- Email: [email protected]
Interaction of normal children with anatomical dolls
C’hlddbuse&
Nqkcr, Vol. IT!, pp. 295-304, Printed in the U.S.A. All rights reserved.
0145-2134/88$3.00+ .oO Copyrigtrt 0 1988 Pe~gamon Press plc
1988
r~E~~I~N OF NORMAL CHILDREN WITH ANATOMICAL DOLLS ABIGAIL B. SIVAN, PH.D., DAVID P. SCHOR, M.D., GINA K. KOEPPL, PH.D. AND LXNND. NOBLE, PH.D. Child Development Clinic, Division ofDevelopmental Disabilities, Department of Pediatrics, University of Iowa, Iowa City, IA
Abstr&ct-A~atorni~~ (sexually exphcit) rag dolls are frequently used in interviews of children suspected of being sexually abused. Abused children have been noted clinically to be aggressive and sometimes sexually precocious in their doll play, and testimony to that effect is often accepted in the caurtroom as pertinent evidence. However, to date, normative data on the play behaviors with the sexually explicit dolls have been unavailable. This pilot study provides empirical information on the play interaction of a relatively large number of normal (nonreferred) children with anatomical dolls. One hundred forty-four children (ages 3 to 8 years) were observed in a playroom containing these special dolls during three conditions: (I) with an adult present, (2) without an adult, and (3) with the dolls undressed. The observations showed that nanreferred children found these dolls no mom interesting than other toys, Little aggression and no explicit sexual activity were observed. In contrast to clinical observation of abused children, the doll play of nonreferred children is unlikely to be characterized by aggression or sexual concerns; thus these behaviors when observed in interaction with these dolls should be taken seriously.
A Technique OFTEN USED in the investigation of sexual abuse is an interview using anatomical (sexually explicit) rag dolls. Although testimony based on the use of these dolls is often accepted in the courtroom, little data exist which provide information about how normal (nonreferred) children freely interact with these dolls. This pilot project was developed to explore the free play interactions of nonreferred children with anatomical dolls and to compare these findings with clinical impressions of abused children. The data presented here were obtained as a part of a larger study. Many health professionals who work daily with children maintain that the play of abused children is substantially different from that of normal (nonreferred) children and testify to that effect in court. Guidelines have been developed for standardizing interviews with sexually abused children f I, 41 and recent research presented ~orn~~n~ the responses of abused and nonabused children f5]. However, the one published ~nves~~~on [3] which has explored the free play of normal children with these dolts is plagued by small sample sizes (n = 20) and broad units of analysis, e.g., sexual activity vs. no sexual activity. Many investigators (social workers and Iaw enforcement officers) have limited knowledge of This research was funded by the Iowa State Criminal and Juvenile Justice Agency as well as the Office of the Vice President for Educational Development and Research, University of Iowa. An earlier version of this paper was presented at the biennial meeting of the Society for Research in Child Develop ment, Baltimore, MD, April 25, 1987. Received for pubiication May 27,1987: final revision received October 15,1987; accepted February 19,198g. Reprint requests may be sent to A. Sivan, Ph.D., Division of Child Psychiitry, Rush-~~b~e~an-St. Center, 1720West Polk Street, Chicago, IL 6Q6i 2. 295
Luke’s Medical
296
A. B. Sivan, D. P. Schor, G. K. Koeppl and L. D. Noble Table 1. Mean (and Range) Age of the Subjects in Months Sex of Interviewer Sex of Child
Boy 71.29 (37-107)
Girl 72.78 (37-107)
Female
Male
51.08 (38-57)
48.75 (37-58)
69.58 (60-79)
70.92 (60-83)
94.08 (85-107)
93.33 (84-106)
47.92 (38-54)
48.25 (37-47)
72.58 (60-83)
74.08 (66-83)
95.42 (87-101)
98.42 (85-107)
71.78 (38-107)
72.79 (37-107)
normal child development and child behavior, and those child specialists who possess ample experience lack specific comparison data which describe the typical interactions (behavioral and verbal) of normal (nonreferred) children with the dolls. Without these comparison data, the use of doll play observations of investigated children may be open to challenge in court. For the above reasons, this pilot study was designed to explore some of commonly held but unresearched assumptions about the dolls. Specifically, it addressed the following questions: 1. In what ways do the children interact with the dolls (i.e., do they exhibit particular themes in their interactions such as caretaking or creative role playing)? 2. Are there differences in the children’s interactions which are systematically related to age or gender? 3. Are there differences in the children’s interactions with the dolls related to the presence or absence of an adult professional? 4. Does the gender of the adult present affect the child’s interactions? 5. For what amount of time do these dolls keep the child’s interest and is this related to age or gender? All these questions have clear clinical implications for professionals who are asked to relate observations of children’s play as a part of investigation and evaluation of the emotional and intellectual competence of the abused victim.
