such samples could distort our sense
of the pathology most victims experience as a result of this abuse.
New studies should take pains to expand the size and diversity
of their samples, and particularly to study victims who have not
sought treatment or been reported. Advertising in the media for
"well-adjusted" victims, as Tsai et al. (1979) did, however, does
not seem an adequate solution, as this injects a different selection
bias into the study.
We favor sampling for sexual abuse victims within the general
population, using whole communities--as in Russell's (in press),
S. Peters's (1984) and Bagley and Ramsay's (1985) designs--or
other natural collectivities (high school students, college students,
persons belonging to a health plan, etc,). Obtaining such samples
may be easier with adult than with child victims. If identified
child victims must be used, care should be taken to sample from
all such identified children, not just the ones that get referred

76 ANGELA BROWNE AND DAVID FINKELHOR
for clinical assessment and treatment and who may therefore
represent the most traumatized group.
Control groups.
Some of the empirical studies cited here did
not have comparison groups of any sort. Such a control is ob-
viously important, even if it is only a group of other persons in
treatment who were not sexually victimized (e.g., Briere, 1984;
Meiseiman, 1978). In some respects, however, this control pro-
cedure may actually underestimate the types and severities of
pathology associated with sexual abuse, because problems that
sexual abuse victims share with other clinical populations will
not show up as distinctive effects. An as yet untried, but we
believe fruitful, approach is to match victims from clinical
sources with other persons who grew up with them: that is,
schoolmates, relatives, or even unvictimized siblings.
Measurement.
Most of the studies we reviewed used fairly
subjective measures of the outcome variables in question (e.g.,
guilt feelings, fears, etc.). We are encouraged by the appearance
of studies such as the Tufts study and Bagley and Ramsay's survey,
which used batteries of objective measures. However, empirical
investigations need to go even further. To test for the specific and
diverse sequelae that have been associated with child sexual abuse,
it would appear that special sexual abuse outcome instruments
now need to be developed. Instruments designed specifically to
measure the aftereffects noted by clinicians might be more suc-
cessful at showing the true extent of pathology related to the
experience of sexual abuse in childhood.
Sexual abuse in deviant subpopulations.
Some ofthe studies
purporting to show effects of child sexual abuse are actually re-
ports of prevalence among specialized populations, such as pros-
titutes (James & Meyerding, 1977; Silbert & Pines, 1981), sex
offenders (Groth & Burgess, 1979), or psychiatric patients (Car-
men, Rieker, & Mills, 1984). To conclude from high rates of
abuse in deviant populations that sexual abuse causes the devi-
ance can be a misleading inference. Care needs to be taken to
demonstrate that the discovered rate of sexual abuse in the deviant
group is actually greater than in a relevant comparison group.
In at least one study of sex offenders, for example, although
abuse was frequent in their backgrounds, even higher rates of
prior abuse were found for prisoners who had not committed
sex crimes (Gebhard, Gagnon, Pomeroy, & Christenson, 1965).
It is important to recognize that such data do not indicate that
sexual abuse caused the deviance, only that many such offenders
have abuse in their backgrounds.
Developmentally specific effects.
In studying the initial and
long-term effects of sexual abuse, researchers must also keep in
mind that some effects of the molestation may be delayed. Al-
though no sexual difficulties may be manifest in a group of college
student victims (as in Fromuth, 1983), such effects may be yet
to appear and may manifest themselves in studies of older groups.
Similarly, developmentally specific effects may be seen among
children that do not persist into adulthood, or that may assume
a different form as an individual matures. The Tufts (1984) study
clearly demonstrated the usefulness of looking at effects by de-
fined age groupings.
Disentangling sources of trauma.
One of the most imposing
challenges for researchers is to explore the sources of trauma in
sexual abuse. Some of the apparent effects of sexual abuse may
be due to premorbid conditions, such as family conflict or emo-
tional neglect, that actually contributed to a vulnerability to abuse
and exacerbated later trauma. Other effects may be due less to
the experience itself than to later social reactions to disclosure.
Such questions need to be approached using careful multivariate
analyses in large and diverse samples, or in small studies that
match cases of sexual abuse that are similar except for one or
two factors. Unfortunately, these questions are difficult to address
in retrospective long-term impact studies, as it may be difficult
or impossible to get accurate information about some of the key
variables (e.g., how much family pathology predated the abuse).
Preoccupation with long-term effects.
