2 assignments
L a w and Human Behavior, Vol. 17, N o . 4, 1993
Anatomically Detailed Dolls Evaluation of Child Sexual Abuse Allegations
and the
Psychometric Considerations*
Linda J. Skinner and Kenneth K. Berry'~
Although anatomically detailed dolls are becoming the most commonly used tool in the validation of sexual abuse allegations, their use is not without significant problems. This evaluation o f the psycho- metric properties of the dolls using the existing empirical literature reveals that there is considerable difficulty in the standardization of the dolls, administration and scoring procedures, and training in the dolls' use. Furthermore, appropriate norms have not been developed. Although good interobserver reliability has been reported, the construct and criterion-related validity has not been demonstrated consistently. The current lack o f an empirically based psychometric foundation does not support the use o f the dolls in validation interviews, nor the admissibility o f doll-based evidence in legal proceed- ings.
With the exponential increase in child sexual abuse allegations, legal and mental health professionals have actively sought new procedures and tests for the eval- uation of suspected abuse cases. Although traditional interview procedures con- tinue to be used, problems in these validation interviews (Berry, 1985) have led to a search for alternative, nonverbal techniques, including the use of drawings,
* The authors greatly appreciate the critical reviews offered by Walter J. Danforth, Ph.D. and the anonymous reviewers and the editorial assistance o f Melissa Giles. An earlier version o f this paper, entitled " P l a y or Reality: Anatomical Dolls in Child Sexual Molestation Investigations," was pre- sented at the annual meeting o f the American Psychological Association in N e w Orleans on August 13, 1989. Requests for reprints should be sent to Linda J. Skinner, Department o f Psychology, University o f Arkansas, Fayetteville, tMR 72701-1201. Chicago School o f Professional Psychology.
399
014%7307/93/0800-0399507.00/0 �9 1993 Plenum Publishing Corporation
T hi
s do
cu m
en t i
s co
py ri
gh te
d by
th e
A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. T
hi s
ar tic
le is
in te
nd ed
s ol
el y
fo r t
he p
er so
na l u
se o
f t he
in di
vi du
al u
se r a
nd is
n ot
to b
e di
ss em
in at
ed b
ro ad
ly .
400 S K I N N E R A N D B E R R Y
puppets, doll houses, stuffed animals, and anatomically detailed (AD) dolls, to augment the interviews. Since their introduction in the late 1970s, AD dolls have b e c o m e "'the assessment tool" (White, 1988, p. 472) in child sexual abuse inves- tigations. Recently the American Psychological Association (APA) Council o f Representatives adopted a statement concluding that AD dolls " m a y be the best available practical solution" (p. 722) to the serious problem o f validating child sexual abuse allegations (Fox, 1991).
T H E N E E D F O R C O N S I D E R A T I O N O F T H E P S Y C H O M E T R I C P R O P E R T I E S O F A D D O L L S
Although m u c h has been written about AD dolls and their use, the psycho- metric properties o f the dolls have not been adequately considered. Six factors underlie the importance o f such a psychometric evaluation o f the use o f AD dolls in the validation o f suspected child sexual abuse.
1. AD Dolls as a Psychological Test. In March o f 1988, APA's Committee on Psychological Testing and Assessment asserted that AD dolls are "'a psychological test and are subject to the standards when used to assess individuals and make inferences about their behavior" (Landers, 1988, p. 25).
2. The Popularity o f the Dolls. AD dolls have become one o f the most used tests to validate allegations o f child sexual abuse. In one survey, 92% o f 212 professionals who assess such allegations reported using the dolls (Conte, Soren- son, Fogarty, & Rosa, 1991).
3. Fulfillment o f Professional Responsibilities. Test developers, distributors, and users are expected to follow the standards outlined in the Standards for Educational and Psychological Testing (APA, 1985). However, because, as some investigators such as Phillippus and Koch (1986) point out, AD dolls were devel- oped independently o f these standards and are distributed in ways that conflict with APA tenets, it becomes even more critical that psychologists using AD dolls consider the dolls' psychometric properties.
4. Courtroom Requirements. Some courts (In re Amber B. and Teela B., 1987; In re Christine C. and Michael C., 1987) have ruled AD doll-based evidence as inadmissible on the basis o f the K e l l y - F r y e rule (Frye v. United States, 1923; People v. Kelly, 1976), the legal standard that requires that the relevant scientific c o m m u n i t y accept the demonstrated reliability and validity o f a device or test used as a basis for expert opinions or conclusions. Not particularly receptive to t h e argument that application o f the K e l l y - F r y e rule is inappropriate in that the use o f dolls has been a " t i m e - h o n o r e d " psychological evaluation tool rather than a n e w scientific test (Snyder, 1988), the courts are demanding a consideration o f t h e psychometric properties o f A D dolls as validation tools.
5. Seriousness o f Decisions. The consequences to the child (Lewis, 1992), the accused, their respective families, and society at large o f decisions made, in part, using AD-doll-based data are far too serious to ignore the scientific evaluation o f the dolls.
T hi
s do
cu m
en t i
s co
py ri
gh te
d by
th e
A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. T
hi s
ar tic
le is
in te
nd ed
s ol
el y
fo r t
he p
er so
na l u
se o
f t he
in di
vi du
al u
se r a
nd is
n ot
to b
e di
ss em
in at
ed b
ro ad
ly .
PSYCHOMETRICS OF ANATOMICAL DOLLS 401
6. Additional Research Needs. A review o f the currently available evidence regarding AD dolls and their psychometric properties when used in the validation o f the child sexual abuse allegations will highlight additional research needs and allow clinical researchers to direct their efforts in ways that m a y improve the use of the dolls.
T E C H N I C A L S T A N D A R D S F O R P S Y C H O L O G I C A L A S S E S S M E N T
Originally intended to facilitate a child's report about sexual abuse and, thereby, increase accurate reporting, the use o f AD dolls in the validation of child sexual abuse allegations is a relatively r e c e n t development and based upon an intuitive and simple belief: Sexually abused children manifest "inappropriate" sexual behavior when playing with AD dolls, and this precocious and/or aberrant sexual play is the result o f sexual abuse or exploitation. The face validity o f these dolls is very strong and convincing to individuals unfamiliar with psychometric concepts. However, face validity, which is generally not considered a legitimate category o f validity, is an inadequate basis for evaluating an assessment tool. The goal o f this article is to evaluate the psychometric properties o f AD dolls using the existing empirical findings and, as a result, address both the use o f the dolls in sexual abuse evaluations and the use o f AD-doll-based evidence in court. The four psychometric properties to be considered are standardization, norms, reliability, and validity.
The S t a n d a r d i z a t i o n o f A n a t o m i c a l l y D e t a i l e d Dolls
Four factors related to the standardization o f AD dolls include materials, administration procedures, scoring, and training for use o f the dolls. Each o f these factors is examined below.
Materials
Ideally, a complete set o f AD dolls includes a mature male with a penis, scrotum, and pubic hair; a mature female with developed breasts, a vagina, and pubic hair; a young male with a penis and scrotum but no pubic hair; and a young female with a vagina but without developed breasts and pubic hair. The dolls should also have oral and anal orifices and separatable fingers, and all openings should a c c o m m o d a t e the mature male doll's penis. However, a frequent criticism o f AD dolls is directed at the heterogeneous collection o f these dolls in use today. F o r example, whereas the majority o f the dolls used by surveyed mental health professionals and child protective workers had vaginal openings, only 17% o f t h e dolls used b y law enforcement respondents had such openings (Boat & Everson, 1988c).
Only about one half o f the dolls used in child sexual abuse investigations are manufactured primarily for that purpose (Boat & Everson, 1988c). Standardiza- tion across the specially manufactured dolls is limited. There are at least 15 different AD doll manufacturers and distributors (White & Santilli, 1988), and no
T hi
s do
cu m
en t i
s co
py ri
gh te
d by
th e
A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. T
hi s
ar tic
le is
in te
nd ed
s ol
el y
fo r t
he p
er so
na l u
se o
f t he
in di
vi du
al u
se r a
nd is
n ot
to b
e di
ss em
in at
ed b
ro ad
ly .
402 SKINNER AND BERRY
established manufacturing standards exist. Thus, these dolls vary considerably both in general appearance and specific characteristics. The remaining dolls in use are genitally neutral dolls that are incompletely detailed (e.g., Barbie dolls), in- completely modified (e.g., breasts or a penis sewn on Cabbage Patch dolls), or home-designed models.
One intentional variation in doll characteristics is the race or cultural back- ground of the dolls. They have been manufactured to include features and colors representing Black and White races and Hispanic, Mexican, and Oriental cultural backgrounds (White & Santilli, 1988). Some practitioners recommend that a child's and the dolls' race/cultural background be matched (Boat & Everson, 1986; Friedemann & Morgan, 1985). However, no systematic research has ad- dressed the importance and impact of such matching.
There has been considerable criticism regarding the accuracy of the size of the genitals of AD dolls (Berliner, 1988; Phillippus & Koch, 1986; McIver, Wake- field, & Underwager, 1989). While such concern may have been applicable to some of the earlier designs, Bays (1990) demonstrated that the primary and sec- ondary sexual characteristics of many commercially manufactured and handmade dolls are not distorted. Extrapolation of breasts and genitalia measurements of 17 sets of AD dolls were consistent with adult human proportions.
