Syndrome Testimony, Base Rates, and What the
Expert Can Ethically Tell the Trier of Fact in Alleged Child Abuse Cases
Jack S. Annon, Ph.D.*
In many alleged child abuse cases, the issues of recantation, delay of
disclosure, and physical, emotional, and behavioral symptoms of abuse are often
described by expert witnesses to the trier of fact. However, it is my firm opinion that there are three important topics that are relevant in
determining what an expert in this area could ethically testify about: a) syndrome testimony, b) base rates, and c) what the expert can and can not
tell the trier of fact.
The issues of "recantation" and "delayed disclosure"
have been popularized by Roland Summit (1983) as two of the five
predictable stages, or events, that a child who has been sexually abused may
allegedly go through. This is outlined in his proposed Child Sexual Abuse Accommodation
Syndrome. This is only one of many such syndromes that have been offered in courts in
recent years (e.g., Parental Alienation Syndrome; Child Sex Abuse
Battered Child Syndrome; Battered Wife Syndrome; Rape Trauma
In order to understand the term syndrome, it may help to clarify and define
some terms. A sign is an observable manifestation of a pathological condition seen
by the examiner (red splotches associated with measles). A symptom is generally
considered a subjective complaint by the individual (reporting a headache).
A disease or
disorder is a group of symptoms or signs that occur together, and implies a specific
cause or pathophysiologic process. On the other hand, a syndrome is a group of signs and symptoms that are based on their
frequent mutual occurrence that may suggest a common underlying course, pattern,
or treatment selection, but do not necessarily imply a specific cause.
Therefore, to start with, a syndrome is less specific than a disorder or a
Myers (1992) suggests that it is important to locate
syndromes on a "continuum of diagnostic certainty." Some
syndromes point with greater certainty to their causes than others do. For
example the Battered Child Syndrome (Kempe, Silverman, Steele, Droegemueller,
& Silver, 1962) points convincingly to abuse from a medical viewpoint, and
has a very high degree of diagnostic certainty (physicians look for injuries to the child that are inconsistent with the
description of how the injuries came about, and they also look for previous
unreported injuries. On the other hand, those syndromes that fall outside the traditional meaning
of syndrome and do not point to a particular cause are considered to be
"non-diagnostic syndromes" and, according to Meyers, should not be
admissible as evidence to prove that a person's symptoms resulted from a
The Child Sexual Abuse Accommodation Syndrome is an excellent example of a
non-diagnostic syndrome. This syndrome, or any components of it, does not detect
sexual abuse, rather it makes the assumption that sex abuse has already occurred
and it explains the possible reactions that an abused child may experience.
Rotgers and Barrett (1996) have pointed out that the syndrome is
"speculation" having neither empirical nor theoretical basis in
Summit (1992) himself has subsequently stated that the syndrome was not
designed to "prove" that abuse occurred, but to explain children's
reactions to sexual abuse. He wrote, "The CSAAS originated, then, not as a
laboratory hypothesis or as a designated study of a defined population. It
emerged as a summary of diverse clinical consulting experience, defined at the
interface with paradoxical forensic reaction. It should be understood without
apology that the CSAAS is a clinical opinion, not a scientific instrument"
(p.156). By his own admission, there is no empirical research support for the
syndrome, or the theory on which it is based, or what percentages of children
experience such events.
Summit also stated that, "Syndrome evidence has become a generic term
for diagnostic medical or psychological testimony which must be closely
scrutinized for scientific reliability, lest the intrinsic authority of the
expert witness improperly prejudice a jury through contrived or eccentric opinion."
And shortly after, "Had I known the legal consequences of the
word at the time, I might have better chosen a name like the child sexual abuse
accommodation pattern, to avoid any pathological or diagnostic
implications" (1992, p. 157).
Myers (1992) sums up this area well when he points out that, if a syndrome is non-diagnostic, it goes without saying that it should not be admissible to establish
the cause of a patient's symptom. Why this has caused so much trouble with the
courts is that examiners do not explain to the judges or juries the distinction
between diagnostic and non-diagnostic syndromes. In a review of 45 studies (Kendall-Tackett, Williams, & Finkelhor,
1993) found that no one symptom characterized a majority of the sexually abused
children. Most importantly, they found that approximately one-third of the victims
had no symptoms whatsoever.
In conclusion, from the above it can be seen that
there is no single
behavioral symptom, or syndrome, that is characteristic or indicative of child
sexual abuse. To the point, recantation or delayed disclosure by itself does not
directly indicate that a child was sexually abused, or that a child was not
[For a more thorough and in depth presentation of this issue see the chapter
Special Issue: Legal Issues; the Witness, Syndrome Testimony, and What the
Examiner Can Tell the Trier of Fact, from my course textbook (Annon, 1996).]
