Police departments and social service agencies witnessed a dramatic increase of criminal cases involving children in the 1980s. This increase, driven by shifts in societal attitudes and the resulting legal and policy reforms, left forensic interviewers searching for optimal ways to elicit accurate testimony from child witnesses. More than a quarter century of research now demonstrates that children can, when interviewed with certain methods, give accurate accounts of previously experienced events.
Most of this forensic developmental research has focused on interview procedures for cases involving alleged child victims of sexual abuse (CSA). Because incidents of CSA rarely involve witnesses or physical evidence, garnering accurate reports in these cases is especially important. Before continuing it should be noted that, although the true prevalence of CSA is unknown, it is well agreed that child abuse is a pervasive societal problem.
History of Forensic Interviews with Children
Prior to the 1970s, children under age fourteen typically were not allowed to give uncorroborated testimony in the courtroom. The courts reasoned that children were neither able to distinguish fact from fantasy nor able to accurately remember past events. In the 1980s, as the reality of the frequency of CSA became more apparent, and as a wave of hysteria swept over the United States regarding mass (and sometimes ritualistic) sex abuse rings, views of children’s competency took a polemic shift. Many child protective workers adopted the belief that children do not make false allegations, and never lie, especially about sexual abuse. Some interviewers adopted the belief that children are reluctant to disclose sexual abuse; some went as far as to consider denial of abuse a sign that the child was abused. Because of these views, some investigators adopted forceful interview styles, doggedly pursuing children with techniques such as bribery, selective reinforcement, repeated suggestive interviewing, peer pressure, and negative stereotyping during interviews lasting several hours (and sometimes over several days or weeks) until the child finally assented to abuse. During the past several decades, however, a corpus of studies demonstrates that interviewer bias manifests in suggestive questioning that leads to false reports from children (see Bruck and Ceci 1999).
Suggestive interviewing extends beyond simply asking children misleading questions. For example, studies have found that negative stereotyping can produce inaccurate accounts in children. Leichtman and Ceci (1995) told children a story about a man, Sam, who was very messy and clumsy. Sam later visited their classroom and engaged in a few neutral actions with the children. When suggestively interviewed about the visit after a ten-week delay, many children made false claims about Sam that were consistent with the stereotype of being clumsy and messy (for example, that he tore pages out of a topic). Studies have also found that selective reinforcement (for example, telling the child he is a good boy only when making abuse disclosures) may lead children to provide false information. The scientific literature indicates that children sometimes agree with false statements about central actions involving bodily touch. While preschoolers are particularly prone to acquiescing to suggestive questions, studies are emerging that find suggestibility effects in older school-aged children.
Despite studies showing that suggestive techniques are detrimental to children’s reports, some interviewers justified their continued use of such techniques by claiming that such methods were necessary to elicit disclosures from reticent children. However, a comprehensive review of the literature revealed that while children often delay disclosing abuse (or never disclose to anyone during childhood), the evidence does not support the notion that abused children commonly deny or recant abuse allegations when asked directly about abuse (London et al. forthcoming). In short, the evidence does not support the practice of interviews characterized by suggestive techniques; rather such methods have been shown to produce erroneous reports.
As a result of research showing the detrimental influences of suggestive interviewing techniques, a number of different teams have developed ”suggestion-free” protocols (for example, NICHD protocol [see Lamb 1994]; Poole and Lamb 1998). These protocols provide detailed accounts of empirically supported interview techniques, which are briefly summarized next.
Empirically Supported Techniques for Interviewing Children
The first interview is crucial in that it provides the first opportunity for the child to provide his or her account of the event. In the absence of suggestive techniques (or some overt motive for the child to employ deception), children as young as age three can provide accurate details about past experienced events.
The interview should start with brief, open-ended rapport building (for example, ”Tell me everything about a recent holiday”). Rapport building helps establish the conversational style of the interview (see Sternberg et al. 1997). The interviewer should explain to children that because the interviewer was not there, he or she does not know what happened, and that the child can respond ”don’t know.” If the interviewer conducts a truth/lie ceremony, it should be brief and he or she should avoid asking children to explain the difference between a truth and a lie. Rather, the interviewer can ask children to identify examples of truths and lies (for example, ”If John said that he was a girl, is that the truth, a lie, or something else?”; see London and Nunez 2002).
Next, the interviewer must introduce the topic of abuse without mentioning a specific event or suspect. Sternberg et al. (1997) found that the following prompt elicited disclosure in 96% of alleged victims: ”Now that we know each other a little better I want to talk about the reason you are here today. I understand that something may have happened to you. Please tell me everything that happened, every detail, from the very beginning to the very end” (p. 1146). The key point is to raise the topic of abuse without suggesting a specific incident or suspect.
Research has found that children give the most accurate reports to open-ended questions (for example, ”Tell me everything you can remember” and ”What else?”). Children should be allowed to make disclosures in their own words. While research has found that open-ended questions lead to the most accurate reports for children, particularly for young children, their open-ended reports are sometimes sparse. The interviewer should be patient and exhaust open-ended prompts. After the child has provided a free narrative, the interviewer should continue to prompt with open-ended questions to get additional information and clarification. More direct questions may be necessary to elicit information from the child. Even so, the interviewer should return to open-ended questions. (For example, an assent to ”Were your clothes on or off?” can be followed with ”Tell me about that.”) Yes/no questions should be avoided as much as possible and only introduced at the end of the interview to clarify statements the child has made. Wh- questions are not necessarily neutral. (For example, the interviewer might ask ”What did he say to you when he touched your penis?” when the child had not made any statements about touching.) The interviewer must be careful not to embed suggestions within a direct question (for example, ”When he took your pants off, were you in the bedroom or the living room?” when the child previously said nothing about having clothes off), but rather only ask direct questions about topics the child has already mentioned.
There is mixed evidence on whether interview props such as dolls and drawings facilitate children’s reports (for a review, see Salmon 2001). If props are used, they should be introduced toward the end of the interview (but only among children who have made disclosures). The use of props should be avoided with preschoolers, because studies suggest that they do not have the cognitive prerequisites to appreciate the props’ symbolic quality. At this time, the safest practice would be to avoid or minimize the use of props during forensic interviews with children.
Best practice guidelines recommend audiotaping and videotaping the interview. Typically, children’s disclosures comprise the sole evidence in CSA cases, so it is crucial for fact-finders to be able to evaluate the quality by which this evidence was collected. Also, by recording interviews, prosecutors can counter unfounded criticisms that shoddy interview techniques were used.
Conducting quality forensic interviews with children requires specialized and continued training. Lamb and colleagues found that intensive training in using highly standardized interview procedures alone was insufficient for maintaining quality interview methods. Additional monthly supervision and individual feedback, however, greatly bolstered the quality of interviewers’ methods.
Forensic interviewers have a weighty job. The stakes of reaching correct decisions are high. Furthermore, once false statements are elicited from children, evidence suggests that children come to believe these false claims, and adults cannot reliably distinguish children’s narratives about true versus false events. In the laboratory, researchers have knowledge of the experimental events; in the real world, however, interviewers, experts, and the jury members do not know whether statements are true or false in most cases.
On a positive note, there is a growing body of empirically supported guidelines for conducting forensic interviews with children. Previously, interviewers were basically being asked to build a house with no tools in their toolbox. These forensic manuals provide the best tools that we have today to promote disclosure among truly abused children while minimizing false reports.