METHOD Subjects Subjects were 144 children (3-8 years old) whose parents agreed to their participation. jects were solicited through advertisements posted in local nursery schools and pediatric child clinics and through news releases. Subjects were stratified by age (3 and 4 years, 5 years, 7 and 8 years), gender of the child, and gender of interviewer into 12 groups children each. This stratification is depicted in Table 1,
Subwelland 6 of 12
Anatomical doll play Table 2. Scores on the PPVT-R months)
297
(Age equivalents
in
Sex of Interviewer Sex of Child
Female 57.08
ROY
83.53 (29-141)
Girl 14.86 1n :39)
Male 60.25
(30-74) 87.42 (66-l 12) 109.92 (89-141)
(29-86) 75.33 (51-98) 111.17 (78-136)
58.33 (39-88)
54.42 (40-77)
83.92
89.92 (68-128)
(62-121) (86-130)
I 17.00 (84- 139)
83.7 I (30-141)
84.68 (29-139)
105.58
Socioeconomic status was to be considered, but attempts to find a low status group in the greater Iowa City area met with failure. All the fathers reported at least a high school education; 63.2% had at least one college degree; and 72.1% held semiprofessional or professional jobs. Only two mothers did not complete high school; 60.5% had at least one college degree; and 43.8% held semiprofessional or professional jobs. Homemakers or students comprised 40.3% of the mothers. No information was gathered concerning marital status of the parents. Procedures
Informed consent was obtained, and all subjects completed the Peabody Picture Vocabulary Test-Revised (PPVT-R) [2] to provide a screening of general intellectual level (IQ equivalent above 85). Table 2 presents data from the PPVT-R showing both the mean score and range of scores for each of the 12 groups. Data collection involved observing and recording the children’s interactions with the dolls under three conditions designed to help the children feel comfortable in the unfamiliar surroundings of a clinic playroom while at the same time broadening the base for observation. The conditions in the current study did not mirror those used in child abuse investigations. Because of time pressures and lack of training, child abuse investigations often involve asking many situation-specific questions after a very short period, if 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. A family of four dolls (Analeka Industries) was presented to the children: two dolls with adult features and clothes (one male, one female) and two smaller dolls with children’s features and clothes (again, one male and one female). The dolls are constructed of cloth and their body parts are proportional. They had removable clothing, their genitalia being visible only when the clothes were removed. Other toys were present in the playroom as well. The toys and furniture were arranged in a standardized manner before each subject entered the playroom. In Condition I(7 minutes), the children were allowed to play with the dolls and other toys
298
A. B. Sivan, D. P. Schor, G. K. Koeppl and L. D. Noble
in the presence of an unobtrusive adult who sat at the side of the room and occupied him/ herself with paperwork while the child explored the playroom. The adult did not interfere with the child’s free interaction and gave minimal answers to any comments addressed to him/her. In Condition II (7 minutes), the adult excused him/herself and left the playroom telling the child s/he would return shortly. This condition allowed the child to interact with the dolls and other toys without an adult present and provided an opportunity to examine whether an adult’s presence was an inhibiting factor in the development of the child’s play activity. Following Condition II, the same adult returned to the playroom to interview the child. S/he asked the child how these dolls differed from other dolls s/he had played with previously. The children were also asked for the names of specific body parts, starting first with neutral parts such as head, neck, arms, etc. and proceeding to those body parts involved in sexual activity. In the process, the dolls were undressed, and the child was asked about all four dolls. The lexicon obtained from this condition will be presented in a separate article. In Condition III (5 minutes), the adult again left the room. The difference between this condition and Condition II was that the dolls were left undressed, perhaps serving as a stimulus for a different kind of play. Parents of the subjects were encouraged to observe the play and the interview through a one-way mirror so that they could appropriately answer questions that might arise after the session. The child and parents were given the opportunity to discuss their observations and any questions they had with one of the investigators at the end of the session. Subjects were paid a sum of $10 for their time and effort. Sessions took place on weekends and evenings in order not to disturb the regular family routines. One of the principal investigators was on call during all sessions in the event that any difficult situations would arise, such as the disclosure of abuse. A beeper was used for contacting the on-call