due tomorrow 11pm

ANRV307-CP03-13 ARI 20 February 2007 19:5

Unwarranted Assumptions about Children’s Testimonial Accuracy Stephen J. Ceci,1 Sarah Kulkofsky,1

J. Zoe Klemfuss,1 Charlotte D. Sweeney,1

and Maggie Bruck2 1 Department of Human Development, Cornell University, Ithaca, New York 14853; email: [email protected] 2 Department of Psychiatry & Behavioral Sciences, Johns Hopkins University, Baltimore, Maryland 21287

Annu. Rev. Clin. Psychol. 2007. 3:311–28

First published online as a Review in Advance on January 2, 2007

The Annual Review of Clinical Psychology is online at http://clinpsy.annualreviews.org

This article’s doi: 10.1146/annurev.clinpsy.3.022806.091354

Copyright c© 2007 by Annual Reviews. All rights reserved

1548-5943/07/0427-0311$20.00

Key Words

suggestibility, open-ended questions, false reports

Abstract We examine eight unwarranted assumptions made by expert wit- nesses, forensic interviewers, and legal scholars about the reliability of children’s eyewitness reports. The first four assumptions mod- ify some central beliefs about the nature of suggestive interviews, age-related differences in resistance to suggestion, and thresholds necessary to produce tainted reports. The fifth unwarranted assump- tion involves the influence of both individual and interviewer factors in determining children’s suggestibility. The sixth unwarranted as- sumption concerns the claim that suggested reports are detectable. The seventh unwarranted assumption concerns new findings about how children deny, disclose, and/or recant their abuse. Finally, we examine unwarranted statements about the value of science to the forensic arena. It is important not only for researchers but also expert witnesses and court-appointed psychologists to be aware of these un- warranted assumptions.

311

A nn

u. R

ev . C

li n.

P sy

ch ol

. 2 00

7. 3:

31 1-

32 8.

D ow

nl oa

de d

fr om

a rj

ou rn

al s.

an nu

al re

vi ew

s. or

g by

T ex

as T

ec h

U ni

ve rs

it y

- L

ub bo

ck o

n 02

/1 5/

08 . F

or p

er so

na l

us e

on ly

.

ANRV307-CP03-13 ARI 20 February 2007 19:5

Contents

INTRODUCTION . . . . . . . . . . . . . . . . . 312 Unwarranted Assumption 1:

Suggestive Interviews Can Be Indexed by the Sheer Number of Leading Questions . . . . . . . . . . 313

Unwarranted Assumption 2: Suggestibility Is Primarily a Problem for Younger Age Groups . . . . . . . . . . . . . . . . . . . . . . . 314

Unwarranted Assumption 3: Multiple Suggestive Interviews Are Needed to Taint a Report; Milder Forms of Suggestion Do Not Produce Tainted Reports . . . . . . . . . . . . . . . 316

Unwarranted Assumption 4: Children’s Spontaneous Reports Are Always Accurate . . . . . . . . . . . 317

Unwarranted Assumption 5: Erroneous Suggestions Ineluctably Lead to Erroneous Reports by Children . . . . . . . . . . . 318

Unwarranted Assumption 6: False Reports Produced by Suggestive Interviewing Are Distinguishable from Accurate Reports . . . . . . . . . . . . . . . . . . . . . . . 320

Unwarranted Assumption 7: Children’s Disclosures of Traumatic Events Are Delayed, Denied, and Often Recanted . . . 322

Unwarranted Assumption 8: Laboratory Research Is Not an Accurate Reflection of Child Witnesses’ Experiences in the Real World . . . . . . . . . . . . . 323

CONCLUSION . . . . . . . . . . . . . . . . . . . . 324

INTRODUCTION

In recent years, an increasing number of chil- dren have entered the legal system to pro- vide testimony in a broad range of cases. In the United States alone, hundreds of thou-

sands of children are deposed, interviewed, and examined each year as part of civil and family court proceedings, abuse/neglect in- vestigations, and other types of criminal in- vestigations. Before the early 1980s, children rarely were permitted to testify in criminal cases1 (see Ceci & Bruck 1993, p. 408). Now, however, it is so common that most English- speaking nations have developed special inter- viewing procedures and techniques to mini- mize children’s discomfort (e.g., video links that allow them to testify remotely, barri- ers between them and the defendant) and in- crease the reliability of their statements (e.g., Home Off. Dept. Health 1992, Natl. Soc. Prev. Cruel. Child. Childline 1993, Smith & Goretsky-Elstein 1993).

Given the recent ubiquity of children’s par- ticipation in forensic matters, it is not surpris- ing that beginning in the 1980s researchers turned their attention to the sensitive issues that arise when children enter the legal arena. This research has made significant contribu- tions to the theoretical and applied science of child development. It has provided new and often surprising insights into the capa- bilities and weaknesses of children’s cognitive, linguistic, social, and emotional development. This research also has dispelled myths or com- mon beliefs about aspects of child develop- ment both in and out of the forensic arena. In this review, we first outline how some of the current research has challenged miscon- ceptions about the reliability and credibility of children’s statements. Given that many of the issues concerning children in the court- room revolve around cases of suspected sex- ual abuse, it is not surprising that having the science accepted in the courtroom has been

1 An anecdote from the John Grisham novel The Last Ju- ror underscores the scientific research. It concerns a cross- examination of a newspaper publisher set in Mississippi in 1970: Attorney: Mr. Traynor, how many cases did you find where children aged five or younger were allowed to testify in a criminal trial? Newspaper Publisher Traynor: None. Attorney: Perfect answer, Mr. Traynor. None. In the his- tory of this state, no child under the age of eleven has ever testified in a criminal trial.

312 Ceci et al.

A nn

u. R

ev . C

li n.

P sy

ch ol

. 2 00

7. 3:

31 1-

32 8.

D ow

nl oa

de d

fr om

a rj

ou rn

al s.

an nu

al re

vi ew

s. or

g by

T ex

as T

ec h

U ni

ve rs

it y

- L

ub bo

ck o

n 02

/1 5/

08 . F

or p

er so

na l

us e

on ly

.

ANRV307-CP03-13 ARI 20 February 2007 19:5

a difficult process. To illustrate this point, we end the article with a misconception of the usefulness of the science to the forensic arena and show how scientists attempt to deal with efforts to keep the current science out of the courtroom.2

Unwarranted Assumption 1: Suggestive Interviews Can Be Indexed by the Sheer Number of Leading Questions

The suggestiveness (and thus the risk of elic- iting false information) of an interview is not adequately reflected by the number of lead- ing questions. Rather, one must consider how the concept of interview bias plays out in the current interview, as well as in all pre- vious interviews. Interview bias characterizes those interviewers who hold a priori beliefs about what has occurred and mold the inter- view to maximize disclosures that are consis- tent with such beliefs. The means by which the bias is communicated to the child goes well beyond the use of misleading questions; other suggestive techniques include provid- ing positive and negative reinforcement (e.g., praising the child for disclosing information consistent with the interviewer’s beliefs, crit- icizing the child or withholding benefits such as trips to the restroom for not disclosing), utilizing peer or parental pressure (e.g., telling the child that his or her friends or parents have already disclosed), creating a negative or ac- cusatory emotional tone (e.g., urging the child to help keep the defendant in jail), and re- peating questions or interviews until the child provides a desired answer.

The following testimony of an expert in a trial illustrates this point. She testified that her questioning was not suggestive because technically speaking she did not ask sugges- tive questions. But as seen from her testimony,

2 In preparing this review, we drew upon several examples from a recent article of ours (Bruck & Ceci 2004), but have gone considerably beyond it.

her approach is characteristic of interviewer bias:

I usually say, “Mama talked about that some- body did some bad touching.” And that’s still pretty open ended. I’m not saying who and I’m not saying exactly what. I’m just introducing the subject. Or I will say, “I see many children, and children come and tell me when bad things happen to them, and I’ve heard other kids tell me when bad things happen. So it’s okay if you want to tell me.” (In the Matter of Riley, Shelby, and Austin Blanchard v. John Blanchard 2001, p. 876)

A number of studies have demonstrated the negative effects of interviewer bias (see Ceci & Bruck 1995 for a review). In one type of study examining interviewer bias, children witness a staged event and are then inter- viewed by an individual who is given misinfor- mation about what has occurred. The inter- viewer is allowed to interview the child in any way he or she deems appropriate; that is, the interviewer is simply told to find out what hap- pened. These studies have found that children who are interviewed by an individual who has been misinformed (or who has a bias) about what has occurred begin to report this misin- formation themselves (e.g., Bruck et al. 1999, White et al. 1997). For example, if the inter- viewer has been misinformed that the child had her knee licked by another child, she ends up getting the child to assent to this false claim (White et al. 1997).