Finally, there is an un-
fortunate tendency in interpreting the effects of sexual abuse (as
well as in studies of other childhood trauma) to overemphasize
long-term impact as the ultimate criterion. Effects seem to be
considered less "serious" if their impact is transient and disap-
pears in the course of development. However, this tendency to
assess everything in terms of its long-term effect betrays an
"adulto-centric" bias. Adult traumas such as rape are not assessed
ultimately in terms of whether they will have an impact on old
age: They are acknowledged to be painful and alarming events,
whether their impact lasts 1 year or 10. Similarly, childhood
traumas should not be dismissed because no "long-term effects"
can be demonstrated. Child sexual abuse needs to be recognized
as a serious problem of childhood, if only for the immediate
pain, confusion, and upset that can ensue.
References
Achenbach, T. M., & Edelbrock, C. (1983).
Manual for the child behavior
checklist.
Burlington: University of Vermont.
Anderson, S. C., Bach, C. M., & Gritiith, S. (1981, April).
Psychosocial
sequelae in intrafamilial victims of sexual assault and abuse.
Paper
presented at the Third International Conference on Child Abuse and
Neglect, Amsterdam, The Netherlands.
Bagley, C., & Ramsay, R. (1985, February).
Disrupted childhood and
vulnerability to sexual assault: Long-term sequels with implications
for counselling.
Paper presented at the Conference on Counselling the
Sexual Abuse Survivor, Winnipeg, Canada.
Bell, A., & Weinberg, M. ( 1981 ).
Sexual preference." Its development among
men and women.
Bloomington: Indiana University Press.
Benward, J., & Densen-Gerber, J. ( 1975, February).
Incest as a causative
factor in anti-social behavior: An exploratory study
Paper presented at the
meeting of the American Academy of Forensic Science, Chicago, IL.
Briere, J. (1984, April).
The effects of childhood sexual abuse on later
psychological.functioning: Defining a "post-sexual-abuse syndrome."
Paper presented at the Third National Conference on Sexual Victim-
ization of Children, Washington, DC.
Briere, J., & Runtz, M. ( 1985, August).
Symptomatology associated with
prior sexual abuse in a non-clinical sample. Paper
presented at the annual
meeting of the American Psychological Association, Los Angeles, CA.
Carmen, E., Rieker, E E, & Mills, T. (1984). Victims of violence and
psychiatric illness.
American Journal of Psychiatry, 141,
378-383.
Constantine, L. (1977).
The sexual rights of children: Implications of a
radical perspective.
Paper presented at the International Conference
on Love and Attraction, Swansea, Wales.
Constantine, L. (1980). Effects of early sexual experience: A review and
synthesis of research. In L. Constantine & E M. Martinson (Eds.),
Children and sex
(pp. 217-244). Boston: Little, Brown.
Courtois, C. (1979). The incest experience and its aftermath.
Victimology:
An International Journal, 4,
337-347.
De Francis, V. (1969).
Protecting the child victim of sex crimes committed
by adults.
Denver, CO: American Humane Association.
DeYoung, M. (1982).
The sexual victimization of children.
Jefferson, NC:
McFarland.

IMPACT OF CHILD SEXUAL ABUSE 77
Farrell, W. (1982). Myths of incest: Implications for the helping profes-
sional Paper presented at the International Symposium on Family
Sexuality, Minneapolis, MN.
Fields, P. J. (1981, November). Parent-child relationships, childhood
sexual abuse, and adult interpersonal behavior in female prostitutes.
Dissertation Abstracts International 42, 2053B.
FinkelhoK D. (1979). Sexually victimized children. New York: Free Press.
Finkelhor, D. (1984). Child sexual abuse: New theory and research. New
York: Free Press.
Finkelhor, D., & Russell, D. (1984). Women as perpetrators of sexual
abuse: Review of the evidence. In D. Finkelhor (Ed.), Child sexual
abuse: New theory and research (pp. 171-187). New York: Free Press.
Friedrich, W. N., Urquiza, A. J., & Beilke, R. (in press). Behavioral prob-
lems in sexually abused young children. Journal of Pediatric Psychology.
Fromuth, M. E. (1983, August). The long term psychological impact of
childhood sexual abuse. Unpublished doctoral dissertation, Auburn
University, Auburn, AL.
Gebhard, P., Gagnon, J., Pomeroy, W., & Christenson, C. (1965). Sex
offenders: An analysis of types. New York: Harper & Row.
Goodwin, J., McCarthy, T., & Divasto, P. (1981). Prior incest in mothers
of abused children. Child Abuse and Neglect, 5, 87-96.
Groth, N. A. (1978). Guidelines for assessment and management of the
offender. In A. Burgess, N. Groth, S. Holmstrom, & S. Sgroi (Eds.),
Sexual assault of children and adolescents (pp. 25-42). Lexington,
MA: Lexington Books.
