ArticlePDF Available

Factors influencing the reliability and validity of statements made by young victims of sexual maltreatment



Reports of maltreatment involving young children have increased in recent years. In many cases, alleged victims and perpetrators are the only sources of information about the incidents concerned. This has prompted many efforts to evaluate the reliability and validity of information provided by young children, who are widely believed to be handicapped by a lack of linguistic facility, poor memories, susceptibility to suggestion, and a tendency to confuse fantasy and reality. The relevant literature on these issues is summarized in this article, in which we then suggest ways of interviewing children so as to obtain the most reliable and informative accounts of events they have experienced. Considerable emphasis is placed on the needs to: (a) elicit accounts from free recall, (b) use directive or leading questions sparingly and only in specific circumstances, and (c) develop systematic procedures for evaluating children's testimony.
Factors Influencing the Reliability and
Validity of Statements Made by Young
Victims of Sexual Maltreatment
National Institute of Child Health and Human Development
Bethesda, Maryland
Private Practice, Phoenix, Arizona
Reports of moltreotment involving young children hove increased in recent yeors. In
many cases, alleged victims and perpetrators ore the only sources of information
about the incidents concerned. This has prompted many efforts to evaluate the
reliability and validity of informotion provided by young children, who ore widely
believed to be handicapped by o lock of linguistic facility, poor memories, suscep-
tibility to suggestion, ond a tendency to confuse fontosy and reality. The relevant
literature on these issues is summorized in this article, in which we then suggest ways
of interviewing children so as to obtain the most reliable ond informotive accounts of
events they have experienced. Considerable emphasis is placed on the needs to: (0)
elicit accounts from free recall, (b) use directive or leading questions sparingly and
only in specific circumstances, and (c) develop systematic procedures for evaluating
children’s testimony.
Interest in children’s testimony has increased dramatically in the past decade
(see, e.g., recent books by Ceci, Leichtman, & Putnick, 1992; Ceci, Ross, &
Toglia, 1989; Ceci, Toglia, & Ross, 1987; Dent & Flin, 1992; Doris, 1991; Perry
& Wrightsman, 1991; Spencer & Flirt, 1990). In part, this interest reflects rapid
increases in the number of reports concerning child maltreatment in many coun-
tries, including the United Kingdom and the United States (American Associa-
tion for Protecting Children, 1986; National Society for the Prevention of Cruelty
to Children, 1989). Much of the research has been interpreted in relation to the
investigation of child sexual abuse-a unique crime because the alleged victims
and perpetrators are often the only available witnesses. Alleged perpetrators may
well be motivated to misrepresent their behavior, particularly in jurisdictions
We are grateful to several colleagues with whom we have discussed many of the issues raised in
this article, especially Frank G. Bolton, Jr., Tascha D. Boychuk, Steven W. Horowitz, David C.
Raskin, and John C. Yuille.
Correspondence and requests for reprints should be sent to Michael E. Lamb, Section on Social
and Emotional Development, 9190 Rockville Pike, Bethesda, MD 20814.
where convictions carry serious penalties, and this has increased the importance
of obtaining and evaluating information provided by children. The same severe
penalties increase the temptation to impugn the credibility of young victims. In
addition, many scholars have suggested that false allegations of sexual abuse
might constitute powerful weapons in the hands of malicious adults, including
those engaged in custody disputes (Green, 199 1; Jones & McGraw, 1987; Jones
& Seig, 1988; Thoennes & Pearson, 1987).
Jurists and legal scholars have increasingly turned to forensic, clinical, and
developmental psychologists and other social scientists for insights into the com-
petency and credibility of young witnesses. Unfortunately, courtrooms in the
United States have become the setting for heated debates in which experts selec-
tively emphasize findings which ostensibly document children’s deficiencies or
capabilities. The adversarial nature of the context in which such findings are
discussed appears to have made many researchers feel obliged to “side” with one
or other of the “camps,” and this tendency may have hindered advances in our
understanding of children’s competence and credibility. In addition, researchers
have tended to focus on issues that have traditionally been of interest to police
investigators and/or experimental psychologists, but may have questionable rele-
vance to the sexual abuse cases to which their results are later generalized. For
example, many studies of children’s competence include identification line-ups
designed to assess the ability to identify suspects, even though laboratory studies
requiring recognition of photographs reveal poor accuracy at all ages (Chance &
Goldstein, 1984), and the identification of perpetrators is seldom at issue in court
because most suspects are known to their alleged victims (see. e.g., Goodman et
al., 1988). More generally, questions have been raised about the “ecological
validity” of much of the research (Doris, 1991) particularly as the incidents
themselves vary widely in nature (e.g., exhibitionism, fondling, forcible rape)
and context (stressfulness, repetition, uniqueness). Broad variations in the uni-
verse of incidents concerned complicate attempts to design studies whose results
elucidate the competencies and deficiencies of young informants. Our goals in
this paper are to summarize our current understanding of the factors that influ-
ence children’s ability to provide accurate information about events they experi-
enced and then use this knowledge to suggest ways in which the greatest amount
of information can be elicited and evaluated.
Few researchers have studied deliberate deception by children (though see
Ceci, Leightman. & Putnick, 1992), but this is obviously an important issue: Like
many adult victims and witnesses, children who report being victims of sexual
maltreatment may be under pressure from others to make, withdraw or modify
their claims and they may be embarrassed to report participation in distasteful
events. The magnitude of the motivation to deceive certainly varies depending on
the specific circumstances and thus must be evaluated carefully when judging the
credibility of any witness, regardless of his or her age, dependency, and relative
naivete. Later in this article, we discuss some characteristics that appear to
distinguish accounts of events that really happened from accounts of events that
were not experienced. For the most part, however, this article is concerned with the
competence of young witnesses who are motivated to be truthful. It is thus
important to distinguish between competence-the ability to provide information
about one’s past experiences-and crrdibiliry-the accuracy or truthfulness of
those accounts. Competence is a necessary but insufficient condition for credibili-
ty, and most of the empirical research has been concerned with defining the extent
of children’s competencies and/or incompetencies; assessments of credibility are
less common, and these are the focus of our final section.
Much of the debate over the competency of young witnesses has focused on
four topics: fantasy, language, memory, and suggestibility. The goal of the pre-
sent report is to summarize what is currently known about children’s competence
by focusing on the extent to which variations in their capacity to remember
events, express themselves clearly and unambiguously, distinguish fantasy from
reality, and resist suggestion may sometimes affect their competence as infor-
mants. We then describe techniques that interviewers might employ when at-
tempting to elicit accounts from young victims and witnesses of sexual abuse in
ways that maximize the quantity and value of the information obtained by taking
into account existing knowledge of children’s abilities, limitations, and tenden-
cies. In the final substantive section, we introduce a promising procedure for
evaluating the validity of children’s allegations.