Another set of studies has examined the ef- fects of combining multiple suggestive tech- niques in eliciting false reports from children. These studies demonstrate that misleading questions asked by a neutral interviewer do not have the same effect as multiple sug- gestive techniques, implying that misleading questions alone are not sufficient to expose an interviewer’s bias. For example, Garven and her colleagues (1998, 2000) examined how the techniques used by investigators in the infamous McMartin Preschool case (State of Calif. v. Buckey 1990) can taint children’s

www.annualreviews.org • Unwarranted Assumptions 313

A nn

u. R

ev . C

li n.

P sy

ch ol

. 2 00

7. 3:

31 1-

32 8.

D ow

nl oa

de d

fr om

a rj

ou rn

al s.

an nu

al re

vi ew

s. or

g by

T ex

as T

ec h

U ni

ve rs

it y

- L

ub bo

ck o

n 02

/1 5/

08 . F

or p

er so

na l

us e

on ly

.

ANRV307-CP03-13 ARI 20 February 2007 19:5

testimony beyond the damage that mislead- ing information alone can cause. In one study (Garven et al. 2000), the researchers asked kindergarten children to recall details from when a visitor named Paco came to their class- room and read a story, gave out treats, and wore a funny hat. Half the children were given interviews that included misleading questions about plausible events (e.g., Did Paco break a toy?) and about bizarre events (e.g., Did Paco take you to a farm in a helicopter?). Be- tween 5%–13% of the children falsely agreed with the misleading questions. A second group of children was also questioned, but these children were given feedback after their an- swers to the misleading questions. “No” re- sponses were negatively evaluated, whereas “yes” responses were positively evaluated. For example,

Interviewer: Did Paco take you somewhere in a helicopter? Child: No. Interviewer: You’re not doing good. Interviewer: Did Paco break a toy? Child: Yes. Interviewer: Great, you’re doing excellent now.

This latter group of children provided the desired but false answer to 35% of the plausi- ble questions and to 52% of the bizarre ques- tions. This study demonstrates that a simple count of misleading/leading questions would not reflect the suggestiveness of the interview. It was the added use of selective reinforcement that provided the child with sufficient infor- mation about the interviewer’s bias—to make “yes” responses for all statements regardless of their plausibility. In a follow-up interview two weeks later, when children were simply asked nonleading questions with no selective reinforcement feedback, the same level of be- tween group differences was obtained. Thus, interviewer bias in a prior interview can pro- duce false reports in a later unbiased/neutral interview. This is an important point to bear in mind when analyzing transcripts of an in- terview: Just because that particular interview

may be neutral, prior interviews may have been suggestive, seeding false claims made in the neutral interview. The bottom line is that the number of leading or suggestive questions deployed in an interview is neither a good in- dex of how suggestive it is, nor a good index of whether prior (nonrecorded) interviews that were more suggestive are responsible for false claims by the child.

Unwarranted Assumption 2: Suggestibility Is Primarily a Problem for Younger Age Groups

The erroneous view that preschool children are the only population vulnerable to sugges- tion can be found in many experts’ testimony. Consider the following example: “Well, in vir- tually all these studies, two and three-year olds do not do well in suggestibility, and the four and five-year olds. . . [d]o pretty well” (ex- pert testimony by prosecution witness In the Matter of Riley, Shelby, and Austin Blanchard v. John Blanchard 2001). “It’s true that the sorts of questioning that were asked of the children are not supported by basic research into sug- gestibility, but these children were all over the age of 6, the cut-off for suggestibility- proneness in scientific studies” (transcript, p. 1441).

This view that only the youngest children are vulnerable to suggestive questioning re- flects the disproportionate attention to the study of preschool children at the end of the twentieth century. This practice was directly motivated by forensic concerns. During the 1980s and 1990s, there were a number of high-profile criminal cases in which preschool children’s horrific claims about sexual abuse by day-care workers, parents, and other unfa- miliar adults were presented to the jury (see descriptions of several of these cases in Ceci & Bruck 1995). Although the case facts showed that these children had been subjected to highly suggestive interviews, at that time there was no relevant body of scientific literature to indicate the risk of these interviewing tech- niques in producing false allegations about

314 Ceci et al.

A nn

u. R

ev . C

li n.

P sy

ch ol

. 2 00

7. 3:

31 1-

32 8.

D ow

nl oa

de d

fr om

a rj

ou rn

al s.

an nu

al re

vi ew

s. or

g by

T ex

as T

ec h

U ni

ve rs

it y

- L

ub bo

ck o

n 02

/1 5/

08 . F

or p

er so

na l

us e

on ly

.

ANRV307-CP03-13 ARI 20 February 2007 19:5

a range of salient events. When researchers began to fill in this empirical void, most of the studies focused on preschoolers, with few examining age-related differences. Those that did include age comparisons usually found ceiling effects for the older children, leading to the conclusion that only preschoolers are suggestible (e.g., Ceci et al. 1987) and that there is little need for concern when older children are subjected to suggestive interview- ing practices.

However, the conclusion that suggestibil- ity is minimal among grade-school children and young adults is discrepant with the find- ing of another body of literature that shows that many of the suggestive techniques used in the child studies also produced tainted re- ports or memories in adults (e.g., see Loftus 2004). Indeed, much of the earliest work on the malleability of memory was conducted with adults, not young children. For example, in a highly cited study by Loftus & Pickrell (1995), adults ranging in age from 18 to 35 were convinced through a variety of sugges- tive techniques that that they had been lost in a shopping mall as young children, and they developed elaborate, albeit false memories of these events. Similarly, Hyman and his col- leagues have conducted a number of stud- ies whereby through suggestive techniques, college-aged students developed false mem- ories of events, such as spilling a punch bowl at a wedding (e.g., Hyman et al. 1995).

By inference then, one might assume that children in middle childhood must also be quite suggestible, given the knowledge of both the younger children and older groups. Re- cent evidence supports this view: Susceptibil- ity to suggestion is highly common in mid- dle childhood, and under some conditions, there are small or even no developmental dif- ferences. For example, Finnila et al. (2003) staged an event (a version of the Paco visit described in Garven et al. 2000) for two age groups of children (four- and five-year-olds, and seven- and eight-year-olds). One week later, half the children were given a low- pressure interview that contained some mis-

leading questions with abuse themes (e.g., “He took your clothes off, didn’t he?”). The other children received a high-pressure inter- view during which the interviewer told them that their friends had answered the leading questions affirmatively. Children were praised for assenting to the misleading questions, and when they did not assent, the question was repeated. In both the low- and high-pressure conditions, there were no significant age dif- ferences, although a significant number (68%) of misleading questions were assented to in the high-pressure condition (see also Bruck et al. 2007, Zaragoza et al. 2001).

Under some conditions, older children are even more suggestible than younger chil- dren are (e.g., Ceci et al. 2007, Finnila et al. 2003, Lindberg 1991, Scullin & Ceci 2001, Zaragoza et al. 2001). For example, in a recent study, researchers administered a suggestibil- ity test to four-year-olds and nine-year-olds (Ceci et al. 2007). Children were read a short story that focused on a series of objects. Later children were given misinformation about the objects in the story. Days later, they were asked to recall the objects that were part of the original story. The direction of age differ- ences in suggestibility was predicted by age differences in children’s semantic represen- tations of the similarity between the actual and suggested object. For example, compared with younger children, older children were much more likely to erroneously report that there was an egg sandwich in a story in which they actually heard about a cheese sandwich. This is because older children find eggs and cheese to be more similar than younger chil- dren do. Similarly, compared with older chil- dren, younger children were more likely to report the false suggestion that there was a soda in the story when there was actually milk. Again, this was because younger children per- ceive milk and soda as more similar than do older children.

These newer findings reshape current views of developmental trends in suggestibil- ity and challenge current conceptualiza- tions of the developmental mechanisms in

www.annualreviews.org • Unwarranted Assumptions 315

A nn

u. R

ev . C

li n.

P sy

ch ol

. 2 00

7. 3:

31 1-

32 8.

D ow

nl oa

de d

fr om

a rj

ou rn

al s.

an nu

al re

vi ew

s. or

g by

T ex

as T

ec h

U ni

ve rs

it y

- L

ub bo

ck o

n 02

/1 5/

08 . F

or p

er so

na l

us e

on ly

.

ANRV307-CP03-13 ARI 20 February 2007 19:5

children’s suggestibility. Specifically, current mechanisms that have been touted for chil- dren’s suggestibility (e.g., theory of mind, social compliance) are commonly known to have developed by the end of the preschool years (Wellman et al. 2001). Clearly, a wider perspective needs to be taken, and skills that develop throughout the childhood years should become the focus of future study (e.g., appreciation of the ramifications of false state- ments, insight into questioner’s motives). The bottom line is that, expert testimony notwith- standing (Bruck & Ceci 2004), all age groups are vulnerable to misleading suggestions, even if preschoolers are disproportionately more vulnerable.

