Autobiographical Memory
Four salient aspects of young children’s recall of experi-enced events are summarized. First, with the emergence of productive narrative language skills around age 2 years, children can remember salient, personally experienced events over long periods of time (Fivush 1993; Peterson and Rideout 1998). Second, although forgetting occurs at all ages, younger children show steeper forgetting curves than older children (Brainerd and Reyna 1995; Brainerd et al. 1990). Third, when asked open-ended questions—
such as “What did you do at school today?”—preschool children often provide little information, although what they do provide is accurate. In this sense their free recall is characterized as accurate but incomplete (Fivush 1993;
Goodman et al. 1987). Fourth, in order to elicit reports from young (preschool) children about their past, inter-viewers often use an array of focused retrieval strategies, such as the use of specific questions (e.g., “Did you have recess at school today?”). However, the use of such strat-egies increases the number of errors in children’s reports (e.g., see Peterson and Bell 1996). These four findings indicate that in the absence of any previous prompting or suggesting by adults, children’s spontaneous
non-prompted statements are likely to be accurate, especially for recently experienced events.
Suggestibility
Children’s reports that emerge as a result of suggestive interviewing can often be inaccurate. They can come to report events that never happened. In our work, we have described the “architecture” and process of suggestive interviewing techniques (Ceci and Bruck 1995). In this characterization, the concept of interviewer bias is the defining feature of many suggestive interviews. Inter-viewer bias is a term used to characterize those inter-viewers who hold a priori beliefs about the occurrence of certain events and, consequently, mold the interview to elicit statements from the interviewee that are consistent with these prior beliefs. One of the hallmarks of inter-viewer bias is the single-minded attempt to gather or accept only confirmatory evidence and to avoid all ave-nues of disclosure that may produce negative or inconsis-tent evidence. Thus, while gathering evidence to support his or her hypothesis, an interviewer may fail to gather any evidence that could potentially disconfirm that hypothesis. The biased interviewer does not ask ques-tions that might provide alternate explanaques-tions for the allegations (e.g., a biased interviewer would not ask, “Did your mommy and daddy tell you that this happened or did you see it happen with your eyes?”). Nor does the biased interviewer ask about events that are inconsistent with the hypothesis or that might prompt the child to answer in a certain way that would call the child’s accu-racy and previous allegations into question (e.g., a biased interviewer would not ask, “Who else besides your teacher touched your private parts? Did your mommy touch them too?”). And the biased interviewer does not challenge the authenticity of the child’s report when it is consistent with the hypothesis (e.g., the biased inter-viewer would not say, “It’s important to tell me only what you saw, not what someone may have told you.” or “Did that really happen?” or “It’s OK to say you don’t remem-ber or you don’t know.”). Biased interviewers also ignore inconsistent or bizarre evidence or else interpret it within the framework of their initial hypothesis (e.g., a biased interviewer who was told by a child that the defendant took him to the moon in a spaceship might try to twist this into a more feasible account of going on a space ride at an amusement park with the defendant). It is impor-tant to note that within this context, a biased interviewer may be a police officer, a therapist, and even a parent. It takes no special skills to be a biased interviewer.
Interviewer bias influences the entire architecture of interviews, and it is revealed through a number of
differ-ent compondiffer-ent features that are potdiffer-entially suggestive.
For example, in order to obtain confirmation of their beliefs, biased interviewers may not ask children open-ended questions such as “What happened?” but quickly resort to a barrage of very specific questions, many of which are repeated and leading. When interviewers do not obtain information that is consistent with their suspi-cions, they may repeatedly interview children about the same set of suspected events until they do obtain such information, sometimes reinforcing responses con-sistent with their beliefs and ignoring information that is inconsistent with their beliefs. “Stereotype induction” is another strategy used by biased interviewers. Here the interviewer gives the child information about some char-acteristic of the suspected perpetrator. For example, chil-dren may be told that a person who is suspected of some crime “is bad” “or does bad things.” Interview bias is also reflected in the use of some techniques that are specific to interviews between professionals and children. One of these involves the use of anatomically detailed dolls and line drawings in investigations of sexual abuse. When interviewers suspect abuse, before the children have made any allegations, they sometimes ask children to show on the dolls how they have been sexually abused.
