Seminar on Child Sexual Abuse
Ralph C. Underwager
V. Interviewing a child who has been exposed to multiple interrogations
A. A professional is often asked to assess a case after others have interviewed the child.
If the initial evaluation and interviews have been conducted by someone else, careful
examination of the procedures is necessary in order to assess possible contamination which
has been brought into the case (Wakefield & Underwager, 1988; White & Quinn,
1988). When children have been subjected to leading and coercive interviews, the
contamination is likely to have altered their recollections so that it becomes extremely
difficult to sort out the truth.
B. When interviewing a child who has been repeatedly exposed to adult social influence,
the goal is to permit and encourage the most reliable statements a child can make. C. A
child who has been interviewed a number of times has learned a certain set of expectations
about adult behavior and about the responses the adults reinforce.
1. Most adult interviewers are very task oriented. They aim to affirm what they believe
happened. They are dominant and active and the children are passive and compliant. The
interviewer is serious, intense, and asserts power and control to keep the child on the
task and efforts by the child to engage in spontaneous play are suppressed and controlled.
2. The oft-interviewed child has learned that when responses are given that please the
adult, there is positive reinforcement. A child gets praised, sometimes given soda,
cookies, access to a special toy, or the ending of the interview.
3. When an evaluation is ordered within the context of an adversarial proceeding, those
who have been interviewing the child may teach the child to fear the defense evaluator.
D. In this situation, the evaluator must get all possible information about the child, the
accused perpetrator, the nature of the accusations, and the process the child has endured
before interviewing the child.
E. It is helpful to have family picture albums, pictures of the accused perpetrator,
favorite toys, and objects from the child's prior experience. These items may be used as
cues to free recall and checks on the emotional responses of the child.
F. Wherever possible, observe the interaction between the child and the accused
perpetrator during the evaluation.
G. The aim is to provide an environment that permits the child to be free from pressure,
free for recall, and thus produce maximally reliable statements. The idea is to break the
mental set, change the expectancies of the child from the earlier interviews, and
establish a new system of relationship for the child.
H. There are two approaches we recommend. The first is genuinely to have fun with the
child and the second is to give power to the child.
I. Specific suggestions for interviewing the frequently-interrogated child. (Decisions
about which of these techniques to use in a particular interview depend upon the
personality of the interviewer and the particular situation.)
1. Start the interview in the waiting room with the very first contact with the child. If
there are adults accompanying the child, do not relate to them but to the child first.
2. Get into a playful mode as quickly as possible. Act expansive, jolly, entertaining so
that the child can see that you are not like all the other adults asking questions.
Children readily pick up that you are kidding and can go along with it.
3. Give power to the child. Let the child lead in the responses to your behavior and then
follow his lead. Shift the conversation back and forth, respecting the child's short
4. Use a Polaroid camera during the interview. Children respond well to photographs and to
5. The use of photographs gives an opportunity to check whether the child's responses
include a visual component. A child who has been repeatedly interviewed may have a learned
response of "bad touch" associated with a name but not a photograph. This
suggests that the account has been learned rather than reflecting a real event.
6. Use unfamiliar toys and objects to check the memory for objects that were present.
7. Permit free recall to see if what may appear to be trivial, peripheral details are
volunteered. When an event occurs in real time, there are always many other things
happening than a central event. We are aware, even if not focusing upon them, of many
stimuli. A cat may meow, a pillow fall off a couch, or an airplane fly overhead. When an
account is being given of a real time event such peripheral details will emerge.
8. When there is a negative emotional response to the accused perpetrator and/or familiar
objects, when visual cues are responded to, and when details commensurate with an actual
experience emerge, the reality of the alleged abuse is suggested V. Criteria for judging
A. Although the justice system makes the determination of the truth or falsity of an
allegation, the mental health professional is often asked to give an opinion.
B. The reports of suspected child abuse have greatly increased in the past 20 years. The
increase has been even greater for reports of child sexual abuse.
1. Approximately 60% to 65% percent of all reports of suspected child abuse turn out to be
unfounded compared to 1975 when only about 35% were unfounded (Besharov, 1991).
