In the late 19th century, Sigmund Freud found that many of the adult patients he treated reported childhood sexual abuse, which Freud found to be causally related to these patients’ symptoms of psychological distress. Overwhelming contention from both scientific and political leaders led Freud to eventually withdraw his findings in this area, although he remained equivocal in later writings. The latter half of the 20th century saw a return to a focus on childhood sex abuse as the etiology of later distress and dysfunction. Fueled by the proliferation of scientific evidence, child advocates refused to back down when confronted by skeptics, and staunchly maintained the attitude that “children don’t lie; if it hadn’t happened, they couldn’t report it.” An attempt to make up for decades of ignorance and rejection of children’s stories of abuse led to an unfettered and unexamined acceptance of every child’s story of abuse as true. By the close of the 1980s, the negative repercussions of this attitude were obvious. Now infamous cases such as California v. Buckey (the McMartin Preschool case, 1990) and New Jersey v. Michaels (1993) highlight the trauma and tragedy that can and do result from children’s stories being uncritically accepted as valid. The State of Montana saw its own version with Montana v Harts (1993).
For a relatively brief period of time, the argument seemed to revolve around whether children would deliberately lie about a personal experience of sexual abuse. An abundance of research has now disassembled the question when it is stated this way. It has been learned and amply documented that the issue of TRUTH when applied to children’s statements is multidimensional. The focus on how children’s statements might differ from adults’ statements has compelled a scientific return to an understanding of child development in moral, cognitive, emotional and social spheres. Many volumes have recently appeared on the suggestibility of children, the creation of false or distorted memories, motivation, and other aspects of truth-telling, all of which attempt to explain why some children’s reports of sexual abuse are not true, even though the child may appear to be sincere. (See Ceci & Bruck, 1995; Doris, 1991; Goodman & Bottoms, 1993; Perry & Wrightsman, 1991).
Statement of the Problem
While some scientists have focused on the main question of how to figure out if a child’s story is true, others have focused on the formulation of investigative techniques that are less likely to facilitate the false authentication of invalid reports, and on the internal dynamics involved in a child’s production of an invalid report. Still others have focused on the social phenomenon characterized by the public’s response to the tremendous rise in the incidence of reported child sex abuse. Due at least in part to the vulnerability of children and the sense of responsibility associated with the need to protect them from harm, it may be the case that people involved in that process overreact to allegations of sexual abuse. This overreaction has been termed “moral panic” by some writers (Edwards & Lohman, 1994). The moral aspect of this phenomenon derives from the sexual component, which is highly charged in our society, as well as from the already charged situation which exists when the protection of children is at issue. It could be reasonable argued that protection of children from harm, particularly sexual abuse, is the moral imperative our society has adopted as the most important. In the arena of child sexual abuse allegations there is, therefore, a particular opportunity for the desire to exhibit “good intentions” to overshadow the need for objective, consistent, and ongoing dynamic tension which exists societally with regards to the question of whether child sexual abuse occurs and, if so, how prevalent it is. This debate also goes on in the realms of law, psychology, and throughout the child protective services network. To the extent that moral panic creates a generic prejudice against those merely accused of child sexual abuse (Vidmar, 1997), it must be accounted for in any competent investigation of allegations. When practitioners are called upon to provide investigative and evaluative procedures in child sexual abuse cases, they are often at a loss as to how to proceed. They may be placed in a double bind, wanting to help, but not knowing how to proceed in a way that is defensible in court. Practitioners may be concerned that treating a false allegation as true can be as traumatizing to non-abused children (who may become convinced through suggestive interviewing that they have been abused), as is treating a true allegation as false (and accusing truly abused children of lying). The literature has unfortunately focused on the disagreements among experts over the various techniques available, rather than on any core similarities.