Unwarranted Assumption 3: Multiple Suggestive Interviews Are Needed to Taint a Report; Milder Forms of Suggestion Do Not Produce Tainted Reports

A third scientific misconception concerns the view that it is difficult to implant memories or to taint reports and, therefore, repeated sug- gestive interviews are required, especially to produce a false report for salient events (e.g., Ceci & Bruck 1995). For example, the expert in the Matter of Ryan D. Smith (2001) testified, “[s]uggestions must be repeated for children to incorporate them into their reports” (tran- script, p. 886).

We have been as responsible as anyone has for this view because often when we de- scribe our own findings, we report that our methodology involved multiple suggestive in- terviews of children over time (e.g., Bruck et al. 1995a, Ceci et al. 1994a, Leichtman & Ceci 1995). Although this is a correct descrip- tion of our own methodology and results, it does not gainsay the many studies that have re- ported that children can incorporate sugges- tions about salient events after a single sugges- tive interview (Bruck et al. 2007, Garven et al. 2000, Thompson et al. 1997). Importantly, the effects of a single interview that produced false reports had powerful and lasting effects:

Children’s initial false responses to sugges- tions, which may have reflected social pres- sure, continued in later interviews even when questioned by different (neutral) interviewers. This enduring pattern may reflect that with time children come to believe that the false suggestion actually happened; in other words, they developed a false belief about a statement they initially realized was false.

Significant tainting and production of false beliefs can also occur with a single mildly sug- gestive interview. For example, in two studies, Poole & Lindsay (1998, 2001) had parents read their child a brief narrative that out- lined the child’s previous encounters with a character known as “Mr. Science” at the re- searchers’ laboratory. Unknown to the par- ent, some of the details in the brief narrative they read to their children were inaccurate and thus were not experienced by their child when they met Mr. Science. Nonetheless, even un- der these mild conditions, significant numbers of children (four- to eight-year-olds) later told an interviewer that they had experienced the suggested events.

In a similar line of research, Principe and her colleagues (Principe & Ceci 2002, Principe et al. 2006) found that children ex- posed to rumors by their peers or by over- hearing adult conversations were as likely to falsely report being part of the event as those who actually did participate. In one study (Principe & Ceci 2002), a group of children participated in a scripted event that contained two highly salient events (the target activi- ties). The second group included classmates of the first group; these children participated in the same scripted event, but they were not shown the two target events. A third group of children was taken from a different school; similar to the second group, they experienced the scripted activity without the target events. The children were then interviewed in either a neutral or a suggestive manner. Children who were classmates of those who participated in the target event made more false claims alleging that they had viewed the target ac- tivities than children who did not view the

316 Ceci et al.

A nn

u. R

ev . C

li n.

P sy

ch ol

. 2 00

7. 3:

31 1-

32 8.

D ow

nl oa

de d

fr om

a rj

ou rn

al s.

an nu

al re

vi ew

s. or

g by

T ex

as T

ec h

U ni

ve rs

it y

- L

ub bo

ck o

n 02

/1 5/

08 . F

or p

er so

na l

us e

on ly

.

ANRV307-CP03-13 ARI 20 February 2007 19:5

target events and were from a different school. These data suggest that there was contami- nation from classroom interactions; children who had not experienced the target events learned of them from their classmates who had, and thus were more likely to assent to false events. In a second study, Principe et al. (2006) found that children who overheard a child talking about a target event (a rabbit that escaped from a magician) were as likely to falsely claim to have seen the event in ques- tion as were peers who actually saw it escape. Moreover, in this study, the effect of sugges- tive questioning did not notably increase their false reports; they were as likely to report falsely if they overheard peers talking about the rabbit, regardless of whether interviewers employed suggestive questions.

So far, we have only focused on the extent to which one exposure to false information can lead to high rates of misreporting about salient events. Some studies directly compare the ef- fects of multiple versus one suggestive in- terview on children’s suggestibility. Evidence suggests that, contrary to common psycho- logical principles, a number of circumstances exist in which one suggestive interview pro- duces the same amount of taint as two or more suggestive interviews. The impact of a second interview depends on the spacing of the in- terviews from the initial events and from the final interview, and also on the strength of the original memory trace (Marche 1999, Melnyk & Bruck 2004).

In light of these considerations, it is re- grettable that expert witnesses create the im- pression that interviewing methods must be repetitive, egregious, and coercive to taint a child’s statement about abuse. In a post- testimonial brief, one expert witness made just such a claim, arguing that short of us- ing what she felt were repetitive, egregiously coercive forms of interviewing, one could ex- pect children’s statements to be highly ac- curate: “[Q]uestioning that is ‘merely’ sug- gestive (e.g., occasional leading questions embedded in an interview with neutral and direct questions) yields a 92–100% accuracy

rate on abuse-relevant or trauma-relevant questions” (Dalenberg 2000, p. 11). In her words,

In the Janitor, Sam Stone, and Garven et al. (in press) studies . . . the authors subject chil- dren to the following procedures: Repeated affirmative statements that the perpetrator is guilty, not just questions about his acts. Refusals to hear and acknowledge the child’s denials of a perpetrator’s guilt . . . statements to the child that she is not doing a good job when she says something that the in- terviewer does not want to hear. Offering of evidence to the child that the perpetra- tor was guilty (such as telling the child that her mother, or other children, knows that the perpetrator is guilty or that the trauma occurred [as in the Garven studies or Mouse- trap studies], or showing the child physical evidence of the crime [as in the Sam Stone study]). Taking the methods of these studies into the sexual abuse arena, the questioner would have affirmatively told R.B. that her father was a sexual abuser, falsely told her that there was physical evidence that she was abused, and then repeatedly, even if she de- nied a type of abuse each and every time she was queried, asked if her father was angry when he abused her, where he went after he abused her, etc. . . . I see no evidence that the interviewers of R.B. went this far.

Unwarranted Assumption 4: Children’s Spontaneous Reports Are Always Accurate

A commonly held belief is that although chil- dren’s prompted statements may be suspect, their spontaneous statements are generally ac- curate, and errors only occur when children are asked specific, often misleading questions. Similar to many of the other assumptions, this too made its way into expert testimony. For example, Dr. Constance Dalenberg testi- fied for the prosecution In the Matter of Riley, Shelby, and Austin Blanchard v. John Blanchard 2001 that “[s]pontaneous statements are

www.annualreviews.org • Unwarranted Assumptions 317

A nn

u. R

ev . C

li n.

P sy

ch ol

. 2 00

7. 3:

31 1-

32 8.

D ow

nl oa

de d

fr om

a rj

ou rn

al s.

an nu

al re

vi ew

s. or

g by

T ex

as T

ec h

U ni

ve rs

it y

- L

ub bo

ck o

n 02

/1 5/

08 . F

or p

er so

na l

us e

on ly

.

ANRV307-CP03-13 ARI 20 February 2007 19:5

likely to be accurate statements” (transcript, p. 39).

It is true that children tend to be more accurate when asked open-ended questions compared with more directed questioning. This has been recognized since Binet’s (1900) earliest studies on children’s suggestibility. However, it is not the case that children’s spontaneous statements are always accurate statements. When children have been ques- tioned suggestively, the suggestions can taint both what they later spontaneously report as well as their answers to specific ques- tions. For example, the children in Poole & Lindsay’s (2001) Mr. Science study, dis- cussed above, were simply asked to describe everything they remembered from interact- ing with Mr. Science after they had been exposed to a misleading narrative by their par- ents about the event. A full 21% of the state- ments children reported were events they had not actually experienced. Furthermore, some of these events included bodily touch, such as Mr. Science putting something yucky in the child’s mouth. In a different study, Poole & White (1993) interviewed children about an event that had occurred two years pre- viously. Following the first event the chil- dren had been exposed to repeated questions, some of which where misleading. Two years later, 39% of the six-year-olds’ (who were four years old at the time of the original event) statements in response to the open-ended re- quest “Tell me what happened” were incor- rect. Older children did not fare much better, with 23% of eight-year-olds’ and 25% of ten- year-olds’ statements being incorrect.

Furthermore, even if children have not been exposed to misleading suggestions, their open-ended recall is still not guaranteed to be accurate. This is especially true if the child is interviewed about a confusing or ambiguous event. For example, Ornstein and colleagues (1998) engaged children in a mock medical exam in which some common features (such as listening to the child’s heart) were omitted and atypical features (such as wiping the child’s belly button with alcohol) were added. When

children were interviewed about the event 12 weeks later, 42% of the four-year-olds and 74% of the six-year-olds spontaneously re- ported at least one of the common features had occurred even though they had not. In another study, Goodman and her colleagues (1994) interviewed children about a painful genital catheterization procedure. Among the children who were three to four years old, 23% of their free-recall statements were in- correct. Finally, Greenhoot (2000) examined five- and six-year-olds’ recall of ambiguous stories, in which the protagonist may have been considered acting prosocially or antiso- cially. Error rates on open-ended recall ranged between 20%–30%.