Procedures for Administration
When first introduced, AD dolls were characterized as an unstructured, pro- jectivelike test requiring interpretation by a clinician (Glasgow & Bentall, 1989). As use of the dolls spread, the need for a systematic structured approach was stressed, and suggestions and recommendations (Boat & Everson, 1988a,b; Jones & McQuiston, 1988; Walker, 1988) and formal structured protocols (Boat & Ever- son, 1986; Friedemann & Morgan, 1985; White, Strom, Santilli, & Quinn, 1987) for the use of AD dolls were forwarded. To date, there are no available data regarding the utilization and popularity of these structured protocols.
Despite these attempts to standardize use of AD dolls, there remain many questions about administrative procedures. White and Santilli (1988) discussed inconsistencies in presentation procedures (i.e., when and how to use them, the number of interviews needed, the presence of other adults, the number of dolls used, and whether the dolls should be presented dressed or undressed). Research is still needed to identify the most workable presentation methods and proce- dures.
The endeavors to standardize administration procedures for the use of AD dolls in validation interviews must address the primary problem of the explicit and implicit communications included in the instructions and question used. Although the evidence is often equivocal, it has been demonstrated that children are sus- ceptible to leading and suggestive questioning (Goodman & Reed, 1986; King & Yuille, 1987; Loftus & Davies, 1984), are sensitive to linguistic changes in ques- tions (Dale, Loftus, & Rathburn, 1978), and respond to bizarre questions (Hughes & Grieve, 1980). Even when interviewers carefully pose their questions in a manner that is age-appropriate for the child, voice intonation and nonverbal corn-
T hi
s do
cu m
en t i
s co
py ri
gh te
d by
th e
A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. T
hi s
ar tic
le is
in te
nd ed
s ol
el y
fo r t
he p
er so
na l u
se o
f t he
in di
vi du
al u
se r a
nd is
n ot
to b
e di
ss em
in at
ed b
ro ad
ly .
PSYCHOMETRICS OF ANATOMICAL DOLLS 403
munication can foster an acquiescent response set. In these interviews, the child can become a participant in an unintended teaching situation in which the inter- viewer molds, shapes, and/or reinforces intentionally o r unintentionally a story o f abuse which m a y or m a y not have a counterpart in reality. Given their eagerness to please, children a r e particularly vulnerable to unwitting adult manipulations. Moreover, children lacking the experience or cognitive ability to understand a perception or experience m a y be easily swayed by the explanations o f adults (Melton, 1985). Thus, a child can be shaped or led into relating incidences o f abuse which m a y never have occurred, and because children can fuse " f a c t " and "~ t i o n " (Flavell, FlaveU, & Green, 1984; Johnson & Foley, 1984; Nelson, 1984), t h e y m a y become unable to determine the veracity o f the abuse allegation. These factors must all be taken into consideration as AD doll administration procedures a r e being developed and subsequently used.
Scoring
To date, no scoring standardization study has been conducted for AD dolls w h e n used in child sexual abuse validation interviews. There is a critical need for an acceptable scoring system.
One study assessing scoring i n t e r ~ e t a t i o n s o f various subgroups o f profes- sionals who use AD dolls highlightsttie~need~or standardized scoring procedures (Boat & Everson, 1988c). Child prota~fi~e ~airkers, mental health professionals, physicians, and law enforcement officers were asked if three specific doll inter- actions o f 3- to 5-year-olds were considered evidence o f sexual abuse. In addition, the descriptions were said to includ@:~U ~ finclude an accompanying verbal description by the child. Classification o f the interactions as indicative o f sexual abuse was dependent upon both professional group membership and the presence o f a verbal description. Furthermore, compared to the o t h e r professionals, the law enforcement officers were significantly less likely to perceive particular c h i l d - d o l l behavioral interactions (i.e., touching the doll's breasts, touching the doll's anal area, and placing prone dolls on top o f each other) as normal, findings consistent with those reported by Weinberg (1989/1990). Finally, there was particular dis- agreement within each professional group about the normalcy o f "avoiding the dolls" or "displaying anxious behavior in the presence o f the dolls." The lack o f agreement in the scoring o f child-doll play was underscored by the lack o f una- nimity on the normalcy o f even one o f the specified behaviors, including " u n - dressing the dolls."
Training in the Use o f AD Dolls
Although training for professionals using AD dolls is critical, it frequently appears to be the exception. While 76% o f surveyed mental health professionals indicated that they had received at least some formal o r informal training (i.e., workshop/course, presentations at a professional meeting, written o r taped in- structions, or discussions with a supervisor), less than 50% o f surveyed law en- forcement officers and child protective workers and 25% o f physicians reported having a n y such training (Boat & Everson, 1988c). Furthermore, less than 20% o f
T hi
s do
cu m
en t i
s co
py ri
gh te
d by
th e
A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. T
hi s
ar tic
le is
in te
nd ed
s ol
el y
fo r t
he p
er so
na l u
se o
f t he
in di
vi du
al u
se r a
nd is
n ot
to b
e di
ss em
in at
ed b
ro ad
ly .
404 SKINNER AND BERRY
the child protective workers and o f the mental health professionals, only 8% o f the physicians, and none o f the law enforcement officers had access to written guide- lines for the dolls' use.
Standardization Summary
Only limited progress has been demonstrated in terms o f standardization o f AD dolls and their use. With the lack o f manufacturing standards and the wide- spread use o f homemade/designed AD dolls, there is considerable variability in the dolls used in validation interviews. Though the greatest progress in the stan- dardization o f AD dolls has been in the introduction o f formal structured interview formats, no information about the popularity or usefulness o f these protocols is available. Moreover, these protocols must address the explicit and implicit mes- sages c o m m u n i c a t e d to a child through the use of the dolls. The development o f these interview formats has not been accompanied by standardized scoring pro- cedures nor programs for training in the use o f AD dolls.
N o r m s
Normative doll play patterns o f both sexually abused children and children with no history o f sexual abuse are needed in order to understand the clinical implications o f a child's interactions w i t h AD dolls. Several studies have at- t e m p t e d to develop the much-needed norms.
Studies with Abused Children
Studies with sexually abused children reveal that explicit demonstrations o f sexual abuse or suggestive sexual play are the predominant clinically significant interactions with the dolls (Britton & O'Keefe, 1991; Jampole & Weber, 1987; Leventhal, Hamilton, Rekedal, Tebano-Micci, & Eyster, 1989; White, Strom, SantiUi, & Halpin, 1986). The nature of the sexual play varies, and the studies have not addressed possible relationships between observed behavior patterns and components o f the actual sexual abuse.
Taking a b r o a d e r perspective, August and Forman (1989) reported greater specificity in observed behavioral patterns. Different response patterns w e r e found for observations o f AD doll interactions o f sexually abused children alone in a r o o m versus in the presence o f an adult. Nonsexualized free play was the primary play pattern observed in the alone condition, whereas avoidant behavior was the most frequent response in the presence o f an adult. Sexual play was m o r e c o m m o n in the alone condition and was the least frequent response in the pres- ence o f an adult. These researchers also raised the possibility that subgroups o f sexually abused children m a y manifest different play behavior patterns (e.g., avoidant vs. aggressive patterns).
Studies with Children with No Sexual Abuse History
Studies exploring AD doll play o f children with no history o f sexual abuse have taken one o f two approaches: (a) focus on all interactions with the dolls o r (b) limit the focus to demonstrations o f sexual intercourse.
T hi
s do
cu m
en t i
s co
py ri
gh te
d by
th e
A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. T
hi s
ar tic
le is
in te
nd ed
s ol
el y
fo r t
he p
er so
na l u
se o
f t he
in di
vi du
al u
se r a
nd is
n ot
to b
e di
ss em
in at
ed b
ro ad
ly .
PSYCHOMETRICS OF ANATOMICAL DOLLS 405
All Doll Interactions. The studies in this category share similar methodolo- gies in that free play with the dolls was the primary procedure used; four of the studies also included identification of body parts and functions segments. Across these studies, manipulation of the dolls' genitals was the most frequently reported doll interaction, and additional doll play included sexual and nonsexual aggres- sion, affection, and avoidance (Dawson & Geddie, 1991; Dawson, Vaughan, & Wagner, 1991; Gabriel, 1985; Glaser & Collins, 1989; Sivan, Schor, Koeppl, & Noble, 1988). The reported incidence of demonstrations of suggestive or clear sexual intercourse ranged from 0% (Sivan et al., 1988) to 9% (Dawson & Geddie, 1991).
In the only study that examined children's reactions to AD versus genitally neutral dolls, a significantly greater number of sexually related behaviors was found with AD dolls (Goodman & Aman, 1990). These behaviors occurred most frequently during a free play period and consisted of manipulation of the genitals and undressing and dressing the dolls; one child demonstrated genital penetration.
The Dawson and Geddie (1991) and Dawson et al. (1991) findings underscore the importance of examining variables that can impact on the behavior of children without (or with) a history of sexual abuse in the development of doll play norms. Using the same structured interview format in both studies, these two groups of investigators found differential incidences of behavioral and verbal interactions with AD dolls with samples of children differing in ethnicity and socioeconomic status (SES).