person. No cases of abuse were identified or reported during the study. Technical Considerations The entire session with the children and the dolls was recorded in the following manner: The children had a wireless FM microphone clipped to their clothing upon entering the playroom. Transmission of the child’s voice was received in an observation room next to the playroom by an FM receiver. The verbalizations of the child were recorded on one channel of a stereo tape recorder; simultaneously, a trained observer in the observation room gave a running account of the child’s behavior (interaction with the dolls or other toys, location in the room, etc.) on a second channel of the same stereo recorder. A mixer was added to the stereo set-up so that a second tape recorder could be used to play a signal tape marking oneminute intervals for the data analysis. The tapes were then transcribed by a secretary for later content analysis. Data Collection and Analysis A prepackaged observational scheme was not chosen for use because the researchers did not want to introduce prior expectations as to the types of interactions which might be observed. Instead, development of a tailor-made observational scheme was done during a prepilot phase in which the technical aspects of the study were tested. The scheme developed used descriptors in the form of phrases containing an actor, a verb, an object, and (if helpful) a body part used to accomplish the action. Descriptions of actions with the dolls were given priority over other actions. Variables such as location and intensity of action were initially considered but finally discarded as neither useful nor reliable. (A list of descriptors used by the two observers can be obtained from the senior author).
Anatomical doll play
299
Two observers were trained to give a running account of the events in the playroom. When an activity was repeated, the phrase “again” could be used. Although many observations were made within the one-minute intervals, no particular time interval was used in making the observational statements. Had a more exact time interval been used, a different coding system using each observation might have been used. Instead, the transcripts were coded in oneminute intervals, and each interval was assigned three codes encompassing activities which occurred during that period of time. The observations were extremely repetitious. Interobserver reliabilities were calculated on duplicate observations in which the two observers both observed the same sessions (buffered by industrial strength earphones to prevent the observers from hearing each other speaking). Reliability coefficients ranged from .80 to .95. Interrater reliabilities were calculated by comparing the coding of randomly duplicated transcripts included in the pool of transcripts. These coefficients ranged from .83 to 1.OOwith a mean of 90.7 and a median of 90.5. Coding of the transcripts was done with the implication that the child was always the actor/ subject of coded phrase or action. Three coding categories were used: 1. Activities/Verbs-( 11) exploration/imagination (touch, hold, moves, inspects); ( 12) imagination (caretaking, feeding, moving cars); ( 13) task orientation (playing by rules, reading); ( 14) dressing/undressing; and ( 15) throws/kicks. 2. Objects-(21) dolls; (22) toys; (23) people (self, others); (24) environment; (25) clothes; and (26) microphone. 3. Other (97) talking (when initiated by child) (98) wandering, sitting or doing nothing; (99) withdrawal (opening door, fear responses). Doll events were easily coded. Most difficulty was encountered in trying to determine when the microphone was interfering, when the child was wandering about the room, as opposed to exploring some aspect of the room itself such as the blackboard. There were some technical problems with the child’s microphone, and at times the child was inaudible. For this reason, no specific analysis was done on the talking children did during the session. However, there were times when the child’s speech made a difference in the coding. For example, if a child was reading a book aloud and the speech was clear, it allowed the coder to consider the activity as “task-oriented” play as opposed to mere “exploration” of the book. Much of the child’s talking was directed to the dolls as characters in role play activities, e.g., “Have some more soup, ” “That’s a good girl,” etc., and many verbalizations were singing or nonsense talk. A second approach to the analysis of the data was an analysis of sequences of behavior. Specifically, the sequence of events in the coded observation protocols was noted to allow examination of probable and nonprobable sequences. These analyses showed that children tended to repeat an activity many times in the course of an observational session, suggesting that the observations of the behavior of the children were not random. Moreover, the fact that observed events tended to repeat themselves many times during a given condition indicated that the observational system itself had strong internal consistency.