As experts independently developed then published formats, guidelines or techniques for conducting the best possible investigation into allegations of child sex abuse, scientific debate over the pros and cons of each major offering proliferated, until it began to seem to the novitiate that there was no agreement at all. Several experts have offered what they considered to be scientifically acceptable procedures for determining the validity of child sex abuse allegations. Paramount are a) Gardner’s (1992) True and False Accusations of Child Sex Abuse and (1995) Protocols for the Sex-Abuse Evaluation; b) Greenberg’s (1990) Conducting Unbiased Sexual Abuse Evaluations; c) Boat and Everson’s (1986) Structured Interview for Using Anatomical Dolls; d) Hindman’s (1987) Sixteen Steps Toward Legally Sound Sexual Abuse Investigations; and e) Raskin and Esplin’s (1991) Statement Validity Analysis. Michael Lamb’s (1994) “Consensus Statement” is a welcome glimmer of hope on the horizon, but is based on the agreement of 20 individuals in attendance at a conference in Switzerland and does not account for the published opinions of absent experts. This article is offered as a caveat against using inappropriate techniques when professionally interacting with a child who is alleging sex abuse, and as a guide to conducting a valid examination of such a child that will be both admissible and defensible in a court of law.
Major Theoretical Approaches
According to Raskin and Esplin (1991), “Statement Validity Analysis” (SVA) is a set of interview techniques and analytical procedures for obtaining and evaluating statements. These procedures compel the evaluator to explore and consider all of the available information and many possible explanations prior to, during, and after the interview. SVA essentially incorporates three procedures: the first is the obtaining of a free narrative by the child who alleges sexual abuse, without using dolls or other props as communication aids. This interview is not therapy and should not be performed by the child’s therapist due to dual role conflicts (Committee on Ethical Guidelines for Forensic Psychologists, 1991; American Academy of Child and Adolescent Psychiatry, 1988). Guidelines are offered as to the formation of alternative hypotheses, when to use cue questions, direct questions and probe questions. The second procedure is the application of Criteria Based Content Analysis (CBCA, Raskin & Esplin, 1991) to the narrative provided by the child and recorded verbatim. CBCA analyzes the narrative statement for general characteristics, specific contents, and motivation-related contents. The third procedure is the application of the Validity Checklist (Raskin & Esplin, 1991) to the entire body of data accumulated through both legal and psychological means relevant to the case. The Validity Checklist consists of four categories of information to be analyzed: a) psychological characteristics of the child; b) interview characteristics of the child and the examiner, c) motivational factors relevant to the child and others involved in the allegations and; d) investigative questions regarding the consistency and realism of the entire body of data. These procedures, taken together, discourage premature conclusions by forcing a systematic consideration of all necessary and available information.
Boat and Everson (1986) have developed a comprehensive set of guidelines on interviewing children who allege sex abuse, using Anatomically Detailed (AD) dolls. This involves a structured interview that begins by assessing cognitive competencies, then uses AD dolls to help children with immature verbal ability communicate what may have happened to them. The American Psychological Association’s Council of Representatives (1994) has recently published a position paper on the use of AD dolls, in which the use of AD dolls is endorsed as a communication and memory aid for children undergoing a sex abuse investigation, but not as a definitive diagnostic test that can say with certainty whether a child has been sexually abused.
Hindman (1987) has published Step By Step: Sixteen Steps Toward Legally Sound Sexual Abuse Investigations. The book provides guidelines for interviewing children who allege child sexual abuse. She recommends use of another of her books, A Very Touching Book (1985), which describes the concepts of “good touch, bad touch and secret touch” to a child who is then asked to relate whether he or she has experienced any “touching trouble” and if so to describe it. Although Hindman states that AD dolls may be used to augment the interview, she offers no specific guidelines as to use of the dolls, and the general guidelines offered differ from the guidelines of Boat and Everson.
Gardner’s (1992, 1995) investigative method is interview-based. He emphasizes the importance of evaluating not only the alleged victim, but also the alleged perpetrator, and the accuser, and reserves the right to bring all three together in the same room for conjoint interviewing. In his 1992 book, he offered 30 “differentiating criteria” for use in assessing the likelihood of sexual abuse. In his 1995 book, he offers an additional 21 criteria derived from direct inquiry and 11 criteria derived from projective testing. He also lists several criteria to be considered when evaluating the alleged victim’s parents, the accused male, the accused female, and the accuser. Gardner explains that there is no scoring scheme or cut-off point to indicate when sexual abuse has occurred, but that his criteria form a continuum of likelihood that a child has or has not been sexually abused: the greater the number of indicators met, the greater the likelihood that a child has been abused.