Demonstrations o f Sexual Intercourse. In the only study that limited its focus to suggestive and/or clear demonstrations of genital, oral, and/or anal intercourse, an incidence rate of 6% was reported for the doll play of 2- to 5-year-old children (Everson & Boat, 1990). Suggestive intercourse demonstrations were most com- mon, particularly in the presence of the interviewer and also in response to the prompt, "What can these dolls do together?" In contrast, clear intercourse dem- onstrations were more common in the absence of the interviewer. Although the majority of the demonstrations involved two dolls, oral intercourse demonstra- tions between a child and a doll were also observed.
The reported 6% incidence rate for demonstrations of intercourse involving AD dolls may be somewhat misleading as none of the 2-year-olds (26% of the sample) engaged in such demonstrations. For the older children, the incidence of suggestible intercourse demonstrations in the presence of the interviewer ranged from 9% for the 4-year-olds to 18% for the 5-year-olds. The incidence of clear intercourse demonstrations when the child was alone ranged from 3% for the 3-year-olds to 12% for the 5-year-olds. Moreover, whereas low SES African- American females (50%) and African-American males across SES levels (33%) manifested clear intercourse demonstrations when alone, only low SES African- American males (27%) displayed such behavior in the presence of the interviewer.
Problems with AD Doll Play Normative Research
Two major problems are characteristic of the "normative" studies conducted to date: (a) the use of an interview format dissimilar to those used in validation
T hi
s do
cu m
en t i
s co
py ri
gh te
d by
th e
A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. T
hi s
ar tic
le is
in te
nd ed
s ol
el y
fo r t
he p
er so
na l u
se o
f t he
in di
vi du
al u
se r a
nd is
n ot
to b
e di
ss em
in at
ed b
ro ad
ly .
406 SKINNER AND BERRY
interviews and (b) the failure to consider variables that m a y have a significant impact on a child's AD doll play (Skinner, Berry, & Giles, 1992).
Differing Methodologies. Sexual abuse validation interviews and " n o r m a - t i v e " studies generally share one c o m p o n e n t - - a doll body parts inventory. How- ever, whereas validation interviews tend to employ direct questions, such ques- tioning is usually replaced with a free play segment in the normative research studies. To date, few researchers have attempted to use a protocol e v e n somewhat similar to those followed in validation interviews. The reasoning offered b y one group o f researchers is revealing:
The conditions in the current study did not mirror those used in child abuse investiga- tions. Because of time pressures and lack of training, child abuse investigations often involve asking many situation-specific questions after a very short period, ff any, of free play with the dolls. We felt that such a protocol for this normative study would inhibit the range of behavior usually exhibited by children in free play with dolls. Moreover, leading questions might create a situation of entrapment, violating provisions for the protection of human subjects. (Sivan et al., 1988, p. 297)
These researchers' concerns about entrapment appear to be valid, given the findings o f M c I v e r et al. (1989), who reported that w h e n intentionally misleading questions intended to elicit sexual and aggressive behavior with AD dolls and similar to those questions used in validation interviews were used, 6 o f 7 children manifested the desired behaviors. Additionally, studies using a structured AD doll interview with sexually abused and nonabused children have not found significant differences in the doll play across the groups (Jensen, Realmuto, Wescoe, & Garfinkel, 1986; Realmuto, Jensen, & Wescoe, 1990). The desire to allow children participating in AD doll normative research to play freely with the dolls in o r d e r to allow the full range o f behavior m a y be well intended. However, the " n o r m s " produced b y this procedure m a y have no relevance for validation interviews, and it is f o r these interviews that norms are greatly needed.
Possible Complicating Variables. Previous research has demonstrated a re- lationship b e t w e e n children's knowledge o f sexuality and the variables o f age and social class (Gordon, Schroeder, & Abrams, 1990a; Wurtele & Muller, 1987). These variables, as well as ethnicity and sex, have also b e e n found in some studies to be related to a child's interaction with AD dolls (Dawson & Geddie, 1991; Dawson e t al., 1991; E v e r s o n & Boat, 1990; Kenyon-Jump, Burnette, & Robert- son, 1991). T h e r e is a need for additional research to determine w h e t h e r these factors should be taken into consideration in t h e development o f AD doll play n o r m s .
Additional factors which m a y need to be considered include a family's ap- proach to sexuality, sex education, personal privacy, and nudity; t h e pregnancy o f a female relative o r the birth o f a sibling; in vivo o r media exposure to sexual behavior; prior victimization; and earlier participation in prevention programs o r police interrogations (Berry, 1985). Some o f these factors have been offered as explanations for the sexualized AD doll play o f children with no history o f sexual assault (Kenyon-Jump e t al., 1991); however, the same sexualized play is inter- preted as reflective o f sexual victimization in children with known sexual assault
T hi
s do
cu m
en t i
s co
py ri
gh te
d by
th e
A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. T
hi s
ar tic
le is
in te
nd ed
s ol
el y
fo r t
he p
er so
na l u
se o
f t he
in di
vi du
al u
se r a
nd is
n ot
to b
e di
ss em
in at
ed b
ro ad
ly .
PSYCHOMETRICS OF ANATOMICAL DOLLS 407
histories (Glaser & Collins, 1989). Theory-driven research aimed at delineating the possible impact these factors have on AD doll play behavior is also needed.
Norms Summary
Presently; no acceptable norms for the use o f AD dolls in the validation o f child sexual abuse allegations exist. Moreover, those studies that have worked toward the development o f these much-needed norms have severe methodological problems that impact considerably on the utility o f the findings reported.
R E L I A B I L I T Y
The issue o f the reliability o f AD dolls in the evaluation o f child sexual abuse allegations has generally been overlooked, a not unexpected consequence given the frequency with which an unstructured or nonstandardized approach to their use has been employed. However, when a specified scoring scale has been used, interobserver reliability data have been reported.
Interobserver Reliability
Kappa coefficients o f agreement ranging from .60 to 1.00, with the exception o f a kappa o f .40 for "aggressive verbalizations," have been reported for the ratings o f trained interviewers conducting independent behavior analyses scoring the occurrence o f 25 specified variables in 15-second intervals of videotaped AD doll interviews (Allen, Jones, & Nash, 1989). Similarly, August and Forman (1989) reported an interobserver reliability measure o f .94 w h e n trained observers used a behavioral checklist to rate the occurrence o f four categories of behaviors in 10-second intervals o f videotapes o f children interacting with the dolls. Using point-by-point agreement calculations to assess the observational reliability o f one third o f AD doll free play observations, Kenyon-Jump et al. (1991) found good reliability for explicit sexual behavior (96%) and " s u s p i c i o u s " behavior (96%). However, the reliability o f touching, rubbing, and inserting was considerably lower (77%). These three studies included sexually abused children and children with no abuse history.
Six additional studies reporting reliability measures included only children without sexual abuse histories. Sivan et al. (1988) reported interobserver reliabil- ities from .80 to .95 for observations o f both structured doll interviews and free play interactions and from .83 to 1.00 for codings o f the transcripts o f the inter- views and interactions. An observational scheme giving priority to actions with the dolls was developed for this project.
Viewing videotapes o f children's structured and free play with AD dolls, Glaser and Collins (1989) reported an overall interobserver reliability measure o f .99 for four categories o f behaviors (i.e., the process o f undressing and dressing the dolls, the children's responses to viewing genitalia, the occurrence o f sexual, aggressive and/or affectionate play, and the quality and quantity o f the child's
T hi
s do
cu m
en t i
s co
py ri
gh te
d by
th e
A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. T
hi s
ar tic
le is
in te
nd ed
s ol
el y
fo r t
he p
er so
na l u
se o
f t he
in di
vi du
al u
se r a
nd is
n ot
to b
e di
ss em
in at
ed b
ro ad
ly .
408 S K I N N E R A N D BERRY
overall reaction to and play with the dolls). The greatest level o f disagreement was reported for the noticing, naming, and touching o f genitalia.
Mean kappa coefficients o f .88 or higher were reported for demonstrations o f both clear and suggestive genital, oral, or anal intercourse (Everson & Boat, 1990). These intercoder agreement scores were based on less than 10% o f the interview sessions conducted.
Acceptable interobserver agreement was reported in two parallel studies in- vestigating the normative doll play o f low-income minority (Dawson & Geddie, 1991) and middle-class White (Dawson et al., 1991) preschoolers. Using the same structured, nonsuggestive interview format, these two groups o f researchers trained observers to rate eight behavioral and eight verbal doll interactions. In both studies, interobserver agreement for a randomly selected subsample o f vid- eotaped interviews was generally above 70%. Dawson and Geddie reported a median o f 87% for o c c u r r e n c e percentage agreement, and Dawson et al. reported a median o f 83%. Across the studies, agreement was greater for the behavioral interactions. The lower agreement for the verbal interactions was attributed to difficulty in understanding the videotaped statements o f the children rather than disagreement about scoring criteria.
Finally, a reliability measure o f .71 as determined by proportion o f agreement scores for 15% o f observed play interactions with both AD and genitally neutral dolls has been reported (Goodman & Aman, 1990). The behaviors considered were limited to those interactions with dolls that have been clinically o r empiri- cally identified as behavioral indicators o f sexual abuse.