The current research in the field of children’s sug-gestibility examines the degree to which these and other interviewing techniques, when used in isolation or in combination, result in tainted and unreliable reports from young children. The following is a summary of the major findings:
1. There are reliable age effects in children’s suggestibil-ity, with preschoolers being more vulnerable than older children to a host of factors that contribute to unreliable reports. Despite these significant age dif-ferences, it is nonetheless important to point out that there is still concern about the reliability of older chil-dren’s testimony when they are subjected to sugges-tive interviews. There is ample evidence that children older than age 6 years are influenced by suggestive questioning about a wide range of events (e.g., Good-man et al. 1989; Poole and Lindsay 2001; Warren and Lane 1995) and that even adults’ recollections are impaired by suggestive interviewing techniques (e.g., Hyman et al. 1995; Loftus and Pickrell 1995). Thus, age differences in suggestibility are a matter of degree.
2. When children incorporate false suggestions into their reports, they often lose the source of the information and make source monitoring errors; that is, they can-not correctly distinguish between suggested and experienced events (Johnson et al. 1993) and,
conse-quently, inaccurately report that they actually wit-nessed the suggested information. These types of errors reflect false beliefs. Although source monitor-ing errors occur at all ages, they are most prominent in children younger than age 6 years (e.g., Ackil and Zaragoza 1995; Parker 1995; Poole and Lindsay 1995, in press).
Suggestibility effects can also result from social factors such as the child’s compliance or acquiescence to the perceived wishes of the interviewer. The child reports what he or she thinks the interviewer wants to hear rather than what actually happened. Finally, sug-gestibility may initially reflect compliance, but with time and additional interviews, the false reports may become false beliefs. That is, what begins as a recog-nized attempt to please an interviewer by making an inaccurate claim can become tomorrow’s false mem-ory, wherein the child actually comes to believe the inaccurate claim.
3. Errors that children make as a result of suggestive techniques involve not only peripheral details but also central events that involve their own bodies. At times children’s false reports can be tinged with sexual con-notations. In research studies, young children have made false claims about “silly events” that involved body contact (e.g., “Did the nurse lick your knee?”
“Did she blow in your ear?”), and these false claims persisted in repeated interviewing over a 3-month period (Ornstein et al. 1992). Young children falsely reported that a man put something “yucky” in their mouth (Poole and Lindsay 1995, 2001). Preschoolers falsely alleged that their pediatrician had inserted a finger or a stick into their genitals (Bruck et al. 1995a) or that some man touched their friends, kissed their friends on the lips, and removed some of the chil-dren’s clothes (Lepore and Sesco 1994). A significant number of preschool children falsely reported that someone touched their private parts, kissed them, and hugged them (Bruck et al. 2000b; Goodman et al.
1990; Goodman et al. 1991; Rawls 1996). In addi-tion, when suggestively interviewed, children will make false allegations about nonsexual events that could have serious legal consequences were they to occur. For example, preschool children claimed to have seen a thief at their day care (Bruck et al.
1997).
4. A range of interviewing techniques can negatively influence the accuracy of children’s reports. These include verbal dimensions of the interview (e.g., the way questions are asked, the number of times ques-tions are repeated, and the use of reinforcement/
punishment), but they also include nonverbal
tech-niques such as the use of dolls, line drawings, and props. Young children who are interviewed with these media tend to make more errors (of commission) than do children who are interviewed without these media (e.g., Bruck et al. 1995a, 2000b; Gordon et al.
1993; Salmon et al. 1995; Steward and Steward 1996).
