حافظه کاذب کودکان و افشای واقعی در برابر پرسش مکرر

The current study was designed to investigate children’s memory and suggestibility for events differing in valence (positive or negative) and veracity (true or false). A total of 82 3- and 5-year-olds were asked repeated questions about true and false events, either in a grouped order (i.e., all questions about a certain event asked consecutively) or in a nongrouped order (i.e., questions about a certain event were interspersed with questions about other events). Interviewer gender was also varied. Individual differences, including attachment style, inhibition, and behavioral adjustment, were examined as potential predictors of memory and suggestibility. Results revealed significant age, valence, and veracity effects on children’s memory reports. Path analysis demonstrated that individual differences in behavioral problems and inhibitory ability predicted children’s provision of inaccurate information. Implications for psychological theory and legal application are discussed.

Over the past few decades, the large number of children coming into contact with the legal system—mostly as a result of abuse cases—has motivated intense scientific effort to understand children’s true and false memory reports. The seriousness of abuse charges and the frequency with which a child victim’s testimony provides the central or sole prosecutorial evidence make issues of children’s eyewitness memory abilities important considerations. Questions about children’s ability to remember and report the truth revolve around both their capability to resist misleading information and their hesitancy to reveal negative experiences. Each of these possibilities poses a significant danger. If children are easily misled by zealous interviewers, then false charges against innocent adults may result. If children fail to reveal negative experiences such as physical or sexual abuse, then child abusers may escape justice and victims may be subjected to further mistreatment. Research has commonly focused on determining the likelihood of the first of these scenarios (Goodman and Schaaf, 1997 and Poole and Lindsay, 2002), with much less research aimed at examining the second scenario, although many would say that the latter is at least equally important, if not more so. Concerns have been raised not only that asking repeated questions within an interview increases the chance of creating false memories in children who have not suffered abuse but also that not asking repeated questions decreases the chances of obtaining accurate disclosures in actual abuse victims. If repeated questioning introduces misinformation, then this interviewing strategy, at least under certain circumstances (e.g., a weak memory or an event that never occurred) (Quas et al., 2007), may lead to false reports in young children (Garven et al., 1998 and Leichtman and Ceci, 1995). However, repeated questions are believed to have beneficial effects as well, including giving a child multiple opportunities to reveal negative information and acting as a measure of consistency and truthfulness. If a child is reluctant to acknowledge a negative event such as abuse, then repeated leading questions may encourage the child to reveal his or her victimization (e.g., Goodman & Clarke-Stewart, 1991). However, inconsistency in children’s answers to repeated questions leads to a perceived lack of credibility in children’s reports (Leippe, Manion, & Romanczyk, 1992). Topic switching to minimize social demands It seems clear that children’s memory reports are at times negatively affected by suggestive interviewing techniques (e.g., Garven et al., 1998). Developmental questions about the fate of children’s memory in the face of suggestive interviewing became a focus of concentration as the “misinformation effect” (Loftus, 1979) was replicated with children (e.g., Ceci et al., 1987 and Schwartz-Kenney and Goodman, 1999; but see Zaragoza, 1991) and as underlying mechanisms were explored (e.g., Holliday, Reyna, & Hayes, 2002). Furthermore, implantation of entirely false memories in research studies (Loftus & Pickrell, 1995) indicated that such implantation was more likely in younger children than in older children and adults (Ghetti and Alexander, 2004 and Pezdek and Hodge, 1999). Nevertheless, the extent to which children experience memory difficulties, and the degree to which they respond based on social demands rather than on the basis of their original memory, is not fully known. Indeed, social demands may play a larger role in children’s memory reports than in adults’ memory reports (Ceci et al., 1987). Studies that support an important role of social demands include those in which misinformation and false memory effects drop significantly when a new interviewer and unbiased questions are introduced (Ceci et al., 1994 and Quas and Schaaf, 2002). To the extent that social demands play a prominent role in influencing children’s memory and suggestibility, child interview techniques that reduce social demands become quite important. In research, and perhaps also in the real world, repeated questions about the same event—be it a true or false event—often are grouped together in an interview (e.g., Bruck, Ceci, & Hembrooke, 1997), and such grouping is likely to