Test-Retest Reliability
T e s t - r e t e s t reliability o f AD dolls m a y not be demonstrable. In an investiga- tion o f the possible impact o f exposure to AD dolls, mothers o f nonabused chil- dren completed a behavioral checklist about 2 weeks after their children inter- acted with the dolls (Boat, Everson, & Holland, 1990). Postinterview sexually focused behaviors w e r e reported by 40% o f the mothers o f 3-year-olds and 70% o f the mothers o f 4-year-olds. It remains to be determined if initial exposure to AD dolls m a y foster increased sexual play in a subsequent interaction with the dolls.
Reliability Summary
The greatest progress in the establishment o f the psychometric properties o f A D dolls has been t h e demonstration o f good interobserver reliability. High in- terobserver reliability has been reported consistently in studies involving both sexually abused and nonabused children.
V A L I D I T Y
Current empirical data on the use o f AD dolls in validation interviews can be interpreted as addressing content, construct, and criterion-related validity.
T hi
s do
cu m
en t i
s co
py ri
gh te
d by
th e
A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. T
hi s
ar tic
le is
in te
nd ed
s ol
el y
fo r t
he p
er so
na l u
se o
f t he
in di
vi du
al u
se r a
nd is
n ot
to b
e di
ss em
in at
ed b
ro ad
ly .
PSYCHOMETRICS OF ANATOMICAL DOLLS 409
C o n t e n t V a l i d i t y
The ideal set o f AD dolls should include the body parts that allow a child to demonstrate all forms o f child sexual abuse (i.e., oral/anal/vaginal penetration by penis, digits, and/or objects, and fondling). Thus, use o f a set o f completely detailed dolls would offer good content validity. Wescott, Davies, and Clifford (1989) suggested that dolls representing a cross-section o f ages (e.g., babies to grandparents) m a y be desirable. In addition, at least one commercial manufac- turer offers male dolls with interchangeable circumcised and uncircumcised pe- nises, which m a y enhance the content validity o f the dolls for selected cases.
Less acceptable levels o f content validity for AD dolls are common, partic- ularly among the homemade/modified dolls. Although greater content validity is likely to be found with the commercially manufactured dolls, even these dolls m a y vary somewhat, primarily in terms o f the presence o f separatable fingers and secondary sex characteristics (e.g., chest hair). The homemade/modified dolls are likely to have the lowest levels o f content validity since body orifices are fre- quently omitted. To date, the relationship between various levels o f content va- lidity and children's play behaviors with AD dolls has not been addressed.
C o n s t r u c t V a l i d i t y
Comparisons Across Groups of Children
The primary procedure employed to obtain evidence o f AD dolls' construct validity as a diagnostic t o o l - - o b s e r v i n g the patterns o f doll play for distinct groups o f c h i l d r e n - - h a s produced mixed results. Research pursuing this line o f evidence falls into three categories: (a) studies including only nonreferred children with no sexual abuse history and n o t drawn from a psychiatric population; (b) studies including both sexually abused children and nonreferred children with no sexual abuse history and not drawn from a psychiatric population; and (c) studies including sexually abused children and children drawn from a psychiatric popu- lation but with no sexual abuse history. This latter group o f studies m a y also include as a third group children with no sexual abuse history and not drawn from a psychiatric population.
Nonreferred Children Only. Two studieshave found that nonreferred children did not manifest doll play patterns similar to those reported for sexually abused children. Sivan et al. (1988) observed that nonabused children appeared to find the dolls somewhat boring. Whereas some o f the 144 3- to 8-year-olds expressed interest in the dolls" genitals, no explicit sexual behavior was seen in the more than 5,000 observed interactions, and less than 1% o f the observations included aggressive behavior. These investigators suggested that unusual activity with AD dolls m a y be considered a defirtitive marker o f sexual abuse.
Agreeing with Sivan et al. (1988) that sexual play with the dolls is meaningful and not a false positive sign o f sexual abuse, Glaser and Collins (1989) observed the doll play o f 91 3- to 6-year-old children. Neither sexually explicit nor aggres- sive interactions with the dolls w e r e common. However, 35% o f the children, primarily 3- and 4-year-olds, avoided the dolls altogether or were reticent to play
T hi
s do
cu m
en t i
s co
py ri
gh te
d by
th e
A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. T
hi s
ar tic
le is
in te
nd ed
s ol
el y
fo r t
he p
er so
na l u
se o
f t he
in di
vi du
al u
se r a
nd is
n ot
to b
e di
ss em
in at
ed b
ro ad
ly .
410 SKINNER AND BERRY
with the dolls, behaviors the investigators identified as reflecting normal social and emotional responses.
Other studies in this category have reported v e r y different results. For ex- ample, Gabriel (1985) o b s e r v e d the AD doll interactions o f 19 2- to 5-year-old children. The children were allowed to engage in free play with the dolls or other available t o y s and w e r e then asked some questions a b o u t the dolls. Three o f the children left the playroom before playing with any o f the toys. O f the remaining 16 children, all but o n e engaged in play that revealed the genitals (56%), had unusual interactions with the genitals (50%), and/or showed an overt interest in the gen- itals (63%).
Using a standardized interview protocol, Herbert, Grams, and Goranson (1987) found that nonabused children, aged 3 to 5, manifested doll play behaviors that are commonly identified as reflective o f sexual abuse. Moreover, when the interviews were evaluated " a s if" an actual sexual a b u s e allegation existed, ap- proximately one half o f the children w e r e classified as probable sexual abuse victims.
As part o f a larger study, B o a t and E v e r s o n (1988b) observed the doll inter- actions o f 209 2- to 5-year-olds. After an opportunity to explore the dolls, each child was given an oral b o d y part and function inventory and then allowed more time for free exploration o f the undressed dolls. In addition, the 3- to 5-year-olds were left alone in the room to play with the dolls. Manual exploration (e.g., touching, rubbing, pinching, and pulling) o f the dolls' genitals and anus was very common across all ages, as was exploration o f the female's breasts, particularly among the younger children. Furthermore, 4% o f the children demonstrated gen- ital intercourse b e t w e e n the dolls. These findings led the researchers to conclude that sexualized play with AD dolls may b e noteworthy but cannot b e considered a definitive indicator o f sexual abuse. However, in a more recent article b a s e d u p o n the same sample, these investigators reported the incidence o f demonstra- tions o f sexual intercourse to b e 6% overall, and the incidence for selected sub- samples o f participants was as high as 29% and 50% (Everson & Boat, 1990), findings that w e a k e n their earlier conclusions about the diagnostic value o f AD dolls.
From a v e r y different direction, Schor and Sivan (1989) examined nonvictim- ized children's labels for specific b o d y parts o f AD dolls. Noting that idiosyncratic terms used b y alleged child sexual abuse victims to refer to sexual b o d y parts, particularly in interactions with AD dolls, are often used as a partial basis to assess the likelihood o f sexual abuse, these researchers asked 144 children, rang- ing in age from 3 to 8 years, their names for specific b o d y parts (i.e., anus, breasts, buttocks, penis, scrotum, and vagina). Generally, the older children had clearer labels for the b o d y parts considered. However, children across all ages tended to have clearer labels for p e n i s , b u t a n u s was difficult for all o f the children. The tendency o f the younger children to offer more varied labels for each o f the b o d y parts, regardless o f the appropriateness o f the labels, led the researchers to sug- gest that younger children m a y b e apt to answer questions even if the responses are meaningless. Since many o f the labels o f all o f the children differed consid-
T hi
s do
cu m
en t i
s co
py ri
gh te
d by
th e
A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. T
hi s
ar tic
le is
in te
nd ed
s ol
el y
fo r t
he p
er so
na l u
se o
f t he
in di
vi du
al u
se r a
nd is
n ot
to b
e di
ss em
in at
ed b
ro ad
ly .
PSYCHOMETRICS OF ANATOMICAL DOLLS 411
erably from those of their parents, it was concluded that the use of an idiosyn- cratic label is not necessarily suggestive of sexual abuse.
Sexually Abused Children and Children with N o Abuse History. The studies in this second category of research offer equivocal support for the construct validity of AD dolls. The results reported by White et al. (1986) support the construct validity of AD dolls. The doll play patterns of 25 nonabused children were reported to differ from those of 25 sexually abused children, with the latter group more likely to comment about specific sexual acts and demonstrate such acts. However, the doll play patterns of the abused children varied somewhat both across and within age groups. Though such doll play pattern variations were found among the sexually abused children, no unusual play patterns were reported among the nonabused children.
Analyzing the AD doll play of 10 matched pairs of sexually abused children and nonabused children in a free play situation, Templeton and Weber (1985) found a significantly higher incidence of sexual play, including demonstrations of vaginal, oral, and anal intercourse, among the victimized children. Similar results were reported by August and Forrnan (1989), who observed 32 5- to 8-year-old females play with AD dolls and asked the children to make up a story about the dolls. The play patterns of the two groups differed significantly in the level of free play, aggression, and sexualized interactions, with the abused children demon- strating higher levels of aggression and sexualized play. In their stories, the two groups differed in level of free play and avoidance, with abused children engaging in more avoidance. Within the abused group, no common play pattern shared by even the majority of children was found.
