Child: Memory and Suggestibility

Amye Warren and Dorothy Marsil wrote an insightful critique regarding children’s suggestibility.  Here’s the abstract:

Why children’s suggestibility remains a serious concern.
Amye R. Warren and Dorothy F. Marsil
65 Law & Contemp. Probs. 124 (2002)
Introduction
After more than two decades of continuous contemporary research on the child as a witness in legal proceedings, a great deal is known about children’s eyewitness memory and suggestibility. Excellent reviews of this research are available,1 and their summaries and conclusions will not be reiterated here. Instead, this article will focus on six areas representing some of the most intractable problems that will require further attention from scientists and practitioners alike. This list of issues is selective and somewhat idiosyncratic, but should serve to illustrate why the current understanding of children’s suggestibility is far from complete. Research on each issue will be highlighted, concentrating primarily on studies published or presented in the past ten years.

The authors identify six areas of concern. 

1. Suggestibility is not limited to preschool children
2. Suggestiveness is not limited to leading questions
3. Suggestibility is not confined to formal interviews
4. It is difficult to identify particular children most susceptible to suggestion
5. It is difficult to train children to resist potentially suggestive questions or to “gate-out” previously suggested information
6. It is difficult to train interviewers to avoid suggestive techniques and to use techniques designed to promote accuracy

This topic paper reviews each of these areas of concern.

1. Each concern is summarized below, with excerpts from the Warren & Marsil article.
2. At the end of each topic, the abstractw for the studies they reference are provided.
3. At the end of each of the topic areas I have also attached additional research abstracts of related studies.
4. Most of the abstracts I am providing you are also available in full-article form. Just let me know which articles you would like, and I’ll email them to you. 

1. Suggestibility is not limited to preschool children

Warren & Marsil:
In recent studies,  researchers are increasingly including older children. The findings indicate that suggestibility generally declines over the school years but that even adolescents can be significantly more suggestible than adults. On the other hand, some studies demonstrate that, under certain conditions, older children and adults can be more suggestible than younger children.

Jennifer Ackil and Maria Zaragoza examined the suggestibility of first graders, third graders, fifth graders, and college students. The subjects viewed a brief video and then heard an experimenter read a summary of the video that included some misleading information. Either immediately afterwards or one week later, participants were given memory tests. Evidence of suggestibility was found for all age groups, but first graders were more susceptible to suggestion than third and fifth graders, who were in turn more susceptible to suggestion than college students. Essentially, the same pattern of age differences was found in the proportion of the suggested items that participants claimed to have actually seen in the video, as opposed to remembering from the summary.

Julie Robinson and Pamela Briggs also showed a film to their participants,  who were four- to five-year-olds, eight- to nine-year-olds, and adults. They then asked a series of questions, some of which were misleading. A day later,  they asked another set of questions. In terms of correct answers to misleading questions, the youngest children’s performance was poorer than that of eight- to nine-year-olds and adults. The latter two groups, however, did not significantly differ. In contrast, when suggested answers to misleading questions were examined separately, the study showed that both age groups of children were more suggestible than adults, but that the children did not differ from one another.

In a similar study, Pamela Coxon and Tim Valentine examined the relative suggestibility of seven- to nine-year-olds, young adults, and elderly adults.  Participants viewed a film and later answered two sets of questions, some of which were misleading. Children were significantly more likely to be misled than were the two groups of adults. Again, the two adult groups did not differ from one another. 

William Cassel, Claudia Roebers, and David Bjorkland showed a film about a bicycle theft to kindergartners, second graders, fourth graders, and adults and then asked some of them misleading questions a week later. Kindergartners were more susceptible to suggestion than all older age groups, who did not significantly differ from one another, except when the suggestions concerned information that was central to the premise of the entire event. This suggests that the importance of the information sought is predictive of a child’s ability to recall it and to resist misleading information about it. When the information is important, younger children may be more likely to recall it, and therefore their recall and ability to resist misleading questions may be equal to that of older children and even adults.

Marc Lindberg, John Keiffer, and Stuart Thomas compared the suggestibility of nine-year-olds, thirteen-year-olds, and seventeen-year-olds. All children watched a videotape that included an apparent act of physical abuse by a mother against her son. Later they were given two memory tests about the film, the first of which included misleading questions. The nine-year-olds were significantly more suggestible than were the thirteen- and seventeen-year-olds about central details of the film, including how many times the mother hit the boy and how much blood came from his nose. Additionally, the youngest children were much more likely than the older two groups to fabricate responses to nonleading questions for which they could not recall the answers. Even the seventeen-year-olds, however, were suggestible, and a third of them confabulated answers.

Finally, another study by Marc Lindberg revealed circumstances in which older children and adults can be more suggestible than younger children. In this study, third graders, sixth graders, and college students were tested on their memories for a film about students taking a test. Prior to the film, some of the participants were misled to believe that the students would cheat. Half of the participants were misled about cheating through questions that followed the film. Although in general the younger children were more suggestible than the older children and college students, the opposite finding occurred for one question involving a student in the film who had asked about the time of day. Apparently the older children and college students, who had more extensive knowledge of cheating, assumed that asking about the time of day was a sophisticated cheating strategy, when in fact it was an innocent question.

Developmental differences in eyewitness suggestibility and memory for source.
Ackil, Jennifer K. and Zaragoza, Maria S.
Journal of Experimental Child Psychology, Vol 60(1), Aug, 1995. Special Issue: Early memory.
Abstract:
Conducted 2 studies to assess potential age-related changes in children’s ability to accurately monitor the source of suggested information either immediately or following a 1-wk delay. 474 1st, 3rd, and 5th graders and college students viewed a short film and then listened to a summary that included information not in the film. Ss were tested on the content immediately or 1 wk later. Results reveal that although all Ss claimed to remember seeing suggested items, the magnitude of this effect varied with age such that 1st graders made more source confusions than 3rd and 5th graders who, in turn, made more confusions than college students. Findings suggest that these age differences are not simply a function of more general age-related memory or performance deficits, but instead reflect developmental differences in the tendency to confuse suggested information for actually witnessed events.

Developmental patterns of eyewitness responses to repeated and increasingly suggestive questions.
Cassel, William S.and Roebers, Claudia E. M.
Journal of Experimental Child Psychology, Vol 61(2), Apr, 1996. pp. 116-133.
Abstract:
Simulated aspects of questioning experiences many witnesses may encounter in a misdemeanor trial. 122 kindergarten (KG), Grade 2, 4, and adult Ss viewed a video of 2 children arguing about the use of a bicycle. One week later Ss were asked for their free recall of the events in the video followed by sets of questions that suggested a correct , an incorrect, or no specific answer. Correct free recall varied with age, with the KG and Grade 2 Ss generally following the lead of the 1st-level questions more so than the older Ss. Older children were as accurate as adults in responding to questions about the central items. Developmental differences were found in responses to repeated suggestive questioning, with KG Ss following misleading questions and changing answers more often than older Ss. On the final, multiple-choice questions, KG Ss were able to provide the correct answer as often as they had to the initial questions, despite intervening errors.

The effects of the age of eyewitnesses on the accuracy and suggestibility of their testimony.
Coxon, Pamela, Valentine, Tim
Applied Cognitive Psychology, Vol 11(5), Oct, 1997. pp. 415-430.
Abstract:
Previous studies have compared the performance of young adult eyewitnesses with that of children or elderly eyewitnesses, but few studies have allowed direct comparison of the performance of all three age groups. The accuracy and suggestibility of accounts of a video recording of a kidnapping were investigated using an experimental eyewitness paradigm. Subjects were drawn from three age groups: children (aged 7–9 years); young adults (aged 16–18 years) and elderly subjects (aged 60–85 years). Subjects’ accuracy in answering non-misleading questions and their susceptibility to misleading information was measured. Both the elderly and child subjects gave fewer correct answers and more incorrect answers to non-misleading questions than did young adults. The elderly subjects gave fewer correct responses but also fewer incorrect responses to non-misleading questions than did child subjects. Children were more suggestible than either elderly or young adults. No significant difference was found in the suggestibility of elderly and young adults. Contrary to the trace strength hypothesis no relationship was found between accuracy of recall and suggestibility.

An interactive approach to assessing the suggestibility and testimony of eyewitnesses.
Lindberg, Marc.
The suggestibility of children’s recollections. Doris, John, (Ed); pp. 47-55; Washington, DC, US: American Psychological Association; 1991. xiii, 193 pp.
Abstract (English):
offer . . . a descriptive taxonomy [of variables affecting suggestibility] and then discuss it in terms of how we can organize the data and complexities in the field within a common terminology the present experiment was designed to manipulate some of the variables as a demonstration of the need to consider these three classes of constructs [memory processes, focus of study, and subject characteristics] in interaction / the subject variable studied was that of age / there were 254 subjects consisting of third graders, sixth graders, and college students.