Groth, N. A., & Burgess, A. W. (1979). Sexual trauma in the life histories
of rapists and child molesters. Victimology: An International Journal
4, 10-16.
Gundlach, R. (1977). Sexual molestation and rape reported by homo-
sexual and heterosexual women. Journal of Homosexuality, 2, 367-
384.
Harrison, P. A., Lumry, A. E., & Claypatch, C. (1984, August). Female
sexual abuse victims: Perspectives on family dysfunction, substance
use and psychiatric disorders. Paper presented at the Second National
Conference for Family Violence Researchers, Durham, NH.
Henderson, J. (1983). Is incest harmful? Canadian Journal of Psychiatry,
28, 34-39.
Herman, J. L. ( 1981). Father-daughter incest. Cambridge, MA: Harvard
University Press.
James, J., & Meyerding, J. (1977). Early sexual experiences and prosti-
tution. American Journal of Psychiatry, 134, 1381-1385.
Landis, J. (1956). Experiences of 500 children with adult sexual deviation.
Psychiatric Quarterly Supplement, 30, 91-109.
Langmade, C. J. (1983). The impact of pre- and postpubertal onset of
incest experiences in adult women as measured by sex anxiety, sex
guilt, sexual satisfaction and sexual behavior. Dissertation Abstracts
International 44, 917B. (University Microfilms No. 3592)
MacFarlane, K. (1978). Sexual abuse of children. In J. R. Chapman &
M. Gates (Eds.), The victimization of women (pp. 81-109). Beverly
Hills, CA: Sage.
Meiselman, K. (1978). Incest. San Francisco: Jossey-Bass.
Miller, J., Moeller, D., Kaufman, A., Divasto, P., Fitzsimmons, P., Pather,
D., & Christy, J. (1978). Recidivism among sexual assault victims.
American Journal of Psychiatry, 135, 1103-1104.
Nelson, J. (1981). The impact of incest: Factors in self-evaluation. In L.
Zakus & E Mahlon (Eds.), Children and sex (pp. 163-174). Boston:
Little, Brown.
Peters, J. J. (1976). Children who are victims of sexual assault and the
psychology of offenders. American Journal of Psychotherapy, 30, 398-
421.
Peters, S. D. (1984). The relationship between childhood, sexual victim-
ization and adult depression among Afro-American and white women.
Unpublished doctoral dissertation, University of California, Los An-
geles, CA.
Ramey, J. (1979). Dealing with the last taboo. Sex Information andEd-
ucation Council of the United States, 7, 1-2, 6-7.
Reich, J. W., & Gutierres, S. E. (1979). Escape/aggression incidence in
sexually abused juvenile delinquents. Criminal Justice and Behavior,
6, 239-243.
Rogers, C. M., & Terry, T. (1984). Clinical intervention with boy victims
of sexual abuse. In I. Stewart and J. Greer (Eds.), Victims of Sexual
Aggression (pp. 1-104). New York: Van Nostrand, Reinhold.
Russell, D. E. H. (in press). The secret trauma: Incest in the lives of girls
and women. New York: Basic Books.
Sandfort, T. ( 1981 ). The sexual aspect ofpaedophile relations. Amsterdam:
Pan/Spartacus.
Sedney, M. A., & Brooks, B. (1984). Factors associated with a history of
childhood sexual experience in a nonclinical female population. Journal
of the American Academy of Child Psychiatry, 23, 215, 218.
Seidner, A., & Calhoun, K. S. (1984, August). Childhood sexual abuse."
Factors related to differential adult adjustment. Paper presented at the
Second National Conference for Family Violence Researchers, Dur-
ham, NH.
Silbert, M. H., & Pines, A. M. (1981). Sexual child abuse as an antecedent
to prostitution. Child Abuse and Neglect, 5, 407-41 I.
Tsai, M., Feldman-Summers, S., & Edgar, M. (1979). Childhood moles-
tation: Variables related to differential impact of psychosexual func-
tioning in adult women. Journal of Abnormal Psychology, 88, 407-
417.
Tufts' New England Medical Center, Division of Child Psychiatry (1984).
Sexually exploited children: Service and research project. Final report
for the Office of Juvenile Justice and Delinquency Prevention. Wash-
ington, DC: U.S. Department of Justice.
Wisconsin Female Juvenile Offender Study (1982). Sex abuse among
juvenile offenders and runaways. Summary report. Madison, WI: Author.
Woods, S. C., & Dean, K. S. (1984). Final report: Sexual abuse of males
research projec t (Contract No. 90 CA/812). Washington, DC: National
Center on Child Abuse and Neglect.
Received October 16, 1984
Revision received July 16, 1985 9