Based upon their observations of 20 3- to 8-year-old children, one half of whom had been sexually abused, Jampole and Weber (1987) concluded that dif- ferent play patterns were manifested by the sexually abused and nonabused chil- dren. Nine of the 10 sexually abused children were reported to display sexual behavior with the dolls, whereas only two of the 10 nonabused children were reported to engage in such interaction with dolls.
Equivocal construct validity support comes from the results reported by Kenyon-Jump et al. (1991). Using a static comparison design with participants matched for age, gender, race, SES, and family status, these researchers ob- served the AD doll free play of 9 sexually abused children and 9 children with no sexual abuse history. None of the 3- to 5-year-old children had prior experience with AD dolls. Comparisons of explicit sexual behaviors (e.g., doll-to-doll oral, anal, or vaginal penile penetration; face-to-face or face-to-back positioning of the dolls accompanied by rocking/thrusting motions; doll-child positioning and thrusting; child masturbation) and of manual stimulation of the dolls' genitals did not reveal significant group differences. Only when explicit and suspicious sexual behaviors (e.g., placing hands between legs, touching own genitals, kissing the dolls" breasts, straddling of the dolls, etc.) were combined did the two groups of children differ significantly from each other. However, the reported differences may be somewhat misleading since 5 children from each participant group en- gaged in the targeted categories of behaviors.
T hi
s do
cu m
en t i
s co
py ri
gh te
d by
th e
A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. T
hi s
ar tic
le is
in te
nd ed
s ol
el y
fo r t
he p
er so
na l u
se o
f t he
in di
vi du
al u
se r a
nd is
n ot
to b
e di
ss em
in at
ed b
ro ad
ly .
412 S K I N N E R A N D BERRY
The results o f the two final studies in this category do not lend support to AD dolls' construct validity. In the first study, both sexually abused and nonabused children, aged 2 to 7, were interviewed with AD dolls under a variety o f condi- tions, including the use o f intentionally leading questions, cues, modeling, and reinforcement, to elicit sexualized behavior (Mclver et al., 1989). No significant differences in the play patterns for the two groups were found. For example, 44% o f the nonabused and 30% o f the abused children spontaneously discussed or touched the dolls' genitals, and 62% o f the nonabused children and 50% o f the abused children either placed the dolls in a sexual position and/or displayed phys- ical aggression toward the dolls. In addition, progressively more sexual and ag- gressive behavior was obtained from the nonabused children by the use o f very subtle questions and cues.
In an attempt to clarify the clinical utility o f AD dolls, Cohn (1991) compared the responses to the dolls o f 35 males and females aged 2 to 6 referred for child sexual abuse and control subjects with no history of sexual abuse and matched for age, race, and sex. The protocol included free play periods both in the p r e s e n c e and a b s e n c e o f t h e i n t e r v i e w e r . Although it was f o u n d that the dolls neither elicited anxiety nor overstimulated the children, it was con- cluded that the doll play o f both groups o f children were similar and not o f clinical value.
Emotionally Disturbed Children and Sexually Abused Children. Only two studies to date have compared the doll play patterns o f sexually abused children and nonabused but emotionally disturbed children, and neither supports t h e con- struct validity o f AD dolls. Using a structured protocol with AD dolls, J e n s e n et al. (1986) videotaped interviews with 6 sexually abused children, 5 children with no history o f sexual abuse, and 4 psychiatrically hospitalized children with no history o f sexual abuse. A panel o f child psychiatrists, attorneys, and social workers was unable to find a n y differences in the doll play across the groups. In fact, the highest rating o f " v e r y suspicious o f a b u s e " was gaven to some o f the nonabused children and the " n o suspicion o f a b u s e " rating was given to some o f the sexually abused children.
In the second study, 11 sexually abused and 12 emotionally disturbed but nonabused children were interviewed in accordance with a structured interview protocol for AD dolls (Allen et al., 1989). Examiners unaware o f the assault history o f each child scored the videotaped interviews for both emotional and behavioral indicants o f sexual abuse. The only significant quantitative o r qualita- tive difference in the doll play between the two groups was that the nonabused children required that the dolls be dressed earlier in the interview than did the sexually abused children, a finding contrary to expectations.
The results o f these studies suggest that even if clinically significant doll play patterns a r e observed, the underlying factor o f these behaviors in sexually abused children m a y not be the sexual abuse experience. Instead, sexually abused chil- dren and emotionally disturbed children m a y share some pathogenic f a c t o r that leads to c o m m o n responses to A D dolls, a possibility not supportive o f the con- struct validity o f AD dolls.
T hi
s do
cu m
en t i
s co
py ri
gh te
d by
th e
A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. T
hi s
ar tic
le is
in te
nd ed
s ol
el y
fo r t
he p
er so
na l u
se o
f t he
in di
vi du
al u
se r a
nd is
n ot
to b
e di
ss em
in at
ed b
ro ad
ly .
PSYCHOMETRICS OF ANATOMICAL DOLLS 413
Construct Validity Design Issue
A fundamental problem underlying AD doll construct validity research is the use o f inappropriate comparison groups. Such studies usually employ comparison groups that are matched on relevant variables (i.e., age, sex, and ethnicity) with the group o f sexually abused children. However, most studies have ignored vari- ables such as sexual awareness/knowledge, family disruption, and prior exposure to AD dolls, factors that some studies have reported to be related to doll play. An additional factor that is likely to be relevant is prior interaction with police, child protective workers, and/or mental health professionals. Studies that fail to match on such variables are likely to produce less than optimal data addressing the issue o f AD doll construct validity.
Construct Validity Threats
There are two major threats to the construct validity o f AD dolls that must be addressed.
A D Doll Stimulus Value. Concerned with the stimulus value or suggestibility o f the dolls, numerous mental health professionals have proposed that the use o f AD dolls m a y not merely facilitate communication about a real experience but also foster the production o f fictional stories (King & Yuille, 1987; Yates & Terr, 1988a, 1988b; Yuille, 1988). Furthermore, Everstine and Everstine (1989) sug- gested that because children have usually played only with genitally neutral dolls, the presentation o f AD dolls m a y be perceived by a child as accompanied by the unspoken message that the adult presenter wants the child to address gender- related or sex-related issues. The communication o f such a message, combined with a child's suggestibility and the prestige o f the interviewer, may increase the likelihood o f misleading o r exaggerated sexualized doll play.
Observations have been offered to refute and to support the position that these dolls " i n v i t e " children to engage in sex play. Both Sivan et al. (1988) and Glaser and Collins (1989) reported that AD dolls were generally incorporated into children's everyday nonsexualized play (i.e., feeding, bathing, and putting the dolls to bed) in much the same w a y as genitally neutral dolls and concluded that AD dolls did not stimulate explicit sexual behavior. However, others have offered support for the suggestibility o f the dolls. For example, both Goranson (1986) and Cohn (1991) reported that essentially all o f their participants " p e n e t r a t e d " e a c h orifice o f the dolls with their fingers, a behavior that many m a y interpret as a demonstration o f intercourse. In addition, Herbert et al. (1987) reported observing suggestibility in t h e children's responses to the interview with the dolls.
Although not directly addressing the issue o f suggestibility o f AD dolls, Ed- wards and Forman (1989) compared the effect o f interview method (i.e., verbal description, using the dolls, and drawings) on the accuracy o f recall o f sexual abuse information. After viewing a film on sexual abuse, each child was asked to report on the presented information. The amount o f incorrect information recalled did not differ across the three conditions, a finding that does not offer support for the suggestibility o f the dolls.
T hi
s do
cu m
en t i
s co
py ri
gh te
d by
th e
A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. T
hi s
ar tic
le is
in te
nd ed
s ol
el y
fo r t
he p
er so
na l u
se o
f t he
in di
vi du
al u
se r a
nd is
n ot
to b
e di
ss em
in at
ed b
ro ad
ly .
414 SKINNER AND BERRY
Two studies have systematically examined the possible suggestibility o f AD dolls. In actual sexual abuse validation interviews, Britton and O ' K e e f e (1991) employed AD dolls with 69 children and genitally neutral dolls with 67 children referred by police or child protective agencies. Sexually explicit doll play was manifested by 68% o f the children using the AD dolls and 72% using the genitally neutral doll. These findings failed to support the hypothesis o f the suggestibility o f AD dolls.
In the second study, Goodman and A m a n (1990) involved 80 children, aged 3 to 5, in a nonsexual social interaction with an adult male and later interviewed each child. None o f the children had a known history o f sexual abuse and, during the subsequent interview, each was asked to reenact the social interaction, using AD dolls, genitally neutral dolls, free recall with visual cues, or free recall without visual cues. Interview questions included intentionally misleading questions and those dealing with possible sexual abuse. Analysis o f the data revealed that the presence o f either set o f dolls during the interview did not result in overall less a c c u r a c y in responses to objective or intentionally misleading sexual abuse- related questions. However, the finding that the younger children did produce less accurate responses to objective questions when dolls were present, particularly AD dolls, suggests that toys may cause stimulus interference. This interference was not manifested in relation to sexual matters. The investigators concluded that while neither type o f doll was found to be suggestible to the children, the dolls also were not found to be effective memory aids. This latter finding is contradictory to the results reported by Price and Goodman (1990) in their investigation o f the use o f a single doll used to reenact a repeated event, and underscores the need for continued research on the possible suggestibility o f AD dolls.