Eyewitness testimony for physical abuse as a function of personal experience, development, and focus of study.
Lindberg, Marc A., Keiffer, John, and Thomas, Stuart W.
Journal of Applied Developmental Psychology, Vol 21(5), Sep-Oct, 2000. pp. 555-591.
Abstract:
329 participants from grade school through high school viewed a video-taped incident that contained, among other things, a mother striking her child in the head and the force of the blow sending him to the floor crying. Suggestions given before and after were compared in terms of several classes of dependent variables. Previous experience with this type of abuse viewed was also studied. A consistent pattern of results emerged, and several conclusions were offered on the boundary conditions for when and how different types of suggestions can enter testimony, when different developmental effects are observed, what types of tests produce the most testimony with the fewest confabulations, and when and how personal experience with the viewed abuse enter into testimonies. It was concluded that clear answers to fundamental questions can be found when Participant × Memory Procedures × Focus of Study Interactions are systematically explored.

Age trends and eye-witness suggestibility and compliance.
Robinson, Julie., Briggs, Pamela
Psychology, Crime & Law, Vol 3(3), 1997. pp. 187-202.
Abstract:
This paper addresses 2 issues: whether there is a developmental trend in suggestibility to misleading post-event information, and whether suggestibility can be reduced by use of part of the cognitive interview. 20 participants from each of 3 age groups (4–5 yr olds, 8–9 yr olds, and adults) watched a filmed event, and half of the participants in each age group were subsequently asked to recall everything they had seen using a method derived from the cognitive interview procedure. Following this, all participants were asked questions about the filmed event, some of which incorporated misleading information. 24 hrs later the witnesses were interviewed again; this time, critical questions were included about the truth of the presuppositions introduced in the initial questionnaire. It was found that although the 8–9 yr olds were more suggestible than adults, the apparent greater suggestibility of very young children (4–5 yrs) could potentially be explained in terms of heightened compliance to the perceived demands of the interviewer. The ‘be complete’ part of the cognitive interview only produced an improvement in performance for the 8–9 yr olds.

Additional Studies related to Suggestibility Across Age Ranges
Flin’s study found that the proportion of inaccurate information from six-and-nine-year-old children doubled from 9% one day after the event to 18% five months after the event, whereas the error rate for adults for the two time periods remained constant (10% vs. 8%). 

This study also found that two years later, 21% of the children attributed actions to one person that had actually been performed by another – an error not made by any of the adults, and one that has obvious legal implications. 

The effect of a five-month delay on children’s and adults’ eyewitness memory.
Flin, Rhona, et al.
British Journal of Psychology, Vol 83(3), Aug, 1992. pp. 323-336.
Abstract:
Investigated whether children can retain accurate memories of events witnessed several months earlier. 134 children (aged 5–6 and 9–10 yrs) and 43 undergraduate students witnessed a staged event and were subsequently interviewed in the days following the event and/or 5 mo later. Results indicate that while all witnesses forgot information over this period, the younger children recalled slightly less information than the older children and the adults. The total amount of incorrect information recalled did not increase over the same period. Two different interviewing techniques were used (cued recall vs enhanced recall), the latter incorporating aspects of the cognitive interview procedure. No differences were found relating to the interview techniques employed.

Two years later: Effect of question repetition and retention interval on the eyewitness testimony of children and adults.
Poole, Debra A., White, Lawrence T.
Developmental Psychology, Vol 29(5), Sep, 1993. pp. 844-853.
Abstract:
Examined witnesses’ memories for an event experienced 2 yrs earlier. Ss in 4 age groups (6-, 8-, and 10-yr-olds and adults; N = 79) answered repeated questions about an ambiguous incident that occurred as part of an earlier study (D. A. Poole and L. T. White; see record 1992-08648-001). Surprisingly, the effects of question repetition were similar to the patterns observed 2 yrs ago. There were important differences in the testimonies of children and adults, however, that were not observed in the initial study: Children were less consistent than adults across sessions on yes–no questions, less accurate in response to open-ended questions, and more likely to fabricate answers to a question about a man’s occupation. Some children also confused the actions of 2 research assistants. These results indicate the need for additional research on qualitative and quantitative changes in children’s testimonies over long delays.

Other people: A child’s age predicts a source’s effect on memory.
Carol, Rolando N., Compo, Nadja Schreiber.
Legal and Criminological Psychology, Vol 22(1), Feb, 2017. pp. 74-87.
Abstract:
Purpose: For decades, researchers have investigated the effects of various interviewing techniques used in child witness interviews. One particular technique yet to be explored fully is Other People, that is, mentioning other witnesses’ alleged statements when interviewing a child. This study thus examined how the source of others’ information affects children’s memory and source monitoring as a function of age (7–18). Method: Children and adolescents (N = 110) watched a 10-min video and were then questioned about the witnessed event 1 week later using a series of yes/no questions. Throughout this series of questions, the source of outside information (peer vs. adult vs. no source) and its veracity (correct vs. incorrect) were manipulated. Results: Findings indicated that an adult source was more detrimental to witness accuracy than a peer source or no source, but this detrimental effect diminished as witness age increased. Source monitoring data mirrored this pattern: As age increased, so did accurate source attributions for information attributed to an adult. Conclusion: Overall findings suggest that information source in conjunction with witness age should be considered when assessing the effect of outside information on child and adolescent memory.

Inviting witnesses to speculate: Effects of age and interaction on children’s recall.
Schreiber, Nadja and Parker, Janat F.
Journal of Experimental Child Psychology, Vol 89(1), Sep, 2004. pp. 31-52.
Abstract:
Inviting speculation has been found to increase children’s false recall. In this study, kindergartners and third graders saw a clown perform actions alone or in interaction with a child. Two weeks later, the speculation group recalled all actions and was asked to speculate on half the actions. The control group recalled all actions without speculating. Four weeks after the show, all children recalled all actions again. The speculation group gave more false answers to the speculated items than the control group. Surprisingly, older children tended to report as many if not more false responses than younger children, regardless of speculation. In the speculation group, there were fewer false answers for interactions than actions, but false answers did not differ across observation types in the control group. Finally, speculation did not affect free and cued recall differentially. 

Constructing rich false memories of committing crime.
Shaw, Julia, and Porter, Stephen.
Psychological Science, Vol 26(3), Mar, 2015. pp. 291-301.
Abstract:
Memory researchers long have speculated that certain tactics may lead people to recall crimes that never occurred, and thus could potentially lead to false confessions. This is the first study to provide evidence suggesting that full episodic false memories of committing crime can be generated in a controlled experimental setting. With suggestive memory-retrieval techniques, participants were induced to generate criminal and noncriminal emotional false memories, and we compared these false memories with true memories of emotional events. After three interviews, 70% of participants were classified as having false memories of committing a crime (theft, assault, or assault with a weapon) that led to police contact in early adolescence and volunteered a detailed falseaccount. These reported false memories of crime were similar to false memories of noncriminal events and to true memoryaccounts, having the same kinds of complex descriptive and multisensory components. It appears that in the context of a highly suggestive interview, people can quite readily generate rich false memories of committing crime.

2. Suggestiveness is not limited to leading questions

The majority of studies on children’s suggestibility has focused on children’s responses to leading and misleading questions. A vast literature now exists on different types of questions and the effects that they appear to have on the accuracy and completeness of children’s testimony. Suggestion, however, takes on many different forms. As with the other topics reviewed, the discussion that follows will be limited to a few recent studies that illustrate some of these lesser-studied forms of suggestion.

Sena Garven and her colleagues recently published two studies on the effects of potentially suggestive techniques used by interviewers in the McMartin Preschool case. In their first study, three- to six-year-olds were interviewed about a prior event with either suggestive questions only or a package of inappropriate interview techniques, including suggestive questions and three additional factors they termed social influence, reinforcement, and invitations to speculate. Social influence involved giving information about the event that other children (“big kids”) had supposedly already provided. Reinforcement involved praising the children’s memory and intelligence if they answered “yes” to leading and misleading questions, or exhorting them to try harder or expressing mild disapproval if they answered “no.” Children were invited to speculate by being asked to think hard and then being asked a leading or misleading question beginning with: “Do you think maybe that X happened?” Children in this package condition were significantly more likely to answer misleading questions with the suggested information than were children only asked suggestive questions, and this was equally true across the age range. Most of the misleading suggestions implied wrongdoing on the part of a male experimenter. Children given the whole package of suggestive techniques made false allegations over three times more often than did children only asked suggestive questions.

In their second study, Garven and her colleagues focused on the effects of reinforcement and the provision of event information supposedly obtained from other children, called co-witness information, on children’s tendencies to accept false suggestions regarding actions of a male experimenter that were both mundane, such as tearing a book, and fantastic, such as taking the child on a helicopter ride. In this study of five- to seven-year-olds, children who received co-witness information made more mundane false allegations than did the children in the control group, but did not differ in terms of fantastic false allegations. In contrast, reinforcement had a strong impact on children’s answers to both types of misleading questions. Thirty-five percent of children who received reinforcement made false mundane allegations, compared to thirteen percent for the controls, and fifty-two percent of the same children made false fantastic allegations, compared to only five percent for controls. In addition, false allegations tended to carry over to a later interview-two to three weeks afterwards – even when no reinforcement was used at that time.