maximize the social demands of the situation. Even very young children may be able to infer that repetition of the same or similar questions can indicate that the questioner is looking for a different answer. This inference might adversely affect children’s sense of competence in the interview, which in itself may have negative consequences for performance. Altering the grouping of questions might have beneficial effects. By switching from topic to topic in an interview, repeated questions may preserve their ability to increase children’s reports of true events while simultaneously failing to increase children’s false reports by reducing the social demands in the repeated questions and bolstering children’s feelings of competence in the interview. “Topic switching” might also affect children’s certainty about their memory. For instance, perhaps children’s “do-not-know” responses increase in the face of repeated questions when the topic is switched, indicating uncertainty about their responses but also comfort or willingness to express that uncertainty. Grouped questions, which may increase social pressure to answer in a particular way, might decrease such responses. In that regard, effects of topic switching on do-not-know responses would be expected aside from age, which could also affect do-not-know responses. Older children are more likely than younger children to give do-not-know responses when asked unanswerable questions, even after warnings about such questions, perhaps reflecting older children’s greater comfort with the social demands inherent in formal interview situations (Beuscher & Roebers, 2005). Furthermore, the use of do-not-know responses has been linked to decreases in false alarms, or greater accuracy of memory reports, when children give do-not-know responses rather than make errors (Nesbitt and Markham, 1999 and Quas et al., 1999). It is possible, then, that topic switching will lead to an increase in do-not-know responses, particularly among older children. This may, in turn, result in fewer false alarms. Thus, it is important to consider the pattern of do-not-know responses as affected by both topic switching condition and age. Event valence and event veracity Children are more likely to assent to positive or neutral false events than to negative false events (for a review, see Bruck et al., 1997). This tendency may underlie, at least in part, why earlier researchers repeatedly found that children are more resistant to untrue abuse-related suggestions (e.g., “He hit you, didn’t he?”) than to numerous other untrue suggestions (Goodman and Aman, 1990 and Goodman et al., 1991). Because findings from child witness research often are seen as relevant to legal cases in which the events are largely of a strongly negative character, it is important to take event valence into consideration. A dearth of research exists, however, on disclosure of positive versus negative true events. In many studies, researchers have used relatively positive events to test children’s memory (e.g., Quas & Schaaf, 2002), yet generalizations across positive and negative events may be inappropriate. For example, patterns of assents and do-not-know responses may vary according to valence and veracity; that is, children may be more likely to assent to positive true events but to provide do-not-know responses when asked about events they do not wish to discuss (i.e., negative events or false events). Studies suggest that children can be hesitant to disclose true events that are embarrassing, that implicate the self or a loved one, and/or that involve transgressions (Bottoms et al., 2002, Saywitz et al., 1991 and Talwar et al., 2004). Given such hesitancy, it has been proposed that leading questioning and/or repeated questioning may be needed to elicit true disclosures of negative experiences from children. Thus, event valence is important to consider when studying memory for true versus false events. Interviewer gender Little is known about how children’s reports may be affected by interviewer gender, yet this is an issue of considerable practical interest to those who interview children. There are several possible effects of interviewer gender on children’s memory. To the extent that males are perceived as wielding more authority, and to the extent that increased perceived authority in interview situations is associated with increased suggestibility (Ceci and Bruck, 1993 and Kuehnle, 1996), children may be more inaccurate when male interviewers repeat misleading questions. This acquiescence to male authority, however, might serve to increase disclosure of true unpleasant events. On the other hand, children of both genders might perform better in interviews conducted by females who may be viewed as more supportive. In the interview environment, supportiveness tends to increase accuracy and decrease suggestibility (Carter et al., 1996, Davies et al., 2000, Davis and Bottoms, 2002 and Goodman et al., 1991; but see Imhoff & Baker-Ward, 1999). Evidence from adults in mental health interviews suggests that both genders may be more responsive and divulge more information to female interviewers regarding sensitive issues such as mental health, finances, and fertility (Axinn, 1991 and Pollner, 1998). A meta-analysis of 205 studies supported