Exposing Junk Science Underlying Expert Testimony in the Guilt-Innocence Phase of a Child Sexual Abuse Trial

Texas Rules of Evidence 702: If scientific, technical, or other specialized knowledge will help the trier of fact to understand the evidence or to determine a fact in issue, a witness who is qualified as an expert by knowledge, skill, experience, training, or education may testify in the form of an opinion or otherwise.

There is a great deal of solid scientific research that can be used as the basis for testimony in the guilt-innocent phase of a child sexual abuse trial. Unfortunately, there is also a great deal of junk science that routinely makes its way into these trials. For example, presently there is no scientific basis for expert testimony about the characteristics of sexually abused children during the guilt-innocence phase of a child sexual abuse trial. If Texas Rules of Evidence Rule 702 were followed in a strict manner, this type of expert testimony, along with many other topics, would be eliminated because this type of expert testimony does not help the trier of fact understand the evidence or determine a fact in issue.

There are two prongs to Rule 702. The first prong of Rule 702 describes what type of opinions an expert can offer. The second prong of Rule 702 describes who qualifies as an expert. Case law views these two prongs as independent, such that an expert could qualify under one prong but not the other. In U.S. v. Frazier (2004), the court explained the independent nature of the two prongs of Rule 702:

Of course, the unremarkable observation that an expert may be qualified by experience does not mean that experience, standing alone, is a sufficient foundation rendering reliable any conceivable opinion the expert may express. As we observed in Quiet Technology, while an expert’s overwhelming qualifications may bear on the reliability of his proffered testimony, they are by no means a guarantor of reliability . . . If admissibility could be established merely by the ipse dixit of an admittedly qualified expert, the reliability prong would be, for all practical purposes, subsumed by the qualification prong? [O]ur case law plainly establishes that one may be considered an expert but still offer unreliable testimony.

Most prosecution and defense attorneys do not know, nor would they be expected to know, how seriously lacking in empiricism some expert testimony really is. This lack of knowledge also affects rulings from the bench. If the attorneys cannot expose the junk science underlying the expert’s testimony, the court has no basis upon which to exclude such testimony.

The purpose of this article is twofold. First, it provides information that attorneys can use to argue, and hopefully exclude, junk science testimony in the guilt-innocence phase of a child sexual abuse trial. If the attorney is unsuccessful in obtaining the court’s cooperation to exclude such testimony, then the second purpose of this paper comes into play. The information herein can be used to impeach experts who use junk science.

The format of the article is simple. An attorney question is posed to a hypothetical expert. The attorney question is designed to identify an area of research that should be excluded from a child sexual abuse trial because of a lack of empirical results that could be used by the trier of fact to determine a fact in issue or understand the evidence. After the question is posed, a summary of the current state of the research will be provided. This summary is designed to be used by an attorney to: (a) argue to eliminate this type of testimony; or (b) use in cross-examination to impeach the opposing expert.

Some experts will offer testimony about studies of sexually abused children seen in mental health clinics, as if to lead the trier of fact into believing that if the alleged victim shows these symptoms, the alleged victim was most likely sexually abused. What the expert doesn’t tell the trier of fact is that there are no known symptoms, or set of symptoms, that all or most sexually abused children exhibit. In other words, just because a child is showing some mental health symptom, the presence of a mental health symptom cannot be used as proof of sexual abuse.

Additionally, the expert almost never tells the trier of fact that a sizeable percentage of sexually abused children never exhibit any symptoms that result from the sexual abuse. Consider the following research findings regarding the lack of reliable indicators of sexual abuse.

When an expert testifies that a symptom indicates a child has been sexually abused, it is based upon the implicit assumption that child sexual abuse always causes some kind of symptoms and all we must do is look hard enough and we will find the symptom proving that the abuse occurred. This implicit assumption is refuted by the scientific research. All children who were sexually abused do not show symptoms related to the sexual abuse. Early research showed that up to 36% of all known victims of child sexual abuse do not develop any symptoms as a result of the sexual abuse (Finkelhor, 1990). More recent research has found that up to 49% of child sexual abuse victims show no clinical symptomatology (Kuehnle & Connell, 2009).

What about the victims of child sexual abuse that do exhibit symptoms? Do they exhibit a specific type of symptom or a recognizable pattern or class of symptoms? No, they do not.

In the early 1990s, Dr. Green collected all the peer-reviewed research he could find concerning the symptoms of children who had been sexually abused (Green, 1993). He set out to determine if he could identify a specific pattern or group of symptoms that were routinely associated with childhood sexual abuse. Dr. Green scrutinized over 100 peer-reviewed studies. Each study had about 20 to 30 children participating in the research, which is to say that he looked at the symptoms of several thousand children. Dr. Green concluded that a wide variety of symptoms result from child sexual abuse and no group or pattern of symptoms was uniquely associated with childhood sexual abuse—i.e., “validation of sexual abuse is hampered by the lack of specific behavioral markers” (p. 890).

The holy grail of symptoms associated with childhood sexual abuse is sexual behavior. Experts have been known to opine that if a child exhibits sexual behavior, it is a sign that the child was most assuredly sexually abused.