An Exposure Effect. The results o f t h e Boat et al. (1990) investigation o f the impact o f exposure to AD dolls on children m a y serve as a basis f o r challenging the construct validity o f the dolls as a diagnostic tool because sexualized play was manifested by children after exposure to the dolls. The authors concluded that the sexualized behaviors manifested by these nonabused children were benign and not o f the nature demonstrated by sexually abused children. However, the meth- odology o f this study did not include reexposure to AD dolls and subsequent observation o f doll play. These results are particularly salient since it is most likely that children identified as having been sexually abused had experienced numerous validation interviews, perhaps involving AD dolls, prior to their research partic- ipation (Steward, Steward, Lloyd, & Lyons, 1991). Thus, is it possible that the observed interactions with AD dolls interpreted as reflecting a history o f sexual abuse m a y have resulted from mere exposure to the dolls. This is a critical issue that must be addressed in the construct validation o f the dolls.
C n ' t e r i o n - R e l a t e d V a l i d i t y
Although the demonstration o f criterion-related validity is critically important for a test serving as a basis for the making o f probability statements, research on this type o f validity for the use o f AD dolls in evaluation interviews is almost nonexistent. Some researchers have alluded to the concurrent validity o f AD doll
T hi
s do
cu m
en t i
s co
py ri
gh te
d by
th e
A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. T
hi s
ar tic
le is
in te
nd ed
s ol
el y
fo r t
he p
er so
na l u
se o
f t he
in di
vi du
al u
se r a
nd is
n ot
to b
e di
ss em
in at
ed b
ro ad
ly .
PSYCHOMETRICS OF ANATOMICAL DOLLS 415
play by making references to relationships between the doll play o f victimized children and other indicators o f sexual abuse. While practitioners report anecdot- ally that the play behavior o f sexually abused children is consistent with other evidence of sexual abuse (e.g., behavioral indicators, verbal reports, medical evidence, and other diagnostic measures), most research studies have not re- ported on the inclusion o f specific additional measures.
Leventhal et al. (1989) have offered the strongest but somewhat equivocal evidence for the concurrent validity o f AD dolls as validation tools in their com- parison o f likelihood o f sexual abuse ratings based upon noninterview data (e.g., the report o f witnesses to the abuse, positive or suspicious medical evidence, a suspicious history, and known sibling abuse, etc.) and based on noninterview and diagnostic interview (with AD dolls) data. These investigators found that o f the 60 children referred for suspected sexual abuse, only 37% received ratings o f definite abuse when only noninterview data were utilized, whereas 70% o f the children received ratings when both noninterview and interview data were employed. There was no noninterview evidence supporting the definitely abused ratings for 38% o f the participants rated as definitely abused on the basis o f interview data. Although such a discrepancy may call into question the concurrent validity o f AD dolls, Leventhal et at. suggested that the discrepancy m a y be a function o f the ability o f AD dolls to elicit more detailed information than noninterview methods o f evidence collecting.
The one study that most directly addressed the issue o f criterion-related validity offers essentially no support for the diagnostic value o f AD dolls (Real- muto et al., I990). The participants included sexually abused children, children with no sexual abuse history, and children being seen at a psychiatric clinic and with no history o f sexual abuse. The sexual abuse status o f the victimized children was corroborated by comprehensive psychiatric evaluations and court and police records. The nonclinic, nonabused children were also participants in a longitudi- nal investigation o f normal child development, whereas the clinic sample con- sisted o f children being evaluated at a psychiatric clinic for depression and hy- peractivity. For these subsamples, a history o f physical or sexual abuse was ruled out by a mental health evaluation.
E a c h o f the children participated in a 20-rain structured interview conducted b y a senior female child psychiatric resident blind to the assault histories o f the children. The resident rated each child's interview according to a 5-point interview indicator scale (White et at., 1986). Overall, only 53% o f the children were correctly classified. The sensitivity (i.e., correct classification o f sexually abused children) o f the doll interviews was found to be 33% while the specificity (i.e., correct classification o f nonabused children) was 67%. These findings led t h e investiga- tors to question the diagnostic value o f the dolls as a single indicator. The finding o f fewer true positives (33%) than false positives (67%) does not support the crite- rion-related validity o f AD dolls as tests in the evaluation o f child sexual abuse allegations. These investigators identified as a major shortcoming o f this study the abbreviated interview t h e y were required to use by the Institutional Re~ciew Board and suggested that a more complete interview m a y have resulted in discriminating results. However, these results are consistent with their earlier findings reflecting
T hi
s do
cu m
en t i
s co
py ri
gh te
d by
th e
A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. T
hi s
ar tic
le is
in te
nd ed
s ol
el y
fo r t
he p
er so
na l u
se o
f t he
in di
vi du
al u
se r a
nd is
n ot
to b
e di
ss em
in at
ed b
ro ad
ly .
416 S K I N N E R A N D B E R R Y
the inability o f a panel to discriminate between the same three t y p e s o f groups o f subjects when watching videotaped doll interviews (Jensen et al., 1986).
Criterion-Related Validity Research Design Issue
As researchers attempt to address the criterion-related validity of A D dolls, particularly the predictive validity, the use of an appropriate design is important. For example, one approach would be to u s e the dolls in the validation interviews o f all children referred for suspected sexual abuse, with follow-up on each case until it is resolved. Subsequently, the A D doll play behavior o f children w h o s e a b u s e was substantiated could then b e compared with that o f children w h o s e abuse accusations w e r e determined untrue. Play data from children whose a b u s e allegations w e r e not clearly resolved would not b e included in the data analysis. This design would control for the extremely important factor o f the circumstances under which AD dolls are used in validation interviews, as well as the investiga- tion and intervention experiences o f the children. Unfortunately, a design such as this has not b e e n utilized to date.
Validity Summary
A correlation exists b e t w e e n the content validity o f AD dolls and the ana- tomical completeness o f the doll set, placing the level o f AD doll content validity under the control o f the user. Available construct validity evidence for the u s e o f AD dolls in validation interviews is weak. O f the 14 studies considered (the 1988 Boat and E v e r s o n and the 1990 E v e r s o n and Boat studies were counted as one b e c a u s e b o t h w e r e b a s e d on the same subjects), 43%, including three studies using structured interview protocols, offered no support for the dolls" construct validity. The other studies offered conflicting and weak evidence. The equivocal construct validity support found in 13% o f the studies suggests quantitative rather than qualitative differences in AD doll play patterns o f abused versus nonabused chil- dren. E v i d e n c e supporting the construct validity o f AD dolls was reported in the remaining 43% o f the studies; however, four of the six supporting studies did not employ a structured interview format similar to those used in validation inter- views. It is important that efforts b e directed at identifying the variables under- lying these discrepant findings and that subsequent research control these factors. The only s t u d y that has addressed the criterion-related validity o f AD dolls di- rectly failed to offer supporting evidence.
P S Y C H O M E T R I C P R O P E R T I E S O F A D D O L L S S U M M A R Y
As this consideration o f A D dolls' psychometric properties and the empirical literature reveals, both progress and the lack o f progress is readily apparent. With regard to standardization factors, the need for standardization o f the dolls has b e e n recognized. Unfortunately, no standards for the manufacturing of t h e dolls exist and considerable variability is c o m m o n across AD doll sets used in valida-
T hi
s do
cu m
en t i
s co
py ri
gh te
d by
th e
A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. T
hi s
ar tic
le is
in te
nd ed
s ol
el y
fo r t
he p
er so
na l u
se o
f t he
in di
vi du
al u
se r a
nd is
n ot
to b
e di
ss em
in at
ed b
ro ad
ly .
PSYCHOMETRICS OF ANATOMICAL DOLLS 417
tion interviews. In contrast, several formal structured interview protocols have b e e n forwarded in response to the need for standardized administration proce- dures. There remains, however, the importance o f continuing to hone these pro- tocols into more workable procedures and, then, ensuring that training in the use o f these administration procedures in an objective manner is available to users o f the dolls. Increased attention to the explicit and implicit messages interview pro- cedures communicate to children and the impact o f these messages on AD doll play is imperative. Finally, while the formal structured interview protocols have attempted to address the need for explicit scoring procedures, no scoring stan- dardization study has been conducted.
Although several studies have identified their primary goal as the develop- ment o f normative data, most o f these studies have relied on AD doll free play rather than questions as used in validation interviews and, consequently, have produced data that have little relevance to the norms needed for the use o f AD dolls in sexual abuse validation interviews. Those few studies employing formal structured interviews have not produced meaningful norms for validation inter- views.
The greatest psychometric stride focuses on the issue o f reliability. Very high interobserver reliability has b e e n consistently reported in studies employing ex- plicit scoring procedures.
The evidence related to the validity o f AD dolls ranges from very p o o r to somewhat positive. The level o f content validity that can b e achieved is dependent upon the completeness o f the AD doll set employed. H o w e v e r , b e c a u s e the ma- jority o f AD doll users do not use the most anatomically complete dolls, the content validity o f the dolls is less than possible. With regard to the construct validity o f the dolls, only t w o studies using a structure i n t e r v i e w format o f the 14 studies that can be interpreted as addressing the issue o f construct validity offered supporting evidence and 6 offered no construct validity evidence. Moreover, in- sufficient research efforts have addressed the construct validity threats o f AD dolls' stimulus value and exposure effect. Finally, no evidence has b e e n forwarded to support the criterion-related validity o f AD dolls as used in child sexual abuse interviews.