greater information being provided to female interviewers (Dindia & Allen, 1992). Furthermore, the gender of children and interviewers may interact to affect children’s accuracy and suggestibility. One study specifically investigated interviewer and child gender in suspected child sexual abuse cases. Lamb and Garretson (2003) examined forensic interviews in a large sample of alleged victims of child sexual abuse. Their results suggested that boys tended to respond similarly to male and female interviewers, whereas girls provided more information to female interviewers. Differences in child responses may have been influenced by variation in questioning styles of male and female interviewers because that study also found that male interviewers questioned boys and girls similarly, whereas female interviewers tended to ask boys more suggestive questions. The accuracy of children’s statements was not examined. Individual differences Knowledge about how individual differences affect children’s ability and willingness to provide accurate answers to questions about both true and false events is important for developmental theory as well as for gaining practical knowledge about children’s susceptibility to suggestive questioning. Knowledge of individual differences can lead to refinement of theoretical formulations of memory development. In forensic settings, such knowledge could be used profitably to tailor children’s interviews individually according to the likely strengths and weaknesses of particular children. The current study was designed to test the contributions of a number of individual difference factors to children’s memory and suggestibility. Based on a complex series of predicted interrelations, a path model was proposed (Fig. 1). Each variable included in the model is discussed next.

This study has uncovered several interviewing, event, and child factors related to the accuracy of children’s reports. Consistent with past research, increasing memory skills and decreasing inaccuracy were evident with children’s age. Our results reveal that children tend to succumb eventually to repeated leading questioning but that, at least for older children, there may be a “safe” low number of repeated questions that can be asked about an unknown event to maximize the likelihood of reporting a true negative event while minimizing the likelihood of assents about a false event. Furthermore, interviewers of both genders are likely able to elicit accurate information from children of both genders. It is important to note that, in the current study, repeated misleading questions about true and false events were asked in a kind but persistent manner, consistent with techniques used in past studies on implantation of false memories. However, we did not go to greater lengths to implant false memories, as have some researchers, such as by telling children that their mothers said all of the events had actually occurred, by telling children that the interview session was just a pretend game (e.g., Ceci et al., 1994 and Ghetti and Alexander, 2004), and all of our questioning took place in one session. It is likely that increased false reports could be elicited with even stronger suggestions and that different factors might influence memory reports given more powerful false memory implantation techniques. Our study does, in any case, offer information relevant to children’s resistance to suggestion, which is likely important in both implantation and suggestive questioning paradigms. The results of the current study could be taken to imply that standardized tests of individual difference factors (e.g., inhibition, behavioral adjustment, attachment) should be administered to children in the legal system as predictors of inaccuracies in children’s eyewitness memory reports. However, results from the path model underscore the complex interplay among many factors, indicating that no single indicator explains children’s inaccuracies. Interactions among these factors (and numerous others not assessed in the current study) across development highlight the serious risk of overemphasizing standardized testing to judge the accuracy of an individual child within a unique event and interview context. Based on the standardized measures we employed, our results suggest, for example, that many children with behavioral maladjustment likely could still provide quite accurate memory reports and that many children without behavioral maladjustment could be quite inaccurate. The current study offers several important findings for theories of children’s memory and suggestibility about personal experiences. Investigations of children’s suggestibility would profit from attending to both accuracy and potential unwillingness to respond (e.g., do-not-know responses) given that both may reveal processes underlying children’s reports. Furthermore, the types of events about which children are questioned can yield different patterns of responses. Also, inhibition skills and other indicators of executive function may be crucial for understanding age improvements in young children’s resistance to suggestion. Finally, there is a predictable interplay between relational styles of parents and children that may relate to children’s behavior problems and, ultimately, their memory errors.