RESULTS The coded observations were analyzed using a number of chi-square tests (SAS package) to address the questions listed earlier. Significant differences were found on all but a few tests. The large number of significant differences may be, in part, an artifact of the large (7,302) number of observations considered. Similarly, some of the chi-square values were very large. For example, comparing doll vs. non-doll events for the three conditions yielded a chi-square of 2,657.9, d’ = 2. For this reason, the findings are presented below in terms of trends
A. B. Sivan, D. P. Schor, G. K. Koeppl and L. D. Noble
300
Table 3. Percentage
of Doll VS. Non-Doll
Events for 144 Children Sex of Child
Boys Srx cflnterviewer.
Aee 3-4 years 5-6 years 7-8 years
Girls
Female
Female
Male
Male
Doll Events
NonDoll Events
Doll Events
NonDoll Events
Doll Events
NonDoll Events
Doll Events
NonDoll Events
I II 111
7.98 11.23 23.24
92.02 88.77 76.76
4.81 6.71 2.88
95.19 93.29 97.12
21.30 31.58 38.19
73.70 68.48 61.81
10.26 12.00 34.57
89.74 88.00 65.43
I
17.01 II.17 16.1 I
82.99 88.83 83.89
14.63 6.52 I.52
85.37 93.48 98.47
35.37 38.68 39.02
64.63 61.32 60.98
13.61 19.35 37.90
86.89 80.65 62.10
7.66 21.10 5.13
92.34 78.90 94.87
10.47 6.49 16.57
89.53 93.5 I 83.43
36.23 28.94 39.66
63.77 7 1.06 60.34
34.05 28.92 44.7 I
65.95 71.08 55.29
Conditions
II III
I II III
and descriptions rather than significance levels. The few nonsignificant findings will be emphasized. During the coding of the data, it was found that seven pre-assigned observational categories were not used. These were imaginative play with people, imaginative play with environment, imaginative play with doll clothes, imaginative play with the’microphone, throwing or kicking people (self or others), throwing or kicking the environment, and throwing or kicking the microphone. Of note was the fact that kicking or throwing the dolls was never observed when the dolls were undressed, and that kicking or throwing the doll clothes was only observed under this condition, i.e., when the dolls were undressed. Throughout this discussion, “conditions” will be used to delineate the three settings in which the observational data was collected. Condition I had an adult present; Condition II had no adult in the room while the child was playing; Condition III was after the interview when the dolls had been undressed and the adult left the playroom. Of most significance is the finding that the dolls occupied little of the children’s interest. Table 3 depicts the percentages of doll as opposed to non-doll events for all 144 children across the three conditions. These percentages are broken down by age and sex of the child as well as gender of the interviewer. Across all the conditions, ages, and genders of both the children and the interviewers, the highest percentage of time the children were occupied with the dolls was 25.6%. These data are presented in Table 4. Only four categories of doll activities were observed in the children’s interactions with the dolls. These were exploration (inspecting, touching, holding, moving, looking); role play
Table 4. Percentage of Doll vs. Non-Doll Events for 144 Children (Collapsed over age and sex of child as well as gender of interviewer) Conditions Events
I
II
III
Doll Non-Doll
17.81 82.19
19.65 80.35
25.60 74.40
Anatomical doll play
301
Table 5. Percentages of Doll and Non-Doll Activities for Boys and Girls Conditions I
III
II
Doll Activity
Boys
Girls
Boys
Girls
Girls
Exploration Role taking Dressing Aggression
6.29 I .64 2.16 .20
10.10 4.23 10.35 .38
5.10 3.15 2.30 .43
10.02 6.17 11.18 .I5
6.53 1.05 3.97
8.00 2.09 29.47
Totals Doll Non-Doll
10.29 89.7 1
25.06 14.96
10.97 89.03
27.53 72.47
11.55 88.45
39.56 60.44
Boys
(feeding, caretaking); dressing; and kicking or throwing (aggression). Exploration was the most often observed category of activity under Conditions I and II, whereas dressing was the most frequently observed category or activity under Condition III when the dolls were left undressed. Under all the observational conditions, aggression (kicking or throwing) toward the dolls had a frequency of less than 1% of the observed interactions. In fact, the frequencies were low enough as to be notable. Under Conditions I and II combined, 16 aggressive events towards the dolls were observed in a total of 5,583 events. As noted earlier, no aggression was observed in Condition III. Feeding was the most popular role taking activity with the dolls. Table 5 depicts the percentages of different types of doll activity as well as the total non-doll activities under the three observational conditions for boys and girls. As is obvious from Table 5, girls played with the dolls more