Another set of studies has examined the effect of interviewers’ distortions or misinterpretations of children’s answers. For example, in one interview, a two- year-old child said, “G.A. touched me.”  Her interviewer responded, “Jesus loves me? Is that what you said?”; the child replied, “Yeah.'” Kim Roberts and Michael Lamb analyzed a sample of sixty-eight sexual abuse interviews involving children between the ages of three and fourteen years, and found 140 instances of interviewer distortions and misinterpretations. Many of these distortions involved the identities of people being discussed in the interview, and many others involved actions. Children actually agreed with the distortions one-third of the time, and explicitly corrected or disagreed with only one-third of them.’ In the absence of corrective feedback, the interviewers continued to use the distorted version of the information during the remainder of the interviews. However, the children later mentioned the distorted information only rarely, and their tendency to do so was unrelated to their initial agreement or disagreement.

Kim Roberts and Michael Lamb analyzed a sample of sixty-eight sexual abuse interviews involving children between the ages of three and fourteen years, and found 140 instances of interviewer distortions and misinterpretations. Many of these distortions involved the identities of people being discussed in the interview, and many others involved actions. Children actuallyagreed with the distortions one-third of the time, and explicitly corrected or disagreed with only one-third of them. In the absence of corrective feedback, the interviewers continued to use the distorted version of the information during the remainder of the interviews. However, the children later mentioned the distorted information only rarely, and their tendency to do so was unrelated to their initial agreement or disagreement.

In actual interviews, it is impossible to determine whether interviewers’ distortions are minor or major errors relative to the reality of the original events. Furthermore, children may fail to disagree with interviewer distortions because the interviewers’ interpretations may in fact be correct. Therefore, Jennifer Hunt and Eugene Borgida designed an experiment in which four- to five-year-olds, nine- to eleven-year-olds, and adults witnessed a videotaped event and then were later interviewed.The interviewers were instructed to distort five of the answers. Overall, participants disagreed with only twenty-two percent of the modifications, and adults were significantly more likely than children to dis agree. In an unbiased follow-up interview, the younger children-four- to five- year-olds-were more likely to incorporate prior interviewer distortions into their answers. Interestingly, information contained in interviewer distortions that were corrected-that is, explicitly disagreed with-was never included in subsequent reports. In other words, when children clearly and confidently recalled information well enough to disagree with their interviewers, the interviewers’ suggested misinformation was never mentioned in the children’s later descriptions or answers to questions about events.

Research has only recently begun to examine alternate forms of suggestion, some of which are explicit and obviously coercive, while others may be subtle and quite unintentional, such as distortions. Although much more research on these types of suggestion is required, it is clear that suggestions may be equally or even more detrimental to children’s testimonial veracity than leading questions. Just as with the research on leading questions, the controversy over these other suggestive techniques is likely to evolve over time, from whether they promote false allegations to how often they are used in typical interviews.

More than suggestion: The effect of interviewing techniques from the McMartin Preschool case.
Garven, Sena, et al.
Journal of Applied Psychology, Vol 83(3), Jun, 1998. pp. 347-359.
Abstract:
Child interviewing techniques derived from transcripts of the McMartin Preschool case were found to be substantially more effective than simple suggestive questions at inducing preschool children to make false allegations against a classroom visitor. Thirty-six children interviewed with McMartin techniques made 58% accusations, compared with 17% for 30 children interviewed with suggestive questions. Social influence and reinforcement appeared to be more powerful determinants of children’s answers than simple suggestive questions. The SIRR model is proposed to explain how false statements may be elicited from children or adults. Categories identified in the SIRR model are suggestive          questions, social influence, reinforcement, and removal from direct experience.

Is that what I said? Witnesses’ responses to interviewer modifications.
Hunt, Jennifer S. and Borgida, Eugene
Law and Human Behavior, Vol 25(6), Dec, 2001. pp. 583-603.
Abstract:
Modifications occur when interviewers contradict statements made by witnesses or imply that witnesses provided information that they (interviewers) did not provide. Because of their suggestive nature, modifications threaten the reliability of investigative interviews. This study investigated developmental differences in witnesses’ responses to modifications during interviews as well as in inclusion of modified misinformation in subsequent answers. 22 preschool (aged 3-5 yrs), 20 elementary school (aged 9-11 yrs), and 37 college (aged 18-50 yrs) students were interviewed about a video presentation. In the experimental conditions, the interviewer contradicted information about the video provided by the participants. Participants then answered two sets of follow-up questions: one immediately following the interview and another 6-8 days later. Results indicated that participants were more likely to ignore modifications than to correct or agree with them. Adult participants were most likely to disagree with modifications. Preschoolers were most likely to incorporate modified misinformation into subsequent answers. Implications of these findings for investigative interviews are discussed.

Co-witness information can have immediate effects on eyewitness memory reports.
Shaw, John S. III. Garven, Sena, and Wood, James M.
Law and Human Behavior, Vol 21(5), Oct, 1997. pp. 503-523.
Abstract:
When questioning a reluctant witness, investigators sometimes encourage the witness by providing information about what other witnesses have said. Three experiments were conducted to test the combined effects of such co-witness information and suggestive questioning on the accuracy of eyewitness memory reports. Exp 1 (N = 102 undergraduates) was analogous to the experience of a witness who receives information from an interviewer or questioner about what other witnesses have already said, whereas Exps 2 (N = 36 undergraduates) and 3 (N = 18 undergraduates) simulated the situation in which a witness receives information directly from a co-witness. In all 3 experiments, when Ss received incorrect information about a co-witness’s response, they were significantly more likely to give that incorrect response than if they received no co-witness information. This effect persevered in a delayed memory test 48 hr after the initial questioning session in Exp 3. Accuracy rates were lowest of all when incorrect co-witness information was paired with questioning that suggested an incorrect response. Theoretical implications for the immediate effects on the accuracy of witnesses’ memory reports and practical implications for the legal system are discussed.

ADDITIONAL STUDIES

Can reinforcement induce children to falsely incriminate themselves?
Billings, F. James, Garvin, Sena, et al.
Law and Human Behavior, Vol 31(2), Apr, 2007. pp. 125-139.
Abstract:
This study examined whether reinforcement can induce children to falsely incriminate themselves. Ninety-nine children in kindergarten through third grade were questioned regarding the staged theft of a toy. Half received reinforcement for self-incriminating responses. Within 4 min reinforced children made 52% false admissions of guilty knowledge concerning the theft, and 30% false admissions of having witnessed it. Corresponding figures for controls were 36 and 10%. Twelve percent of children admitted to participating in the theft, but the effect of reinforcement was only marginally significant. The findings indicate that reinforcement can induce children to falsely implicate themselves in wrongdoing.

Allegations of wrongdoing: The effects of reinforcement on children’s mundane and fantastic claims.
Garven, Sena, et al.
Journal of Applied Psychology, Vol 85(1), Feb, 2000. pp. 38-49.
Summary:
Garven, J. M. Wood, R. S. Malpass, and J. S. Shaw (1998) found that the interviewing techniques used in the McMartin Preschool case can induce preschool children to make false allegations of wrongdoing against a classroom visitor. In this study, 2 specific components of the McMartin interviews, reinforcement and cowitness information, were examined more closely in interviews of 120 children, ages 5 to 7 years. Children who received reinforcement made 35% false allegations against a classroom visitor, compared with 12% made by controls. When questioned about ‘fantastic’ events (e.g., being taken from school in a helicopter), children receiving reinforcement made 52% false allegations, compared with 5% made by controls. In a second interview, children repeated the allegations even when reinforcement had been discontinued. The findings indicate that reinforcement can swiftly induce children to make persistent false allegations of wrongdoing.

‘What else could he have done?’ Creating false answers in child witnesses by inviting speculation.
Schreiber, Nadja, et al.
Journal of Applied Psychology, Vol 86(3), Jun, 2001. pp. 525-532.
Abstract:
In 2 experiments (N = 111 children), a suggestive technique for interviewing child witnesses called ‘inviting speculation’ was examined. Children were presented with atypical actions for common objects in a clown show. One week later, the children were asked to speculate (e.g., ‘What else could he have done with the knife?’) in a between-subjects design on all or none of the items (Experiment 1) and in a within-subjects design on part of the items (Experiment 2), thereby getting highly probable speculations (e.g., ‘to cut’). After a 3-week delay, the experimenters found more highly probable but not more other false answers for the experimental items (Experiment 2). After a 5–6-month delay, the rate of (unspecified) false answers increased compared with the baseline (Experiments 1 and 2). The short-term effect is explained by a speculation-as-misinformation assumption, whereas the long-term effect is explained by the use of a metastrategy.