2. This increase in unfounded accusations results from the broadening of the reporting
laws along with the great attention given to child abuse. 3. This dramatic increase in
unfounded reports overloads the system and prevents help from reaching children who
actually are being abused. The protective service agencies are making mistakes on both
4. The more we try to reduce the number of sexually abused children that are missed, the
more we will misidentify children as sexually abused when they are not. C. Definitional
1. The disagreement over the proportion of false allegations is partially due to differing
definitions of a false allegation. Criteria such as as substantiation by child protection,
a finding by the justice system, the opinion of a mental health or law enforcement
professional, or some combination of these appear in the literature.
2. The terms substantiated and unsubstantiated create confusion. Unsubstantiated is not
the same as false and substantiated does not necessarily mean the allegation is true.
3. The category of false allegations sometimes includes all cases which cannot be
substantiated. At other times it is limited to cases in which the accuser is purposefully
deceiving and situations of good faith misunderstanding are excluded. When an accusation
of sexual abuse is false, this does not mean that it was a deliberately fabricated.
4. If actual abuse is defined in terms of substantiated cases and false allegations are
limited to deliberate fabrications, a study will report a small number of false
allegations. There will be more if a false allegation is defined as all cases that are not
substantiated. There will be a still larger number if the criteria is the justice system's
finding of abuse, since not all allegations substantiated by social services result in a
finding of abuse by the court.
D. Suggested criteria
1. A spontaneous disclosure made by a young child without evident adult influence is more
likely to be true.
a. Young children almost never initiate false allegations without influence from an adult.
b. This may be less true of older children and teenagers who have been exposed to
prevention programs or the publicity about sexual abuse.
2. Some writers believe allegations that turn out to be false often involve very young
children (Schaefer & Guyer, 1988; Everstine & Everstine, 1989, Wakefield &
Underwager, 1991b, 1994a). 3. The probability of the behaviors alleged.
a. Knowledge of the behavior of actual abusers is important for evaluating whether an
allegation is true in cases where the accused denies the allegation and there is no
corroborating evidence. If the alleged behaviors are extremely improbable, then it is less
likely that the allegation is true.
b. Investigators and therapists often appear to be ignorant of the real, actual pattern of
sexual contact between children and adults. It is crucial to think very realistically and
practically about what would actually have to happen in order for the alleged behavior to
have actually taken place.
c. Tollison and Adams (1979) describe the general behaviors engaged in by the pedophile as
Pedophiliac behavior may involve caressing a child's body, manipulating a child's
genitals, or inducing a child to manipulate an adult's genitals. Occasionally, the
behaviors also include penile penetration (partial or complete--vaginal or anal), oral
sex, and any practice utilizing the sexual parts or organs of a child so as to bring the
person in contact with the child's body in any sexual manner. Pedophiliac acts may be
homosexual or heterosexual in nature and may include touching, caressing, masturbation,
oral-genital contact, and intercourse, as well as pedophilic exhibitionism, voyeurism,
rape, sadism, and masochism....Physical violence to the child occurs in only 2 percent of
instances... (page 326).
d. In homosexual pedophilia, the most common contact is masturbation-done to rather than
by, the boy. This is followed by fellatio. Anal intercourse is less common.
e. Erickson (1985) and Erickson, Walbek, & Seely, (1988) report that vaginal and anal
penetration is rare in young children and is extremely painful. It results in injuries and
laceration, not stretching of the involved organ. The molester must have a strategy for
muffling the child's screams.
f. Vaginal penetration is more likely with an older child and is more common in clinical
samples compared to community samples (Wakefield & Underwager, 1994b).
f. Bribery is more common than threat. Threatening a child is the way to assure the
quickest disclosure when the threatening agent is not present.
g. Aggression and violence are not usually part of the behavior. Sadistic, bizarre, or
homicidal forms of abuse occur but are extremely rare.
h. There is no evidence for satanic ritual abuse conspiracies (see section below on this)
i. In incest, a grooming process is often involved (Erickson et al., 1988; Christiansen
& Blake, 1990).
j. In false cases, the allegations may initially be very vague and not easily amenable to
being verified or refuted.
