The Empirical Basis of the Forensic Evaluation Protocol

Faller (1996) divided the primary controversies in child sexual abuse interviews into four categories: "(a) the ability of the interviewer to conduct a competent interview, (b) the competence of the child to describe actual events, (c) interview structure and process and (d) decision making about the likelihood of sexual abuse" (p. 86). Each of these controversies were considered and incorporated into the forensic evaluation model.

The Ability of the Interviewer to Conduct a Competent Interview

Some detractors suggest that interviewers foster false allegations and use leading and suggestive practices (e.g., Ceci, Huffman, Smith, & Loftus, 1994). Others claim child interviewers are poorly trained and educated and have questionable motives for working in the field (e.g., Gardner, 1991). Although the actual motives of individuals are virtually impossible to validate scientifically, Gardner (1992) provided a criticism suggesting that interviewers may be abuse survivors working out their own issues.

A paucity of scientific data exists related to interviewers' previous abuse history. In a survey of 656 social workers, pediatricians, psychiatrists and psychologists, 17% of the respondents reported a history of sexual abuse (Jackson & Nuttall, 1993). Kelly (1990) surveyed 228 police officers, child protective workers and nurses who attended an educational program on child abuse, and found a 13% incidence of child sexual abuse history. In another study, Howe, Herzberger and Tennen (1988) surveyed mental health service providers at the masters and doctoral levels. Eight percent of their 101 respondents reported childhood sexual abuse.

Taken together, the results of the three survey studies suggest that sexual abuse survivors are not over represented in the identified professions. In fact, the percentages are generally below the statistical estimates of sexual abuse prevalence in the general population. Finkelhor (1994) has reported as a result of summarizing 19 prevalence studies that a solid, albeit possibly conservative, estimate based upon the available research of prevalence among sexual abuse histories is 20% for U.S. females and 5-10% for U.S. males. Comparisons of abuse prevalence in the survey research with Finkelhor's estimates suggest that sexual abuse survivors may actually avoid working in the child protection field.

All three studies (Howe et al., 1988; Jackson & Nuttall, 1993; Kelly, 1990) looked at clinicians' responses using an analog methodology. They used vignettes and asked subjects to make judgments about them. Jackson and Nuttall found that clinicians who reported childhood abuse history were more likely to believe abuse allegations than were colleagues who did not have an abuse history. Kelly found that subjects who reported childhood sexual abuse attributed more responsibility to the offender than did the non-victims. Howe et al. found that those reporting childhood abuse histories viewed the physically abusive parental acts depicted in the vignettes as more severe and more likely to be harmful than those with no reported history of abuse. None of these outcomes suggest that professionals are working out their own abuse issues, but they do suggest that abuse survivors may have increased empathy for the child and a firmer belief that abusers are in the wrong, orientations that are not inappropriate for those working in the field of child maltreatment.

The criticism that evaluators use leading and suggestive practices was directly addressed in the creation of the forensic evaluation protocol. All practices used within the protocol are research-based, and designed to obtain accurate information that can either support or not support allegations of sexual abuse. The tools and techniques are specifically designed to be non-leading and to be used in a developmentally appropriate fashion. A forensic evaluation is considered successful when it yields sufficient quality and quantity of information to be able to help validate or invalidate suspicions of abuse. According to Reed (1996), "[t]he primary purpose of investigations of suspected child maltreatment should be to arrive at valid conclusions about the 'truth' of the matter" (p. 104). This is an important orientation stressed to those evaluators trained on the use of the model. This orientation encourages critical evaluation of children's statements and counters the claims by some critics that abuse investigators are "looking for abuse under every rock."

The Competence of the Child to Describe Actual Events

Reluctance

Reluctance is commonplace and difficult to overcome in suspected child sexual abuse cases. In a laboratory study (Saywitz, Goodman, Nicholas, & Moan, 1991) children exhibited reluctance to acknowledge even socially sanctioned genital touching by a doctor. Although children's reluctance and embarrassment in discussing sexual material pose a challenge to interviewers, these factors also create a deterrent against false allegations. Lyon (1995) points out that those pressures that discourage true abuse reports, also operate to discourage false reports. In fact, research has shown children are far more likely to deny or fail to report abuse that has occurred (Lawson & Chaffin, 1992, Faller, 1988).

As discussed above, the NCAC forensic evaluation protocol was designed to be conducted over time to give children the needed safety and non-pressured pace, in accordance with data showing that some children tend to disclose over time (Sorenson & Snow, 1991). The model was specifically designed to help reluctant children overcome their reluctance, fear, embarrassment, and avoidance coping. Thus, it begins with sessions of rapport building, developmental evaluation and social and behavioral assessment before directly addressing the more threatening topics pertaining to abuse.

Memory issues

The complex issues of memory acquisition, storage and retrieval have been widely studied in the laboratory. One salient finding in the research literature is that preschoolers need different cues for retrieval than do school age children (Fivush, 1993). Preschool children do not do well on free-recall tasks, and require specific external cues to direct their attention to specific interview topics. The NCAC protocol employs abuse specific, non-leading questioning procedures to inquire about such topics as care routines, substance abuse, domestic violence and physical and sexual abuse.

