Since the beginning of the 1980s, there have been a number of legalcases in which young childrenhave provided uncorroborated testimony involving sexual abuse. Although it seemed from the evidence that the children in many ofthese cases were subjected to anumber of suggestive interviews,the primary issue in deciding guilt or innocence was the degree to which such interviews could actually bring children to make serious allegations.

Until recently, scientific data provided little insight into this forensic issue. Specifically, although there were a number of studies showing that young children are more suggestible than adults (reviewed by Ceci & Bruck, 1993), these studies were limited to examinations of the influence of single misleading suggestions on children's recall of neutral, and often uninteresting, events. In other words, the conditions of the studies were not similar to the conditions that brought children to court.

Recommended Reading

Ceci, S.J., & Brack, M. (1995). (See References)

Poole, D.A., & Lindsay, D.S. (in press). Assessing the accuracy of young children's reports: Lessons from the investigation of child sexual abuse. Applied and Preven-\tative Psychology.

Interviewer bias and suggestive intervieuwing techniques

We have proposed that interviewer bias is the central driving force in the creation of suggestive interviews. Interviewer bias characterizes an interviewer who holds a priori beliefs about the occurrence of certain events and, as a result, molds the interview to elicit from the interviewee statements that are consistent with these prior beliefs. 

One hallmark of interviewer bias is the single-minded attempt to gather only confirmatory evidence and to avoid all avenues that may produce disconfirmatory evidence. Thus, a biased interviewer does not ask questions that might provide alternate explanations for the allegations (e.g., "Did your mommy tell you, or did you see it happen?"). 

Nor does a biased interviewer ask about events that are inconsistent with the interviewer's hypothesis (e.g., "Who else beside your teacher touched your private parts? Did your mommy touch them, too?"). And a biased interviewer does not challenge the authenticity of the child's report when it is consistent with the interviewer's hypothesis. When a child provides inconsistent or bizarre evidence, it is either ignored or interpreted within the framework of the biased interviewer's initial hypothesis.

A number of studies highlight the effects of interviewer bias on the accuracy of children's reports (reviewed in Ceci & Bruck, 1995).

In some studies, children are engaged in a staged event. Later, naive interviewers, who did not witness the event, are given either accurate or false information about the event and then told to question the children. Interviewers who are given false information are unaware of this deliberate deception, which is carried out to create a "bias." 

In other studies, children are asked to recall a staged event by an experimenter who intentionally conveys a bias that is either consistent or inconsistent with the staged event. In both types of studies, when questioned by interviewers with false beliefs, children make inaccurate reports that are consistent with the interviewers' biases.

According to our model, interviewer bias influences the entire architecture of interviews, and it is revealed through a number of different component features that are suggestive. We briefly describe some of these in this section.

In order to obtain confirmation of their suspicions, biased interviewers may not ask children open-ended questions, such as "What happened?" but instead resort to a barrage of specific questions, many of which are repeated, and many of which are leading.

This strategy is problematic because children's responses to openended questions are more accurate than their responses to specific questions. This finding has been reported consistently since the beginning of the century (e.g., see Ceci & Bruck, 1995) and is highlighted in a recent study by Peterson and Bell (1996), who interviewed children after they visited an emergency room for a traumatic injury.

Children were first asked openended questions (e.g., "Tell me what happened"), and then asked more specific questions (e.g., "Where did you hurt yourself?" or "Did you hurt your knee?"). The children were most likely to report the important details accurately in response to open-ended questions (91% accuracy); errors increased when children were asked specific questions (45"/o accuracy).

Forced-choice questions (e.g., "Was it black or white?") also compromise the reliability of children's reports because children tend not to respond, "1 don't know" (e.g., see Walker, Lunning, & Eilts, 1996), even when the question is nonsensical (Hughes & Grieve, 1980).

Not only does accuracy decrease when children are asked specific questions, but there is increased risk of taint when young children are repeatedly asked the same specific questions, either within the same interview or across different interviews (e.g., Poole & White, 1991). Under such circumstances, young chiidren tend to change their answers, perhaps to provide the interviewer with the information that they perceive he or she wants.

Some interviewers convey their bias by asking leading questions and providing information about the alleged target events. When these techniques are repeated across multiple interviews, children's reports may become tainted.

For example, in one study (Bruck, Ceci, Francoeur, & Barr, 1995), 5-year-old children visited their pediatrician and received an inoculation. One year later, they were interviewed four times about salient details of that visit. Children who were repeatedly interviewed in a neutral, nonleading manner provided accurate reports about the original medical visit. In contrast, children who were repeatedly given misinformation about some of the salient details were very inaccurate; not only did they incorporate the misleading suggestions into their reports (e.g., falsely claiming that a female research assistant, rather than the male pediatrician, inoculated them), but they also reported nonsuggested but inaccurate events (e.g., falsely reporting that the female research assistant had checked their ears and nose).