This is an important concept that many expert witnesses fail to take into
consideration when testifying in this area. Quite often, a witness testifies
that, if a given child displays a number of signs, symptoms, or behaviors (such
as "sexualized behavior" or "recantation" or "delayed
recall", etc.), then the child has most likely experienced sexual abuse.
How many signs or symptoms are necessary before a syndrome can be concluded,
and what is the prevalence of the syndrome in abused versus non-abused children?
To understand the difficulty of making conclusions based on such signs or
behaviors it is important for the expert to understand the concept of base
rates that unfortunately very few professionals understand.
For illustrative purposes, let's assume that there is a particular state in
which there are 220,000 children. Let's hypothetically assume that 20,000 of
these children have been shown to be sexually abused. Let's take another step
further and in examining these children, we find that 75% of the abused children
display symptoms of nightmares, bed-wetting, depression, and sexualized
Notice that 75% of the 20,000 children means that 15,000 of the abused
children have displayed this combination of signs and symptoms. At this point,
we examine an allegedly abused child and find this combination of signs and
symptoms in the child. Because 75% of our abused group displayed them, we might
erroneously believe that there is a 75% probability that this child we have
examined was also sexually abused.
However, let's take it one step further; remember we are talking about 20,000
out of the 220,000, which leaves 200,000 children who presumable were not
abused. Assume that we examine these, non-abused children and find that only a
small 10% of these children exhibit the symptoms of nightmares, bed-wetting,
depression, and sexualized behavior. At first glance it appears obvious that
when 75% of the abused children show such signs, and only 10% of the non-abused
children show the signs, then there must be a higher probability that the child
we are examining has been sexually abused.
However, notice that 10% of the non-abused population of 200,000 still means
that 20,000 children now are in this non-abused group with the same symptoms as
the abused group. If we take the number of children who have experienced the
symptoms in the abused situation (which is 15,000), and add them to the number
of non-abused children, (which is now 20,000) we now have a pool of 35,000
children with the symptom combination.
What is most important to note here is if we now mix up this pool so that we
cannot identify anyone, then randomly pick a child out, of the pool, there is a higher
probability that we will select a child from the non-abused population,
as compared to the abused population, and yet believe that we have an abused child.,
though only 10% of the non-abused population has the symptoms, and 75% of the
abused population does.
these figures could be changed in any way. The point is, while we may have a
lot of information from research and clinical practice concerning rates for
abused children, it is extremely important to obtain base rates from the general
population of non-abused children before we can offer any firm opinion as to the
probability of whether a child has or has not been abused.
This should be kept in mind for any examiner who is looking for any signs,
symptoms, or syndromes in any given population. This is why such studies as Friedrich, et al.(1991) that gathered normative sexual
behavior on presumably non-abused children are so important. Equally important are Friedrich
et al.'s (1992) subsequent studies where they compared
sexual behavior between abused and non-abused children.
It can be seen in these studies that some behaviors that have been put forth,
including sexualized behavior, as indicators of sexual abuse are also found in
non-abused children. On the other hand, there are also some behaviors found in
abused children that are very rarely found in non-abused children. It is very
important for any expert witness to keep base rates in mind if they are going to
make any kind of evaluative statements about what a given sign, symptom, or syndrome
indicates. [For further information on this subject refer to Annon, 1996.]
What the Expert Can and Can Not tell the Trier of Fact
In the past, the Frye (1923) general acceptance standard for expert
testimony was not used that much in mental health cases, which relied a lot on
the clinical experience of the individual offering testimony. However, in 1993
with the advent of the Daubert v. Merrill Pharmaceuticals opinion by the United
States Supreme Court, this situation began to change. This opinion again caused
quite a stir for forensic and mental health experts. Some felt it would
open the gates to all types of evidence, while others suggested it would lock the
gates to a
lot of junk science.
The first commentaries on the shift was by Underwager and Wakefield (1993), who pointed out that this finding replaced the
Frye test with the principal of "falsification" as a determinant of scientific knowledge.
their opinion that if the concept of falsification is properly understood and followed by the judges,
it would most likely render inadmissible testimony based on such concepts or
theories as the Child Sexual Abuse Accommodation Syndrome, and other similar
One of the characteristics that was pointed out by the Supreme Court was the
concept of falsifiability. In other words, do credible research studies produce
results that mean the theory cannot be held, or continue to be maintained.