How sexual abuse interviews go astray: Implications for prosecutors, police, and child protection services.
Wood, James M and Garven, Sena
Child Maltreatment, Vol 5(2), May, 2000. pp. 109-118.
Abstract:
This article argues that child sexual abuse interviews can go astray in two different ways: (1) improper interviewing has the potential to elicit false allegations from children, and (2) clumsy interviewing does not typically produce false allegations, but may have other negative consequences, particularly for child victims. The article clarifies the distinction between the two kinds of bad interviewing and suggests that clumsy interviewing is the more common of the two. The potential negative consequences of both improper and clumsy interviewing are described, along with implications for prosecutors, police, and child protection services. In the authors’ opinion, improper interviewing can probably be eliminated rather easily, but clumsy interviewing may be considerably more resistant to change.

3. Suggestibility is not confined to formal interviews

Most of the research on the suggestibility of child witnesses has focused on fairly explicit suggestions made in the course of formal interviews with children about prior events. This is understandable, given that one can better document and control the formal interview process. Much of children’s exposure to suggestive influences, however, probably occurs outside the formal interview setting. Suggestions may be made by parents, other adults, or other children prior to the first formal investigative interview or between repeated forensic or clinical interviews.

Such outside suggestive influences have only recently begun to receive research attention. Debra Poole and Stephen Lindsay designed the innovative procedure of having parents provide misleading information to their children prior to two neutral, nonleading interviews. Children, ages three to eight years, participated in an experimental session with Mr. Science, who conducted some science demonstrations. Immediately afterward, they were neutrally interviewed. Approximately three months later, parents were sent books to read to these children at home. The books were about a visit to Mr. Science and described events that had occurred during the child’s own visit, as well as events that had not occurred. Although the children were not told that the events in the story accurately described their own visits to Mr. Science, they also were not told that the stories contained false information. Within a few days after parents read the stories, and then again after a month-long delay, children were interviewed neutrally, with the interviews progressing from free recall to direct, yes or no questions about experienced (true) and nonexperienced (false) events. The key findings of this study were that twenty-one percent of the events reported in free recall three and one-half months after the event, and ten percent four and one-half months after the event, were false because they had not been actually experienced.’ The rates of reporting false information in free recall did not differ by age across this range from three to eight years. More than one-third of the children falsely answered direct questions about nonexperienced touching after three and one-half months.  By four and one-half months, older children’s false acquiescence dropped, as it had been at least a month since they were exposed to the misinformation, but three- to four-year-olds’ false acquiescence increased, such that over half of them reported false information about nonexperienced touching. These findings raise concerns about the potential of false information to carry over into children’s responses during later nonleading interviews.

Michelle Leichtman and Stephen Ceci took another approach to supplying  preschoolers with misleading information outside of the interview context. Prior to a preschool visit by a man called Sam Stone, children were presented negative stereotypes of him. Sam was depicted as clumsy and prone to breaking objects. Children in this “stereotype only” condition then experienced Sam’s visit, during which he did not break or harm anything, and were neutrally questioned once a week for the next four weeks about the visit.’ In a fifth and final, neutral interview, more than one-third of children responded falsely to a direct question about Sam Stone’s behavior, such as “Did he rip the book?,” in a direction consistent with the negative stereotype, despite the fact that these children had not been asked any misleading questions. Similarly, in the Lindberg study described above, prior to viewing a video depicting a mother hitting her son, grade schoolers and college students were given one of three descriptions of the video. ‘ In one, they were told that the mother was mean and had been previously arrested for child abuse (“mean morn”); in the second, they were told that the son’s behavior had been very problematic at home and school, and that the mother had already tried many gentle forms of discipline (“bad boy”); and in the third, they were given no biasing information (neutral). The form of instructions affected many of the responses to both suggestive and nonsuggestive questions. For example, when asked how hard the boy deserved to be hit, those in the bad boy condition thought he should have been hit harder than did those in the neutral condition, who in turn thought he should be hit harder than did those in the mean mom condition.

Even when false information is never presented to children, whether within or outside of interview questions, false reports may emerge nonetheless. When children’s memories become distorted without external influence, it is sometimes termed “autosuggestibility.” In one form of autosuggestibility, children and adults may distort information to make it more congruent with their prior knowledge. In the Lindberg study, older children and adults were more likely to answer a nonleading question with false information than were younger children.  The younger children were not as suggestible because they did not possess sophisticated, elaborate knowledge of cheating.’ In another example of autosuggestion due to prior knowledge, Stephen Ceci and his colleagues presented a tape-recorded story to seven- and ten-year-olds. The stories featured familiar or unfamiliar television characters; in some versions, the familiar characters were described in ways that clashed with their usual portrayals, such as a superhero who was too weak to carry a can of paint.’ Children were asked to rate the characters and their behavior either immediately after the story or three weeks later. When asked immediately, children correctly recalled the character descriptions and behavior, but three weeks later, children’s ratings were distorted in the direction of their prior knowledge of the characters’ typical attributes and behaviors. On the other hand, when story characters were completely unfamiliar, these types of distortions did not occur even after the delay. Therefore, Ceci and colleagues concluded that when information conflicts with prior knowledge, prior knowledge over time will distort or perhaps even supplant that information.

These results highlight the role of time in the emergence of autosuggestibility. Especially after a delay, memories of events may become inaccurate even without outside interference. For example, in a study by Debra Poole and Lawrence White, four-, six-, and eight-year-old children and adults experienced an event and were repeatedly neutrally questioned about it one week later and then again two years later. When examining the total amount of inaccurate information reported in response to open-ended questions, they found no age differences after a week-children averaged seven percent inaccurate information compared to six percent for the adults. After two years, however, children were much more likely than adults to report inaccurate information with error rates of twenty percent versus seven percent, respectively.  Children’s inaccuracies ranged from fairly minor, such as incorrect physical descriptions ofthe persons involved in the event, to major, where the actions of one person were attributed to the other.

Suggestions outside of the formal interview context, just like those within, take a variety of forms. Suggestions can come from parents, from other children, and even from within children themselves, and can range from completely false, nonexperienced events to distortions of experienced events. Very little is known about suggestions outside the interview setting because of a myriad of practical and ethical difficulties in studying them. One issue that has received no research attention is the effect of suggesting to a child that a real event never happened. Such a suggestion could be quite different from threats not to disclose experienced events or encouragement to keep actual events secret, because in both of the latter cases the events themselves are not denied. Although researching all types of outside-interview suggestions is difficult, suggestions involving denials of true events could prove especially problematic from an ethical standpoint.

Abstracts

Autosuggestibility in memory development.
Brainerd, C. J. and Reyna, V. F.
Cognitive Psychology, Vol 28(1), Feb, 1995. pp. 65-101.
Abstract:
Five experiments examined a form of autosuggestibility (AS) in which 288 5-yr-old and 288 8-yr-old children’s answers to memorytests (MTs) were shifted in the direction of their illogical solutions (inclusion illusions) to reasoning problems. In Exps 1 and 2, illogic-consistent shifts were identified in Ss’ memories of the numerical inputs on class-inclusion problems. Magnitudes of the shifts declined with age and appeared to be due to the intrusion of inappropriate gist on memory probes rather than retroactive interference from illogical reasoning. Exps 3–5 examined a model of how gist intrusion causes AS. In this model, children retrieve and process inappropriate gist when MTs supply cues that are inadequate to permit access to verbatim memories.

Children’s long-term memory for information that is incongruous with their prior knowledge.
Ceci, Stephen J., Caves, Richard D. and Howe, Michael J.
British Journal of Psychology, Vol 72(4), Nov, 1981. pp. 443-450.
Abstract:
In the present experiment, 67 7- and 10-yr-olds listened to a tape-recorded story. In one version, familiar TV characters were described as having attributes that were incongruous with those that the Ss thought them to have (e.g., the Six-Million-Dollar Man was said to be too weak to carry a can of paint). Immediately after hearing the story, half of the Ss were asked to rate the characters on the basis of their descriptions and behavior in the story. At this point, there was no evidence of systematic distortions in remembering; however, Ss who were asked to rate the incongruous characters in the story 3 wks later displayed considerable shifts in memory in the direction of their pre-experimental knowledge of the character’s attributes. Long-term memory for unfamiliar characters in a parallel version of the story showed no such shift, indicating that memory distortion for incongruous informationwas influenced by children’s prior knowledge, and was not the result of random forgetting.

‘The effects of stereotypes and suggestions on preschoolers’ reports’: Correction.
Leichtman, Michelle D. and Ceci, Stephen J.
Developmental Psychology, Vol 31(5), Sep, 1995. pp. 758.
Abstract:
Reports an error in the original article by M. D. Leichtman and S. J. Ceci(Developmental Psychology, 1995[Jul], 31[4], 568–578). On pages 572 and 573, the artwork for Figures 3 and 4 was reversed. (The following abstract of this article original appeared in record 1995-40033-001). Children’s (N = 176) reported memories of a strange man’s visit were studied. Three- to 6-yr-olds were interviewed repeatedly after the event in 1 of the following conditions: (a) control, in which no interviews contained suggestive questions; (b) stereotype, in which children were given previsit expectations about the stranger; (c) suggestion, in which interviews contained erroneous suggestions about misdeeds committed by the stranger; and (d) stereotype plus suggestion, in which children were given both pre- and postvisit manipulations. Results from open-ended interviews after 10 weeks indicated that control participants provided accurate reports, stereotypes resulted in a modest number of false reports, and suggestions resulted in a substantial number of false reports. Children in the stereotype-plus-suggestion group made high levels of false reports. All experimental conditions showed dramatic developmental trends favoring older children.