Interviewers can also use subtle verbal and nonverbal cues to communicate bias. At times, these cues can set the emotional tone of the interview, and they can also convey implicit or explicit threats, bribes, and rewards for the desired answer. Children are attuned to these emotional tones and act accordingly.

 In one study, for example, children were asked to recall the details of a visit to a university laboratory that had occurred 4 years previously (Goodman, Wilson, Hazan, & Reed, 1989). At the 4-year follow-up interview, the researchers deliberately created an atmosphere of accusation by telling the children that they were to be questioned about an important event and by saying, "Are you afraid to tell? You'll feel better once you've told." Although few children remembered the original event from 4 years earlier, a number of the children assented to suggestive questions implying abuse; some children falsely reported that they had been hugged or kissed, or that they had had their picture taken in the bathroom, or that they had been given a bath.

Thus, children may give incorrect information to misleading questions about events for which they have no memory, if the interviewer creates an emotional tone of accusation.

Stereotype induction is another possible component of a suggestive interview. For example, if a child is repeatedly told that a person "does bad things," then the child may begin to incorporate this belief into his or her reports. 

A study of preschool children illustrates this pattern (Leichtman & Geci, 1995). On a nuniber of occasions, the experimenters told the children about their "clumsy" friend Sam Stone, whose exploits included accidentally breaking Barbie dolls and ripping sweaters. Later, Sam came to the children's classroom for a short, accident-free visit. The next day, the teacher showed the children a torn book and a soiled teddy bear. Several weeks later, a number of these 3- to 4-year-old children reported that Sam had been responsible for these acts; some even claimed that they had seen him do these things. Children who had not received the stereotype induction rarely made this type of error.

Techniques that have been especially designed for interviewing children about sexual abuse may be potentially suggestive. For example, anatomically detailed dolls are commonly used by professionals when interviewing children about suspected sexual abuse. It is thought that the use of the dolls overcomes language, memory, and motivational (e.g., embarrassment) problems. 

However, the existing data indicate that the dolls do not facilitate accurate reporting. In some cases, children are more inaccurate with the dolls, especially when asked to demonstrate certain events that never happened (e.g., Gordon et al., 1993). Thus, dolls may be suggestive if children have not made any allegations but are asked by an interviewer who suspects abuse to demonstrate abuse with the dolls.

Our recent studies provide evidence for this hypothesis (Bruck, Ceci, & Trancoeur, 1945; Bruck, Ceci, Francoeur, & Renick, 1995).

Three- and 4-year-old children had a medical examination during which some of them received a routine genital examination. After the children were interviewed about the examination, they were given an anatomical doll and told, "Show me on the doll how the doctor touched your genitals." Approximately 50% of the children who had not received a genital examination falsely showed touching on the doll.

Furthermore, when the children who had received a genital examination were asked the same question, a number of them incorrectly showed that the do ctor had inserted a finger into their genitals; the pediatrician had never done this. Next, when the children  in the study were given a stethoscope and a spoon and asked to show what the doctor did or might do with these instruments, some children incorrectly showed that he used the stethoscope to examine their genitals, and some children inserted the spoon into the genital or anal openings or hit the doll's genitals. None of these actions had occurred. 

We concluded that these false actions were the resuli of implicit suggestions that it was permissible fo show sexualized behaviors. Also, because of the novelty of the dolls, children were drawn to insert fingers and other objecfs into fheir cavities.

Guided imagery is another inferviewing technique that is potentially suggestive. Interviewers sometimes ask children to try to remember if or prefend that a certain event occurred and then to create a menfal picture of the event and to think about its details. Because young children sometimes have difficulty distinguishing between memories of actual events and memories of imagined events (e.g., Parker, 1995; Welch-Ross, 1995), when asked to pretend about or imagine certain events, children may later come to report them as real and believe them to be so. 

This hypothesis is supported by studies in which young children were repeatedly asked to think about real as well as imaginary events, creating mental images each time they did so.

 In one of these studies (Ceci, Loftus, Leichtman, & Bruck, 1994), children increasingly assented to false events with each successive interview. When these children were told after 11 sessions fhat some of the imagined events had not happened, most of fhe children who had previously assented to false beliefs continued fo hold onto their false statements. 

These data indicate that a number of the children had actually come to believe that they had experienced fhe false events.