During the Middle Ages, children were not considered to be competent witnesses
until they reached their mid-teens (Ceci et al., 1987) but over time the defini-
tion of competence has shifted from one based on chronological age to one based
upon demonstrated proficiency. In most jurisdictions today, children of all ages
can be deemed competent witnesses, but their credibility is often doubted; they
have even been equated with intoxicated and mentally ill adults (American Juris-
prudence, 1976, p. 670). In legal settings, competence is defined by the capacity
to distinguish between reality and fantasy as well as between truth and falsehood.
In practice, decisions and doubts concerning the adequacy of child witnesses
usually focus on the ability to distinguish between fantasy and reality, linguistic
incapacities, the fallibility of children’s memory, and the extent to which children
are suggestible. In this section, we review the literature on each of these issues.
The widely held belief that young children may fantasize about or fabricate
allegations of a sexual nature is frequently used to cast doubt on their testimony.
Johnson and Foley (1984) suggest that this belief stems from a general concep-
tion of the child as an “incompetent human,” anecdotal material, and over-
generalized extrapolations from developmental theories rather than from empiri-
cal studies. In fact, few researchers have addressed the question-When are
children capable of “distinguishing between memories for actually experienced
events and the products of their imaginations and thoughts”‘? (Johnson & Foley,
1984, p. 38).
In general, children over 6 years of age appear similar to adults in their ability
to specify the origin of events, discriminating between events of internal and
external origin (Johnson & Foley, 1984; Lindsay & Johnson, 1987). According to
Johnson and her colleagues, externally derived memories (i.e., those concerning
actual events) are characterized by more detail (particularly spatial, temporal,
and sensory detail) than internally derived memories (i.e., the products of fanta-
sy or imagination) which are more schematic in nature and contain more refer-
ences to cognitive processes (see also Undeutsch, 1982). Because of their
metacognitive skills and experience, however, adults have an advantage over
children when it is difficult to differentiate between internally and externally
derived memories. For example, an adult who vividly “remembers” a purple
gorilla will have less difficulty than a child would have in determining whether
the memory pertains to a dream or a real event because s/he “knows” that purple
gorillas do not exist.
Although children appear to be less accurate than adults when asked to dis-
criminate between actions they committed and actions they imagined committing
(Foley, Johnson, & Raye, 1983), we know of no research evaluating children’s
ability to distinguish between memories of another person’s actions and fanta-
sies/“imaginations” of what another person did. Studies addressing this question
would be of greatest relevance to accounts of sexual maltreatment, for which it
may be necessary to determine whether children can distinguish between memo-
ries of an action committed by another person and fantasies about an action
committed by another person.
Lawyers and forensic experts commonly confuse fantasy, distorted recollec-
tions, deceit or falsehood, and instances in which children are the unwitting or
witting tools of manipulative or vengeful adults attempting to fabricate allega-
tions using children as accomplices. Fabrication and false reporting are unrelated
to the ability to discriminate between imagined and real events; their occurrence
speaks to credibility rather than competence.
It is, in fact, unusual for children to invent or fabricate the central details of
sexual offenses (Goodman & Hegelson, 198.5; Undeutsch, 1982). When fabrica-
tion is believed to have occurred, the interviewers often appear to have behaved in
ways that make children more susceptible to adult influence (Goodman &
Hegelson, 1985). Interviewers should thus be aware that fabrications may occur
but they should not confuse them with an (in)ability to distinguish reality from
Language and Communicative Abilities
In his novel, Double Negative, David Carkeet (1980) described an intriguing
case in which a preverbal child provided crucial testimony concerning a murder
he witnessed. Unfortunately, the appraisal of such testimony is still the province
of novelists rather than investigators, but an increasing number of interviewers
have been forced to obtain information from inarticulate young children. The
vocabularies of young children are often much more limited and less descriptive
than those of adults. Adjectival and adverbial modifiers are especially likely to be
absent in their accounts, which tend to be extremely brief and sparse (Marin,
Holmes, Guth, & Kovac, 1979). Individual differences are large and develop-
mental change rapid, however, and it is thus valuable for investigators to use
samples of children’s conversations to judge children’s abilities. The challenge
confronting investigators is to obtain accounts that are sufficiently rich in de-
scriptive detail to permit an understanding of the children’s testimony. The more
impoverished the child’s language, the greater the likelihood that his/her state-
ments will be misinterpreted or that the child will misinterpret the interviewer’s
questions and purposes (King & Yuille, 1987; Perry & Wrightsman, 1991; Walk-
er, 1993).
Young children’s accounts are also brief because they cannot draw upon an
array of past experiences that provide a network of associations to enrich their
descriptive accounts with analogies or metaphors (Johnson & Foley, 1984). The
same lack of clear associations appears to foster loose associations and digres-
sions when recounting incidents. Interviewers should obviously be aware of
children’s tendency to digress and remain ready to refocus their attention without
being suggestive.
The accuracy of children’s accounts may also be influenced by the linguistic
style of the investigator. In a detailed linguistic analysis of the testimony pro-
vided by a 5-year-old child who might have witnessed the murder of three
children, Walker (1993) highlighted three ways in which interviewing style can
influence children’s apparent communicative competence: (1) by using age-
inappropriate words and expressions, (2) by constructing syntactically complex
sentences (e.g., she noted one 48-word question), and (3) by being ambiguous
(e.g., “Who else if anybody?‘). Perhaps we should pay more attention to the
competence of investigators than the competence of child witnesses! It is cer-
tainly inappropriate to assume that “because the child was giving [interviewers] a
response of some kind, she was giving them an answer to their questions”
(Walker, 1993, p. 67).