Taken together, these studies show that al- though children’s free-recall and spontaneous statements are generally more accurate than their responses to directed questions, free re- call is by no means error free.

Unwarranted Assumption 5: Erroneous Suggestions Ineluctably Lead to Erroneous Reports by Children

In the literature on children’s suggestibility, it is rare to find a study in which 100% of the children are influenced by 100% of the false suggestions. Although some children easily and readily succumb to suggestive inter- views, others remain steadfast in their denial that the suggested events occurred. Mount- ing evidence from different paradigms and methodologies leads to the conclusion that suggestibility proneness is jointly determined by contextual (factors intrinsic to the inter- view) and individual (factors intrinsic to the child) differences. Thus, suggestive interviews need not inevitably lead to false reports by children.

For example, Peterson and her colleagues (2004) have demonstrated that in certain con- texts (in which an event actually occurred, was strongly encoded, and probably reinforced by family retellings), the provision of misinfor- mation can actually lead to superior memory

318 Ceci et al.

A nn

u. R

ev . C

li n.

P sy

ch ol

. 2 00

7. 3:

31 1-

32 8.

D ow

nl oa

de d

fr om

a rj

ou rn

al s.

an nu

al re

vi ew

s. or

g by

T ex

as T

ec h

U ni

ve rs

it y

- L

ub bo

ck o

n 02

/1 5/

08 . F

or p

er so

na l

us e

on ly

.

ANRV307-CP03-13 ARI 20 February 2007 19:5

accuracy. This is because the strong mem- ory stands in stark contrast to the misinfor- mation, thus reinforcing and refreshing it. These researchers conducted an experiment to find out if presenting misleading sugges- tions to children after they experienced an emergency medical procedure (e.g., setting a broken bone, suturing a laceration, treat- ment for a dog bite, or treatment for crushed fingers) would taint their recall of that ear- lier event. The misleading suggestions were presented one year after the medical proce- dure; children were tested one week later and one year later (i.e., two years after the medi- cal procedure). Contrary to the reasonable as- sumption that the mere presence of mislead- ing suggestions would inevitably lead to lower accuracy and incorporation of false details, Peterson et al. found that none of the mis- leading suggestions contaminated the chil- dren’s reports, especially one year later. How can we reconcile this finding with other find- ings showing that the passage of time makes it likely that children will incorporate such mis- leading information into their long-term re- calls (e.g., Bruck et al. 1997)? Two aspects to this finding merit attention.

First, the children in Peterson et al.’s study had accurate memories of their medical pro- cedures. Not only were these salient events, but they were rehearsed via family discussions during the intervening year, in the absence of any attempt to taint their recall. Therefore, when one year elapsed and the researchers provided the first misleading suggestions, they were flying into a headwind, trying to over- write what by then had become strong rep- resentations. Imagine trying to alter a child’s memory of his or her own name; you could not do it unless you worked on the child relent- lessly for long periods. Children’s emergency- room experiences may not have been as deeply entrenched as their names, but they were un- doubtedly quite entrenched.

Second, many studies showing incorpora- tion of false suggestions in children’s reports may have been documenting report inaccu- racy rather than alteration of the underlying

memory, a suggestion made by Peterson and her colleagues as well. In other words, chil- dren may report misinformation for a variety of reasons, including some that have noth- ing to do with their memory of what actually occurred. For example, they may tell the in- terviewer the false details that had been sug- gested because they believe this is what the interviewer expects, rather than because they really believe what they are reporting.

One example from an actual case illus- trates both points. In a Jordan, Minnesota, case, Andy Meyers, an 11-year-old child, was removed from the family home and placed across town in an emergency foster placement because his father was suspected of abusing him (see Feher 1987). He steadfastly denied that his parents, or anyone else, had abused him. However, after three months of almost daily interrogations, he finally reported that he had been abused. He claimed that his par- ents held orgies in the woods and killed babies and then dumped them in a nearby river. His charges, along with those of 39 other children against 24 adults (many of whom were their parents), formed the basis of a long number of indictments that eventually were dropped because of the suggestive manner in which the children had been interviewed. Approxi- mately nine years after making these charges, Andy described his disclosures to a journalist as follows:

Andy: I finally just said (to investigators) “fine, yeah that happened.” Journalist: Why did you say “yes” that day? Andy: I have no idea. Probably cause I was just sick of being badgered. I didn’t think I was ever going home. I mean, I figured if this is gonna be the way life is, I might as well make it a little more tolerable for myself. (Feher 1987, p. 232)

Concerning the role of individual dif- ferences, numerous reviews have been pub- lished in recent years showing that various factors differentiate children’s suggestibility proneness, such as personality and cognitive variables (see Bruck & Melnyk 2004, Bruck

www.annualreviews.org • Unwarranted Assumptions 319

A nn

u. R

ev . C

li n.

P sy

ch ol

. 2 00

7. 3:

31 1-

32 8.

D ow

nl oa

de d

fr om

a rj

ou rn

al s.

an nu

al re

vi ew

s. or

g by

T ex

as T

ec h

U ni

ve rs

it y

- L

ub bo

ck o

n 02

/1 5/

08 . F

or p

er so

na l

us e

on ly

.

ANRV307-CP03-13 ARI 20 February 2007 19:5

et al. 1997, Roebers & Schnieder 2001). For example, there seems to be fairly consis- tent evidence that children with more ad- vanced language abilities are more resis- tant to suggestion, whereas children who are more creative are less resistant to sugges- tion. Other psychological factors such as in- telligence, memory abilities, theory of mind, and attachment status have been implicated in distinguishing children who are more ver- sus less likely to fall prey to suggestive in- fluences, although research with these other factors is more equivocal (Bruck & Melnyk 2004).

Recently, Gilstrap & Ceci (2004) have ar- gued that the degree to which children as- sent to false information in an interview is best predicted by the child’s own disposition rather than by the interviewer’s use of sugges- tive techniques. Specifically, they showed that when interviewing children about a game- like staged event, the best predictor of a child’s rate of incorporating false sugges- tions into his or her report was his or her prior behavior rather than the interviewer’s prior behavior. For example, lagged sequen- tial analyses showed that it was possible to predict a child’s subsequent answer directly from his or her prior answer, effectively skip- ping the intervening adult questioning. Al- though the interviewer’s use of misleading questioning still had deleterious effects on the child’s accuracy, the results demonstrated that a child’s answers were driven at least as much by his or her personal style as by the in- terviewer’s form of questioning. Researchers and expert witnesses need to be aware of these possibilities, although clearly this find- ing needs to be replicated under more realistic conditions.

One may conclude from this body of lit- erature that although suggestive influences can have a negative impact on the accu- racy of a child’s statement, these influences do not guarantee inaccuracy in all situations and for all children and, conversely, the ab- sence of these influences does not guarantee accuracy.

Unwarranted Assumption 6: False Reports Produced by Suggestive Interviewing Are Distinguishable from Accurate Reports

Expert witnesses often state that it is easy to detect false reports that are the result of sug- gestion because such children merely parrot the words of their interrogators or, conversely, that true reports integrate affect into their ac- counts (e.g., showing appropriate emotions). Consider two claims made by an expert wit- ness in Oregon v. Ryan Smith, a criminal case involving alleged sexual abuse:

Children can only provide information about body sensations if they have experi- enced the event. If children are suggested in- formation and later repeat this information but include with it sensory information then their reports must be accurate. The child’s accusations could not be the result of sug- gestions that he’d been sodomized because he reported that he could hardly walk back to his classroom. Children would not incorpo- rate this type of detail into their false reports. Their inclusion is another sign the state- ment is valid. Children who are lying merely parrot suggestions and are easier to detect because they are rehearsed. These children were going beyond anything an adult might have suggested. (transcript, p. 886)

When children have been coached to talk about sexual abuse, they talk about it in a different way than when it is genuine. Often there is no affect involved, they aren’t show- ing an emotional reactions. They don’t show any distress or anything. None of the chil- dren in this case talked spontaneously about details to give validity to their allegations. (transcript, p. 819)

However, there is no support for this asser- tion; in fact, most research to date suggests the contrary.

First, when children are suggestively in- terviewed, their subsequent narratives include false reports that go beyond what was sug- gested to them but are consistent with the

320 Ceci et al.

A nn

u. R

ev . C

li n.