The research on childhood sexual behavior does not support this claim. Dr. Michael Lamb (1994) authored a position paper resulting from the work of a conference of European, Canadian, and American researchers. Dr. Lamb summarized the position of the international conference of experts as follows:

No specific behavioral syndromes characterize victims of sexual abuse. Sexual abuse involves a wide range of possible behaviors which appear to have widely varying effects on its victims. The absence of any sexualized behavior does not confirm that sexual abuse did not take place any more than the presence of sexualized behavior conclusively demonstrates that sexual abuse occurred. (p. 153)

The reader probably noticed that the research articles reported in this section come from the early ’90s. Science has a way of establishing a fact and moving onto new areas of research. However, from time to time, researchers do take stock of the state of the research. With regard to the issue of scientifically based indicators of child sexual abuse, in 2009 Dr. Connell offered the following summary of the research:

One of the most important findings from the past several decades of research addressing the effects of child sexual abuse (CSA) is that no single sign or symptom, including aberrant sexualized behavior, characterizes the majority of sexually abused children. As Kuehnle (2002) wrote: Child sexual abuse is an event or a series of events, not a psychiatric disorder. The view of sexual abuse as a trigger that sets off an internal process in the child that surfaces as predictable behavioral and emotional symptoms, does not have an empirically based foundation. The array of symptoms exhibited by sexually abused children is also demonstrated by other types of maltreated and traumatized children, while a substantial percentage of sexually abused children (21% to 49%) are found to be asymptomatic. (Kuehnle & Connell, 2009; page 129)”

The current state of the research is very well established: there is no reliable way for an expert, or a jury, to use signs or symptoms to determine if a child has been sexually abused. Some experts try to sidestep the scientific literature and instead claim to be relying upon their “knowledge, skill, experience, training, or education” (Rule 702). Even if an expert does not rely upon a specific methodology, the manner in which the expert uses his or her education, training, and experience is subject to the Daubert standard, “In determining whether an expert’s testimony is reliable, the Daubert factors are applicable in cases where an expert eschews reliance on any rigorous methodology and instead purports to base his opinion merely on ‘experience’ or ‘training’” (U.S. v. Frazier, 2004).

Even before we address the matter of whether an expert can look at a victim statement and determine if sexual abuse has occurred, we must first address the issue of whether this is even a scientifically sound course of action. At issue is whether a state­ment that something has occurred (the outcry) can be used to prove that the event occurred (sexual abuse). If this sounds tautological, it is. Current research identifies this as tautology as “double dipping.”

Dr. David Faust and his colleagues (Kuehnle & Connell, 2009) explained that double dipping occurs when an indicator is used during intake to determine if a child needs to be evaluated, and it is also used during the sexual abuse evaluation as the basis for substantiating that the sexual abuse occurred. These researchers warned that any value an indicator of sexual abuse had was neutralized when it was used as the basis of the referral. For that reason, the outcry must not be re-used as proof of sexual abuse during the evaluation.

Let’s take a look at how double dipping can be misleading. Suppose that an outcry of sexual abuse was used as the basis for making referrals for an interview at a Child Advocacy Center. As you know, some children who make an outcry of sexual abuse are, in fact, sexually abused, and some children who outcry have never been sexually abused. Assume that all children who made an outcry of sexual abuse were referred for a forensic interview at a Child Advocacy Center. Let’s also assume all children who made an outcry, those who were sexually abused, and those who were not sexually abused made outcries during the forensic interview. At the end of the interview process, all children who made an outcry during the forensic interview are classified as sexually abused because they are consistent. The result: All children who were sexually abused would be correctly classified as sexually abused and all children who were never sexually abused would be incorrectly classified as sexually abused.

The foregoing is a classic example of double dipping, and it invariably results in the misclassification of children who were never sexually abused. The outcry made prior to and during the forensic interview are the same thing (i.e., they are both outcries). An outcry during the forensic interview cannot be used to prove an outcry prior to a forensic interview. When the same indicator, “outcry of sexual abuse,” is used during the referral and evaluation process, it cannot separate sexually abused children from children who were not sexually abused.

Double dipping also occurs when the indicator variables used for referral and screening are not identical but redundant, with the degree of redundancy reducing the potential additional value of the second variable proportionally (Kuehnle & Connell, 2009). If a variable is totally redundant with another variable, then it would be like using the same variable twice.

The harm of double dipping is manifold. First, double dipping guarantees misclassification of children who were never sexually abused as children who were sexually abused. Second, an expert who is unaware of the problems caused by double dip­ping might have unwarranted confidence in their findings. When these experts are on the witness stand, they could mislead the trier of fact. Third, double dipping can cause a premature end to the investigative process, resulting in a premature and erroneous conclusion, without considering all the possibilities.

Some experts claim to avoid the double dipping problem by stating that the outcry per se is not the thing that caused them to believe that sexual abuse occurred. These experts claim that they can look at child’s video or written statement and see signs that the sexual abuse occurred (e.g., sensory detail, logical sequence, detailed information, and so forth). The expert who claims this ability is claiming an ability that the scientific research says none of us possess.