Interviewers often implicitly ask children to negate adult statements (e.g., “Is
is not true that . .?“). Because the ability to actively negate a statement pro-
posed by an adult emerges quite late in linguistic development (Pea, 1980) these
types of questions may be problematic. In addition, active voice sentences are
understood sooner than the passive voice sentences that lawyers often employ,
and questions including double negatives (e.g., “Is it not true that your Mom
wasn’t at home?“) are difficult to comprehend (Perry & Wrightsman, 1991). To
minimize confusion, therefore, children should be encouraged to use their own
words when describing or explaining what they have witnessed or experienced
(Dent, 1982; Saywitz, 1988). If aspects of a child’s account are unclear, it is best
to ask her or him to explain it “in another way” rather than ask the child to verify
the interviewer’s “restatement” (Perry & Wrightsman, 199 I ). Although sugges-
tive questions should be avoided (see later on), furthermore, young children are
particularly likely to benefit from attempts to structure their recall using prompts
and cue questions (Dent, 199 1; Johnson & Foley. 1984; Oates & Shrimpton,
1991). Later in this article, we review ways in which direct but non-suggestive
questions can be used to help children retrieve information without contamina-
tion. Finally, investigators should avoid asking questions that children are not
capable of answering. Whether or not the penis went “inside the pee-pee” is a
question that someone unfamiliar with intercourse may not be able to answer
accurately. Children may perceive pressure on the outside of the vagina as
painful and misinterpret this sensation (Katchadourian & Lunde, 1975).
In debates over the competency of young witnesses, deficiencies in the ability to
remember and describe experiences are often alleged. Unfortunately, however,
many studies of children’s memory were conducted under conditions that are
very different from those which characterize the contexts of sexual abuse and
courtroom testimony (Ceci & Bronfenbrenner, 1991; Ceci et al., 1987; Doris,
I99 I ; Ornstein, Larus, & Clubb, 1992) and thus questions about their “ecologi-
cal validity” must be raised. In order to evaluate these findings, it is necessary to
specify which memory capacities are tapped in specific studies (Tulving. 1985)
and how these abilities arc differentially influenced by such factors as age, stress,
and the delay between experience and recall.
Although children clearly cm remember incidents they have experienced, the
relationship between age and memory is influenced by a variety of factors (Say-
witz, 1988). Younger children appear more susceptible than older children and
adults to various memory distortions which increase in magnitude as the extent of
delay between the experienced event and its recall increases (Brainerd & Orn-
stein, 199 1). According to information-processing theorists, the human brain
organizes information about the world in the form of schemata-interconnected
items which tend to be remembered in conjunction with one another (Chi & Ceci.
1987). Within these schemata, familiar and routine events are often organized
into “scripts,” whereas information about familiar people is organized into “ste-
reotypes” (Nelson, 1986; Nelson 6i Gruendel, 198 I; Shank & Abelson, 1977).
Scripts are representations of “averaged” or “typical” events rather than memo-
ries of particular incidents, whereas stereotypes summarize memories and judg-
ments about a person or category of people. As children develop, they formulate
scripts for routine events like brushing their teeth, going to bed, and going to
McDonald’s, and they construct stereotypic representations of the people with
whom they interact. Scripts are useful because they help individuals to focus on
and remember the important features of repetitive events or sequences while
enabling them to ignore less central elements. Scripts and stereotypes have
disadvantages too, however. For example, general knowledge about a class of
events can influence the memory of specific information, as when 5 and 6-year-
olds incorrectly remembered the gender of a character who played a nontradition-
al gender role (Martin & Halverson, 1983) or when 5- and 7-year-olds embel-
lished their restatements of stories with items and events that were part of their
own scripts (McCartney & Nelson, 1981). In addition, whenever events recur
with any regularity, both children and adults tend to blur distinctions among
incidents and establish script memories or stereotypes. This implies that accounts
of complex and recurrent incidents of sexual abuse may be more vulnerable to
distortions than accounts of simple or single events.
Accounts based on script memories are likely to contain fewer distinctive
details than are memories of discrete incidents (Nelson & Gruendel, 198 l), and
the passage of time between experience and recall increases the tendency to rely
on scripts (Myles-Worsley, Cromer, & Dodd, 1986). According to Undeutsch
(1982) accounts of actual experiences differ in both quality and richness of detail
from those not based on experience. In addition, Davidson (1991) found that
children remembered unusual events better than specific events that were congru-
ent with their general or script memories. One goal of the interview procedure
suggested in this paper is to elicit descriptions of specific incidents rather than
composite accounts of multiple incidents.
As children grow older, the length, informativeness, and complexity of their
scripts and recall memories increase, but the basic structure remains the same
(Davies, Tarrant, & Flin, 1989; Nelson & Gruendel, 1981; Saywitz, 1988). In
general, young children tend to provide briefer accounts of their experiences than
do older children and adults, but their accounts are quite accurate (e.g., Good-
man & Reed, 1986; Johnson & Foley, 1984; Marin et al., 1979; Oates &
Shrimpton, 1991). Although errors of omission are more common than errors of
commission among both adults and children (Oates & Shrimpton, 1991; Stew-
ard, 1993), both children and adults may add erroneous information based on
script memory to accounts of specific events (because she usually uses tooth-
paste, for example, a child may remember doing so even when she did not do
so). The tendency to do this generally declines with age (Collins, 1970; Collins
& Wellman, 1982; Collins, Wellman, Keniston, & Westby, 1978), but in a recent
study, Lindberg (199 1) found that sixth graders and college students made more
false assumptions about cheating than third graders did, apparently because the
older subjects had more experience and better established scripts for cheating
than did the younger children. Saywitz (1988) has suggested that script-based
errors can be reduced by pre-interview counseling or instruction, and the inter-
view technique described below is designed to help interviewers elicit incident-
based rather than script-based accounts.
Considerable controversy persists concerning the effects of increased arousal
or stress on the accuracy of children’s memory. Early in the century, Yerkes and
Dodson (1908) reported that moderate levels of stress improved the performance
of mice on complicated tasks, whereas extreme stress overwhelmed the mice and
impeded their performance. Evidence concerning the association between mem-
ory and stress in humans is substantially less definitive, however. Deffenbacher
(1983) concluded that when “forensically-relevant” (i.e., high) levels of stress
were involved, stress is associated with diminished accuracy, but the relevance of
this conclusion to children’s testimony is often disputed (e.g., Goodman, Hirsch-
man, Hepps, & Rudy, 1991). Some of the enduring controversy doubtless re-
flects differences in methodology and type of stress, as we show later.