P sy

ch ol

. 2 00

7. 3:

31 1-

32 8.

D ow

nl oa

de d

fr om

a rj

ou rn

al s.

an nu

al re

vi ew

s. or

g by

T ex

as T

ec h

U ni

ve rs

it y

- L

ub bo

ck o

n 02

/1 5/

08 . F

or p

er so

na l

us e

on ly

.

ANRV307-CP03-13 ARI 20 February 2007 19:5

suggestions (e.g., Bruck et al. 1995a, 2002). For example, Leichtman & Ceci (1995) found that when children were given false sugges- tions that a person was clumsy, they later generalized this information to make false claims that this individual spilled, ripped, soiled, or broke things during his visit to their classroom.

Second, subjective ratings of children’s re- ports after suggestive interviewing reveal that children who make false reports after being suggestively interviewed appear highly cred- ible to trained professionals in the fields of child development, mental health, and foren- sics (e.g., Ceci et al. 1994a,c; Leichtman & Ceci 1995); these professionals cannot reliably discriminate between children whose reports are accurate from those whose reports are in- accurate as a result of suggestive interview- ing techniques. The children who provided the false reports spoke sincerely and provided accounts laden with emotion and perceptual details.

Even when experts attempt to apply more systematic means to distinguish true from false reports, their decisions are not reli- able. Statement validity assessment, and more specifically criterion-based content analysis (CBCA), has been touted as one way to dis- tinguish true from false reports. In CBCA, ex- perts code the witness’s statement for the pres- ence of specific contents, such as the presence of peripheral details or mentions of the per- petrator’s state of mind. Each of the CBCA criteria is expected to appear more often in true reports than in false reports; thus, every instance that a criterion appears in a child’s report increases the likelihood that the re- port is valid (Raskin & Esplin 1991, Steller & Koehnken 1989). For example, a report that includes reproductions of dialogue, super- fluous details, spontaneous corrections, and mentions of the perpetrator’s state of mind would be more likely to be considered true than a statement that did not include these characteristics.

There is some limited evidence that CBCA can identify truthful statements from inten-

tional lies (Vrij 2005); however, CBCA can- not distinguish true statements from state- ments that were developed as the result of sug- gestive questioning. For example, Stromwell et al. (2004) found no significant differences in CBCA scores for real versus imagined events; CBCA scores correctly predicted 44% of real events and 64% of imagined events (see also Erdmann et al. 2004).

Moving beyond the specific criterion of CBCA, results of several studies (Bruck et al. 2002, Powell et al. 2003, Principe & Ceci 2002, Scullin et al. 2002) reveal that linguis- tic markers, such as the amount of elaboration or temporal connectivity, do not consistently differentiate true from false narratives that emerge as a result of repeated suggestive inter- views. In the Bruck et al. study (2002), analysis of children’s narratives of true and false events showed that narratives of the false events actu- ally contained more embellishments (includ- ing descriptions and emotional terms) and de- tails than their narratives of the true events. The false narratives also had more sponta- neous statements than the true narratives and included more temporal connectors.

Third, one cannot make judgments of re- liability based on features of the child’s nar- rative that are known to be true. That is, one cannot conclude that everything in the child’s report about abuse by his or her pediatrician must be accurate because he or she correctly described the layout of the pediatrician’s of- fice. This is because when children are sugges- tively interviewed, they often interweave the false suggestions with accurate details about their lives. For example, Hershkowitz (2001) reported the case study of a 10-year-old girl who, as a consequence of her mother’s sug- gestive questioning, falsely reported being the victim of a rape. In reality, the girl had been the victim of a man exposing his genitalia to her. She interwove the details of the actual expo- sure incident into her detailed false narrative of a rape.

These findings show that reliability and credibility are orthogonal dimensions. Cred- ibility refers only to the degree to which the

www.annualreviews.org • Unwarranted Assumptions 321

A nn

u. R

ev . C

li n.

P sy

ch ol

. 2 00

7. 3:

31 1-

32 8.

D ow

nl oa

de d

fr om

a rj

ou rn

al s.

an nu

al re

vi ew

s. or

g by

T ex

as T

ec h

U ni

ve rs

it y

- L

ub bo

ck o

n 02

/1 5/

08 . F

or p

er so

na l

us e

on ly

.

ANRV307-CP03-13 ARI 20 February 2007 19:5

witness is motivated to tell the truth but not to the accuracy of the witness’ statement (Ceci et al. 1995, Kohenken 1989, Poole & Lindsay 1998). Children may appear highly credible (or their interviews may have the character- istics of credible narratives), yet their reports may be unreliable. Children can also be re- liable yet not appear credible in the eyes of judges and jurors. Accordingly, one cannot use perceived credibility to judge reliability. However, in the forensic arena, it is crucial for those who make judgments about credi- bility (i.e., the court) to make these judgments based on the reliability of the report, in other words, details on the degree to which sugges- tive interviewing was used to elicit the child’s statements.

Unwarranted Assumption 7: Children’s Disclosures of Traumatic Events Are Delayed, Denied, and Often Recanted

A highly influential assumption is that sexu- ally abused children do not readily disclose abuse because of shame, guilt, and fear. Con- sequently, there are long delays between abuse and disclosure, and often there is never any disclosure. It has repeatedly been asserted that, when questioned, children will initially deny their abuse, but with repeated question- ing, they will slowly divulge its details. It is further claimed that these disclosures are fre- quently recanted, but with additional support their disclosures will be reinstated. Summit (1983) proposed the most popular embodi- ment of this model, which is termed the child sexual abuse accommodation syndrome.

Because models of the child sexual abuse accommodation syndrome were not based on empirical data but on clinical intuitions and unsystematically collected observations, we recently reviewed the literature to determine their empirical support (London et al. 2005). We identified ten studies in which adults with documented histories of childhood abuse were asked to recall their disclosures in child- hood. Results were consistent; on average

only 33% of the adults remembered disclos- ing the abuse during childhood. These data support the view that sexually abused children are silent about their victimization and de- lay disclosure for long periods of time. These data are also important in that they support the contention that prevalence data on sex- ual abuse (which are based on reports to offi- cial agencies) greatly underestimate the extent of the problem because so much abuse is not reported.

Although these studies are informative on the issue regarding the delay of reporting, the data are silent on issues of denial and recanta- tion by sexually abused children because the adults in these studies were never asked, “As a child, did anyone ever ask you or question you about abuse?” However, a second set of studies provides some relevant data on this point. We identified 17 studies that exam- ined rates of denial and recantation by sexually abused children who were, in fact, asked di- rectly about being abused when they were as- sessed or treated at clinics. The rates of denial at assessment interviews were highly variable (4% to 76%) as were the rates of recantation (4% to 27%). We found that the methodolog- ical adequacy of each study (sampling proce- dures, validation of sexual abuse) was related to denial and recantation rates, with the high- est rates being associated with the weakest studies. For the six methodologically superior studies, the average rate of denial was only 14%, and the average rate of recantation was only 7%. Thus, sexually abused children do not usually deny or recant the details of their abuse. In other words, although these children do not readily disclose their abuse, if they are asked about it in a structured interview, they will tell, and few will recant.

Even if it is conceded that children will not deny when asked, this does not address the issue of how to elicit disclosures. Even though there are known risks of using lead- ing or specific questions, perhaps these are necessary to elicit reports or details from sexually abused children who are frightened, ashamed, or feeling too guilty to reveal their

322 Ceci et al.

A nn

u. R

ev . C

li n.

P sy

ch ol

. 2 00

7. 3:

31 1-

32 8.

D ow

nl oa

de d

fr om

a rj

ou rn

al s.

an nu

al re

vi ew

s. or

g by

T ex

as T

ec h

U ni

ve rs

it y

- L

ub bo

ck o

n 02

/1 5/

08 . F

or p

er so

na l

us e

on ly

.

ANRV307-CP03-13 ARI 20 February 2007 19:5

victimization. A series of studies conducted by Lamb, Sternberg, and colleagues has recently challenged this claim. This team constructed a structured interview protocol and then trained interviewers of suspected child abuse victims in its use. The major characteristic of the protocol is to encourage the child to provide detailed life-event narratives through the guidance of open-ended questions (e.g., “Tell me what happened.” “And what hap- pened next?” “And you said it happened at the store. Tell me about the store.”). The use of specific questions is allowed only after exhaus- tive free-recall prompting. Suggestive ques- tions are highly discouraged. In their latest study, Lamb et al. (2003) examined the inter- views of 16 trained police officers with 130 children (four to eight years old), all of whom had made allegations of sexual abuse. They found that 83% of all allegations/disclosures were elicited through free-recall questions (78% for preschoolers), and 66% of all chil- dren identified the suspect through open- ended question (60% for preschoolers). These data dispel the belief that children need to be bombarded with suggestive techniques to elicit details of their traumatic events; in fact, children can provide detailed informa- tion through open-ended prompts, and if a child denies abuse when asked directly, there is no scientifically compelling evidence that the child is in denial. Abused children usually disclose when directly asked.