The inability of professionals to look at a videotape and de­termine if a child was sexually abused was first confirmed in research done by Dr. Stephen Ceci and his colleagues (1994). Dr. Ceci showed twelve professionals ten videotapes of children. Five children were talking about things that actually happened to them. Five children were talking about false memories that the researchers had suggested to the child. The professionals who reviewed the videotapes were researchers and clinicians who worked in the area of child sexual abuse. The professionals were allowed to view, rewind, and review the videotapes as many times as they liked. At the end, the professionals were asked to identify which child was talking about a true memory and which child was talking about a false memory. The professionals did no better than flipping a coin—they were right about 50 percent of the time, or wrong about 50 percent of the time.

As you will note, Dr. Ceci’s study only involved twelve professionals. If this seems like a small sample, it is. Consequently, Dr. Ceci and his colleagues repeated this study with more than 1500 judges, researchers, and mental health professionals (Ceci and Bruck, 1995). The results have remained constant (i.e., 50 percent chance of getting it right).

Sometimes an expert will say that it was not the review of a videotape that led him to believe the child was sexually abused. Rather, the expert claims that a review of the child’s transcribed statement was the basis of the conclusion that the child showed signs of sexual abuse. However, there is research to refute the claim that transcribed statements can discriminate children who were sexually abused from children who were not.

Dr. Margie Bruck and her colleagues (1997) did a study in which she subjected children to misleading interviews. She had children describe two actual events and two fictitious events. Then she and her researchers interviewed the children four more times. Two of the interviews were designed to solidify mem­ories of the fictitious events by having children visualize the false events while the interviewer reinforced the child for recalling the fictitious event. There was even an interview in which the children told a puppet about the fictitious event. The final interview was with a new interviewer the children had never seen.

Dr. Bruck and her researchers used a highly sophisticated methodology of content analysis to code the child’s transcribed statements made during the last interview, looking for variables that were presumed to discriminate between true and false reports—including number of details, spontaneous reminiscences, consistency, contradiction, and narrative cohesion. The results: There were no significant differences between true and false stories told by the children (i.e., there is no way for a group of trained experts skilled in content analysis to identify truthful and contrived stories).

Finally, let’s consider research that was done with children who were actual sexual abuse victims. Dr. George Realmuto and his colleagues (1990) asked highly trained psychiatrists to interview children and determine which child had been sexually abused and which child had not been sexually abused. The psychiatrists correctly identified 53% of the children as either sexually abused or not sexually abused.

Dr. Realmuto’s research based upon evaluations of actual victims of child sexual abuse showed that professionals performed no better than flipping a coin when trying to differentiate sexual abuse victims from children who were never sexually abused. The chance rate found by Dr. Realmuto was the same rate that Dr. Ceci and Dr. Bruck found. Overall, it is safe to say that there is no scientific proof that a professional can reliably distinguish between sexually abused children and children who have never been sexually abused.

Some experts still offer testimony based upon a method that they developed to discriminate sexually abused children from children who were never sexually abused, despite the overwhelming, pervasive, and widely accepted opinion that it is not possible to do so. The experts who offer such testimony often claim to have their own methodology, which they have constructed and relied upon in their clinical practice. In Coble v. State (2010), the Texas Court of Criminal Appeals stated that there was no place in the courtroom for “idiosyncratic” methodologies. In that case, a psychiatrist testified in the punishment phase of a capital trial using his own “idiosyncratic” method of risk assessment. The Texas Court of Criminal Appeals determined that idiosyncratic methods do not meet the criteria for expert testimony, and the expert’s testimony was deemed to be “unreliable.”

There is research showing that if an individual has been charged with a crime, some jurors are inclined to believe that individual committed the crime (Narby, Cutler, & Moran, 1993). As appalling as that may seem, even more appalling is testimony by an expert supporting this myth, especially in the area of child sexual abuse. The research regarding base rates of confirmed cases of child sexual abuse shows that only about half of these cases were deemed to be credible.

Dr. David Jones and Dr. Melbourne McGraw (1987) analyzed all the cases (total = 576) of child sexual abuse that were reported in Denver, Colorado, in one year (1983). Of these sexual abuse cases, 53% of the reports were classified as founded.

Dr. Jones participated in a study that replicated his 1987 study (Oates, Jones, Denson, Sirontak, & Krugman, 2000). In that study, the researchers examined all the child sexual abuse allegations reported by child protective services in Denver, Colorado, for a one-year period (1992). Of the 551 sexual abuse allegation cases, only 43% were substantiated. The rest were deemed to be not sexual abuse, inconclusive, or intentionally false.

Sometimes experts get overly concerned with false allegations as opposed to unfounded allegations. False allegations are a subset of unfounded allegations. It is important to keep in mind that the trier of fact is not merely concerned with false allegations, but whether or not the current allegation is true beyond a reasonable doubt. Therefore, the trier of fact in a child sexual abuse trial should be concerned with the rate of unfounded allegations.

The issue of burden of proof should be kept in mind when considering the research on unfounded child sexual abuse cases. The burden of proof in these cases, which were Child Protective Services cases, was “Reason to Believe.” This burden of proof is less stringent than “Probable Cause for Arrest.” The Reason to Believe burden of proof is substantially less than the burden of proof used in a criminal trial. Thus, it might be reasonable to argue that if the burden of proof used in criminal trials was used by CPS, the rate of unfounded child sexual abuse cases would be much greater than 50%.