Two years later: Effect of question repetition and retention interval on the eyewitness testimony of children and adults.
Poole, Debra A., White, Lawrence T.
Developmental Psychology, Vol 29(5), Sep, 1993. pp. 844-853.
Abstract:
Examined witnesses’ memories for an event experienced 2 yrs earlier. Ss in 4 age groups (6-, 8-, and 10-yr-olds and adults; N = 79) answered repeated questions about an ambiguous incident that occurred as part of an earlier study (D. A. Poole and L. T. White; see record 1992-08648-001). Surprisingly, the effects of question repetition were similar to the patterns observed 2 yrs ago. There were important differences in the testimonies of children and adults, however, that were not observed in the initial study: Children were less consistent than adults across sessions on yes–no questions, less accurate in response to open-ended questions, and more likely to fabricate answers to a question about a man’s occupation. Some children also confused the actions of 2 research assistants. These results indicate the need for additional research on qualitative and quantitative changes in children’s testimonies over long delays.

Children’s eyewitness reports afterexposure to misinformation from parents.
Poole, Debra Ann and Lindsay, D. Stephen
Journal of Experimental Psychology: Applied, Vol 7(1), Mar, 2001. pp. 27-50.
Abstract:
This study examined how misleading suggestions from parents influenced children’s eyewitness reports. Children (3 to 8 years old) participated in science demonstrations, listened to their parents read a story that described experienced and nonexperienced events, and subsequently discussed the science experience in two follow-up interviews. Many children described fictitious events in response to open-ended prompts, and there were no age differences in suggestibility during this phase of the interview. Accuracy declined markedly in response to direct questions, especially for the younger children. Although the older children retracted many of their false reportsafter receiving source-monitoring instructions, the younger children did not. Path analyses indicated that acquiescence, free recall, and source monitoring all contribute to mediating patterns of suggestibility across age. Results indicate that judgments about the accuracy of children’s testimony must consider the possibility of exposure to misinformation prior to formal interviews.

4. It is difficult to identify particular children most susceptible to suggestion

In studies such as that by Leichtman and Ceci, where a sizable minority or, in some cases, a majority of children are led to allege events that did not occur, researchers and practitioners alike have been divided on the meaning and practical implications of the results. Should we focus on the substantial proportion of children who succumbed to suggestive influences or on those who somehow stood their ground and resisted the repeated and combined forces of various kinds of suggestions from an authoritative adult? Which type better represents the average child? Is there any way to predict whether any individual child witness fits into one category or the other?

In an effort to answer these questions, scores of studies have been conducted over the past decade to investigate relations between suggestibility and various cognitive, personality, and social measures, as well as bio-psycho-social factors such as gender and ethnicity.  Some of the many social and personality factors examined in relation to suggestibility in recent studies include: compliance, dependence, self-esteem, temperament-specifically inhibition and adaptability-and parents’ and children’s attachment styles-categories of parent-child relationship security. Some connections between cognitive factors and susceptibility to suggestion seem fairly obvious, including intelligence, memory ability, event-relevant knowledge levels, and source-monitoring ability. Children who cannot source-monitor well may not be able to keep track of which pieces of information came from their own experience of the original event and which came from external sources. Children with greater event-relevant knowledge, on the other hand, may show increased or decreased suggestibility relative to children who possess less event-relevant knowledge. Although they may remember more accurate information, children who know more about the theme of an event may infer information that was not presented.

Despite, or perhaps because of, the ever-growing list of factors associated with susceptibility to suggestion, there are several major problems for those who hope to develop diagnostic tests of an individual child witness’s suggestibility. First, the single best predictor of suggestibility in children remains their chronological age. Age is a better predictor because it encompasses develop ment in so many relevant areas, such as increased moral maturity, increased memory monitoring ability, and decreased social compliance with authority. Second, the research on individual differences in children’s suggestibility is replete with inconsistent and often weak findings. For example, intelligence is sometimes related to suggestibility, and other times not. Finally, there are different forms of suggestion, and a child’s performance on even a single type of suggestibility task varies as a result. In other words, not only is it difficult to predict a child’s suggestibility from a measure of cognitive ability or personality, it is also difficult to predict that child’s performance on one suggestibility measure from another suggestibility measure.

Expanding on this latter point, Johann Endres points out that suggestibility  in one context may not relate to suggestibility in others because suggestibility is a multidimensional construct and varies depending on memory strength for that specific content, type of suggestion, and the like.  Thus, even a very suggestible child may be able to resist certain misleading questions. Likewise, the fact that a child yields to any given misleading question does not allow inferences about his performance on other questions.

In one recent study by Michael Brady and colleagues, this lack of predict ability from one question to another was empirically demonstrated.  Children between the ages of three and seven years were shown a videotaped event.  Afterwards, they were asked several yes or no questions about the video and then two “suggestibility check” questions unrelated to the videotaped event, including one regarding the name of the interviewer’s dog. Children’s performances on these suggestibility check questions had absolutely no relationship to their actual performance on target event questions.

Research on individual differences has revealed numerous variables that may affect children’s susceptibility to suggestion. The practical implications of these findings remain unclear. The optimistic position, as described by Jodi Quas and her colleagues, is that identification of the individual difference factors that predict suggestibility may eventually allow legal professionals to take special precautions when interviewing children who are at the greatest risk of incorporating suggestions into their reports. But individual difference research cannot now, and perhaps never should be, used “retrospectively to indicate whether a particular child’s report is or is not true, or as a litmus test to determine whether a given child is even suitable to be a witness.

Abstracts

Young children’s responses to yes–no questions: Patterns and problems.
Brady, Michael S., Poole, Debra A., Warren, and Amye R. Jones, Heather R.
Applied Developmental Science, Vol 3(1), 1999. Special Issue: New research on child witnesses: I. pp. 47-57.
Abstract:
Concern about the accuracy of children’s responses to yes–no questions has created controversy regarding the appropriateness of these questions for forensic interviews. To evaluate response patterns, 56 children (aged 3–7 yrs) were twice asked a set of yes–no questions, either in standard or in a modified, forced-choice format, about a videotaped event. Younger children were less accurate and consistent than were older children. Unlike the older children, the younger children were less accurate on questions that adults rated as probing central information compared to those involving more peripheral details. Question format did not alter children’s accuracy, their tendency to answer ‘I don’t know,’ or their consistency across repeated questions. No clear response biases were observed for the majority of children regardless of question format, and accuracy was equivalent on ‘yes-correct’ and ‘no-correct’ questions. Consistency and answers to suggestibility check questions were not predictive of performance. Because multiple mechanisms underlie errors on yes–no questions, the goal of postdicting the accuracy of children’s responses remains elusive.

External and internal sources of variation in the creation of false reports in children.
Bruck, Maggie, . Ceci, Stephen J., and Melnyk, Laura
Learning and Individual Differences, Vol 9(4), 1997. Special Issue: II. Children’s False Memories.
Abstract:
Describes several ‘families’ of variables that may account for reliable variation in children’s suggestibility. Specifically, factors are discussed that are external to the organism (e.g., various forms of biased interviewing such as visualization inductions, accusatory tone, repeated yes/no questioning, repeated interviews, stereotype induction, interviewer status, anatomically detailed dolls) that could explain why at any age studied, large suggestibility effects are produced in some situations but not in others. Research on factors that are internal to the organism that may be at the source of individual differences in suggestibility-proneness (e.g., IQ memory strength, relevant content knowledge, psycho-social factors) are also discussed. Finally, a framework is postulated in which multiple and complex interactions among cognitive, social, personality, and biological factors converge to make some children and some situations more or less suggestible than others. Suggestibility research through the 20th century is also reviewed.

Repetitions, warnings and video: Cognitive and motivational components in preschool children’s suggestibility.
Endres, Johann. and Claudia Erben, Christina
Legal and Criminological Psychology, Vol 4(Part 1), Feb, 1999. pp. 129-146.
Abstract:
Studied the effect of situational factors on the suggestibility of preschool children in 2 experiments using the Bonn Test of Statement Suggestibility as a dependent measure. This test comprises 3 subscales which are made up of different question formats (e.g., misleading yes-no questions, wrongly disjunctive alternative question). The 2 studies tested expectations about different effects of situational and individual factors on these 3 subscales. In Exp 1, a 2 × 2 factorial design with stimulus presentation (single vs repeated reading of the story) and warning (standard condition vs explicit warning of misleading questions) was used with 92 children aged 4–7 yrs. Exp 2 used a 1-factorial design with 3 conditions (standard, warning and video recording) and a non-suggestive follow- up questioning phase; Ss were 60 children aged 5–6 yrs. Both better memory due to repeated presentation and the warning led to a reduction of errors in answers to suggestive questions, with differing patterns of effects for these experimental factors. Video recording did not affect Ss’ performance. In the non-suggestive follow-up interview the Ss were able to correct most, but not all, of their previous errors. The 3 suggestibility scales appear to include cognitive and social-motivational components.