Unwarranted Assumption 8: Laboratory Research Is Not an Accurate Reflection of Child Witnesses’ Experiences in the Real World

The issues and themes in this section do not come from basic researchers in the field of children’s suggestibility but instead reflect the concerns and beliefs of expert witnesses and professionals who advocate for sexually abused children. The central theme of the ar- gument is that much of the science (some of which is reviewed above) is not ecologically

valid because the science either overestimates children’s suggestibility or underestimates the capabilities of the child witness. Research ac- cused of overestimating children’s suggestibil- ity has been referred to as new wave to distin- guish its investigators from prior cohorts who found children to be less suggestible (Lyon 1999).

The essence of the first argument is that laboratory studies conducted by new-wave researchers yield high rates of suggestibility in children because these researchers employ extreme or coercive tactics that are seldom found in real-world interviews of children. For example, Lyon (1999) wrote,

In sum, “leading” questions are certainly common in investigative interviews. Yet the new wave’s research goes far beyond ask- ing leading questions in assessing children’s suggestibility. The limited observational ev- idence available indicates that the new-wave methods are not, in fact, common investiga- tive techniques. The experimental evidence indicates that such techniques are largely re- sponsible for the impressive demonstrations of suggestibility that the new wave has pro- duced. (Lyon 1999, p. 15)

Following Lyon’s views, Dr. Dalenberg’s testimony for the prosecution In the Matter of Riley, Shelby, and Austin Blanchard v. John Blanchard (2001) reflected her view that in the absence of extremely coercive interview- ing tactics, children’s statements are highly accurate:

The children are 95–98% accurate typically—often 100% accurate with abuse-related, misleading questions. When they become inaccurate is in these coercive studies . . . in which the children are asked six or seven leading questions in a row, contradicted every time they say something positive, frowned at when they say some- thing positive, shown evidence against the individual, told that other people know the guy is guilty. Those kinds of extremely

www.annualreviews.org • Unwarranted Assumptions 323

A nn

u. R

ev . C

li n.

P sy

ch ol

. 2 00

7. 3:

31 1-

32 8.

D ow

nl oa

de d

fr om

a rj

ou rn

al s.

an nu

al re

vi ew

s. or

g by

T ex

as T

ec h

U ni

ve rs

it y

- L

ub bo

ck o

n 02

/1 5/

08 . F

or p

er so

na l

us e

on ly

.

ANRV307-CP03-13 ARI 20 February 2007 19:5

Table 1 Some rates of suggestive utterances by forensic interviewers

Study Mean number of

suggestive utterances % of total utterances

Lamb et al. (1996) 14 8.75 Sternberg et al. (1996) 9.5 9.2 Sternberg et al. (1997) 8.75 8 Sternberg et al. (1999) 7.6 6.5 Hershkowitz (1999) 11 8.8 Orbach et al. (2000) 9.1 10 Overall mean 9.99 8.54

coercive circumstances are necessary to get children to—to do this. (transcript, p. 39)

Thus, the insinuation among some who come to court is that studies of children’s sug- gestibility are unrealistic because they utilize high-pressure, coercive tactics not found in actual forensic interviews. Proponents of this argument, however, do not present any data to back their claim that the techniques used in the laboratory do not mirror those used in real-world forensic interviews. In fact, in the real world, children are often exposed to similarly highly suggestive questioning (see Ceci & Friedman 2000 for evidence for this claim). Table 1 reports the average sugges- tions for a number of field studies of so- cial workers and law-enforcement profession- als. As shown, typically between 6.5% and 10% of all field utterances can be classified as suggestive questions. This translates into approximately 10 suggestions per interview. This is more than many laboratory studies have employed that are associated with high rates of suggestibility errors. Thus, research that utilizes frequent suggestive questioning is appropriate for generalization to field-based interviews.

Moreover, most scientific studies produc- ing high rates of suggestibility have not em- ployed coercive or high-pressured techniques such as positive or negative reinforcement or peer pressure. In the majority of these stud- ies, as shown above, children are suggestible even when the interviewer is neutral and sup-

portive. This bears noting because if children succumb to noncoercive questioning that is commonly employed in research, they can be expected to succumb even more so in response to the types of coercive questioning observed in some field-based interviews.

Rather than arguing whether the condi- tions of a specific study reflect practices in the real world (which are highly variable), it is more logical to argue from the facts of the interviews from a specific case. If a child ini- tially denied abuse and then only gradually disclosed some details after multiple inter- views with many suggestive elements, then it is entirely appropriate to use the suggestibil- ity studies of the new wave. For ethical rea- sons involved in research with children, these studies of children’s suggestibility most likely underestimate how a combination of motives, strong suggestions, threats, and inducements that have been documented in actual foren- sic interviews might lead a child to make an inaccurate report. Thus, the value of the sci- ence depends on the specifics of the case to which it is applied. Blanket statements about the practical significance of a study or the rel- evance of scientific findings are meaningless in the abstract.

CONCLUSION

Despite the substantial advances that develop- mental forensic researchers have made to both application and theory, there remain miscon- ceptions that border on urban legends. In this review, we unveil eight unwarranted assump- tions that in each case have made their way into expert testimony in court cases. We show that the sheer number of leading questions does not indicate the suggestiveness of an in- terview, and interviewer bias, which may con- tribute an interviewer’s utilization of other suggestive factors, must be taken into account. In addition, young children are not alone in their susceptibility to suggestive techniques; older children and adults can be quite sus- ceptible as well. In addition, extremely coer- cive techniques are not necessary to produce

324 Ceci et al.

A nn

u. R

ev . C

li n.

P sy

ch ol

. 2 00

7. 3:

31 1-

32 8.

D ow

nl oa

de d

fr om

a rj

ou rn

al s.

an nu

al re

vi ew

s. or

g by

T ex

as T

ec h

U ni

ve rs

it y

- L

ub bo

ck o

n 02

/1 5/

08 . F

or p

er so

na l

us e

on ly

.

ANRV307-CP03-13 ARI 20 February 2007 19:5

tainted reports, and furthermore, even free re- call is not error free in both younger and older children. At the same time, children will not always report false information if they have been exposed to suggestive techniques; con- textual and individual factors are important and must be taken into consideration. How- ever, when a child does produce an erroneous report as a result of suggestive techniques, these reports often appear highly credible, and there is no scientifically established means to distinguish these false reports from true ones. There is also no compelling evidence to sup- port the belief that victims of abuse deny their abuse when they are initially questioned about it, nor do they usually recant, as some have claimed. Thus, arguments in favor of relent- lessly pursuing reticent interviewees find no scientific support. Finally, we argue that sci- entific findings may be quite pertinent to real- world cases of child sexual abuse and should not be ignored by the courts, expert witnesses, or child advocates.

The literature presented above may seem ambiguous or even self-contradictory at some points. For example, on the one hand we show that children may develop false reports in the absence of extreme suggestive techniques, but on the other hand, we review studies showing that extreme techniques do not ineluctably lead to false reports. How should these am- biguities and discrepancies be reconciled to apply the scientific research in actual foren- sic settings? We suggest practitioners, inter- viewers, and expert witnesses must take into account the qualifying role of context, and

further research must consider how the child may interact in a given context. A child’s re- port, its degree of embellishment of central versus peripheral details, and its narrative co- herence and consistency are the product of the full history of the child’s involvement with interviewers as well as the child’s own par- ticular psychological characteristics. In this regard, it is important to note that tainting misinformation can occur prior to the first officially recorded interview. For example, a child may be interviewed informally by a par- ent, teacher, social worker, or mental health professional before any formal interviews take place. Even hearsay reports and suggestions from informal sources (e.g., overheard con- versations, rumors, and even interactions with others who overheard suggestions) are capa- ble of tainting a child’s report, and when this occurs, the tainted report may appear indis- tinguishable from true reports.

In summary, although tainting of reports is neither easy nor inevitable, it is also not that difficult to accomplish, especially at the hands of a biased interviewer. Researchers and expert witnesses alike must refrain from the temptation to assert specialized knowledge that permits them to determine the reliability of a child’s claim, based solely on that child’s responses in a neutral interview that may have followed unrecorded suggestive ones. Fur- thermore, forensic professionals should avoid characterizing a child’s testimony as veridi- cal based solely on group norms and averages without giving full consideration to the con- text of the child’s testimony.