When the author was learning criminal profiling, he worked with a profiler trained at the FBI Behavioral Science In­sti­tute at Quantico, Virginia. One of the things that the veteran profiler emphasized was, “When you hear hoof beats, think horses, not zebras.” In North America, there are more horses than zebras. So, if you hear hoof beats and you are in North America, if you must take an educated guess about what type of animal is making the sound, choose horses, not zebras. Odds are, you will be correct more times than not.

The same principle applies to investigations of child sexual abuse. There are more children who were never sexually abused compared to the minority of children who were sexually abused. Regardless of the symptom that is under consideration, given the relatively small percentage of children who are sexually abused, if you had to guess whether a symptom was a sign of sex­ual abuse or not, chose no sexual abuse. Odds are, you will be correct more times than not.

To really appreciate this, consider what we know about the base rate of sexual abuse in childhood. To begin with, we must start with an accurate estimate of the base rate. We know that sexual abuse is often not reported, or if it is, the sexual abuse is reported years after it happens. So, we can’t look at the number of child sexual abuse cases that are reported. This would be an underestimate. The best way to determine the base rate of sexual abuse in childhood is to interview adults and ask them about their childhood experiences.

Dr. Kevin Gorey and Donald Leslie (1997) synthesized the findings of 16 surveys of adults who were questioned about sexual abuse in childhood. Many of these adults never revealed their history of sexual abuse until the researchers interviewed them. Based upon the research results, Gorey and Leslie estimated that by the age of 18, 17% of the females and 8% of the males had been sexually abused.

It should be noted that the base rates of 17% and 8% apply to males and females by the age of 18. But all of the complainant children involved in child sexual abuse trials are younger than age 18. Consequently, these base rates are actually an over-estimate of any child belonging to an age group younger than age 18.

We know from the work of Dr. David Finkelhor the median age of sexual abuse is age nine (Finkelhor, Hotaling, Lewis, & Smith, 1990). In other words, half of the children who are going to be sexually abused are sexually abused prior to age nine and half of the children are abused after age nine. Based upon their research, it appears that by the time a child completes the eighth grade, about 19% of all children will have been sexually abused, 13% of the girls and 6% of the boys.

Let’s take a look at how base rates affect the ability to identify sexually abused children. Imagine a school with 1,000 children in grades K through 8. Applying the base rate of childhood sexual abuse to the 1,000 children in this school, 19 percent, or 190 children, will have been sexually abused. There will be 810 children who never experienced sexual abuse. See the table below, which depicts the manner in which the base rate of childhood sexual abuse was used to determine the number of boys and girls who were sexually abused.

Let’s also say we believe some indicator of sexual abuse, Symptom X, happens 3 times more often among sexually abused children than among children who were not sexually abused. For example, 15% of the sexually abused children will show Symptom X but only 5% of the children who were never sexually abused will show Symptom X. The result is that 29 sexually abused children will show Symptom X, and 41 children who were never sexually abused would show Symptom X.

  • (190 sexually abused children) x 20% (% who show Symptom X) = 29 children
  • (810 children who were never sexually abused) x 5% (% who show Symptom X) = 41 children

As can be seen from the example, there are a total of 70 children who exhibited Symptom X. If jurors used Symptom X to identify children who were sexually abused, they would be correct 41% of the time (29/70 = 41%). Using Symptom X to identify sexually abused children means jurors would be wrong 59% of the time (41/70 = 59%).

Rule 702 requires that an expert help the trier of fact determine a fact in issue or understand the evidence. The foregoing mathematical proof shows that expert testimony about symptoms of child sexual abuse could reduce the accuracy to a rate of 41%. If the jurors wanted to be more accurate, they could ignore the expert testimony and flip a coin. The accuracy rate would be closer to 50–50.

Accuracy is the linchpin in determining admissibility of evidence under Rule 702. In Weatherred v. State (2000), the Texas Court of Criminal appeals reinforced the requirement that expert testimony must help the trier of fact be more accurate: “Under Rule 702, the proponent of scientific evidence must show, by clear and convincing proof, that the evidence he is proffering is sufficiently relevant and reliable to assist the jury in accurately [emphasis added] understanding other evidence or in determining a fact in issue.”

So, let’s return to the original question: Are there reliable indicators of childhood sexual abuse? No, the odds don’t favor such a contention. Odds are that any indicator of childhood sexual abuse will be more common among children who were never sexually abused. In other words, when you hear hoof beats, think horses, not zebras.

Ask any parent and they will tell you what the research says. A child can be going along and doing quite well, and then for no apparent reason the child will enter into a period of acting out. The basis of the acting out doesn’t have to be sexual abuse, or any problem. Change like this is so common that researchers who created norms for measuring childhood development define childhood in the following manner: “childhood is an upward spiral with periods of greater equilibrium alternating with periods of disequilibrium” (Poole and Wolfe, 2009).

In plain English, children can be doing fine and for no rea­son, other than natural development, the child can begin having all kinds of acting-out problems. This has direct implications for using acting-out symptoms to identify child sexual abuse victims: “behaviors that are typical of a large percentage of non-abused children have little value for determining if a child has been sexually abused” (Poole and Wolfe, 2009).