I M P L I C A T I O N S O F T H E P S Y C H O M E T R I C A N A L Y S I S O F A D D O L L S
Given the known psychometric properties o f AD dolls, there are t w o signif- icant problem areas for psychologists using these dolls in the validation o f child sexual abuse allegations. The first o f these areas focuses on the inappropriate substitution o f personal data for scientific data. With the lack o f standardized materials, administration procedures, and scoring, and with the a b s e n c e o f nor- mative doll play patterns, each doll user is apt to develop personal administration and scoring guidelines and norms that not only m a y not generalize across prac- titioners but may also reveal more about the practitioner than the child being assessed. A practitioner must protect against operating from the potentially dan- gerous position o f equating personal experience with scientific evidence.
T hi
s do
cu m
en t i
s co
py ri
gh te
d by
th e
A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. T
hi s
ar tic
le is
in te
nd ed
s ol
el y
fo r t
he p
er so
na l u
se o
f t he
in di
vi du
al u
se r a
nd is
n ot
to b
e di
ss em
in at
ed b
ro ad
ly .
418 SKINNER AND BERRY
Secondly, given the equivocal evidence for the validity of the dolls, mental health professionals may need to rethink the proposition that sexual play or knowledge in young children is an indicator of sexual abuse. The lack of discrim- inative vaf:J~ty evidence is not restricted to the analysis of AD doll play patterns. For example, Allen et al. (1989) failed to find significant differences between sexually abused and nonabused children on the Child Sexual Behavior Inventory (CSBI), a checklist of 41 sexualized behaviors and their frequency. Though re- porting significant differences in the frequencies of behaviors included in the CSBI, Friedrich (1991) found that both children with and without a history of sexual abuse were reported to manifest each of the specified sexual behaviors. Furthermore, increased sexual behavior for both groups of children was found to be related to sexual (i.e., family nudity) and nonsexual experiences (i.e., amount of life stress and parental violence). In addition, Gordon, Schroeder, and Abrams (1990b) reported that sexually abused children and children with no history of sexual abuse did not differ in nondoll sexualized play and sexual knowledge (i.e., gender differences, pregnancy and birth, nonsexual body parts and functions, adult sexual behavior, genitals and functions, and abuse prevention). These find- ings strongly suggest that sexually abused and nonabused children may not differ as expected in terms of sexual behavior, play, and/or knowledge. Yates (1991) has pointed out the significant need for research on sexual behavior in children with no history of sexual abuse.
The lack of sufficient evidence supporting the psychometric properties of AD dolls calls into question the use of those dolls in the validation of child sexual abuse allegations. Being ethically responsible for ensuring that tests employed meet adequate psychometric standards, it appears that a psychologist who uses AD dolls in sexual abuse validation interviews is operating in a professionally risky and, perhaps, indefensible manner. Moreover, given that validity is the principle issue underlying the admissibility of psychological evidence in the court- room (Ziskin & Faust, 1988) and the inadequate evidence for the construct and criterion-related validity of AD dolls, evidence collected using AD dolls should not be admitted in court in child sexual abuse cases at this time. Presently, the use of dolls in validation interviews fails to meet scientific test criteria and, conse- quently, it can only be concluded that AD dolls should not be used as the basis for expert opinions or conclusions.
R E F E R E N C E S
Allen, M. E., Jones, P. D., & Nash, M. R. (1989, August). Detection o f sexual abuse among emo- tionally disturbed preschoolers: Unique effects o f sexual abuse as observed on doll play and other standard assessment procedures. Paper presented at the annual meeting of the American Psy- chological Association, New Orleans.
American Psychological Association. (1985). Standards f o r educational and psychological testing. Washington, DC: Author.
August, R. L., & Forman, B. D. (1989). A comparison of sexually abused and nonsexually abused children's behavioral responses to anatomically correct dolls. Child Psychiatry and Human De- velopment, 20, 39-47.
T hi
s do
cu m
en t i
s co
py ri
gh te
d by
th e
A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. T
hi s
ar tic
le is
in te
nd ed
s ol
el y
fo r t
he p
er so
na l u
se o
f t he
in di
vi du
al u
se r a
nd is
n ot
to b
e di
ss em
in at
ed b
ro ad
ly .
PSYCHOMETRICS OF ANATOMICAL DOLLS 419
Bays, J. (1990). Are the genitalia of anatomical dolls distorted? ChildAbuse & Neglect, 14, 171-t75. Berliner, L. (1988). Anatomical dolls. Journal oflnterpersonal Violence, 3, 468-470. Berry, K. K. (1985, August). Problems in the investigation of child sexual abuse allegations. In K. K.
Berry (Chair), latrogenic effects o f child sexual abuse investigation and intervention. Symposium conducted at the annual meeting of the American Psychological Association, Los Angeles.
Boat, B. W., & Everson, M. D. (1986). Guidelines f o r interviewing young children in sexual abuse investigations. Unpublished manuscript, Department of Psychiatry, University of North Carolina, Chapel Hill.
Boat, B. W., & Everson, M. D. (1988a). Interviewing young children with anatomical dolls. Child Welfare, 67, 337-352.
Boat, B. W., & Everson, M. D. (1988b). Research and issues in using anatomical dolls. Annals o f Sex Research, 1, 191-204.
Boat, B. W., & Everson, M. D. (1988c). Use of anatomical dolls among professionals in sexual abuse evaluation. Child Abuse & Neglect, 12, 171-179.
Boat, B. W., Everson, M. D., & Holland, J. (1990). Maternal perceptions of nonabused young chil- dren's behaviors after the children's exposure to anatomical dolls. Child Welfare, 69, 389-400.
Britton, H. L., & O'Keefe, M. A. (1991). Use of anatomical dolls in the sexual abuse interview. Child Abuse & Neglect, 15, 567-573.
Cohn, D. S. (1991). Anatomical doll play of preschoolers referred for sexual abuse and those not referred. Child Abuse & Neglect, 15, 455-466.
Conte, J., Sorenson, E., Fogarty, L., & Rosa, J. D. (1991). Evaluating children's reports of sexual abuse: Results of a survey of professionals. American Journal o f Orthopsychiatry, 61, 428-437.
Dale, P. S., Loftus, E. E, & Rathburn, E. (1978). The influence of the form of the question on the eyewitness testimony of preschool children. Journal o f Psycholinguistics Research, 7, 269-277.
Dawson, B., & Geddie, L. (1991, August). Low-income children's behavior with sexually anatomically detailed dolls. Paper presented at the annual meeting of the American Psychological Association, San Francisco.
Dawson, B., Vaughan, A. R., & Wagner, W. G. (1991). Normal responses to sexually anatomically detailed dolls. Journal o f Family Violence, 7, 135-152.
Edwards, C. A., & Forman, B. D. (1989). Effect of child interview method on accuracy and com- pleteness of sexual abuse information recall. Social Behavior and Personality, 17, 237-247.
Everson, M. D., & Boat, B. W. (1990). Sexualized doll play among young children: Implications for the use of anatomical dolls in sexual abuse evaluations. Journal o f the American Academy o f Child and Adolescent Psychiatry, 29, 736-742.
Everstine, D. S., & Everstine, L. (1989). Sexual trauma in children and adolescents: Dynamics and treatment. New York: Brunner/Mazel.
Flavell,-J. H., Flavell, E. R., & Green, E C. (1984). Development of the appearance-reality distinc- tion. Cognitive Psychology, 15, 95-120.
Fox, R. E. (1991). Proceedings of the American Psychological Association, Incorporated, for the Year 1990. American Psychologist, 46, 689-726.
Friedemann, V. M., & Morgan, M. K. (1985). Interviewing sexual abuse victims using anatomical dolls: The professionals" guidebook. Eugene, OR: Migima Designs.
Friedricli, W. N. (1991, August). Assessing sexual behavior in sexually abused children. In G. S. Goodman (Chair), Assessment, diagnosis, and support o f child sexual abuse victims. Symposium conducted at the annual meeting of the American Psychological Association, San Francisco.
Frye v. United States, 293 E 1013 (D.C. Cir., 1923). Gabriel, R. M. (1985). Anatomically correct dolls in the diagnosis of sexual abuse of children. The
Journal o f the Melanie Klein Society, 3, 46-51. Glaser, D., & Collins, C. (1989). The response of young, non-sexually-abused children to anatomically
correct dolls. Journal o f Child Psychology and Psychiatry, 30, 547-560. Glasgow, D., & BentaU, R. P. (1989). What do expert witnesses in child sexual abuse think they are
doing?-"Diagnosis", and the sexually accurate doll "'test" as professional myths. The Liverpool Law Review, 11, 43-57.
Goodman, G. S., & Aman, C. (1990). Children's use of anatomically detailed dolls to recount an event. Child Development, 61, 1859-1871.
T hi
s do
cu m
en t i
s co
py ri
gh te
d by
th e
A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. T
hi s
ar tic
le is
in te
nd ed
s ol
el y
fo r t
he p
er so
na l u
se o
f t he
in di
vi du
al u
se r a
nd is
n ot
to b
e di
ss em
in at
ed b
ro ad
ly .