Over the course of the past few years, researchers have conducted a number of
studies designed to examine children’s abilities to recall anxiety-provoking
events. To simulate some of the characteristics of sexual abuse incidents, re-
searchers have focused on such putatively stressful events as having a stranger
visit a kindergarten classroom, routine dental or pediatric examinations, veni-
puncture, and inoculation. The results have been inconsistent, however: Some
researchers (Goodman, Bottoms, Schwartz-Kenney, & Rudy, 1991; Goodman,
Hirschman, Hepps, & Rudy, 1991; Ochsner & Zaragoza, 1988; Steward, 1993)
believe that stress increases children’s accuracy, whereas other researchers
(Oates & Shrimpton, 199 I ; Ornstein, Gordon, & Larus, 1992; Peters, 1987,
1991; Peters & Hagan, 1989; Vandermaas, 1991) report that arousal either re-
duces accuracy or has no effect. In most of these studies, unfortunately, many
children experienced little stress and a primary dependent measure was identi-
fication or recognition of individuals in a photographic line-up. As mentioned
earlier, this measure may tell us relatively little about the effects of stress on the
ability to recall central elements of events that have been experienced by the
child. Steward (1993), however, employed a variety of procedures to assess
children’s memory, noting that children’s ratings of distress were correlated with
the completeness and accuracy of their descriptions, after a 6-month delay, of
medical examinations they had experienced.
In a series of studies, Pynoos and his colleagues (Pynoos & Eth, 1984; Pynoos
& Nader, 1988a. 1988b) have examined children’s memories of traumatic inci-
dents (such as rape or murder) that they witnessed. Although some errors were
found, the children studied usually reported the observed incidents accurately.
Regardless of variations in the amount of detail recalled, the central elements
were invariably remembered. Interestingly, many of the memory distortions were
related to the child’s physical proximity to the violent events: Distance appeared
to be a protective factor associated with more complete recollection. Researchers
have not yet studied the effects on performance of stress at the time of recall.
Children are certainly more likely to remember personally meaningful and
salient items and events than meaningless items (see Omstein, Larus, & Clubb,
1992, for a review), but this does not necessarily mean that their memory for
incidents of maltreatment is enhanced. First of all, not all incidents of sexual
abuse are painful or traumatic, and thus the potentially facilitative effects of
arousal on the process of encoding information cannot be assumed. Second, the
context in which the child is asked to retrieve information about the experienced
event-during interviews with a child protection service (CPS) worker, a police-
man, an attorney, or a judge--may be stressful experiences regardless of whether
or not the target event was stressful (Goodman, Taub, Jones, England, Port,
Rudy, & Prado, 1992). Third, different types of memory (e.g., recall, recogni-
tion, reconstructive memory) may be tapped at different stages, and these differ-
ent capacities or processes may be differentially affected by arousal levels.
Even when events are remembered, the process of retrieval is itself compli-
cated (Tulving, 1985). In particular, when a considerable amount of time passes
before retrieval, some portions of the memory may be more difficult to access.
Flin and her colleagues (Flin, Boon, Knox, & Bull, 1992) reported that partici-
pants of all ages reported less information 5 months after an “event” than they
had initially reported, and that 6-year-old children reported less information than
did 9-year-old children and adults. Interestingly, the amount of incorrect infor-
mation did not increase over time. Memory is a constructive process, however:
Like adults, children actively work on memory traces in order to retrieve and
organize them. This reconstructive process can obviously be influenced by the
ways in which others question the child, as we discuss in the next subsection.
Thus, when children are repeatedly interviewed, as is often the case when sexual
abuse has been alleged, this is likely not only to consolidate the memory (facili-
tating subsequent recall) but also to shape it (Omstein, Larus, & Clubb, 1992).
Whatever the vagaries and strengths of children’s memories, the competency of
child witnesses is often doubted on the grounds that children are too susceptible
to influence by misleading questions (Ceci & Bruck, 1993). Suggestibility ap-
pears to reflect both memory impairment and the acceptance of misleading
information from an interviewer, and although the issue is most frequently dis-
cussed in relation to child witnesses, suggestibility is, in fact, a much broader
problem. Psychologists like Elizabeth Loftus ( 1979) have demonstrated that
adults are easily manipulated by suggestive questioning, at least when the inci-
dents that they have been asked to remember are of little personal significance.
Research concerning children’s suggestibility has revealed a mixed and con-
fusing picture (Ceci, Ross, & Toglia, 1987b). In a series of studies, Goodman
and her colleagues (Goodman & Aman, 1990; Goodman, Aman, & Hirschman,
1987; Goodman, Bottoms, Schwartz-Kenney, & Rudy, 1991; Goodman, Rudy,
Bottoms, & Aman, 1990; Goodman, Wilson, Hazan, & Reed, 1989) showed
that children as young as 3 to 4 years old were seldom misled by questions such
as “Did he keep his clothes on‘?“, “Did he kiss you?“, and “He took your clothes
off, didn’t he‘?” which suggested actions quite different from those that were
witnessed or experienced. Other results might have been obtained if the actions
had been more ambiguous and the suggestions more plausible, however. As
Steller (1991, p. 107) noted in a sharp critique, the suggestive questions in these
studies “were unrealistic in content and had nothing to do with the event ob-
served by the children prior to the interview.” In other experimental settings, in
fact, preschoolers appear especially susceptible to suggestion (Ceci, Ross, &
Toglia, 1987a, 1987b; King & Yuille, 1987; Ornstein, Gordon, & Larus, 1992;
Toglia, Ceci, & Ross, 1989). In addition, Ceci and his colleagues (1987a, 1987b)
showed that preschoolers were particularly susceptible to a form of suggestion-
postevent contamination-which involves the incorporation into later reports of
incorrect details suggested to them between the time of the incident and the time
of the interview. Age trends in susceptibility among school-aged children are less
clear, with some researchers reporting that suggestibility continues to decline
through the early grades (Cohen & Harnick, 1980; King & Yuille, 1987), where-
as others report children over 8 or 9 years of age to be less suggestible, or no
more suggestible, than adults (Duncan, Whitney, & Kunen, 1982; Marin et al..
1979). However, there is agreement that suggestions are less likely to be effective
when they pertain to central or salient details (Dodd & Bradshaw, 1980; King &
Yuille, 1987) or to appearances rather than activities and the sequence of events
(Dent & Stephenson, 1979). Unfortunately, little research has been conducted on
suggestibility regarding memories of incidents that traumatized or affected indi-
viduals profoundly. In a small laboratory study evaluating children’s memories of
inoculations, Goodman, Hirschman, Hepps, and Rudy (1991) found that chil-
dren who were more distressed were less suggestible than children who appeared
less stressed by the inoculations.