4. Female child sexual abusers (see Wakefield & Underwager, 1991a).
a. Sexual abusers are primarily male.
b. There are widely different circumstances in which females may engage in behavior that
is defined as "child sexual abuse" and the circumstances that lead women to
sexually abuse children can often be differentiated from those causing men to do so.
i.One example of this is sexual abuse which occurs in conjunction with a dominant male and
in which the woman plays a secondary role.
ii.Another is found by the retrospective surveys of college men in which many of the boys
reported that they had engaged in the incidents voluntarily and did not feel victimized.
c. Some of the recent literature which discusses female perpetrators is likely to have
included cases of false accusations which gives a misleading picture of both the frequency
with which females abuse children and the characteristics of such women. An example of
this is the Finkelhor day care study (Finkelhor, Williams, & Burns, 1988; Finkelhor,
Williams, Burns, & Kalinowski, 1988).
5. Characteristics of the child's statement
a. There is general agreement that valid accounts of abuse will have appropriate details
given the child's age, especially affective and contextual details and the child's affect
will be appropriate to the report (deYoung, 1986; Faller, 1988; Jones & McGraw, 1987;
Jones & Seig, 1988; Köhnken & Steller, 1988; Sink, 1988a Raskin &
1991; Rogers, 1990; Undeutsch, 1988)
b. If there is strong hatred expressed toward the accused that is based upon trivial and
vague reasons, this may be the result of learning from the accusing parent rather than
from actual abuse (Gardner, 1987).
c. A child who is very eager to talk about the abuse may have learned that adults reward
such talk (Wakefield & Underwager, 1988).
d. Significant contradiction and variation in the story across time, especially when the
account shows that the child has no visual image but is responding to verbal cues,
supports the possibility of the child learning the story from adults.
e. However, Ceci and Bruck (1995) found that professionals were unable to differentiate
between videotapes of children describing real events and children describing events they
had learned about through suggestive questioning. Adult judgments or opinions about the
accuracy of children's accounts may be no better than chance. 6. Personality
characteristics of the persons involved.
a. A history of sexual abuse in the childhood experience of an adult involved in an
accusation may produce anxiety, sensitivity, suspiciousness, and readiness to
overinterpret innocuous events.
b. On the other hand, an experience of having been sexually abused may mean that an adult
can pick up accurate signs in the behavior of children and the perpetrator.
c. In divorce and custody cases, a personality disorder may increase the possibility of an
allegation being false (Wakefield & Underwager, 1990). d. The personality patterns of
the person accused.
i. If it cannot be demonstrated that an accused person has the pathology associated with
most child sexual abusers, the likelihood of a false accusation increases.
ii. A "normal" personality based on an MMPI or other assessment techniques does
not mean that the individual could not be a sexual abuser. A significant minority of child
sexual abusers have normal MMPIs and appear relatively normal.
iii. The presence of psychological problems does not prove that the abuse is real.
iv. Consideration of the personality characteristics of the accused is particularly
important when the allegations are of highly deviant behaviors.
v. Unusual or infrequent sexual behaviors in the life of the accused does not necessarily
increase the likelihood of being a child molester. An example is transvestite behaviors
which do not increase the probability of being a child abuser.
7. The behavior of the professionals involved:
a. Conclusions or opinions based on personal experience alone which ignore the research
data are most likely wrong (Dawes, 1989, 1994).
b. There are no scientific data supporting symbolic sign interpretations of drawings,
coloring books, play, and dolls (Underwager & Wakefield, 1995).
c. When the professional has done a careful investigation, used accepted procedures, and
considered alternative explanations, the conclusions are more likely to be accurate.
d. The tendency to make unwarranted assumptions and the failure to consider multiple
hypotheses and to gather as much uncontaminated information as possible can lead to
premature decisions (Mandel, Lehman, & Yuille, 1994). In cases of false accusations
professionals often immediately conclude that an accusation is true (Blush & Ross,
e. Investigators often do not follow the procedures established by law or by internal
regulations. The presence of procedural errors by the decision makers increases the
likelihood of a mistake being made.
8. A discernible "language game" has developed within the system for dealing
with sexual abuse accusations.
a. Examples are the surplus meaning attached to words and phrases like "touch"
or "hurt," "private parts," "bad touch," or "felt
yucky." b. When children and adults show verbal behaviors that can be identified as a
shared language game, there has been some learning and a mutual acceptance of the tacit
assumption defining the activity. c. Actual sexual abuse occurring within the customary
language of most people does not require a different language game in order to do it or
talk about it.
9. There are situational factors that increase the probability that an accusation is
a. Accusations arising within a divorce/custody dispute.
b. Accusations involving a disturbed adolescent.
c. Accusations about sexual abuse in day care centers.