420 SKINNER AND BERRY
Goodman, G. S., & Reed, R. S. (1986). Age differences in eyewitness testimony. Law and Human Behavior, 3, 295-306.
Goranson, S. E. (1986). Young child interview responses to anatomically detailed dolls: Implications for practice and research in child sexual abuse. Unpublished doctoral dissertation, University of British Columbia, Vancouver, Canada.
Gordon, B. N., Schroeder, C. S., & Abrams, J. M. (1990a). Age and social-class differences in chil- dren's knowledge of sexuality. Journal o f Clinical Child Psychology, 19, 33-43.
Gordon, B. N., Schroeder, C. S., & Abrams, M. (1990b). Children's knowledge of sexuality: A com- parison of sexually abused and nonabused children. American Journal o f Orthopsychiatry, 60, 250-257.
Herbert, C. P., Grams, G. D., & Goranson, S. E. (1987). The use of anatomically detailed dolls in an investigative interview: A preliminary study o f "'non-abuse" children. Unpublished manuscript, Department of Family Practice, University of British Columbia, Vancouver, BC.
Hughes, H. M., & Grieve, R. (1980). On asking children bizarre questions. In M. Donaldson, R. Grieve, & C. Pratt (Eds.), Early childhood development and education. Oxford: Basil Blackwell.
In re Amber B. and Teela B., 191 Cal App.3rd 682 (1987). In re Christine C. and Michael C., 191 Cal. App.3rd 676 (1987). Jampole, L., & Weber, M. K. (1987). An assessment of the behavior of sexually abused and nonsex-
ually abused children with anatomically correct dolls. Child Abuse and Neglect, 11, 187-192. Jensen, J. B., Realmuto, G., Wescoe, S., & Garfinkel, B. D. (1986, October). Are there dbrferences in
the play with anatomically correct dolls: Abused versus non-abused children. Paper presented at the American Academy of Child Psychiatry, Washington, DC.
Johnson, M. K., & Foley, M. A. (1984). Differentiating fact from fantasy: The reliability of children's memory. Journal o f Social Issues, 40, 33-50.
Jones, D. P. H., & McQuiston, M. G. (1988). Interviewing the sexually abused child. London: Gaskell.
Kenyon-Jump, R., Burnette, M. M., & Robertson, M. (1991). Comparison of behaviors of suspected sexually abused and nonsexually abused preschool children using anatomical dolls. Journal o f Psychopathology and Behavioral Assessment, 13, 225-240.
King, M. A., & Yuille, J. C. (1987). Suggestibility and the child witness. In S. J. Ceci, M. P. Toglia, & D. E Ross (Eds.), Children's eyewitness memory (pp. 24-35). New York: Springer-Verlag.
Landers, S. (1988, June). Use of "detailed dolls" questioned. APA Monitor, pp. 24-25. Leventhal, J. M., Hamilton, J., Rekedal, S., Tebano-Micci, A., & Eyster, C. (1989). Anatomically
correct dolls used in interviews of young children suspected of having been sexually abused. Pediatrics, 84, 901-906.
Lewis, J. E. (1992, August). Sexually traumatized children? False positives from use of"anatomical'" dolls. Paper presented at the annual meeting of the American Psychological Association, Wash- ington, DC.
Loftus, E. F., & Davies, G. M. (1984). Distortions in the memory of claildren. Journal o f Sociallssues, 40, 51-67.
Mclver, W., Wakefield, H., & Underwager, R. (1989). Behavior of abused and non-abused children in interviews with anatomically correct dolls. Issues in Child Abuse Accusations, 1, 39-48.
Melton, G. B. (1985). Sexually abused children and the legal system: Some policy recommendations. American Journal o f Family Therapy, 13, 61-67.
Nelson, K. (1984). The transition from infant to child memory. In E. Moscovitch (Ed.), Infant mem- ory: Its relation to normal and pathological memory in human and other animals (pp. 103-130). N e w York: Plenum Press.
People v. Kelly, 17 Cal.3d 18, 53 (1976). Philllppus, M. J., & Koch, G. V. (1986). How to evaluate sexuality in children and how to avoid using
"anatomically correct dolls." Trial Talk, 35, 372-373. Price, D. W. W., & Goodman, G. S. (1990). Visiting the wizard: Children's memory for a recurring
event. Child Development, 61, 664-680. Realmuto, G. M., Jensen, J. B., & Wescoe, S. (1990). Specificity and sensitivity of sexually anatom-
ically correct dolls in substantiating abuse: A pilot study. Journal o f the American Academy o f Child and Adolescent Psychiatry, 29, 743-746. .
T hi
s do
cu m
en t i
s co
py ri
gh te
d by
th e
A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. T
hi s
ar tic
le is
in te
nd ed
s ol
el y
fo r t
he p
er so
na l u
se o
f t he
in di
vi du
al u
se r a
nd is
n ot
to b
e di
ss em
in at
ed b
ro ad
ly .
PSYCHOMETRICS OF ANATOMICAL DOLLS 421
Schor, D. E, & Sivan, A. B. (1989). Interpreting children's labels for sex-related body parts of ana- tomically explicit dolls. Child Abuse & Neglect, 13, 523-531.
Sivan, A. B. Schor, D. E, Koeppl, G. K., & Noble, L. D. (1988). Interaction of normal children with anatomical dolls. Child Abuse and Neglect, 12, 295-304.
Skinner, L. J., Berry, K. K., & Giles, M. K. (1992, June). Anatomically detailed dolls: The lack o f normative play patterns f o r validation interviews. Paper presented at the annual meeting of the American Psychological Society, San Diego.
Snyder, N. (1988). Who tells the truth about sexual abuse? California Lawyer, 8, 10-11. Steward, M. S., Steward, D. S., Lloyd, D., & Lyons, E (1991, August). Potential uses and abuses of
computer-assisted interviewing. In G. S. Goodman (Chair), Assessment, diagnosis, and support o f child sexual abuse victims. Symposium conducted at the annual meeting of the American Psychological Association, San Francisco.
Templeton, S., & Weber, C. (1985). An assessment o f the behavior o f sexually abused and nonsexually abused children with anatomically correct dolls. Unpublished manuscript, Louisiana State Uni- versity, Baton Rouge.
Walker, L. E. D. (1988). New techniques for assessment and evaluation for child sexual abuse victims: Using anatomically "detailed" dolls and videotape procedures. In L. E. D. Walker (Ed.), Hand- book in sexual abuse o f children: Assessment and treatment issues (pp. 175-197). New York: Springer~
Weinberg, L. S. (1990). Factors associated with professionals' degree of confidence that sexual abuse occurred (Doctoral dissertation, California School of Professional Psychology--Los Angeles, 1989). Dissertation Abstracts International, 50, 3181B-3182B.
Wescott, H., Davies, G. M., & Clifford, B. R. (1989). The use of anatomical dolls in child witness interviews. Adoption & Fostering, 13, 6-14.
White, S. (1988). Should investigatory use of anatomical dolls be defined by the courts? Journal o f Interpersonal Violence, 3, 471-475.
White, S., & Santilll, G. (1988). A review of clinical practices and research data on anatomical dolls. Journal o f Interpersonal Violence, 3, 430-442.
White, S., Strom, G., Santilli, G., & Halpin, B. (1986). Interviewing young sexual abuse victims ~4th anatomically correct dolls. Child Abuse and Neglect, 10, 519-529.
White, S., Strom, G., Santilli, G., & Quinn, K. (1987). Guidelines for interviewing preschoolers with sexually anatomically detailed dolls. Unpublished manuscript, Case Western Reserve University School of Medicine, Cleveland, Ohio.
Wurtele, S. K., & Muller, C. L. (1987). Children's conception of sexual abuse. Journal o f Clinical Child Psychology, 16, 184-191.
Yates, A. (1991). Differentiating hypererotic states in the evaluation of sexual abuse. Journal o f the American Academy o f Child and Adolescent Psychiatry, 30, 791-795.
Yates, A., & Terr, L. (1988a). Anatomically correct dolls: Should they be used as the basis for expert testimony? Journal o f the American Academy o f Child and Adolescent Psychiatry, 27, 254-257.
Yates, A., & Terr, L. (1988b). Issue continued: Anatomically correct dolls: Should they be used as the basis for expert testimony? Journal o f the American Academy o f Child and Adolescent Psychi- atry, 27, 387-388.
Yuille, J. C. (1988). The systematic assessment of children's testimony. Canadian Psychology, 29, 247-262.
Ziskin, J., & Faust, D. (1988). Coping with psychiatric and psychological testimony (4th ed.). Los Angeles, CA: Law & Psychology Press.
T hi
s do
cu m
en t i
s co
py ri
gh te
d by
th e
A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. T
hi s
ar tic
le is
in te
nd ed
s ol
el y
fo r t
he p
er so
na l u
se o
f t he
in di
vi du
al u
se r a
nd is
n ot
to b
e di
ss em
in at
ed b
ro ad
ly .
- lhb_17_4_0.pdf
- lhb_17_4_1.pdf
- lhb_17_4_2.pdf
- lhb_17_4_3.pdf
- lhb_17_4_4.pdf
- lhb_17_4_5.pdf
- lhb_17_4_6.pdf
- lhb_17_4_7.pdf