The focused questioning techniques are based upon the work of Walker (1994) and others, to be developmentally appropriate and non-suggestive. Abuse is approached obliquely, without direct leading. While detractors would describe almost any focused question as leading, for the purposes of practice and research using the Forensic evaluation protocol, a leading question means a question in which the answer is suggested, for example, "He touched your privates, didn't he?" or "Didn't he make you touch his pee-pee?"

School aged children can respond to more open-ended techniques for memory retrieval, therefore, the Cognitive Interview (Saywitz, Geiselman and Bornstein, 1992), and Narrative Elaboration (Saywitz, Snyder, & Lamphear, 1996) procedures are employed in the model when questioning older children given there has been an acknowledgment by the child of an abuse incident. In laboratory studies, these techniques increase accuracy and quantity of detail by providing memory retrieval cues without being leading or suggestive.

Suggestibility

Suggestibility has been defined as "the degree to which one's 'memory' and/or 'recounting' of an event is influenced by suggested information or misinformation" (Reed, 1996, p. 106). Suggestibility concerns take three forms: (1) coaching or shaping by caregiver to make false allegations, (2) coaching or shaping to cause the child to keep the secret or interpret abuse as something innocuous and (3) inadvertent interviewer influence. Lyon (1997) described a recent series of laboratory research as the "new wave" of suggestibility research (e.g., Leichtman & Ceci, 1995, Ceci & Bruck, 1993). This research applies in different ways to the three forms of suggestibility.

Coaching to report false abuse

Ceci stated during a television broadcast of "Nightline" that his research was designed to be used in understanding false allegations coached by adults, which he believes are a small percentage of cases (Hayes & Hill, 1996). In his studies, children are exposed to misleading, suggestive interviewing practices, and sometimes (although not the majority of the time) preschoolers say they have seen or experienced something they have not.

In the "Sam Stone" study (Leichtman and Ceci, 1995), a stereotype induction is used, and then children are purposely misled with embedded suggestions. The researchers' hypotheses that some children can potentially be coerced or manipulated into false allegations, and that preschoolers are more suggestible than older children are supported by their findings. An extremely high degree of purposeful misleading over long periods of time occurs in these studies. Ceci commented that they had to work very hard at leading children to get their reported effects (Hayes & Hill, 1996).

This type of paradigm may be ecologically valid when one looks at the dynamics of the small percentage in which caregivers coach false allegations (prevalence studies generally estimate these comprise less than 8% of reported cases). These caregivers have time and access to offer repeated suggestions, and they have the power over the child to coerce false statements. Individuals trained in the use of the forensic evaluation protocol are taught to be alert for symptoms of coaching and shaping to report false abuse, particularly in young children, and to specifically assess for these potential factors during an evaluation.

Coaching to avoid reporting true abuse

Coaching and shaping are more frequently used to keep children from reporting true abuse or to cause recantation. Abuse perpetrators have motivation to purposely shape a child's perceptions and statements by repeatedly exposing a child to suggestions that abuse did not occur. They also have access, time and a relationship with which to impose coercive influence on the child to remain silent or to retract true statements of abuse. These pressures on the child may be exerted over months or sometimes years. In these cases, the Ceci paradigm of coercive misleading over long periods of time is also applicable. Children can be misled by caregivers to keep silent or to make up elaborate explanations to keep their caregivers out of trouble. Even with an extended evaluation model, overcoming this challenge is difficult.

Interviewer influence

The Ceci research is less ecologically valid when applied to an interview setting in which the interviewer has very little time, access and relationship with the child. In fact, the studies were designed not for use in attacks on legitimate interviewers, but to learn about cases in which false reports are coached by caregivers. According to Ceci, "the majority of interviews done with kids by front line workers are well done." (Hayes & Hill, 1996). Ethical standards would clearly prohibit legitimate interviewers from purposely and overtly misleading children. Professional interviewing training invariably focuses on techniques to avoid coercion and suggestibility effects. In the NCAC forensic evaluation protocol, the risk of suggestibility is minimized by limiting direct non-leading questioning of preschoolers to two or three sessions during the forensic evaluation protocol. Interviewers in this study are trained to be alert for signs that a child may be trying to please, and to be proactive in avoidance of coercion.

The suggestibility studies do illustrate the obvious need to avoid bribes, coercion and stereotype induction with children. A balanced, reasonable response is needed. Fortunately, in actual practice, available interviewing techniques range from open-ended to focused and non-leading, and can be tailored to the developmental level and situational factors of the child. Strong support exists in the literature that children can give accurate accounts when not asked leading questions. There is little support for the contention that an interviewer could inadvertently alter a child's memory or ability to recount events within two to three interviews using non-leading techniques accurately.