LITERATURE CITED

Binet A. 1900. La Suggestibilité. Paris: Schleicher Freres

Bruck M, Ceci SJ. 2004. Forensic developmental psychology: unveiling four common miscon- ceptions. Curr. Direct. Psychol. Sci. 13:229–32

Bruck M, Ceci SJ, Francoer E. 1999. The accuracy of mothers’ memories of conversations with their preschool children. J. Exp. Psychol. Appl. 5:89–106

Bruck M, Ceci SJ, Francouer E, Barr R. 1995a. “I hardly cried when I got my shot”: influencing children’s reports about a visit to their pediatrician. Child Dev. 66:193–208

www.annualreviews.org • Unwarranted Assumptions 325

A nn

u. R

ev . C

li n.

P sy

ch ol

. 2 00

7. 3:

31 1-

32 8.

D ow

nl oa

de d

fr om

a rj

ou rn

al s.

an nu

al re

vi ew

s. or

g by

T ex

as T

ec h

U ni

ve rs

it y

- L

ub bo

ck o

n 02

/1 5/

08 . F

or p

er so

na l

us e

on ly

.

ANRV307-CP03-13 ARI 20 February 2007 19:5

Bruck M, Ceci SJ, Francoeur E, Renick A. 1995b. Anatomically detailed dolls do not facilitate preschoolers’ reports of a pediatric examination involving genital touching. J. Exp. Psychol. Appl. 1:95–109

Bruck M, Ceci SJ, Hembrooke H. 2002. Nature of true and false narratives. Dev. Rev. 22:520–54 Bruck M, London K, Landa R, Goodman J. 2007. Autobiographical memory and suggestibility.

Dev. Psychopathol. 19:73–95 Bruck M, Melnyk L. 2004. Individual differences in children’s suggestibility: a review and

synthesis. Appl. Cogn. Psychol. 18:947–96 Bruck M, Melnyk L, Ceci SJ. 1997. External and internal sources of variation in the creation

of false reports in children. Learn. Individ. Differ. 9:289–316 Ceci SJ, Bruck M. 1995. Jeopardy in the Courtroom. Washington, DC: Am. Psychol. Assoc. Ceci SJ, Bruck M. 1993. The suggestibility of children’s recollections: an historical review and

synthesis. Psychol. Bull. 113:403–39 Ceci SJ, Bruck M, Rosenthal R. 1995. Children’s allegations of sexual abuse: forensic and

scientific issues: a reply to commentators. Psychol. Public Policy Law 1:494–520 Ceci SJ, Crotteau-Huffman M, Smith E, Loftus EW. 1994a. Repeatedly thinking about non-

events. Conscious. Cogn. 3:388–407 Ceci SJ, Friedman RD. 2000. The suggestibility of children: scientific research and legal im-

plications. Cornell Law Rev. 86:34–108 Ceci SJ, Loftus EF, Leichtman M, Bruck M. 1994b. The possible role of source misattributions

in the creation of false beliefs among preschoolers. Int. J. Clin. Exp. Hypn. 42:304–20 Ceci SJ, Loftus EW, Leichtman M, Bruck M. 1994c. The role of source misattributions in the

creation of false beliefs among preschoolers. Int. J. Clin. Exp. Hypn. 62:304–20 Ceci SJ, Papierno PB, Kulkofsky S. 2007. Representational constraints on children’s suggestibil-

ity. Psychol. Sci. In press Ceci SJ, Ross D, Toglia M. 1987. Suggestibility of children’s memory: psycholegal implications.

J. Exp. Psychol. Gen. 116:38–49 Erdmann K, Volbert R, Bohm C. 2004. Children report suggested events even when inter-

viewed in a nonsuggestive manner: What are the implications for credibility assessment? Appl. Cogn. Psychol. 18:589–611

Feher B. 1987. The alleged molestation victim, the rules of evidence, and the constitution: Should children really be seen and not heard? Am. J. Crim. Law 14:227, 231–33

Finnila K, Mahlberga N, Santtilaa P, Sandnabbaa K, Niemib P. 2003. Validity of a test of children’s suggestibility for predicting responses to two interview situations differing in their degree of suggestiveness. J. Exp. Child Psychol. 85:32–49

Garven S, Wood JM, Malpass R, Shaw JS. 1998. More than suggestion: consequences of the interviewing techniques from the McMartin preschool case. J. Appl. Psychol. 83:347–59

Garven S, Wood JM, Malpass RS. 2000. Allegations of wrongdoing: the effects of reinforce- ment on children’s mundane and fantastic claims. J. Appl. Psychol. 1:38–49

Gilstrap L, Ceci SJ. 2004. Reconceptualizing children’s suggestibility: bidirectional and tem- poral properties. Child Dev. 75:1–14

Goodman GS, Quas JA, Batterman-Faunce JM, Riddlesberger MM, Kuhn J. 1994. Predic- tors of accurate and inaccurate memories of traumatic events experienced in childhood. Conscious. Cogn. 3:269–94

Greenhoot AF. 2000. Remembering and understanding: the effects of changes in underlying knowledge on children’s recollections. Child Dev. 71:1309–28

Hershkowitz I. 1999. The dynamics of interviews involving plausible and implausible allega- tions of child sexual abuse. Appl. Dev. Sci. 3:86–91

326 Ceci et al.

A nn

u. R

ev . C

li n.

P sy

ch ol

. 2 00

7. 3:

31 1-

32 8.

D ow

nl oa

de d

fr om

a rj

ou rn

al s.

an nu

al re

vi ew

s. or

g by

T ex

as T

ec h

U ni

ve rs

it y

- L

ub bo

ck o

n 02

/1 5/

08 . F

or p

er so

na l

us e

on ly

.

ANRV307-CP03-13 ARI 20 February 2007 19:5

Hershkowitz I. 2001. A case study of child sexual false allegation. Child Abuse Negl. 25:1397–411 Home Off. Dept. Health. 1992. Memorandum of Good Practice on Video Recorded Interviews with

Child Witnesses for Criminal Proceedings. London: Her Majesty’s Station. Off. Hyman IEJ, Husband TH, Billings JF. 1995. False memories of childhood experiences. Appl.

Cogn. Psychol. 9:181–97 In the Matter of Riley, Austin, and Shelby Blanchard v. John Blanchard. July 2001. Calif. Sup. Ct.,

CK40772/CK40775 In the Matter of Ryan D. Smith. May 30, 2001. State of Oregon Appelate Ct., #98–0325-JA;

A106007 Kohenken G. 1989. Behavioral correlates of statement credibility: theories, paradigms, and

results. In Criminal Behavior and the Justice System: Psychological Perspectives, ed. H Wegener, F Leosel, J Haisch, pp. 271–89. New York: Springer-Verlag

Lamb ME, Sternberg KJ, Orbach Y, Esplin PW, Stewart H, Mitchell S. 2003. Age differences in young children’s responses to open-ended invitations in the course of forensic interviews. J. Consult. Clin. Psychol. 71:926–34

Lamb ME, Hershkowitz I, Sternberg KJ, Esplin PW, Hovav M, et al. 1996. Effects of inves- tigative style on Israeli children’s responses. Int. J. Behav. Dev. 19:627–37

Leichtman MD, Ceci SJ. 1995. The effects of stereotypes and suggestions on preschoolers’ reports. Dev. Psychol. 31:568–78

Lindberg M. 1991. An interactive approach to assessing the suggestibility and testimony of eye- witness. In The Suggestibility of Children’s Recollections, ed. J Doris, pp. 47–55. Washington, DC: Am. Psychol. Assoc.

Loftus EF. 2004. Memories of things unseen. Curr. Direct. Psychol. Sci. 13:145–47 Loftus EF, Pickrell JE. 1995. The formation of false memories. Psychiatr. Ann. 25:720–25 London K, Bruck M, Ceci SJ, Shuman D. 2005. Disclosure of child sexual abuse: What does

the research tell us about the ways that children tell? Psychol. Public Policy Law 11:194–226 Lyon TD. 1999. The new wave in children’s suggestibility research: a critique. Cornell Law

Rev. 86:1–84 Marche TA. 1999. Memory strength affects reporting misinformation. J. Exp. Child Psychol.

73:45–71 Melnyk L, Bruck M. 2004. Timing moderates the effect of repeated suggestive interviewing

on children’s suggestibility. Appl. Cogn. Psychol. 18:613–31 Natl. Soc. Prev. Cruel. Child. Childline. 1993. The Child Witness Pack. London: Nat. Soc. Prev.