In theory at least, susceptibility to misleading suggestions should vary de-
pending on the child’s motivation to be completely accurate and/or comply with
the interviewer’s implicit or explicit agenda (King & Yuille, 1987). Children feel
obliged to answer adults’ questions no matter how bizarre (Hughes & Grieve,
1980) and they may assume that the repetition of a question implies that the
initial answer was unsatisfactory (Ceci & Bruck, 1993). Several recent studies
have explored how interviewer style affects children’s suggestibility. Goodman,
Bottoms, Schwartz-Kenney, and Rudy (1991) reported that 3- to 7-year-olds
were equally resistant to suggestions by “nice” and more neutral interviewers.
Goodman et al. ( 1989) reported that 7- and IO-year-old children were sur-
prisingly likely to accept suggestions made “in an atmosphere of accusation” 4
years after the event being recalled (Goodman & Clarke-Stewart, 1991). Ceci et
al. ( 1987a, 1987b) reported that preschoolers were more likely to accept sugges-
tions made by an adult than by a 7-year-old confederate.
Regardless of the resolution of the various controversies concerning children’s
suggestibility, most researchers agree that the manner in which children are
questioned can have profound implications for what is “remembered,” and this
increases the importance of careful interviews (Brainerd & Omstein, 1991; Foley
& Johnson, 1985; Saywitz, 1988). Misleading or suggestive questioning can
manipulate both young and old witnesses. Such questions are most likely to be
influential when the memory is not rich or recent, when the questions themselves
are so complicated that the witness is confused, and when the interviewer ap-
pears to have such authority or status that the witness feels compelled to accept
his or her implied construction of the events.
Fantasy, memory strategies and deficiencies, suggestibility, and communication
abilities importantly affect the accounts provided by young children of their
experiences. Many important questions remain unanswered, but some conclu-
sions can be offered with a degree of certainty. Not surprisingly, children can
remember important details of incidents that they have observed or experienced.
And although their accounts can be manipulated in certain circumstances, sensi-
tive interviewers who are aware of children’s capacities and deficiencies can
avoid many of the problems posed by questions that force children to operate at
or beyond the limits of their capacities.
Linguistic and memorial difficulties do not make children incompetent wit-
nesses, but an understanding of their capacities and limitations should influence
the ways in which children are interviewed and the ways in which their accounts
are interpreted. Both children and adults can be informative witnesses, despite
general and age-related variations in memory, suggestibility, and linguistic abili-
ty. In the next section, we suggest how children can be interviewed so as to
maximize the amount and quality of information obtained from them.
The increased awareness of sexual abuse has fostered the development of a
variety of techniques for interviewing children (e.g., Boat & Everson, 1988;
Jones & McQuiston, 1988; Myers, 1990). There are great variations across
jurisdictions regarding which professionals are mandated to interview children
and which interview techniques are employed. Unfortunately, interviewers are
often thrust into investigative roles without adequate training or knowledge of
children’s abilities, tendencies, and limitations, and many errors are commonly
made. Most scholars agree that the way children are interviewed can have a
profound impact on the type of information obtained (Brainerd & Ornstein,
1991; Ceci, 1991; Foley & Johnson, 1985; Perry & Wrightsman, 1991; Walker,
1993). The purpose of this section is to introduce and suggest ways in which
popular interviewing techniques might be revised to accommodate the expanding
knowledge of children’s memory, language, fantasy, and suggestibility as re-
viewed earlier in this paper.
Investigative interviews (as distinct from therapeutic interviews) are designed
to obtain as much information as possible from children and formulate hypothe-
ses about the alleged incidents. To these ends, interviewers must motivate chil-
dren to be as informative as possible while framing their questions in ways that
do not encourage children to fantasize, embellish details, or claim to remember
things that either did not occur or have been forgotten (see Dent, 1991; Warren,
Hulse-Trotter, & Tubles, 1991). The goal of the therapeutic session is usually
quite different, focusing less on accurate recall of the facts and more on the
impact of events. Because the goals of investigation and therapy can conflict
with one another, it has been suggested that the roles of therapist and investigator
should be separated and that the investigation should be completed prior to
beginning therapy (Raskin & Yuille, 1989). Otherwise, interviewers run the risk
of contaminating the child’s recall with therapeutic techniques. Sometimes, how-
ever, allegations of sexual abuse arise in the context of therapy, and the therapist
is thus forced to investigate allegations of sexual abuse.
Children are most likely to provide accurate information when they are placed
in a comfortable situation in which they: (a) do not feel pressure to satisfy the
expectations of an adult interviewer (Perry & Wrightsman, 1991), (b) are al-
lowed to describe incidents with minimal suggestive questioning by interview-
ers (Dent, 1982; Spencer & Flirt, 1990). and (c) can devote their full attention
without being distracted by toys, dolls, and other props (Raskin & Esplin,
The interview procedure recommended here emphasizes reliance on verbal
recall rather than nonverbal reenaction. We are aware that nonverbal means of
communication may be very informative (e.g., Steward, 1994), but we are also
impressed by the problems of interpretation, suggestion, and fantasy associated
with extensive reliance on nonverbal cues and props. Props have been shown to
stimulate script-based rather than incident-based accounts (Steward, I994), thus
reducing the accuracy of the information provided by the child. Sometimes,
however. children (especially young children) have difficulty discussing sexual
acts. If attempts to engage the child in verbal recollection fail, it may be neces-
sary to use nonsuggestive props like puppets, dolls, and/or a doll house (Raskin
& Yuille, 1989). Although the introduction of these props may distract the child
and encourage script- rather than incident-based memory, they may be necessary
if the child is reluctant to provide information.
Although most professionals recommend that interviews with alleged sexual
abuse victims be recorded on audio- or videotape, there are some who are
opposed to this. Among the advantages cited are: the value of having an exact
record of the child’s and the interviewers’ statements, thereby allowing inter-
viewers to pay closer attention to the child’s statements; a reduced need to
reinterview the child; and the elimination of distracting note-taking. Among the
disadvantages cited are: concerns about confidentiality, defense attornics’ ten-
dencies to focus attention on minor inconsistencies, and the fact that some people
are uncomfortable with being taped (Myers, 1992; Spencer & Flirt, 1990). We
believe that the advantages far outweigh the disadvantages: Nothing is more
compelling than a clear allegation in the child’s own words.