Experts in child sexual abuse trials most often mislead the trier of fact when testifying about three symptoms that they contend indicate sexual abuse: sexualized behavior, enuresis (i.e., wetting), and Posttraumatic Stress Disorder. Here is what the scientific research says about the utility of each of these symptoms as indicators of sexual abuse.

Sexualized Behavior. Sexual play and masturbation are quite common throughout childhood. Contrary to reports that sexual behavior occurs in infants and toddlers, but disappears around the time the child goes to school, the research shows that school-age children continue to be sexual. They just get better at hiding their sexual behavior the older they get. Here is a breakdown of normal sexual behaviors at different ages (Poole and Lamb, 2009):

The table above summarizes some of the research showing sexual behavior is common throughout childhood. This research takes on significance when it is compared to another research finding: Less than 50% of sexually abused children exhibit sexualized behavior, and sexualized behavior can also be created by family problems, physical abuse, life stress, and psychiatric disturbance (Poole & Wolfe, 2009).

Expert testimony about sexual behavior does nothing to assist the trier of fact determine a fact in issue or understand the evidence. As noted in U.S. v. Bahena (2000), courts should “screen out evidence that is unreliable and would have a tendency to confuse or mislead the jury.”

Enuresis. Enuresis is bed wetting and wetting clothing. Enuresis is common among children, with only half of 3½-year-olds staying dry during the day and only half of four-year-olds staying dry at night. Of course, as a child ages, problems with enuresis decrease but do not disappear.

The research says that there is no specific cause for enuresis. The available research shows an astounding diversity of causes of enuresis, including medical issues, car accident, bullying at school, and embarrassment about toileting at school. Given that a minority of children are sexually abused, sexual abuse is rarely the cause of enuresis.

Posttraumatic Stress Disorder. An expert should never be allowed to offer testimony about Posttraumatic Stress Disorder (PTSD) during the guilt-innocence phase of child sexual abuse trials because it addresses the ultimate issue (i.e., the expert must assume that the sexual abuse occurred in order to make a diagnosis of PTSD).

The diagnostic criteria for PTSD can be found in the Diagnostic and Statistical Manual, Fifth Edition (American Psychiatric Association, 2013). Criterion A (1) for the PTSD diagnosis states, “The person has been exposed to actual or threatened death, serious injury, or sexual violence.”

Experts and fact witnesses offering testimony that a child has PTSD allows them to say that the child meets Criterion A—the child was sexually abused. In essence, this expert witness is saying not only did the sexual abuse occur; it also had a profound effect on the child. The expert cannot talk about a child having PTSD without talking about Criterion A (1), which invades the jury’s responsibility for determining if the sexual abuse occurred.

Aside from the obvious legal problem with testimony about PTSD, there are good empirical reasons why this testimony should be excluded because it would mislead the trier of fact. Researchers have found that between 63% and 80% of child sexual abuse victims cannot be diagnosed with PTSD (Finkelhor, 1990). So, if the trier of fact is supposed to use testimony about PTSD to resolve a fact issue (i.e., is the defendant guilty of child sexual abuse), the trier of fact would be misled 63–80% of the time that the expert testifies about PTSD.

Many researchers and clinicians agree that the PTSD diagnosis is:

well suited to traumas such as war shock and rape and prob­ably sexual abuse that occurs under violent circumstances. However, much sexual abuse does not occur under conditions of danger, threat, and violence. Many abusers, misusing their authority or manipulating moral standards, act with the child’s trust. Sometimes the fact of having been abused is recognized only in retrospect as children learn more about appropriate conduct . . . Sexual abuse cannot be subsumed or explained with the framework of PTSD. Moreover, to think of victims of sexual abuse as primarily suffering from PTSD will lead us to miss some of its most serious effects (Finkelhor, 1999; pp. 328–329).

Dr. Stuart Greenberg and Dr. Daniel Shuman (1997) opined that therapists should not testify as experts because of irreconcilable conflicts between the roles of treating mental health professionals and expert witness mental health professionals. These authors identified ten different conflicts between the roles of a therapist and testifying expert.

In order to provide good treatment, the treating professional must believe the client and place the client’s needs as paramount. In a word, the treating professional must be the client’s advocate. To do less would be unethical. On the other hand, a testifying expert owes allegiance to the trier of fact, which excludes the possibility that this professional can be an advocate for the client. For this reason, there is no way to reconcile the roles of treating mental health professional and expert witness mental health professional.

If treating mental health professionals should not serve a role in legal proceedings, does it mean that the complainant in a child sexual abuse case should not receive treatment? No, that is not the upshot of delineating the irreconcilable role conflict. What it means is that the treating professional can provide treatment but should not testify. But this does not fully resolve the conflict in roles. The treating professional should be sensitive to legal status of the client and avoid the kinds of treatment that could contaminate the legal case.

Based upon their extensive work in the area, Dr. Ceci and Dr. Bruck have written:

On the basis of what we now know, it would be imprudent to use fantasy inductions, imagery play, and memory work during therapy sessions conducted before the completion of forensic interviews. These practices can be saved for after the legal resolutions. Prior to it, therapy should be restricted to working on every day coping strategies that cannot be challenged by the defendant’s counsel as creating false memories. This would seem to be a reasonable compromise, one that provides needed mental health support to the child while minimizing potentially suggestive practices (Ceci and Bruck, 1995, p. 289).