The other concept of falsifiability is that the theory or concept in
question must be capable of being falsified. Could it be definitively or
conclusively falsified? Wakefield and Underwager noted that much of the expert testimony in cases of alleged sexual
abuse is unfalsifiable. They maintained that this type of testimony should be declared
inadmissible, as it is
not helpful to the trier of fact. For example, if the accused proclaims
innocence, then the individual is in denial. If the accused shows little or no
emotion when being questioned, then he must be in denial, or sociopathic, or a
manipulator. On the other hand, if the accused becomes emotional or
confronted, then he is overwhelmed by guilt and disgust by his behavior.
Most to the point are statements by so-called experts about the alleged victim
that can not be falsified. For example, the issue of immediate disclosure or
delayed disclosure is equally supported. If the child "discloses"
abuse immediately after the alleged abuse occurred, the child is seen as
overwhelmed by the abuse, or is too fearful to maintain the secret, or has found
him or herself in a "safe" situation within which to disclose the abuse.
On the other hand, if the child "discloses" abuse years after it
allegedly happened, it is assumed to be typical of abused children
that they will delay disclosure. As can be seen neither one of these positions
can be falsified or disproved.
Also, to the point, it is said that abuse is supported when the child
initially acknowledges alleged abuse, but then recants or retracts an earlier
statement. Though recanted or retracted, the abuse allegation is still true
because the child is under pressure to recant and is scared. Recantation is
described as "typical" of children who have been abused and,
therefore, cannot disprove abuse. Thus, another opinion that cannot be falsified
(see Wakefield and Underwager, 1993 for a more comprehensive discussion of this
The fact is, most children who initially allege abuse do not retract their
statements, and there are a smaller number that do retract. Some who retract
their statements have been abused, and some that retract have not been abused.
What is the base rate for these situations, and how do you determine which
statement is true? Retraction by itself does not mean that abuse did occur, or
that it did not occur.
Unfortunately, the Daubert gate-keeping function did not keep such opinions out of
our courts. In Hawaii, many "expert witnesses" are allowed to testify
to many issues that have no empirical research support and are not capable
of being falsified. The usual reason is that their testimony is not considered
"scientific" but is based on "technical" or other
"specialized knowledge" (for further information on this subject see,
However, in the wake of the recent 1999 Kumho Tire Company. v. Carmichael
United States Supreme Court decision, this may finally change. The court's
decision that the Daubert gatekeeping obligation also applies to testimony based
on technical and other specialized knowledge, as well as scientific knowledge,
may now truly screen out "junk science" and the testimony of many so
called "mental health experts" with little awareness of the issues of
syndrome testimony, base rates, and falsifiability.
This sums up the database that expert witnesses in this area should be aware
of and knowledgeable about before the proffer themselves as expert witness in
alleged child abuse cases.
Annon, J. S. The Forensic Detection of Deception, Distortion, and Malingering
in the Witness, Victim, Defendant, and Patient, Course One: Forensic Aspects of
Memory and Recall in Children. Adolescents. and Adults. A distance course of the
American College of Forensic Examiners, (1996).
Daubert v. Merrell Dow Pharmaceuticals, Inc., 113 S.Ct. 2786 (1993)
Kempe, C. H., Silverman, F. N., Steele, B. F., Droegemueller, W., & Silver,
H. K. (1962). Battered-child syndrome. Journal of the American Medical Association, 181, 17-24.
Friedrich, W. N., Grambsch, P., Broughton, B., Kuiper, J., & Beilke, R.
L. (1991). Normative sexual behavior in children. Pediatrics,
Friedrich, W. N., Grambsch, P., Damon, L., Hewitt, S., Koverola, C., Lang,
R., Wolfe, V., & Broughton, D. (1992). The child sexual behavior inventory:
Normative and clinical comparisons. Psychological Assessment,
Frye v. United States, 293 F. 1013 (D.C. Cir. 1923)
Kendall-Tackett, K. A., Williams, L. M., & Finkelhor, D. (1993). Impact of
sexual abuse on children: A review in synthesis of recent empirical studies. Psychological Bulletin,
Kumho Tire Co. v. Carmichael, 119 S.Ct. 1167 (1999).
Myers, E. B. (1992). Evidence in Child Abuse and Neglect, 2nd edition, v. 1()
& 2, New York: John
Wiley & Sons.
Summit, R. C. (1983). The child sexual abuse accommodation syndrome. Child
Abuse and Neglect, 7, 177-193.
Summit, R. C. (1992). Abuse of the child sexual abuse accommodation syndrome.
Journal of Child Sexual Abuse, 1, 153-167.
Underwager, R. & Wakefield, H. (1993). A paradigm shift for expert
witnesses. Issues in Child Abuse Accusations,
* Jack S. Annon is a clinical and forensic psychologist at 680 Ainapo
Street, Honolulu, Hawaii, 96825. [Back]