Prior to the Interview
Because the purpose of the investigative interview is to evaluate a number of
hypotheses related to the child’s allegations, interviewers should prepare them-
selves by gathering as much information as possible about the alleged incident
and about the child’s capacities and propensities (Myers, 1992; Raskin & Yuille,
1989). This information can help the interviewer clarify a variety of alternative
hypotheses concerning what might have happened to the child and the possible
motivations for either false or truthful accounts (Green, 199 1; McGovern, 199 1;
Perry & Wrightsman, 1991). Obtaining information about the case prior to the
investigation does not imply that the interviewer has made a decision about the
validity of the child’s allegation. The purpose of the investigative interview is
“hypothesis testing” not “hypothesis confirming.” Background information
about the case is crucial if interviewers are to focus their questions so that the
interview can be concluded in a timely fashion.
Interviewers may also want to pay attention to informal conversations, involv-
ing the child and familiar adults, in order to gauge the child’s level of linguistic
competence. This information can later help the interviewer to determine wheth-
er and when rapport has been established, to frame questions using developmen-
tally appropriate language, and to evaluate the child’s allegation relative to his or
her description of neutral topics (Raskin & Yuille, 1989; Saywitz, 1988; Walker,
1993). Foreknowledge of the child’s linguistic abilities facilitates preparation and
may protect interviewers from their own impatience (Nurcombe, 1986). Inex-
perienced interviewers often become exasperated when children lack fluency and
provide brief statements or responses. They are then tempted to ask many direct,
leading, and suggestive questions, which not only reduce the amount of informa-
tion obtained, but may also lead witnesses to impeach themselves (Dent, 1982;
Dent & Stephenson, 1979). Others familiar with the child may also provide
personal information about the child (his or her friends, siblings, pets, favorite
foods or activities, etc. so on) which can be used to establish rapport.
Setting the Stage
The interview begins as soon as the child and interviewer meet one another,
preferably alone (Jones & McQuiston, 1988; Raskin & Esplin, 199 la) and
without subsequent interruption (Saywitz, 1988). There are several tasks to be
achieved, however, before the substantive portions of the interview can begin. In
the initial phase, the interviewer should define his or her job in simple terms that
are not confusing and are chosen so as to motivate the child to be informative
(Raskin & Yuille, 1989; Saywitz, 1988). For example, the investigator might
My name is Ms. X. and my job is to talk to children to try to find out the truth about
what happened to them.
I’m the kind of doctor who helps moms and dads and kids. and I understand that
your family needs some help. In order to help you. though. 1 need to find out the
truth about what happcncd. Your job is to help me lcarn the truth about what
After defining their roles and emphasizing the importance of telling the truth,
the interviewer must next convey to the child that she or her is interested in
receiving full and detailed descriptions of events, not simply yes or no responses
to focused questions (Raskin & Yuille, 1989). As Walker (1993) illustrated, a
preponderance of yes/no questions in an interview makes it difficult to evaluate a
child’s understanding of the questions; children often feel they riced to provide
responses to adult questions (Hughes & Grieve, 1980) even if they are incorrect.
To establish a narrative response set, we recommend that the child be asked to
describe some recent meaningful event, such as her or his last birthday party.
Consider the following interaction:
Ccc. how old arc you. then? You’re five’? Wow. you‘re a big girl. When was your
birthday? Can you tell mc everything about your birthday party’?
Reviewing the birthday party and encouraging the child to “really tell r)~cr-~-
thing” about the birthday party graphically illustrates that the interviewer expects
to hear detailed narrative accounts. If the child provides a very brief account or
only describes part of the birthday party the interviewer can encourage the child
to provide more information (e.g.. Saywitz. 1988) by asking the child to elabo-
You mentioned something about going to the swimming pool for your party. Can
you tell me everything that happcncd from the time you got to the pool?
You mentioned some birthday prcscnts. Tell me everything you rcmcmber about the
presents your friends gave you.
By encouraging the child to describe the birthday party in more detail, the
interviewer not only “trains” the child to provide detailed accounts (e.g., of the
sexual allegation), but can also evaluate the child’s linguistic, expressive, and
descriptive capacity informally.
The Substantive Interview
Once rapport has been established, the substantive portion of the interview can
begin. Eighty years ago, Pear and Wyath (1914) suggested that unconstrained
narratives constitute the best ways of obtaining information from children about
events they have experienced. More recently, Dent (1982, 1986; Dent & Ste-
phenson, 1979) among others (e.g., Oates & Shrimpton, 1991) found that free
recall yielded fewer details, but fewer errors, than responses to either general or
specific questions tapping recognition memory. A German forensic psychologist,
Udo Undeutsch (1982, 1989) proposed that descriptions of events that actually
happened differ in content and quality from descriptions of events that were not
actually experienced. Proceeding from these empirical and forensic findings, it
appears that the most accurate accounts are typically obtained by children provid-
ing unfettered, open-ended descriptions of the events in question. The substan-
tive portion of the interview thus begins when the interviewer says:
Now it’s time to talk about something else. 1 understand there are some problems in
your family. Tell me about them in your own words.
One can follow pauses in the child’s account with “And then what happened?”
until the child clearly comes to the end of his or her narrative. Because of the
gravity of the allegations in cases of sexual abuse, it is important to get descrip-
tions of actual incidents or events (Raskin & Esplin, 1991a, Saywitz, 1988), not
simply composite accounts (script memories) that represent what usually or
typically happened to the child. Consequently, when the child describes a pattern
of behavior with repeated incidents, it is often helpful to focus the child’s atten-
tion on particular occasions that can be identified as “the very first time,” “the
very last time,” or by reference to certain distinctive cues that were provided by
the child in her initial account (Raskin & Esplin, 1991a; Saywitz, 1988). Thus,
for example, an interviewer might say: “You told me a moment ago that he was
wearing green underpants. Tell me more about that time” or “Tell me more about
the very first time this happened.” Such cue questions are not suggestive but
facilitate recall (Dent, 1991; Raskin & Esplin, 1991a). Likewise, interviewers
can often get children to provide more detailed accounts of their experiences by
feigning confusion and/or requesting clarification (Raskin & Yuille, 1989). For
I’m confused about some of the things you said before. Maybe you can help me by
describing the time at the pool again.