Ceci and his colleagues (Toglia, Ross, Ceci and Hembrook, 1992) showed that the suggestibility effect was drastically reduced when an interviewer was perceived by the child as less knowledgeable about facts. Error rates decreased from 33% to 18% when the interviewer presented him/herself as less knowledgeable. Evaluators using the forensic evaluation model are trained to clearly impart to the child that they do not know what happened. To reinforce this further, evaluators are taught to establish early in the evaluation and reiterate as necessary the following set of rules designed to empower the child (Saywitz, Geiselman and Bornstein, 1992):

1. If you don't know the answer, don't guess, just say you don't know.
2. If you don't want to answer, it's ok to say so.
3. If you don't understand something, let me know, and I will say it a different way.
4. If a question is asked more than once, you don't have to change your answer, just tell me what you remember, the best you can.

Unusual or improbable information in the disclosure

Some children's statements contain unusual or improbable information that defies a common sense of reality. Everson (1997) proposed explanations for these elements in children's abuse accounts including: deliberate attempts by the perpetrator to confuse the child, threat incorporation, traumagenic memory distortion, the child's coping mechanisms, developmental limitations, interviewer errors, leading techniques, errors due to misused media and deceptive processes on the part of the child.

Dahlenberg (1996) studied cases in which fantastic elements were present. She notes: "fantastic elements occurred most frequently in the accounts of children known to have been abused, and indeed were most common among children known to have suffered severe abuse. These findings directly counter the hypothesis that fantastic elements in children's accounts of abuse give reason to discredit the entire account" (page 8). Dahlenberg suggests the bias that can be produced by hearing fantastic elements should be countered by investigating the source and meaning of the elements.

The Forensic Evaluation Critical Analysis Guide, (discussed in detail in chapter 8) which was designed to assist the decision making process during a forensic evaluation, encourages the user to explore on several levels for the meaning of unusual or improbable elements in children's statements. Embedded in the guide is a systematic process of evaluating the disclosure in terms of several categories described by Everson, including developmental and emotional factors, motivational factors and alternative explanations. Everson suggests that interviewers may overly dismiss disclosures due to these unusual elements and/or are too defensive when encountering the elements during interviewing or court. An evaluator need not be as defensive in the presence of such elements if he or she conducts a careful analysis of them in the context of other aspects of the disclosure and of the child's functioning. This multi-level analysis should lead to increased accuracy in the decision whether or not to dismiss the disclosure due to the unusual elements, and increase precision in communicating with other professionals and juries about the child's statements.

Interview Structure and Practice

How many interviews?

There are negatives associated with both too many and too few interviews. If decisions are made with too few interviews, the system may miss identifying abuse and place or leave a child at continued risk. If too many interviews are conducted, the associated risk is the potential of influencing the child. Excessive numbers of interviews can also be costly, reducing the number of children who can reap the benefits of an extended evaluation.

The forensic evaluation protocol has been refined based on practice and research. It was begun as a 12-session model and reduced to eight based on results of a year of practice. Incorporating findings of a two-year multi-site research project on the model, it was reduced to a 6-session model (1 session with the non-offending caregiver and 5 with the child).

How to Use Media

A variety of tools are used during forensic evaluation. Anatomical dolls, standardized and free hand anatomical drawings, free style drawing, regular dolls and puppets as well as other tools are incorporated into practice to help children in communication. Information obtained using tools and props is not interpreted by the evaluator, but is simply used to facilitate the child's communication of his or her experience. The props are simply used as a means to an end--helping the child to communicate clearly. APSAC guidelines for use of anatomically detailed dolls (APSAC, 1995) are followed.

Decision-making About the Likelihood of Sexual Abuse

Much has been written about how to assess interview and other data (Faller, 1988; Pence and Wilson, 1994; Raskin and Yuille, 1989; Yuille, 1988). Some factors consistently examined are interview findings, external factors such as medical evidence, and linguistics of the child's account. The final step of the forensic evaluation protocol is to examine the credibility of the results using the Forensic Evaluation Critical Analysis Guide. The guide is not an empirically normed scale. It is a desk guide, designed to help the evaluator in analyzing the results of a forensic evaluation. It is intended to be used as a tool during decision making regarding the disclosure. The elements in the guide are drawn in part from the literature on statement validity analysis. The elements are provided as a framework for analysis of the evaluation outcome, and they fall into eight categories:

1. Confirming qualities of statements.
2. Specific details obtained.
3. Developmental factors.
4. Emotional content.
5. Behavioral checklist results (Friedrich, 1990, Achenbach, 1988, and Briere, 1996).
6. Corroborative information/confirmatory factors.
7. Motivational factors.
8. Alternative explanations.

The presence or absence of any one or more of the factors does not dictate the results. The primary usefulness of a guide such as this is the discipline for critical analysis of children's disclosure developed by the user. Evaluators who become accustomed to this type of analysis carry the awareness into the interviewing setting and incorporate the thought processes into practice. Such thinking is inconsistent with a "witch hunt" mentality.

A pilot study on the forensic evaluation protocol was conducted over a two-year period. A subsequent study evaluated the efficacy of the model for an additional two years at multiple sites.