Conclusions and qualifications

There is ample evidence fhaf children older than 6 years of age are suggestible about a wide range of evenfs (e.g., Goodman et al., 1989; Poole & Lindsay, 1996; Warren & Lane, 1995) and fhaf adulfs' recollections are impaired by suggestive interviewing techniques (e.g., Hyman & Penfland, 1996; Loftus & Pickrell, 1995).

Second, although there are consistent findings of age differences across studies, there are nevertheless individual differences. Some preschoolers are very resistant to interviewers' suggestions, whereas some older children will immediately fall away to the slightest suggestion. 

Researchers are a long way from understanding the source of these individual differences but are beginning to assess fhe association between suggestibility and a number of cognifive characteristics (e.g., knowledge base, memory), psychosocial factors (e.g., compliance, self-esteem), and interviewing techniques (e.g., the use of various suggestive componenfs).

Third, contrary to previous claims that children are suggestible only about peripheral details (e.g.. Melton, 1992), the newer studies show that children are also suggestible about central events. These central events may involve bodily touching that may have sexual connotations. Thus, in some suggestibility studies, children falsely claimed fhat a nurse licked their knees, a scientist put something "yucky" in their mouths, a pediatrician inserted a spoon into their genitals, and a man kissed their friends on the lips and removed some of fhe children's clothes.

Fourth, fhe number of suggestive interviewing techniques (which reflects the degree of interviewer bias) can account for variations in suggestibility estimates across and within studies. If a biased interviewer uses more than one suggestive technique, there is a greater chance for taint than if he or she uses just one technique.

 For example, we (Bruck, Ceci, & Flembrooke, in press) constructed interviews that combined a variety of suggestive techniques (visualization, repeated questioning, repeated misinformation) to elicit children's reports of true events (helping a visitor in the school, getting punished) and false events (helping a woman find her monkey, seeing a thief taking food from the day care). After fwo suggestive interviews, most children in this study had assented to all events, a pattern that continued to the end of the experiment.

Fifth, the procedures used in most studies do not allow one to determine if the children's false reports reflect false belief (false memory) or merely knowing compliance to the interviewer's suggestion. There may be a time course for the emergence of these different states. Children may start out knowingly complying to suggestions, but with repeated suggestive interviews, they may come to believe the suggestions and incorporate them into their memories.

There are a few studies that show that when suggestions are repeated to children over time, a number of the children do develop false beliefs (e.g., Ceci et al., 1994; Leichtman & Ceci, 1995; Poole & Lindsay, 1996). 

Furthermore, if the suggestive interviews cease for a period of time, these false memories fade (e.g., Huffman, Crossman, & Ceci, 1996; Poole & Lindsay, 1996).

Sixth, children who have undergone repeated suggestive interviews appear highly credible. When highly trained professionals in the fields of child development, mental health, and forensics view videotaped interviews of these subjects, they cannot reliably discriminate between children whose reports are accurate and children whose reports are inaccurate as the result of suggestive interviewing techniques (see Leichtman & Ceci, 1995). 

We have attempted to isolate the linguistic markers that might differentiate true narratives from false narratives that emerge as a result of repeated suggestive interviews (Bruck et al., in press). We have found that with repeated suggestive interviews, false stories quickly come to resemble true stories in terms of 

• the number of details, 
• the spontaneity of utterance, 
• the number of details not previously reported (reminiscences), 
• inconsistency across narratives, 
• the elaborativeness of the details, and 
• the cohesiveness of the narrative. 

It is only the greater consistency of narratives of true events that differentiates them from narratives of false events. Thus, suggestive interviewing procedures can result in highly credible but inaccurate witnesses.

Finally, although we have focused here on the conditions that can compromise reliable reporting, it is also important to acknowledge that a large number of studies show that children are capable of providing accurate, detailed, and useful information about actual events, including traumatic ones (for reviews, see, e.g., Fivush, 1993; Goodman, Batterman-Faunce, & Kenney, 1992). 

What characterizes these studies is the neutral tone of the interviewers, the limited use of leading questions (for the most part, if suggestions are used, they are limited to a single occasion), and the absence of any motive for the children to make false reports.

When such conditions are present, it is a common (although not universal) finding that children are relatively immune to suggestive influences, particularly about sexual details. When such conditions are present in actual forensic or therapeutic interviews, one can have greater confidence in the reliability of children's allegations. It is these conditions that one must strive for when eliciting information from young children.

References

Bruck, M., Coti, SJ,, & Francoeur, T., (1995, March). Anatomically detailed dolls do not facilitate preschoolcrs' reports of touching. Paper presented at biannual meeting of the Society for Research on Child Development, Indiandapolis, IN.