The first Child Advocacy Center (CAC) was created in 1985 by Robert “Bud” Cramer, who was the district attorney in Madison County, Alabama. The first CAC was called “The Little House,” and it was located in Huntsville, Alabama.

When District Attorney Cramer became Congressman Cramer, he was instrumental in establishing the National Children’s Alliance, which provided training, support, technical assistance, and leadership to CACs throughout the United States. A visit to the NCA website would tell you that there are more than 700 CACs receiving support from the NCA.

Congressman Bud Cramer has been quoted as saying: “As it was initially conceived, a primary goal of the CAC mode was to increase successful prosecution of child sexual abuse. A secondary goal was to conduct more child-friendly interviews in settings other than intimidating police stations, medical settings, or social services offices.”

Every organization is free to choose its goals, and there is nothing wrong with CACs choosing prosecution as its primary goal. There is, however, something wrong with a CAC center portraying the forensic interviews it conducts as objective, when everything about a CAC has a pro-prosecution bias.

Consider the work of Dr. Gary B. Melton and Dr. Robin J. Kimbrough-Melton (2006), who examined the effect that working in a CAC had upon mental health professionals. These researchers observed that the multidisciplinary structure of CACs, where mental health professionals joined forces with investigative staff (who were focused on prosecution), may have caused the mental health professionals to lose their objectivity and be drawn into the advancement of a specific agenda (i.e., prosecution of sexual offenders, rather than retaining a neutral stance on the question of alleged abuse).

As previously stated, an agency is free to choose its own goals. However, if CACs were to broaden their goals to include objectivity as a goal, the composition of the CAC multidisciplinary team would have to include defense attorneys, investigators for the defense, and forensic mental health experts who had a command of the scientific research regarding child sexual abuse. If this were to happen, there would be an increased opportunity for the results of CAC interviews to be forensic (i.e., objective).

Props such as anatomical dolls, anatomical diagrams, and drawings by children are often used during forensic interviews. Whatever the intended function of these props, all props must meet an uncompromising standard. The props should help children remember and recount what happened without affecting the quality or accuracy of the information produced (Pipe & Salmon, 2009). Unfortunately, no prop has been proven to meet this standard.

All props used during forensic interviews have been determined to cut both ways, which is to say the props can facilitate memory and verbalization in some children and can cause other children to make false statements, even false allegations of touching. Consider some of the findings from the scientific research:

  • Anatomically Detailed Dolls: Dolls have been deemed to be problematic for two reasons. First, they have been considered to be inherently suggestive (Ceci & Bruck, 1993). Second, correct use of the dolls during a forensic interview require the child to have specific cognitive abilities, which they often lack. For example, Dr. Judy DeLoache and Dr. Donald Marzolf (1995) did an experiment in which they asked children to place a sticker on a doll where a sticker had been placed on their bodies. Error rates for different age children were as follows: age four—8% error rate; age three—29% error rate; age two and half—59% error rate. Dr. Karen Theirry and her colleagues (2005) examined 175 child sexual abuse interviews and concluded that information obtained using dolls was inferior to information obtained when the interviewer did not use a doll.
  • Body Diagrams: Anatomical line drawings of adult and child bodies are often used in child sexual abuse interviews. Four studies have examined the use of body diagrams as an aid to the report of touch using analog event paradigm: “All four of the studies show that while body diagrams do help children report touch that occurred, they also lead to reports of touch that did not occur, particularly when used in combination with specific questions about touch”(Pipe & Salmon, 2009).
  • Children’s Drawings: Sometimes children are asked to make drawings during a forensic interview. As found with other nonverbal techniques, the effects of drawing are not uniform. Sometimes drawing can facilitate recall and discussion of past events, and sometimes drawing can result in children presenting false events as if the events were true (Pipe & Salmon, 2009).
     For example, Dr. Deryn Strange and his colleagues (2003) asked children to draw events that did not happen. Dr. Strange found that children who were asked to draw fictitious events were more likely to claim that those events happened as compared to children who did not draw them.
     Dr. Julien Gross (2006) and his colleagues did a similar study. They asked children to draw false events about a visit to a police station. They discovered that drawing false events led to five- to ten-year-olds reporting false information as if it were true.
     Perhaps the most prototypical study was done by Dr. Margie Bruck and her colleagues (2000). She found that drawing was associated both with more accurate information and increased acceptance of false reminders. This study epitomized the research: Drawing leads some children to be more accurate and some children to create false memories, and there is no way to look at the drawings and know which drawings pertain to real or fictitious events.


It is quite common for an expert to meet the qualification prong of Rule 702 (knowledge, skill, experience, training, or education). It is equally common that once an expert meets the qualification prong, little attention is paid to the accuracy prong of Rule 702 (if scientific, technical, or other specialized knowledge will help the trier of fact to understand the evidence or to determine a fact in issue).

As discussed in Rosen v. Ciba-Geigy Corp. (1996), “Under the regime of Daubert a district judge asked to admit scientific evidence must determine whether the evidence is genuinely scientific, as distinct from being unscientific speculation offered by a genuine scientist.” In other words, just because an expert meets the qualification prong, it does not mean the expert automatically meets the accuracy prong of Rule 702.