Although open-ended questions are most likely to encourage accurate ac-
counts of events children have experienced, these accounts may not provide
Figure 1. Relationship between type of question and accuracy of response.
enough information to evaluate the hypotheses about the allegations of abuse,
and thus more focused follow-up questions may be needed. Like us, a number of
professionals have noted the advantages of beginning with open-ended questions
and if necessary progressing to more focused (i.e., direct or leading) questions
(e.g., Myers, 1992; Saywitz, 1988; Spencer & Flin, 1990). As represented in
Figure 1, open-ended questions are more likely to yield accurate information
than are direct, leading, or suggestive questions. The former tap recall memory,
whereas the latter place demands on recognition memory. Whenever it is neces-
sary to use a direct or leading questions (i.e., one that focuses the child’s
attention on certain events, people, or places) it should be followed by an open-
ended question designed to elect a free narrative about the topic to which the
interviewer directed the child’s attention (Raskin & Esplin, 1991a). For example,
the directive utterance “You said something about the living room” can be fol-
lowed by “Tell me everything that happened there.” The line on the inverted
pyramid in Figure 1 indicates the threshold which investigators should only cross
(by asking more focused questions) with extreme caution; any dip below this line
should always be followed by a question or comment that takes the child and the
interviewer back to the level of free narrative and out of the realm of more
focused questions.
The distinction between direct or leading and suggestive questions is also
important (Lamb, Hershkowitz, Stemberg, Esplin, Hovav, Manor, &
Yudilevitch, in press). Direct or leading questions focus the child’s attention on
specific places or incidents, but do not suggest specific responses. Children are
clearly susceptible to suggestive questions (see earlier) which should thus be
avoided if at all possible or posed at the end of an interview. Suggestive questions
should be followed by open-ended questions whenever they are used (Spencer &
Flin, 1990).
Once a complete account of the incident has been provided by the child and
the interview appears to be at an end, it is often helpful to pause and review the
initial hypotheses in light of the information provided by the child. Did the child
mention any details that were unexpected? Do the initial hypotheses still appear
adequate? Does the child’s account clearly support one hypothesis more than
another? Following a brief pause for the interviewer to mentally review the
hypotheses and information, the interview can resume. Many productive inter-
views are ruined by the impatient inability to pause at appropriate occasions
during the interview.
At the end of the interview, the interviewer might shift to a neutral topic and
then thank the child for his or her assistance. If she or he may need to reinterview
the child, this fact should simply be stated as a possibility, without a request for
the child’s permission.
Following the interview, a review of the tape-recorded account will help the
investigator to pull the case together conceptually, and determine whether or not
further interviewing is warranted.
As mentioned earlier, Undeutsch (1982) suggested that accounts of incidents that
really happened differ in quality and content from accounts of incidents that did
not happen. Guided by this general and seemingly robust observation, re-
searchers have sought to formalize this “Undeutsch Hypothesis” so that it is
possible to systematically evaluate the credibility of children’s accounts. Steller
and Koehnken (1989) and Raskin and Esplin (199 1 a) specified 19 criteria that
can be used to conduct a “Criterion Based Content Analysis” (CBCA) of chil-
dren’s statements. In our research, trained raters review a verbatim transcript of
Criteria for Evaluatina Children’s Statements
General Characteristics
1. Logical structure and coherence
2. Unstructured production with spontaneous digressions
3. Quantity of details, especially regarding time, place, persons, and events
Specific Contents
4. Embedding of events in temporal and spatial context
5. Descriptions of interactions with sequences of actions and reactions
6. Reproduction of conversation
7. Unexpected complications or interruptions
Peculiarities of Content
8. Unusual details that are meaningful
9. Superfluous or peripheral details
10. Accurately reported details misunderstood
II. References to other sexually toned events occurring outside the specific
12. References to one’s own feelings or thoughts during the incident
13. Attributions of thoughts, feelings, or motivations to the perpetrator
Motivation-Related Contents
14. Spontaneous corrections or revisions
15. Admissions of memory deficits
16. Raising doubts about one’s testimony or credibility
17. Self-deprecation or assumption of blame
18. Pardoning or excusing the perpetrator
Note. This table was adapted from Steller and Koehnken (19891, Steller (1989), and
Raskin and Esplin (1991a).
the child’s account, decide whether or not each of the criteria’ or characteristics
is present, and then assign a score to the child’s statement, indicating how many
criteria were present. The criteria are listed and briefly defined in Table 1.
As the table indicates, the criteria are of four types. The first three refer to the
general characteristics of the statement, particularly its coherence and infor-
mativeness. Statements that are illogical or completely lacking in detail may not
be suitable for further analysis. The next four criteria refer to the spec$c con-
tents of the child’s statement. These criteria index the extent to which the account
contains rich details that are unlikely to be included in a fabricated account, and
that locate the incident in its everyday context. The third category (peculiurities
ofcontent) comprises six criteria that tend to occur less frequently, but are highly
informative when they do occur because they involve rich, minor, or irrelevant
‘Criterion 19 (details characteristic of the offense) did not really tap the Undeutsch Hypothesis,
and thus has been dropped from the list of content criteria in our current research. It is not listed in
Table I Some of the other criteria are also being reevaluated. but are listed in the table becauac they
are still widely viewed as part of the CBCA system.
details that children would be unlikely to fabricate and adults would be unlikely
to coach. The five motivation-related criteria indicate whether the child is strong-
ly motivated to “sell” his story, perhaps by denying any possible deficiency in his
or her account or behavior.
Research on CBCA
In a recent study, Horowitz, Lamb, Esplin, Boychuk, Reiter-Lavery, and Krispin
( 1992) showed that trained raters showed high levels of agreement with one
another in determining whether or not most of these characteristics or criteria
were present; the three raters were also highly reliable over a 3-month interval,
and their high levels of agreement did not vary depending upon the age of the
child or the degree to which the account was valid or not. It thus seems that the
CBCA system provides a reliable means of quantitatively evaluating the Un-
deutsch Hypothesis.
In order to evaluate the validity of a procedure like CBCA, researchers need
to determine whether accounts of incidents that actually happened are assigned
higher scores on assessments like the CBCA checklist than accounts of incidents
that did not happen. Two field studies conducted in Phoenix, Arizona (Boychuk,
1991; Raskin & Esplin, 199 la, 1991b) suggest that CBCA is a promising tech-
nique for discriminating real events from those that did not occur. In conducting
studies such as these, however, a major difficulty inevitably arises: How can
researchers determine whether or not the incident actually occurred? Because
judicial disposition is itself influenced in some way by the child’s statement, it
does not constitute an independent validation of the child’s allegations, and it is
thus necessary for research purposes to consider only independent validating
information (Wells & Loftus, 199 1). Raskin and Esplin (199 1 b) thus based their
discrimination between doubtful and certain cases on the results of polygraphic
and medical examinations, confessions, and eyewitness accounts. They reported
no overlap between the distribution of scores assigned to statements known to
involve “confirmed” and “doubtful” incidents. Horowitz et al. (1991) have pro-
posed a multifaceted procedure in which the results of medical examinations,
confessions, polygraphic examinations, witness statements, and other circum-
stantial or physical evidence are combined to establish, with varying degrees of
certainty, the probability that the alleged event or events actually occurred, and
this may be helpful in future research on this topic.