Individual differences in children’s and adults’ suggestibility and false event memory.
Quas, Jodi A., Qin, Jianjian, Schaaf, Jennifer, and Goodman, Gail S.
Learning and Individual Differences, Vol 9(4), 1997. Special Issue: II. Children’s False Memories. pp. 359-390.
Abstract:
Presents an overview of the emerging area of research concerning individual differences in children’s memory, suggestibility, and false event reports. Recent research on children’s false event memories is discussed, as well as research and theory concerning sources of individual differencesin children’s memory and suggestibility, including both cognitive (e.g., understanding of dual representations, source monitoring, imaginativeness, and event knowledge), and social-personality (e.g. attachment styles and temperament, parent-child communication, and sequelae of maltreatment) influences. Implications of these sources for children’sfalse event reports are highlighted. Finally, the authors examine how individual differencefactors proposed to mediate adults’ false memories relate to those that may mediate children’s false memories. Implications for legal professionals are discussed.

5. It is difficult to train children to resist potentially suggestive questions or to “gate-out” previously suggested information

Given the previously documented difficulties in identifying children who are  especially susceptible to suggestive questioning, one of the next logical steps is  to teach all children to resist suggestions that they may encounter in the future  or to separate prior suggestions from original memories of experienced events. A good deal of research has been dedicated to inducing resistance to suggestion in children, but less attention has been paid to undoing the damage of suggestions already provided. One of the earliest studies in which researchers attempted to buffer children against suggestive questioning presented a story to seven-year-olds, twelve-year-olds, and college students, and asked them to recall it. Half of the participants were then told that they were about to be asked questions, some of which were potentially difficult or tricky, and that they should “only answer with what they really remembered. This warning was equally effective in reducing suggestibility across the age range, but the reduction, in comparison to the control group, was small at approximately five percent.

Other recent studies have utilized warnings or preinterview training to reduce children’s suggestibility. In one study, a story was presented to four-through seven-year-old children. The story was followed by free recall and then, for half of the children, by an explicit warning about tricky questions. This warning included a specific example of a question to which the children could not know the answer because the information was not provided in the story, along with the correct answer-“I don’t know.” This explicit warning reduced the children’s suggestibility by about twelve percent.

Somewhat older children were shown to profit from pre-interview training in a series of two studies. In the first, children between the ages of nine and thirteen went on a school field trip to an interactive science center. Approximately a week later, children were interviewed about the trip, and some of the  interview questions were misleading. Before conducting the interviews, half of the children were trained to resist suggestive questions.'” The training included instructions not to guess or make up answers, to respond “I don’t know” or to correct the interviewer’s mistakes when appropriate, and practice with several misleading questions and feedback on answers. Pre-interview training reduced suggestibility by thirteen percent. It also, however, reduced correct answers to nonleading questions by increasing “don’t know” responses. A second study was conducted in which the pre-interview training was revised to emphasize and practice providing correct answers when known. The revision appeared to be successful in reducing suggestibility by eighteen percent without reducing correct answers for the nine- to eleven-year-olds who were trained prior to the interviews.

The most successful training program for reducing suggestibility was developed by Karen Saywitz and Susan Moan-Hardie. Seven-year-old children witnessed an event in their classrooms. Two weeks later, half of the children were given extensive “resistance” training, whereas the remaining children were encouraged to try hard and to do their best. The resistance training involved several components above and beyond those used in the training study described earlier. For example, children were engaged in discussions about reasons a child might go along with an adult interviewer’s false suggestion and were taught new response strategies such as stopping and thinking before answering. Other response strategies included using self-statements’ to promote confidence in their own memories and to suppress inappropriate responses. They also practiced their new skills by watching a video and being interviewed about it afterwards; feedback was provided for their responses to each question. A review session was conducted the following day, prior to the target interview. The reduction in suggestibility due to training, compared to the control group, was twenty-six percent.’ Unfortunately, as in the prior study, there was a corresponding reduction in correct responses to correctly leading questions. Thus, a second study was conducted with revised training procedures, placing less emphasis on “don’t know” responses and providing more practice and reinforcement in reporting the correct answer when known. ‘ The revised training package resulted in a significant reduction in suggestibility without decreasing accuracy on correctly leading questions.

Despite these encouraging results, it would be premature to declare preinterview training a panacea for the problem of children’s suggestibility for several reasons. First, none of these training packages has been attempted, let alone succeeded, in children younger than seven years old, the age group that presents the biggest challenge. Second, although suggestibility was reduced, it was by no means eliminated. Third, these training packages must be carefully devised in order to ensure their effectiveness in reducing suggestibility without also decreasing accurate and complete reporting to nonleading questions. It remains to be determined whether it would be possible or feasible for practitioners to conduct these kinds of extensive and carefully crafted training interventions prior to investigative interviews. Finally, these sorts of interventions have been designed to inoculate children against future suggestions, but children may have already encountered suggestions prior to their first formal forensic interview. Thus, recent studies have begun to investigate the possibility of helping children to separate experienced from suggested information.

One such study was reported by Debra Poole and Stephen Lindsay. In a follow-up to their Mr. Science study, 67 children experienced an event and were later given misinformation about it from a book their parents read to them. Prior to a final interview, children were trained to “source-monitor,” that is, to separate what they had experienced (personal experience source) from what they had only heard about (other source). This training was successful with six- to eight-year-old children, in that they now correctly attributed some of the false information they had previously reported to the book they had heard rather than personal experience. However, there was no evidence that the training was effective in reducing reports of nonexperienced events in three- to  five-year-olds.

A different technique for reducing the effects of previously suggested false information was studied by Mary Lyn Huffman and colleagues. Children between the ages of four and six years participated in an event at their preschools with a visitor named Sam Stone and were interviewed a week later. Some of the interview questions involved nonexperienced events whereas others embedded misleading or correctly leading information about actually experienced events. Two days following this interview, children were reinterviewed with specific but nonleading questions. These final interviews began in one of three different ways. One-third of the children were first asked three questions about the difference between truth and lying. Another third were askedmany more elaborate questions about truth and lying and were asked to judge three vignettes about other children telling the truth or lying. The final group of children did not discuss truth or lying before the questioning began. All children were asked to tell the truth about Sam Stone’s visit. Children who had been asked the lengthiest series of questions about truth and lying were significantly more accurate in their responses to the final neutral interview than were the other two groups of children. A follow-up study by one of the researchers, however, failed to replicate this effect. 

Another method for reducing contamination from prior suggestions was investigated by two research groups with different outcomes. In both studies, either the Cognitive Interview or a standard interview was conducted after exposure to misleading questions. In one study, eight- to nine-year-old children witnessed a videotaped event and were then asked misleading questions.

Next, they were reinterviewed with either the cognitive or standard interview, and then another set of misleading questions was administered. Relative to a standard interview, the cognitive interview increased accuracy and decreased false answers to subsequent misleading questions. However, one-third of all children reported false information even in free recall, and suggestibility appeared to increase with the second set of misleading questions. Thus, the cognitive interview did not reduce suggestibility in any absolute way, but only relative to the standard interview. In a second study examining the potential of  the cognitive interview for reducing children’s suggestibility, five- to seven-year- old and nine- to eleven-year-old children were first exposed to a videotaped event and then to misleading information about it, followed three days later by either the cognitive or standard interview. The cognitive interview did not prove to be superior to a standard interview in terms of children’s tendency to intrude suggested answers in free recall and answers to direct questions.

Finally, Michelle Leichtman and Stephen Ceci “challenged” children who had reported suggested information with “Did you really see that or just hear about it?” and a countersuggestion, “You didn’t really see him do that, did you?” In their most suggestive condition involving repeated misleading questioning plus the induction of a negative stereotype, initially seventy-two percent of the children responded falsely to a direct question about Sam Stone’s behavior. With the first challenge, this rate dropped to forty-four percent, and with the countersuggestion it fell further, to twenty-one percent.  Thus, these challenges were effective in eliminating two-thirds of the false allegations.

A growing literature is documenting a variety of techniques to reduce the  impact of suggestions once they have already occurred. These research efforts are in their infancy, however, and the successes appear inconsistent across studies and age groups. Moreover, as with the techniques designed to increase resistance to suggestions before they occur, these techniques do not eliminate suggestibility, and they may be impractical or unethical for use in the field. For example, if a child reported that someone had inserted a stick into her vagina, it would be completely inappropriate and probably counterproductive to obtaining any further truthful disclosures to say, “That didn’t really happen, did it?”