Bruck, M,, Ceci, S.J., Francoeur, E., & Barr, K.J. (1993). "I hardly cried when I got my shot!": Influencing children's reports about a visit to their pediatrician. Child Developnent. 66, 193-208.

Bruck, M., Ceci, S.J,, Francoeur, F., & Kenick, A, (1995). Anatomicaliy detailed dolls do not facilite the preschoolers' reports of a pediatric exiimination involving genital touching. Journal of Experimental Psychology Applied, 1, 95-109.

Bruck, M., Ceci, S.J., & Hembrookc, H. (in press). Children's reports of pleasant and unpleasant events. In; D. Read & S, lindsay (Eds.), Recollections of trauma: Scientific research and clinical practiee. New York: Plenum Press.

Ceci, S.J., & Bruck, M. (1993). The suggestibility of the child witness: A historical review and synthesis. Psychological Bulletin, 115, 403-439.

Ceci, S.]., & Bruck, M. (1995). Jeopardy in the courtroom: A scientifie analysis of children's testimony. Washington, DC: American Psychological Association.

Ceci, S.J., Loftus, E.W,, Leichtman, M , & Bruck, M. (1994). The role of source misattributions in the creation of false beliefs among preschoolers, Internalional Journal of Clinical and Erperiinential Hypnosis, 62. 304-320.

Fivush, R. (1993). Developmental perspectives on autobiographical recall. In:  G.S. Goodman & B. Bottoms (Eds.), Child victims and child witnesses: Understanding and improving testimony (pp. 1.- 24). New York: Guilford Press.

Goodman, G.S., Batterman-Faunce, J.M,, & Kenney, R. (1992). Optimizing children's testimony: Research and social policy issues concerning allegations of child sexual .ibuse. In: D. Cicchetti & S. Toth (Eds.), Child abuse, child developnient, and social policy (pp. 139- 166). Norwood, NJ: Ablex.

Goodman, G.S., Wilson, M.E., Hazan, C , & Reed, R.S. (1989, April). Children's testimony nearly four years after an event. Paper presented at the annual meeting of the Eastern Psychological Association, Boston,

Gordon, B., Ornstein, P.A,, Nida, R., Follmer, A,, Creshaw, C , & Albert, G. (1993). Does the use of dolls facilitate children's memory of visits to the doctor? Applied Cognitive Psycholgy, 7, 459-474.

Huffman, M.L., Crossman, A., & Ceci, S. (1996, March). An investigation of the long-term effects of source misattribution error. Are false memories permanent? Poster presented at the biannual meeting of the American Psychology-Law Society, Hilton Head, SC.

Hughes, M , & Grieve, R. (1980). On asking children bizarre questions. First Language, I, 149-160.

Hyman, I.E., & Pentldnd. J. (1996). The role of mental imagery intihe creation of false childhood memories, journal of Memory and Language, .W, 101-H7.

Leichtman, M.D., & Ceci, S.J. (1995). The effects of stereotypes and suggestions on preschoolers' reports. Developmental Psychology, 31, 368-578.

Loftus, E.F,, & Pickrell, J.E. (1995). The fomiation of false memories. Psychiatric Annals, 25, 720-725.

Melton, G, (1992). Children as partners for justice: Next steps for developmentalists. Monographs of the Society for Research in Child Developinent, 57(5, Serial No. 229), 153-159.

Parker, J. (1995). .Age differences in source monitoring of performed and imagined actions on immediate and delayed tests,.Journal of Experimental Child Psychology, 60, 84-101.

Peterson, C , & Belf, M. (1996), Children's memory for traumatic injury. Child Development, 67, 3045-3070.

Poole, D.A., & Lindsav, D.S. (1996, June). Effcets of parental suggestions, intcrviewing techchniqucs, and age on young  children's reports, paper presented on the NAIO  Advanced Study Insititute, Port de Bourgenay, France.

Poole, D.A. & White, L., (1991), Effects of question repetition on the witness testmimony of childrcn aind adults. Developmental Psychology, 27, 975-986

Walker, N., Lunning, S. & Eilts, J.,  (1996, June). Do children respond accurately to forced choice choice questions? Paper prresenyted tot the NATO Advance Study Institue  Recollections of Tiauma; Scientific Research and Clinicai Pacticc, Talont Saint Hilare, France.

Warren, A.R., & Lane, P. (1995). The effect off timing and type of questioning on eyewitness  accuracy aind suggestibilily. In: M. Zaragoza (Eds.), Memory and testimony in the child windness (pp, 44-60). Thousand Oaks, CA: Sage Publications.

Welch-Ross, M. (1945). Developmental changes in preschoolers' ability to distinguish memories of performed, pretended, and imagined actions. Cognitive Development, 10, 421-441.