Although these studies suggest that the CBCA procedure can discriminate
effectively between truthful and nontruthful accounts, the procedure has also
been criticized. Wells and Loftus (1991), for example, criticized the represen-
tativeness of the 40 cases included in Raskin and Esplin’s study; they emphasized
the importance of using independent case facts (i.e., not judicial dismissal or
lack of prosecution) to establish the plausibility of the allegations, questioned the
evaluation of interrater reliability, and asked about the association between chil-
dren’s age and their CBCA scores. As noted earlier, Raskin and Esplin (1991b)
reanalyzed their data so as to ensure that independent case facts were used to
assess plausibility, and interrater reliability is now known to be high (Horowitz et
al., 1992). Because Boychuk (1991) and Raskin and Esplin (1991a, 1991b) only
included allegations of highly improbable and highly probable incidents, how-
ever, we do not yet know whether the CBCA procedure can help discriminate
incredible from accurate accounts in a broad range of cases. Nor have investiga-
tors systematically studied the relationship between age and CBCA scores.
We are currently conducting research on the validity and reliability of the
CBCA system using three large and more representative samples of children’s
allegations. Cases are selected on the basis of the victims’ ages and the type of
incidents and are validated using multiple criteria independent of the child’s
statement and disposition of the case. One study involves a random national
sample of all alleged incidents of specific types investigated in Israel over a
2-year period; the others involve investigative agencies in the United States. Our
studies include many statements obtained by a variety of professionals (including
policemen, social workers, and psychologists) from children ranging in age from
3 to 12 years. Multiple independent criteria are used to confirm the probability or
doubtfulness of the events described by the child (Horowitz et al., 199 l), and
independent raters evaluate the strength of the corroborative information and the
children’s statements. We hope that these ongoing studies will provide further
insight into the usefulness of the CBCA procedure and suggestions about ways in
which the procedure could be modified in order to increase its precision in
evaluating statements by children of different ages and backgrounds. The tech-
nique is, of course, an experimental tool that is still being developed and refined.
Except as a heuristic device, it would be premature to use it in forensic settings to
quantify veracity or credibility.
In this article, we have surveyed the empirical literature on children’s memory,
communicative capacities, and the ability to distinguish between fantasy and
reality. Like many other researchers, we question the ecological validity of many
of the studies reviewed and have identified areas where additional research is
especially necessary. Drawing on the relevant empirical evidence, we have also
outlined an interview technique which can be used to elicit free narrative ac-
counts from children about their experiences, emphasizing the importance of
obtaining accounts from children in their own words, and the dangers implicit in
the use of suggestive questions. In an ongoing research program, we are cur-
rently exploring the effectiveness of different interview strategies and styles in
hopes of learning how to maximize the value of interviews with children (see,
e.g., Lamb et al., in press). In this article, we also introduced an investigative
technique (the CBCA system) for evaluating the quality of information provided
by children, and we describe our ongoing research focused on conditions which
improve or hamper children’s accounts and the evaluation of these reports.
Throughout the paper, we show how basic research must be interpreted with
caution when it is used to advance our understanding of real-world “applied”
American Association for Protecting Children. (1986). Highlights of Oficial Child Neglect and
Abuse Reporting. 1984. Denver: Author.
American jurisprudence (1976). Rochester, NY: Lawyers Cooperative.
Boat, B., & Everson, M. (1988). Use of anatomically correct dolls among professionals in sexual
abuse evaluations. Child Abuse and Neglect, 12. 17 I - 186.
Boychuk, T.D. (1991). Criteria-based content analysis of children’s statements about sexual abuse:
A field-based validation study. Unpublished doctoral dissertation, Arizona State University.
Brainerd, C., & Omstein. P.A. (1991). Children’s memory for witnessed events: The developmental
backdrop. In J. Doris (Ed.), The suggestibiliry qfchildren’s recollections (pp. 10-20). Wash-
ington, DC: American Psychological Association.
Carkeet, D. (1980). Double negative. New York: Penguin.
Ceci, S.J. (1991). Some overarching issues in the children’s suggestibility debate. In J. Doris (Ed.),
The suggestibility of children’s recollections (pp. I-9). Washington, DC: American Psycho-
logical Association.
Ceci, S.J., & Bronfenbrenner, U. (1991). On the demise of everyday memory: The rumors of my
death are greatly exaggerated. American Psychologist. 46, 27-31.
Ceci, S.J., & Bruck, M. (1993). Suggestibility of the child witness: A historical review and synthe-
sis. Psyhological Bulletin, 113, 403-439.
Ceci. S.J., Leichtman, M.D., & Putnick, M.E. (Eds.). (1992). Cognifive and socia/,factor.s in early
deception. Hillsdale. NJ: Erlbaum.
Ceci, S.J., Ross, D.F., & Toglia, M.P. (1987a). Suggestibility of children’s memory: Psycholegal
issues. Journal of Experimental Psychology: General, 116, 38-49.
Ceci, S.J., Ross, D.F., & Toglia, M.P. (1987b). Age differences in suggestibility: Narrowing the
uncertainties. In S.J. Ceci, M.P. Toglia, & D.F. Ross (Eds.), Children’s eyewitness memory
(pp. 79-91). New York: Springer-Verlag.
Ceci, S.J., Ross, D.F., & Toglia, M.P. (Eds.). (1989). Perspectives on children’s testimony. New
York: Springer-Verlag.
Ceci, S.J., Toglia, M.P., & Ross, D.F. (Eds.). (1987). Children’s eyewitness memory. New York:
Chance, J.E., & Goldstein, A.G. (1984). Face-recognition memory: Implications for children’s
eyewitness testimony. Journal of Social Issues. 40. 69-85.
Chi. M.T.H.. & Ceci. S.J. (1987). Content knowledge: Its role, representation, and restructuring in
memory development. In H.W. Reese (Ed.), Advances in child development and behavior
(Vol. 20, pp. 91-142). New York: Academic.
Cohen, R.L., & Hamick, M.A. (1980). The suggestibility of child witnesses to suggestion. LaM: and
Human Behavior. 4. 201-210.
Collins, W.A. (1970). Learning of media content: A developmental study. Child Development, 4/.
Collins, W.A., & Wellman, H. (19