Abstracts

Memory-enhancing techniques for investigative interviewing: The cognitive interview.
Fisher, Ronald P., and Geiselman, R. Edward
Springfield, IL, England: Charles C Thomas, Publisher; 1992. xi, 220 pp.
Abstract:
Despite the obvious importance of eyewitness information in criminal investigation, police receive surprisingly little instruction on how to conduct an effective interview with a cooperative eyewitness (Sanders, 1986). . . . Reflecting this lack of formal training, police often maintain a less-than-rigorous attitude toward this phase of investigation. . . . It is not surprising, therefore, that police investigators often make avoidable mistakes when conducting a friendly interview and fail to elicit potentially valuable information. The intent of this book is to provide the police interviewer (INT) or any other investigative INT with a systematic approach so that he can elicit the maximum amount of relevant information from cooperative eyewitnesses (E/Ws).

The language of this book is couched in terms of police investigations, primarily because our research has been conducted with police participants. However, since the Cognitive Interview is based on general principles of cognition, it should be useful to anyone conducting an investigative interview, whether a police detective, fire marshal, state-, defense-, or civil attorney, private investigator, etc.

The Cognitive Interview has evolved over the past several years and reflects a multidisciplinary approach. We have relied heavily on the theoretical, laboratory research in cognitive psychology (hence the name ‘Cognitive Interview’) that we and other psychologists have conducted over the past thirty years. . . . The Cognitive Interview . . . is an eclectic approach, making use of ideas found across a variety of people, research approaches and disciplines.

‘What colour is your pet dinosaur?’ The impact of pre-interview training and question type on children’s answers.
Gee, Susan, Gregory, Marian, and Pipe, Margaret-Ellen
Legal and Criminological Psychology, Vol 4(Part 1), Feb, 1999. pp. 111-128.
Abstract:
Evaluated the effects of question type and of brief pre-interview training, involving instructions and practice, on the number of correct answers and errors given by children in a structured interview. 157 children aged 9–13 yrs were interviewed about a visit to a science center with both misleading and non-misleading open and closed questions. The children also rated their confidence in each of their answers. Half the children received pre-interview training designed to discourage compliance and guessing. In Study 1, pre- interview training decreased commission errors to misleading questions, but also decreased the number of correct responses to non-misleading questions. In Study 2, a revised training package decreased errors for misleading questions without impacting on correct responses. Brief pre-interview interventions can reduce children’s compliance with misleading questions in experimental situations. Both studies provided some support for the cognitive processing hypothesis that the confidence–accuracy relationship will be stronger for open than for closed questions. 

Cognitive interviewing procedures and suggestibility in children’s recall.
Hayes, Brett K., and Delamothe, Katrina
Journal of Applied Psychology, Vol 82(4), Aug, 1997. pp. 562-577.
Abstract:
In this study the authors examine the effects of procedures adapted from the cognitive interview of R. E. Geiselman, R. P. Fisher, D. P. MacKinnon, and H. L. Holland (1985) on children’s recall following exposure to misleading suggestions. Children aged 5–7 years and 9–11 years saw a videotaped story and were presented with misleading or neutral information concerning story details. All were later given free- and cued-recall tests preceded by standard interview instructions or instructions that reinstated the encoding context and encouraged exhaustive reporting. Increased recall accuracy was found following cognitive interview instructions. Both age groups were susceptible to misleading suggestions, but susceptibility was unaffected by interview type. The authors discuss the implications for interviewing child witnesses. 

Discussing truth and lies in interviews with children: Whether, why, and how?
Huffman, Mary Lyn, Warren, Amye R., and Larson, Susan M.
Applied Developmental Science, Vol 3(1), 1999. Special Issue: New research on child witnesses:
Abstract:
Two studies examined discussions of truth and lying during interviews with children. In Study 1, truth–lie discussions (TLDs) during 132 actual sexual abuse interviews were analyzed, focusing on the types of questions asked and their developmental appropriateness. TLDs, which were fairly common for all ages of children interviewed (ages 1 yr 10 mo to 14 yrs 8 mo), typically involved asking children closed-ended questions and did not differ in quality or form by age of child interviewed. Study 2 compared the typical TLDs (found in Study 1) to either no discussion or a more elaborate discussion in their effects on 67 preschoolers’ (aged 4 yrs 6 mo to 6 yrs 9 mo) reports of an interactive event. Children given the extended TLD were significantly more accurate than those questioned following a typical or no TLD. The results suggest that discussing truth and lying with young children is effective only if the discussion is more elaborate than those typically conducted in forensic interviews.

Reducing the potential for distortion of childhood memories.
Saywitz, Karen J., and Moan-Hardie, Susan
Consciousness and Cognition: An International Journal, Vol 3(3-4), Sep-Dec, 1994. Special Issue: The recovered memory/false memory debate. pp. 408-425.
Abstract:
Conducted 2 studies involving 102 2nd graders to test the efficacy of an innovative procedure (IP) designed to reduce distortion and enhance communication of accurate childhood memories. In both studies, Ss participated in a staged activity and were randomly assigned to either an intervention or control treatment condition. Two weeks later, 50% of Ss participated in an IP designed to increase resistance to misleading questions, while the other 50% participated in control sessions and were given motivating instructions. Memory for the staged activity was tested in an interview with an unfamiliar authority figure. Ss who participated in the IP made significantly fewer errors in response to misleading questions than did Ss in the control groups, thus showing diminished acquiescence. Results were accomplished without generating additional errors on the other question types. 

Inducing resistance to suggestibility in children.
Warren, Amye, Hulse-Trotter, Katherine, and Tubbs, Ernest C.
Law and Human Behavior, Vol 15(3), Jun, 1991. pp. 273-285.
Abstract:
30 7-yr-olds, 30 12-yr-olds, and 39 adults (mean age 30 yrs) were administered G. H. Gudjonsson’s suggestibility scale, which consists of a story followed by 20 questions, 15 of which are misleading. After Ss were told that their answers were not all correct, the questions were readministered to look for shifts. Approximately half of Ss in each age group had been warned that the questions were difficult or tricky and that they should only answer with what they confidently remembered. Younger children recalled less of the story and were more likely to acquiesce to leading questions than older children and adults. Children also changed more of their answers on the 2nd questioning. Recall was negatively correlated with both acquiescence to leading questions and likelihood of changing answers, even within age groups. Warnings significantly reduced the effect of misleading questions across all age groups. 

It is difficult to train interviewers to avoid suggestive techniques and to use techniques designed to promote accuracy

Given the difficulties of identifying particularly suggestible children and of training children to resist suggestive influences, it is important for interviewers to avoid the use of suggestive techniques. Although this seems obvious, the interviewers studied in many different countries tend to over-use closed-ended, specific, and potentially leading questions and other “risky” practices. For example, one study examined seventy-two interviews conducted by experienced interviewers in Sweden. Despite universal recommendations to begin interviews with general, open-ended or “invitational” questions that promote fairly spontaneous, narrative responses, thirty-five of these interviews (forty-nine percent) began with a suggestive question. Throughout the interviews, the interviewers relied on suggestive and “option-posing” (forced-choice) questions, which accounted for fifty-three percent of the interviewers’ utterances, and elicited fifty-seven percent of the information from children. Similarly, fifty- three percent of the utterances of a comparison sample of United States interviewers and thirty-five percent of Israeli interviewers’ utterances were suggestive or option-posing. Only six percent of the Swedish interviewers’ utterances were invitational. Corresponding figures for Israeli and United States interviewers were similar, at two percent and five percent, indicating remarkable consistency across cultures with different interview selection and training procedures.

Other studies of United States interviewers have documented similar problems. An analysis of forty-two United States sexual abuse interviews found that general, open-ended questions account for ten percent or fewer of all interviewer questions, and that specific, yes-or-no-format questions account for two-thirds of all questions. In addition, interviewers sometimes (twenty-nine percent of the time) completely fail to establish rapport and often (seventy-one percent of the time) fail to establish interview ground rules by telling children that they should feel free to correct the interviewers and to answer that they do not remember or do not understand questions.

Can interviewers’ performance be improved such that they reduce their reliance on risky techniques and increase their use of recommended techniques?  Several recent training evaluation programs suggest that this is more difficult than it seems.

Jan Aldridge and Sandra Cameron examined the effectiveness of a week- long intensive interviewer training program in the United Kingdom. No differences between trained and untrained interviewers were observed in terms of their use of open, specific, and leading questions. Moreover, there were very few requests for free reports, such as invitational, in either group, and over half of the questions in both groups were specific and leading.

In a similar evaluation study, the effectiveness of a ten-day interviewer training institute was examined by comparing pre- to post-training videotaped interviews with children about previously experienced staged events. Analyses of the use of open-ended, specific, and leading and misleading questions indicated no significant changes due to training. On a more encouraging note, interviewers were significantly more likely to establish the ground rules in the interviews they conducted after training (forty-five percent) than before training (five percent). Additionally, interviewers’ knowledge of the relevant scientific literature increased due to training, even though their new knowledge did not appear to translate directly into new interview practices.

In England, Graham Davies, Helen Westcott, and Noreen Horan assessed  interviews conducted by interviewers trained in techniques outlined by the Memorandum of Good Practice. Thirty-six interviews were examined, most of which had been conducted in the mid- to late 1990s, during which time training was more widely available than before. Despite the fact that all of the interviewers had been trained in Memorandum-recommended practices, only two percent of their questions were open-ended. Again, some encouraging results were found as well, in that only three percent of these interviewers’ utterances were classified as leading.