Put another way, once an expert meets the qualification prong, the expert is not free to testify willy-nilly based upon “knowledge, skill, experience, training, or education.” There is considerable case law that argues against such testimony:

  • “But it is the basis of the witness’ opinion, and not the witness’ qualifications or his bare opinions alone, that can settle an issue as a matter of law; a claim will not stand or fall on the mere ipse dixit of a credentialed witness” (Burrow v. Arce, 1999).
  • “An expert’s simple ipse dixit is insufficient to establish a matter; rather, the expert must explain the basis of his statements to link his conclusions to the facts” (Earle v. Ratliff, 1999).
  • “Nothing in either Daubert or the Federal Rules of Evidence requires a district court to admit opinion evidence that is connected to existing data only by the ipse dixit of the expert” (Flores v. Johnson, 2000).

Many of the problems associated with expert testimony in child sex abuse trials could be avoided if a three-step process is employed. First, the attorney should contact the expert and ask if the expert could testify in a specific case about specific issues. When making this request, the attorney should send the expert all relevant discovery. Second, the expert should conduct a review of the scientific literature, not the clinical literature (more about that in a moment). If the expert finds overlap between the case and the scientific literature, the expert should inform the attorney about the overlap. If the attorney finds the potential testimony helpful, the attorney should arrange for the expert to testify. Third, the expert should testify based upon a hypothetical, keeping in mind that a hypothetical must be based upon facts in evidence. Otherwise the expert will not be helping the trier of fact. When testifying, the expert should merely describe and explain how certain scientific studies overlap with the case, leaving it up to the trier of fact to use the expert’s testimony about the research to understand the evidence or determine a fact in issue.

This approach to expert testimony has the potential to work well if the expert is testifying about research literature, not clinical literature. Research literature meets the criteria established by Rule 702 and Daubert. Clinical literature does not.

Research literature can be identified in the following man­ner. The authors published in a peer-reviewed journal for research. The authors reviewed the literature, used a specific scien­tific methodology, used inferential statistical analysis, and discussed results of inferential statistics.

Clinical literature, such as a therapy workbook, is little more than the ipse dixit of the author, and there is no scientific methodology, measuring subjects, inferential statistical analysis, and discussion of inferential statistical results.

Perhaps the easiest way to differentiate clinical and research literature concerns one of the Daubert criteria: known error rate. In a research study, the authors are required to report an error rate—e.g., 95% confidence interval. In clinical literature, the author assures readers that the offering is accurate based upon “knowledge, skill, experience, training, or education,” which case law in Texas has determined to be an inadequate basis for expert testimony.

Finally, it is worth noting that some books are good compendiums of scientific research that could serve as the basis of scientifically based testimony meeting Rule 702 and Daubert criteria. Examples of books that could readily serve as the basis for expert testimony in child sexual abuse case include Ceci’s book, Jeopardy in the Courtroom, and Kuehnle’s book, The Evaluation of Child Sexual Abuse Allegations.


American Psychiatric Association (2013). Diagnostic and Statistical Manual: Fifth Edition—Text Revision. Washington, DC: American Psychiatric Association.

Bruck, M., Hembrooke, H. & Ceci, S (1997). Children’s reports of pleasant and unpleasant events. In J. D. Read & D. S. Lindsay (Eds.) Recollections of Trauma: Scientific Evidence in Clinical Practice (pp. 199–213). New York: Plenum.

Bruck, M., Melnyk, L., and Ceci, S. J. (2000). Draw it again Sam: The effects of drawing on children’s suggestibility and source monitoring ability. Journal of Experimental Child Psychology, 77, 169–196.

Burrow v. Arce, 997 S.W. 2d 229, 235 (Tex. 1999).

Ceci, S. J., & Bruck, M. (1993). Suggestibility of the child witness: A historical review and synthesis. Psychological Bulletin, 113, 403–439.

Ceci, S. J. & Bruck, M. (1995). Jeopardy in the Courtroom: A Scientific Analysis of Children’s Testimony. Washington, DC: American Psychological Press.

Ceci, S. J., Loftis, E., Leichtman, M. D., and Bruck, M. (1994). The possible role of source misattributions in the creation of false beliefs among preschoolers, International Journal of Clinical and Experimental Hypnosis, 42, 304–320.

Coble v. State, 330 S.W.3d 253 (Tex. Crim. App. 2010).

Daubert v. Merrell Dow Pharmaceuticals, 509 U.S. 579 (1993).

DeLoache, J. S., &. Marzolf, D. (1995). The use of dolls to interview young children: Issues of symbolic representation. Journal of Child Psychology, 60, 155–173.

Earle v. Ratliff, 998 S.W. 2d 882, 890 (Tex. 1999).

Flores v. Johnson, 210 F. 3d 456, 465–466 (5th Cir. 2000).

Kuehnle, K. & Connell, M. (2009). The Evaluation of Child Sexual Abuse Allegations: A Comprehensive Guide to Assessment and Testimony. New Jersey: John Wiley & Sons, Inc.

Finkelhor, D. (1990). Early and long-term effects of child sexual abuse: An update. Professional Psychology: Research and Practice, 21, 325–330.