than boys under all the conditions (Condition I, chi-square = 115.9, df= 1, p < .OOl; Condition II, chi-square = 107.1, df = 1, p < .OO1; Condition III, chi-square = 177, df = 1, p < .OO1). Girls were also more influenced by the changing circumstances of Condition III and engaged in more dressing behaviors with the dolls when the dolls were presented undressed. The gender of the interviewer had an effect on the children’s interactions with the dolls. In the first two conditions, children played with the dolls more when a female interviewer had been present in Condition I. Table 6 depicts this finding (Condition 1, chi-square = 20.8, df = 1, p < .OOl; Condition II, chi-square = 43.0, df = 1, p c.001; Condition III, chi-square = 3.1, ns). In contrast to the gender variables, the age ofthe child had a differential effect on the interactions between the children and the dolls. Younger children interacted significantly less with the dolls than did older children under the first two conditions, and in Condition III, no significant differences were found among the three age groups: 3-4 year olds, 5-6 year olds and
Table 6. Percentage of Doll vs. Non-Doll
Events for Female and Male Interviewers
Conditions I
II
III
Events
Female
Male
Female
Male
Female
Male
Doll Non-Doll
20.97 79.03
14.71 85.29
24.63 75.37
14.14 85.86
27.30 72.10
23.57 76.43
A. B. Sivan, D. P. Schor, G. K. Koeppl
302
and L. D. Noble
Table 7. Percentage of Doll vs. Non-Doll Events for Three Age Groups Conditions I
II
III
Ages
Doll Events
Non-Doll Events
Doll Events
Non-Doll Events
Doll Events
Non-Doll Events
3-4 5-6 7-8
11.17 20.06 22.57
88.83 79.94 77.43
16.39 20.34 21.53
83.61 79.65 78.47
25.27 24.12 27.06
74.73 75.88 72.94
7-8 year olds (Condition I, chi-square = 52.2, df = 2, p < .OOl; Condition 11,chi-square = 7.2, & = 2, p < .05; Condition III, chi-square = 1.4, df = 2, ns). These findings are depicted in Table 7. Examination of individu~ categories under the various conditions yields several trends. Under the first two conditions (i.e., before the dolls were undressed), the most frequent activity, it-regardless of age or sex of the child or gender of the adult, was exploration of the other toys in the playroom. This remained the most popular activity for boys in the third condition as well. Girls, however, switched their focus to dressing the dolls under the third condition. In addition, a notable increase in wandering around the room was noted for both boys and girls in the third condition. A notable trend was seen in the youngest age group (3-4 years old); these children were observed to refuse to stay alone in the playroom during the third condition. These responses comprised 5.1% of their total responses during the third condition. In addition, an increase in wandering and exploration of the microphone by the older children was also noted in the third condition. A sequential analysis of the data was done by calculating the probability that one event would follow another specified event. Again, the data were studied by examining the differences among the conditions, the age and gender of the child, and the gender of the adult interviewer. In all conditions, those categories which included over 5% of the observations were found to be preceded by themselves; in other words, the event most likely to precede a given event was the same event itself. Doll activity was preceded by doll activity; role playing was preceded by role playing. This finding indicated that the observation system was not random; children tended to repeat an activity many times in the course of an observational session. Moreover, these probabilities were not insignificant; for all children, the probabilities for events comprising over 5% of the total observations ranged from 69.43 to 9 1.7 1. Similar values were obtained when the sample was broken down in the twelve groups.
DISCUSSION Anatomical (sexually explicit) dolls are not particularly interesting toys for normal children. In fact, given the choice between what appeared to be many ordinary toys, these novel dolls were found to have marginal attractiveness. Moreover, although the sexual body parts of these dolls were inspected by many children, role playing of explicit sexual behavior was not observed. Contrary to the assertion sometimes made about the dolls in judicial proceedings, these dolls did not appear to stimulate discussions or activities of a sexual nature. Aggression, in general and toward the dolls in particular, was a rare event. Less than 1% of the observed interactions were aggressive. Of particular interest was the observation that when the dolls were left undressed, no incidents of aggression towards the dolls were observed.