5. The “mix” of suggestive interviewing techniques in conjunction with 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 a tainted interview than if he or she uses just one technique. For example, the effects of stereo-type induction paired with the repeated use of mis-information had greater detrimental effects on the accuracy of young children’s reports than using only repeated misinformation or stereotype induction (Leichtman and Ceci 1995). In another study (Bruck et al. 1997), we constructed highly suggestive inter-views that combined a variety of suggestive tech-niques (visualization, repeated questioning, repeated misinformation) to elicit children’s reports of true events (helping a visitor in the school; getting pun-ished) and false events (helping a woman find her monkey; seeing a thief taking food from the day care).
After two suggestive interviews, most children had assented to all events, a pattern that continued to the end of the experiment. Thus, it appears that combin-ing multiple suggestive techniques results in higher levels of inaccurate reports than commonly observed when only a single suggestive technique is used.
6. Real-world interviewers unfortunately appear to use many of the techniques that laboratory researchers have demonstrated to be deleterious to children’s accuracy, such as strongly suggestive questions, invo-cation of peer pressure, and introducing information not previously provided by the child. For example, Lamb and his colleagues (1996) have examined inter-views with alleged sexual abuse victims by investiga-tors in Israel and in the United States. A small minor-ity of the interviewers’ utterances invited open-ended responses from the child. A much larger proportion of their utterances were leading or suggestive (e.g., Lamb et al. 1996; Sternberg et al. 1999).
7. Suggestive interviewing affects the perceived credi-bility of children’s statements. Subjective ratings of children’s reports after suggestive interviewing reveal that these children appear highly credible to trained professionals in the fields of child development, men-tal health, and forensics (e.g., Leichtman and Ceci
1995, Ceci et al. 1994a, 1994b); these professionals cannot reliably discriminate between children whose reports are accurate from those whose reports are inaccurate as the result of suggestive interviewing techniques. Another line of study suggests that even well-trained professionals cannot reliably differenti-ate between true and false reports of sexual abuse (Realmuto et al. 1990), even when they are provided with extensive background information about the case (Horner et al. 1993a, 1993b)
The major reason for this lack of accurate discrim-ination between true and false reports is perhaps due to the fact that suggestive techniques breathe authen-ticity into the resulting false reports. Children’s false reports are not simple reflections or monosyllabic responses to leading questions. Under some condi-tions, their reports become spontaneous and elabo-rate, going beyond the suggestions provided by their interviewers. For example, in the study by Bruck et al.
(1997), children’s false reports contained the prior suggestion that they had seen a thief take food from their day care, but the reports also contained nonsug-gested details such as chasing, hitting, and shooting the thief (also see Bruck et al. 1995b).
In sum, the existing scientific literature suggests that when children’s false statements emerge from suggestive interviews, there is no “Pinocchio” test that can be used even by the most qualified profes-sionals to definitively ascertain whether or not the event occurred.
8. Even though suggestibility effects may be robust, the effects are not universal. Thus, even in studies with pronounced suggestibility effects, there are always some children who are highly resistant to suggestion, and there are also some adults who are highly suggest-ible. Further, although suggestibility effects tend to be most dramatic after prolonged and repeated inter-viewing, some children incorporate suggestions quickly, even after one short interview (Garven et al.
1997; Thompson et al. 1997). Despite the variation in suggestibility, social scientists have not identified a particular cognitive, personality, or temperament style that can predict on an individual level which child will or will not fall sway to suggestive interview techniques. To date the strongest determinant is age.
9. Finally, although we have concentrated on the condi-tions 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, some of which are traumatic (e.g., see
Fivush 1993 and Goodman et al. 1992 for a review).
What characterizes these studies is the neutral tone of the interviewer, the limited use of misleading questions (for the most part, if suggestions are used, they are limited to a single occasion), and the absence of any motive for the child to make a false report.
When such conditions are present, it is a common (although not universal) finding that children are much more immune to suggestive influences, partic-ularly 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 we must strive for when eliciting information from young children.