Cruel. Child. Orbach Y, Hershkowitz I, Lamb ME, Epslin PE, Horowitz D. 2000. Assessing the value of

structured protocols for forensic interviews of alleged sexual abuse victims. Child Abuse Negl. 24:733–52

Ornstein PA, Merritt KA, Baker-Ward K, Furtado E, Gordon BN, Principe G. 1998. Children’s knowledge, expectation, and long-term retention. Appl. Cogn. Psychol. 12:387–405

Peterson C, Parsons T, Dean M. 2004. Providing misleading and reinstatement information a year after it happened: effects on long-term memory. Memory 12:1–13

Poole DA, Lindsay DS. 1998. Assessing the accuracy of young children’s reports: lessons from the investigation of child sexual abuse. Appl. Prevent. Psychol. 7:1–26

Poole D, Lindsay DS. 2001. Children’s eyewitness reports after exposure to misinformation from parents. J. Exp. Psychol. Appl. 7:27–50

Poole DA, White LT. 1993. Two years later: effects of question repetition and retention interval on the eyewitness testimony of children and adults. Dev. Psychol. 29:844–53

Powell MB, Jones CH, Campbell C. 2003. A comparison of preschooler’s recall of experienced versus nonexperienced events across multiple interviewers. Appl. Cogn. Psychol. 17:935–52

www.annualreviews.org • Unwarranted Assumptions 327

A nn

u. R

ev . C

li n.

P sy

ch ol

. 2 00

7. 3:

31 1-

32 8.

D ow

nl oa

de d

fr om

a rj

ou rn

al s.

an nu

al re

vi ew

s. or

g by

T ex

as T

ec h

U ni

ve rs

it y

- L

ub bo

ck o

n 02

/1 5/

08 . F

or p

er so

na l

us e

on ly

.

ANRV307-CP03-13 ARI 20 February 2007 19:5

Principe GF, Ceci SJ. 2002. “I saw it with my own ears”: the influence of peer conversations and suggestive questions on preschoolers’ event memory. J. Exp. Child Psychol. 83:1–25

Principe GF, Kanaya T, Ceci SJ, Singh M. 2006. Believing is seeing: how rumors can engender false memories in preschoolers. Psychol. Sci. 17(3):243–48

Raskin DC, Esplin PW. 1991. Statement validity assessment: interview procedures and content analysis of children’s statements of sexual abuse. Behav. Assess. 13:265–91

Roebers CM, Schnieder W. 2001. Individual differences in children’s eyewitness recall: the influence of intelligence and shyness. Appl. Dev. Sci. 5:9–21

Scullin M, Kanaya T, Ceci SJ. 2002. Measurement of individual differences in children’s sug- gestibility across situations. J. Exp. Psychol. Appl. 8:233–46

Scullin MH, Ceci SJ. 2001. A suggestibility scale for children. Personal. Individ. Differ. 30:843– 56

Smith BE, Goretsky-Elstein S. 1993. The Prosecution of Child Sexual and Physical Abuse Cases. Washington, DC: Am. Bar Assoc.

State of Calif. v. Buckey. 1990. Sup. Ct., Los Angeles County, #A750900 Steller M, Koehnken G. 1989. Criteria-based statement analysis. In Psychological Methods in

Criminal Investigation and Evidence, ed. DC Raskin, pp. 217–45. New York: Springer Publ. Sternberg KJ, Lamb ME, Hershkowitz I. 1996. Child sexual abuse investigations in Israel.

Crim. Justice Behav. 23:322–37 Sternberg KJ, Lamb ME, Hershkowitz I. 1997. Effects of introductory style on children’s

abilities to describe experiences of sexual abuse. Child Abuse Neglect 21:1130–41 Sternberg KJ, Lamb ME, Esplin PW, Baradaran L. 1999. Using a scripted protocol in inves-

tigative interviews: a pilot study. Appl. Dev. Sci. 3:70–76 Stromwell LA, Bengtsson L, Leander L, Granhag PA. 2004. Assessing children’s statements: the

impact of a repeated experience on CBCA and RM ratings. Appl. Cogn. Psychol. 18:653–68 Summit R. 1983. The child sexual abuse accommodation syndrome. Child Abuse Neglect 7:177–

93 Thompson WC, Clarke-Stewart KA, Lepore SJ. 1997. What did the janitor do? Suggestive

interviewing and the accuracy of children’s accounts. Law Hum. Behav. 21:405–26 Vrij A. 2005. Criteria-based content analysis: a qualitative review of the first 37 studies. Psychol.

Public Policy Law 11:3–41 Wellman HM, Cross D, Watson J. 2001. Meta-analysis of theory-of-mind development: the

truth about false belief. Child Dev. 72:655–84 White TL, Leichtman MD, Ceci SJ. 1997. The good, the bad, and the ugly: accuracy, inac-

curacy, and elaboration in preschoolers’ reports about a past event. Appl. Cogn. Psychol. 11:S37–54

Zaragoza M, Payment K, Kichler J, Stines L, Drivdahl S. 2001. Forced confabulation and false memory in child witnesses. Presented at Bienn. Meet. Soc. Res. Child Dev. Minneapolis, MN

328 Ceci et al.

A nn

u. R

ev . C

li n.

P sy

ch ol

. 2 00

7. 3:

31 1-

32 8.

D ow

nl oa

de d

fr om

a rj

ou rn

al s.

an nu

al re

vi ew

s. or

g by

T ex

as T

ec h

U ni

ve rs

it y

- L

ub bo

ck o

n 02

/1 5/

08 . F

or p

er so

na l

us e

on ly

.

AR307-FM ARI 2 March 2007 14:4

Annual Review of Clinical Psychology

Volume 3, 2007Contents

Mediators and Mechanisms of Change in Psychotherapy Research Alan E. Kazdin � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 1

Evidence-Based Assessment John Hunsley and Eric J. Mash � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 29

Internet Methods for Delivering Behavioral and Health-Related Interventions (eHealth) Victor Strecher � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 53

Drug Abuse in African American and Hispanic Adolescents: Culture, Development, and Behavior José Szapocznik, Guillermo Prado, Ann Kathleen Burlew, Robert A. Williams, and Daniel A. Santisteban � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 77

Depression in Mothers Sherryl H. Goodman � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 107

Prevalence, Comorbidity, and Service Utilization for Mood Disorders in the United States at the Beginning of the Twenty-first Century Ronald C. Kessler, Kathleen R. Merikangas, and Philip S. Wang � � � � � � � � � � � � � � � � � � � � � 137

Stimulating the Development of Drug Treatments to Improve Cognition in Schizophrenia Michael F. Green � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 159

Dialectical Behavior Therapy for Borderline Personality Disorder Thomas R. Lynch, William T. Trost, Nicholas Salsman, and Marsha M. Linehan � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 181

A Meta-Analytic Review of Eating Disorder Prevention Programs: Encouraging Findings Eric Stice, Heather Shaw, and C. Nathan Marti � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 207

Sexual Dysfunctions in Women Cindy M. Meston and Andrea Bradford � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 233

Relapse and Relapse Prevention Thomas H. Brandon, Jennifer Irvin Vidrine, and Erika B. Litvin � � � � � � � � � � � � � � � � � � � 257

vii

A nn

u. R

ev . C

li n.

P sy

ch ol

. 2 00

7. 3:

31 1-

32 8.

D ow

nl oa

de d

fr om

a rj

ou rn

al s.

an nu

al re

vi ew

s. or

g by

T ex

as T

ec h

U ni

ve rs

it y

- L

ub bo

ck o

n 02

/1 5/

08 . F

or p

er so

na l

us e

on ly

.

AR307-FM ARI 2 March 2007 14:4

Marital and Family Processes in the Context of Alcohol Use and Alcohol Disorders Kenneth E. Leonard and Rina D. Eiden � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 285

Unwarranted Assumptions about Children’s Testimonial Accuracy Stephen J. Ceci, Sarah Kulkofsky, J. Zoe Klemfuss, Charlotte D. Sweeney, and Maggie Bruck � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 311

Expressed Emotion and Relapse of Psychopathology Jill M. Hooley � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 329

Sexual Orientation and Mental Health Gregory M. Herek and Linda D. Garnets � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 353

Coping Resources, Coping Processes, and Mental Health Shelley E. Taylor and Annette L. Stanton � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 377

Indexes

Cumulative Index of Contributing Authors, Volumes 1–3 � � � � � � � � � � � � � � � � � � � � � � � � � � � 403

Cumulative Index of Chapter Titles, Volumes 1–3 � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 405

Errata

An online log of corrections to Annual Review of Clinical Psychology chapters (if any) may be found at http://clinpsy.AnnualReviews.org

viii Contents

A nn

u. R

ev . C

li n.

P sy

ch ol

. 2 00

7. 3:

31 1-

32 8.

D ow

nl oa

de d

fr om

a rj

ou rn

al s.

an nu

al re

vi ew

s. or

g by

T ex

as T

ec h

U ni

ve rs

it y

- L

ub bo

ck o

n 02

/1 5/

08 . F

or p

er so

na l

us e

on ly

.