Greenberg (1990) offers suggestions for conducting unbiased investigations of alleged child sexual abuse victims. He focuses on the interview format, but allows certain toys to be used as stimuli for verbalizations and behavior. He speaks to the role and context of the forensic investigator, the importance of evaluating children’s competencies before judging the content of their speech, the use of counterbalanced questioning about the alleged victimization and the alleged perpetrator, questioning about sequelae of the alleged abuse, and achieving closure at the end of the interview.
Wells and Loftus (1991) state that “Statement Validity Analysis” has inadequate empirical support and may lack ability to partition individual and age-related differences in linguistic abilities from validity-related differences. Gardner (1992) opposes the use of AD dolls, calling them a source of “psychological grief” to children. Others, including Underwager and Wakefield (1990) posit that AD dolls are sexually suggestive. Contrary to the other methods, Gardner’s method emphasizes interviewing the alleged perpetrator whenever possible, and often jointly with the alleged victim, before rendering an opinion as to whether sexual abuse of a child has occurred.
Bruck and Ceci (1993/1995) emphasize the myriad ways in which children’s competence and accuracy of information can be sabotaged by incompetent investigative procedures and interview techniques, while Lamb, Sternberg & Esplin (1995) offer ways in which the conscientious investigator can enhance the quality, accuracy, and amount of information obtained from children who allege sex abuse. In addition, Lenore Terr (1994) emphasizes the accuracy of children’s memory, while Elizabeth Loftus (1994) eloquently argues the fallibility of children’s memory. Memory itself cannot be judged to be accurate or inaccurate unless the investigator’s interviewing style and techniques are sound.
Case Synopsis: Montana v. Harts
In Montana v. Harts (State of Montana v. Harts, 1993), the State’s child protective services workers had performed the first several evaluations of two children, ages 3 and 5, who had alleged sexual abuse by their great-grandparents, ages 78 and 81, who had no documented history of previous criminal behavior or of sexually inappropriate behavior. Child protective services workers and police rewarded the children with praise when they provided affirmative responses to their questions. When one child reported something that the other could not at first remember, pressure applied until the child could remember it. No effort was made to verify the physical possibility or impossibility of the allegations. The children were sent to a therapist who insisted that the children elaborate on this abuse by asking them to “draw a picture of your rectum” and “draw a picture of how you feel about Pa’s genitals.” These drawings were submitted to the county attorney over the next two years as evidence of abuse. Other grossly inappropriate “therapeutic techniques” were also used to extract confirmation from the children that bizarre and violent sexual abuse had been perpetrated against them. The 5-year-old boy was put into treatment sessions with a 9-year-old boy who was a confirmed sexual abuse victim. The therapist typically saw the children in her home for up to 6 hours at a time. The therapist forbade anyone to talk to the children, including their grandparents, unless they promised not to express any doubt as to the children’s allegations, or unless the therapist was present. When the children tried to say that their reports were “just dreams” or had never been true, these statements were discounted. The therapist asked the children to draw something and when they did not, she produced the drawing, labeled it as the child’s, and sent it to the county attorney. The therapist collected crime victim’s compensation funds for the children’s treatment, long before an unbiased investigation was performed, reinforcing the necessity from the therapist’s perspective that the crime be confirmed even if it had not occurred. Through disorderly and biased procedures, these children were induced, albeit unintentionally, to report ever more heinous acts of sexual abuse against them.
Deposition testimony indicated that upon re-evaluation the children’s statements did not meet credibility criteria when the procedure, “Statement Validity Analysis” (Raskin & Esplin, 1991) was applied. Re-evaluation also indicated that initial evaluation procedures had been faulty. The county attorney filed a brief to quash this challenging testimony at trial, maintaining that SVA and the other procedures described above were inadmissible as expert testimony due to major disagreements and lack of consensus among experts in the field.