The most encouraging findings to date on interviewer training effectiveness were reported by Yael Orbach and her associates. Their training program differed in many critical ways from those evaluated in the studies discussed previously. First, they trained interviewers to use a very structured protocol, similar to a “script.” Thus, rather than being encouraged to use open-ended questioning, interviewers were given sample open-ended questions to ask and the order in which they should be posed. Second, the initial intensive training was followed with monthly individual sessions and group discussions that included utterance-by-utterance reviews of videotaped interviews. Interviews conducted by trained interviewers using the protocol were compared to non-protocol interviews conducted previously by the same interviewers. The protocol interviews were significantly improved in terms of the percentage of open-ended prompts, as well as how long interviewers waited before asking their first potentially suggestive (option-posing) question. Protocol interviews also yielded more information from open-ended prompts and less from more specific, focused questions than did non-protocol interviews. There were, however, no differences between protocol and non-protocol interviews in terms of the total amount of abuse-relevant information elicited.

Extant research thus demonstrates that interviewer training is effective in reducing problematic questioning techniques only when training is both intensive and extensive, and only when it includes practice, individualized feedback, and follow-up.  Across the United States, interviewer guidelines vary dramatically, and there is little documentation of current interview practices.   Improving the quality of investigative interviews with children will, therefore, require greater commitment of and cooperation between researchers and practitioners, and far greater investments of effort and resources on the part of the research and legal communities alike.

Abstracts

Investigative interviews of child witnesses in Sweden.
Cederborg, Ann-Christin, et al.
Child Abuse & Neglect, Vol 24(10), Oct, 2000. pp. 1355-1361.
Abstract:
Examined the structure and informativeness of interviews with 4- to 13-yr-old alleged victims of sexual abuse in Sweden. 72 alleged victims of sexual abuse were interviewed by 6 experienced officers from 1 police district in Sweden. Evaluation focused on the structure of the interviews, the distribution and timing of the investigators’ utterance types, and the quantity and quality of the information provided by the children. Content analysis revealed that the interviewers relied primarily on option-posing and suggestive questions—together, these comprised 53% of their utterances—when interviewing the alleged victims. As a result, most of the details (57%) obtained from the children were elicited by option-posing and suggestive utterances. Only 6% of the interviewers’ utterances were open-ended invitations, and these elicited only 8% of the information obtained. 

Interviewing child witnesses: Questioning techniques and the role of training.
Aldridge, Jan and Cameron, Sandra
Applied Developmental Science, Vol 3(2), 1999. Special Issue: New research on child witnesses:
Abstract:
Evaluated the effect of a 1-wk intensive training course on police and social worker forensic interviewing with children (mean age 6.2 yrs) and investigated the actual types of questions employed by interviewers. Analysis of 19 videotaped interviews was used to compare trained and untrained interviewers on a series of rating scales designed to assess interviewer performance. The number of requests for free reports and the number of open, specific, leading, and nonleading questions used were obtained. No differences were found in performance between trained and untrained interviewers on any rated behaviors with both trained and untrained interviewers rating poorly. Specific and leading questions were found to occupy over half the total number of questions used by both sets of interviewers, and few free report requests were used. That is, interviewers mostly asked the types of questions least likely to obtain valid and reliable information from children, with no evident variation from this pattern within the trained group. It is concluded that these findings suggest that interviewers use inappropriate questioning strategies to obtain information even after training and rely heavily on specific rather than open questioning.

The impact of questioning style on the content of investigative interviews with suspected child sexual abuse victims.
Davies, Graham M., et al.
Psychology, Crime & Law, Vol 6(2), 2000. pp. 81-97.
Abstract:
Explored the influence of question type and interviewer style on the quantity and quality of responses offered by children in interviews for suspected sexual abuse. The analysis covered 36 investigative interviews conducted by police officers in England under the Memorandum of Good Practice with children aged 4–7, 8–11 and 12–14 yrs. The dependent measures were the temporal length of children’s answers and the number of criteria derived from Criteria Based Content Analysis (CBCA) they contained. Only 2% of all questions were open-ended, but just 3% were judged leading. Open questions were most effective with 12–14 yr olds but not with younger children who provided more information in response to specific yet not leading or closed questions. Longer answers containing more CBCA criteria were associated with interviews containing many affirmative utterances and verbal affirmations and a brief rapport. The implications of these findings for interview theory and practice are briefly discussed. 

Assessing the value of structured protocolsfor forensic interviews of alleged child abuse victims.
Orbach, Yael, Hershkowitz, Irit, Lamb, Michael E., Esplin, Phillip W., Horowitz, Dvora
Child Abuse & Neglect, Vol 24(6), Jun, 2000. pp. 733-752.
Abstract:
Examined the effectiveness of the National Institute of Child Health and Human Development (NICHD) investigative protocol, a flexibly structured protocol incorporating a wide range of strategies believed to enhance retrieval with child witnesses. Six forensic investigators were trained to use the NICHD protocol while conducting feedback-monitored simulation interviews. The protocol’s was evaluated by comparing 55 protocol interviews (PRIs) with 50 prior interviews by the same investigators, matched with respect to characteristics likely to affect the richness of the children’s accounts. The comparison was based on analysis of investigators’ utterance types, distribution, and timing, and quantitative and qualitative characteristics of information produced. PRIs contained more open-ended prompts than non-PRIs did. More details were obtained using open-ended invitations and fewer were obtained using focused questions in PRIs than in non-PRIs, although total number of details elicited did not differ significantly. In both conditions, older children provided more details than younger children did. 

Assessing the effectiveness of a training program for interviewing child witnesses.
Warren, Amye R., et al.
Applied Developmental Science, Vol 3(2), 1999. Special Issue: New research on child witnesses: Part II. pp. 128-135.
Abstract:
Assessed the effectiveness of a training program for those who conduct investigative interviews with young children. 27 18–53 yr old experienced interviewers attended a 10-day training institute designed to provide knowledge and skills for improving investigative interviews with young children. Ss completed pre- and posttraining surveys assessing their knowledge of the scientific evidence regarding memory, suggestibility, and other aspects of children’s ability to provide accurate accounts of events during interviews. They also conducted pre- and posttraining interviews with preschool children about 2 previously experienced events. Ss’ knowledge about children’s abilities and the scientific basis of various interviewing protocols increased significantly after the training. However, training did not have a significant impact on interviewers’ questioning styles or the amount of accurate information elicited from the children. Results indicate that successfully translating knowledge into practice requires multiple opportunities for skill practice and feedback.

Various Studies

The Danger of Value-Laden Investigation in Child Sexual Abuse Cases: Are Defendants’ Constitutional Rights Violated When Mental Health Professionals Offer Testimony Based on Children’s Hearsay Statements and Behaviors,”
DeSarbo, Lynne Celander. “
University of Pennsylvania Journal of Constitutional Law vol. 2, no. 1 (December 1999): p. 276-317.
Introduction
In criminal cases involving child sexual abuse, an analysis of the extent to which mental health professionals (“MHPs”) should be able to introduce hearsay statements  generated from investigative or therapeutic interviews with  children depends upon a critical examination of the Constitution’s Confrontation Clause and various hearsay exceptions. The Confrontation Clause, as well as the Federal  Rules of Evidence, restrict the introduction of hearsay testimony. Despite this general distrust of hearsay statements and the particular difficulties associated with child witnesses, child hearsay testimony may be the best or only evidence available to a prosecutor in a criminal sexual abuse  case. As a result, MHPs are frequently called upon to testify to children’s hearsay statements and behaviors based on  their interview with those children.

Tell me again and again: Stability and change in the repeated testimonies of children and adults.
Poole, Debra Ann, and White, Lawrence T.
Memory and testimony in the child witness. Zaragoza, Maria S., (Ed); Graham, John R., (Ed); Hall, Gordon C. N., (Ed); Hirschman, Richard, (Ed); Ben-Porath, Yossef S., (Ed); pp. 24-43; Thousand Oaks, CA, US: Sage Publications, Inc; 1995. xii, 298 pp.
Abstract:
Review studies that have examined the eyewitness performance of children and adults across repeated questions [during interviews] / consider the impact of a cross-session repetition, or multiple interviews / review . . . within-session repetition / [describe] the empirical or theoretical bases for predicting that repetition will enhance or degrade testimony / compare those predictions with findings obtained from studies of eyewitness performance / [discuss] unresolved issues and suggestions for future research

Research on children’s suggestibility:  Implications for the Investigative Interview
Amye R. Warren and Lucy S. Mcgough
23 Crim. Just. & Behavior 269 (1996), 35 pages, 269 to 303
Abstract
The authors review research on children’s suggestibility as it applies to the investigative sexual abuse interview. They focus on identifying the optimal conditions for securing an account of a child’s remembered experience that will be equally or more reliable than that of an adult. The discussion is divided into four major sections, corresponding to the questions of how, when, where, and by whom a child witness should be interviewed to diminish potential distortions and enhance the trustworthiness of the child’s remembered account.