Finkelhor, D., Hotaling, G., Lewis, I. A., & Smith, C. (1990) Sexual abuse in a national survey of adult men and women: Prevalence, characteristics, and risk factors. Child Abuse and Neglect, 14, 19–28.

Green, A. H. (1993). Child sexual abuse: Immediate and Long-Term Effects and Intervention, Journal of the American Academy of Child Adolescent Psychiatry, 32, 890–902.

Gorey, K. M. & Leslie, D. R. (1997). The prevalence of child sexual abuse: Integrated review and adjustment for potential response and measurement biases. Child abuse and Neglect, 21, 391–398.

Jones, D. & McGraw, J. M. (1987). Reliable and fictitious accounts of sexual abuse in children, Journal of Interpersonal Violence, 2, 27–45.

Greenberg, S. A. & Shuman, D. W. (1997). Irreconcilable conflict between therapeutic and forensic roles. Professional Psychology: Research and Practice, 28, 50–57.

Gross, J., Hayne, H., and Poole, A. (2006). The use of drawings in interviews with children: A potential pitfall. In J. R. Marrow (ed.). Focus on Child Psychology Research. NY: Nova.

Lamb, M. E. (1994). The investigation of child sexual abuse: An interdisciplinary consensus statement. Expert Evidence, 2, 151–163.

Oates, R. K., Jones, D. P. H., Denson, D., Sirotnak, A., Gary, N., Krugman, R. D. 2000 Erroneous Concerns about Child Sexual Abuse. Child Abuse and Neglect, 24, 149–157.

Melton, G. B. & Kimbrough-Melton, R. J. (2006). Integrating assessment, treatment, and justice: Pipe dream or possibility. In S. N. Sparta & G. P. Koocher (eds.). Forensic Mental Health Assessment of Children and Adolescents. NY: Oxford University Press.

Narby, D. J., Cutler, B. L. & Moran, G. (1993). A meta-analysis of the association between authoritarianism and juror’ perceptions of defendant culpability. Journal of Applied Psychology, 78, 34–42.

Poole, D. & Wolfe, M. (2009). Normative Sexual and Nonsexual Behaviors that May be Confused with Symptoms of Sexual Abuse. In Kathryn Kuehnle and Mary Connell (eds.). The Evaluation of Child Sexual Abuse Allegations: A Comprehensive Guide to Assessment and Testimony. New Jersey: John Wiley & Sons, Inc.

Pipe, M., & Salmon, K. (2009). Dolls, drawings, body diagrams, and other props. In Kathryn Kuehnle and Mary Connell (eds.). The Evaluation of Child Sexual Abuse Allegations: A Comprehensive Guide to Assessment and Testimony. New Jersey: John Wiley & Sons, Inc.

Realmuto, G., Jensen, J., & Wescoe, S. (1990). Specificity and sensitivity of sexually anatomically correct dolls in substantiating abuse: A pilot study. Journal of the American Academy of Child and Adolescent Psychiatry, 29, 743–746.

Rosen v. Ciba-Geigy Corp., 78 F. 3d 316 (7th Cir. 1996).

Salmon, K. & Pipe, M. (2000). Recalling an event one year later: The impact of props, drawing, and prior interviewer. Applied Cognitive Psychology, 14, 184–220.

Strange, D., Garry, M. & Southerland, R. (2003). Drawing out children’s false memories. Applied Cognitive Psychology, 17, 607–619.

Theirry, K. L., Lamb, M. E., Orbach, Y. & Pipe, M. E. (2005). Developmental differences in the function and use of anatomical dolls during interviews with alleged sexual abuse victims. Journal of Consulting and Clinical Psychology, 73, 1125–1134.

U.S. v. Bahena, 223 F.3d 797, 809 (8th Cir. 2000).

U.S. v. Frazier, 387 F.3d 1244, 1261 (11th Cir. 2004).

Weatherred v. State,15 S.W. 3d 540, 542 (Tex. Crim. App. 2000).

Dr. Matthew L. Ferrara
Dr. Matthew L. Ferrara
Matthew L. Ferrara, PhD, is a licensed psychologist and Licensed Sex Offender Treatment Provider in private practice in Austin, Texas. He is responsible for creating all the major sex offender treatment programs operated by the State of Texas, including the Sexually Violent Predator Program. Along with Professor Mark Stafford, Texas State University, he conducted research that led to the current risk assessment deregistration criteria and methodology. He has trained all the deregistration evaluation specialists in Texas. For more information about deregistration visit Dr. Ferrara’s website: www.deregistertexas.com

Matthew L. Ferrara, PhD, is a licensed psychologist and Licensed Sex Offender Treatment Provider in private practice in Austin, Texas. He is responsible for creating all the major sex offender treatment programs operated by the State of Texas, including the Sexually Violent Predator Program. Along with Professor Mark Stafford, Texas State University, he conducted research that led to the current risk assessment deregistration criteria and methodology. He has trained all the deregistration evaluation specialists in Texas. For more information about deregistration visit Dr. Ferrara’s website: www.deregistertexas.com

Previous Story

Nondisclosure Update

Next Story

Richard Racehorse Haynes

Latest from Features