A review and analysis of the literature identified eight core similarities among the major approaches to such evaluations. Expert testimony was offered that these eight similarities, used as the foundation for the investigation of the sex abuse allegations, do meet evidence admissibility requirements. The judge allowed the challenged testimony, ruling that investigative procedures utilizing these eight core similarities were scientifically acceptable and admissible as evidence. Prior to trial a Statement Validity Analysis was performed enabling further expert testimony to the effect that the children probably had not been abused, but had been led to believe that they were, based on suggestive, coercive, and biased investigative and therapeutic conduct. The Judge ultimately ruled that the alleged sexual offenders were “not guilty.”
Areas of Consensus
Analysis of these major investigative formats indicates that, in spite of the many criticisms of each method of investigation, there is substantial overlap among them. Eight similarities are either explicitly or implicitly a part of the major published approaches to the clinical investigation of alleged child sexual abuse. Several months after testimony was given in the Hart case detailing these eight similarities, Michael Lamb (1994) published an interdisciplinary consensus statement on the investigation of child sex abuse allegations that reviews areas of consensus among 20 experts from Europe, North America and the Middle East. Conference attendees were able to agree on seven areas of consensus. Table 1 is a compilation of items on both the “Hart” list and the “Lamb” list.
(Insert Table 1 here)
Areas of Agreement Among Experts in Child Sex Abuse Investigations
1) The investigator must carefully examine his or her own emotions and possible biases regarding child sex abuse before undertaking to interview children who allege sex abuse, so as not to unwittingly project those biases into the assessment of the child’s allegations.
2) A well-trained and experienced forensic interviewer and not the child’s therapist should conduct the investigation. This will yield more informative and accurate accounts by children.
3) The greatest accuracy is obtained by eliciting a free narrative from the child in response to open-ended questions. Inaccuracies increase with the level of suggestiveness and coerciveness of the interview techniques used.
4) A structured interview technique with which the interviewer is familiar should be used, not a freeform or unstructured interview.
5) The interview and child’s behavioral responses should be recorded, preferably by videotape, but at least by audiotape with detailed notes. Note-taking alone should be reserved for cases involving special circumstances.
6) Preschool age children are prone to suggestiveness and fantasy/reality confusion, and thus require special skills by the interviewer, including a concession that such confusion is possible, and knowledge of developmental differences.
7) Some measure of a child’s ability to distinguish between truth and falsehood must be taken and all such measures are not equal. For example, asking a child to “tell the truth” about the color of your sweater and the child says ‘red’ and the sweater is red, does not mean that the child can reliably distinguish between truth and falsehood in all applications.
8) Some children’s statements will be false and must be distinguished from true statements by the application of a structured technique, and not by “gut feeling” or “hunches” unsupported by articulable criteria.
9) Behaviors that mimic sexual activity are seen more frequently in abused children than in non-abused children.
10) Children’s accounts that contain peripheral details and are logically embedded in a rich context are more likely to represent a true occurrence.
11) Tools and props such as anatomically detailed dolls, puppets, or human figure drawings may be useful when interviewing children under age 5 or older children who are uncommunicative.
12) Although medical exams commonly do not show evidence of sexual abuse, they should be performed and documented in every case as soon as possible after the allegation, by a highly trained specialist using multiple techniques and sophisticated equipment.
Training Public Officials
Doris et al. (1995) recently emphasized the importance of formal training for child protective services workers. Public officials are not to blame for their relative ignorance in the area of child sex abuse investigations. Several have pointed out that psychologists first hammered them for a decade to take child sex abuse allegations seriously, and that they are now criticized for their well-intentioned (albeit misguided) efforts to advocate for children who cannot advocate for themselves. The criticism, however, is not against advocacy for children, but against poor quality of training and of investigative procedures. Poorly trained workers and investigators commit errors based on lack of knowledge of the fundamentals delineated here, giving rise to legal and personal fiascos such as the Hart case discussed briefly herein. Whichever scientific method of investigation of child sex abuse allegations is chosen as the method to be taught to child protective services workers, or others, the fundamental procedures delineated herein should form the minimal foundation for a clinical investigation.