Anatomical doll play
303
As might be expected, girls found the dolls more attractive and played with them more than did boys. However, the finding that the presence of a female interviewer stimulated children’s play with the dolls was unexpected. It is interesting to speculate whether children see females as more nurturant and therefore identify with the role or whether in some way the female interviewers themselves subtly suggest doll play more than do the male interviewers. Should this finding be replicated, investigating agencies might want to consider using only females in those investigations in which doll interviews or doll play are utilized. The youngest children (3-4 year olds) were observed to play less with the dolls when the dolls were dressed than did other children. During the study, these young children sought to communicate with the adult when s/he was in the room with the child. Once the adult left the room, these children behaved in a manner more similar to that of the older children. The youngest children also had a tendency to leave the room or to refuse to stay alone after the doll interview. In general, they appeared to be less comfortable and more unsure of the expectations for their behavior under the different conditions. It was hoped that a sequential analysis of the data would provide some insight into the antecedents of particular behaviors such as aggression, wandering, or leaving the room. Instead, these analyses provided validity information on the observation system itself. Since this pilot study involved only middle class subjects, the degree to which its findings apply to other socioeconomic and cultural groups needs to be determined. Replication of the study is suggested for other groups of subjects and might best be done with only minor changes in methodology. Likewise, a larger, more systematic study exploring the differences between the doll play of nonabused and abused children might follow; data from such a study could provide information which would be invaluable in the training of child health workers involved in the investigation of abuse incidents. The authors are now considering the possibility of developing a projective evaluative technique based on free play/observation of interactions of children with the dolls. The noninvasive nature of such a technique would provide many advantages over the often criticized interview methodologies. The results of the present study suggest that in most ways children’s reactions to anatomical dolls are similar to their reactions to other toys. Thus, unusual behavior observed in interactions with these dolls should be taken seriously. In direct contrast to clinical observations of the play of abused children, observations of these 144 nonreferred children suggest that the doll play of nonabused children is, for the most part, free of aggression and sexual concerns.
REFERENCES 1. BOAT, B. W. and EVERSON, M. D. Using anatomical dolls: Guidelines for interviewing young children in sexual abuse investigations. Department of Psychiatry, University of North Carolina, Chapel Hill, NC (1986). 2. DUNN, L. M. and DUNN, L. M. Peabody Picture Vocabulury Test-Revised. American Guidance Service, Circle Pines, MN (1981). 3. JAMPOLE, L. and WEBER, M. K. An assessment of the behavior of sexually abused and nonsexually abused children with anatomically correct dolls. ChildAbuse& Neglect 11:187-192 (1986). 4. WHITE, S., STROM, G. A. and SANTILLI, G. Clinical protocol for interviewing preschoolers with sexually anatomically correct dolls. School of Medicine, Case Western Reserve University, Cleveland, OH (mime0 l/11/ 86). 5. WHITE, S., STROM, G. A., SANTILLI, G. and HALPIN, B. M. Interviewing young sexual abuse victims with anatomically correct dolls. ChildAbuse & Neglect IO:5 19-529 (1986). R&sum&-Des poupkes en Ctoffe, anatomiques (explicites du point de vue sexuel) sont souvent utiliies lors d’interviews d’enfants sou~onnk d’avoir &z maltraitb sexuellement. On a observk que ces enfants maltraitks ont tendance cliniquement B se montrer agressifs et parfois prk&ment sexuels dans leurs jeux avec la poupke; souvent dans les tribunaux un tkmoignage de cette sorte est accepti comme une preuve suffisante de la mat&alit& du d&lit. Cependant,
304
A. B. Sivan, D. P. Schor. G. K. Koeppl and L. D. Noble
il manque jusqu’i ce jour des don&es fondamentales SW le compo~em~nt de jeu dimontre par des enfants normaux lorsqu’ils jouent avec des poupkes sexuellement explicites. La presente etude apporte une information empirique sur le jeu interactif d’un relativement grand groupe temoin d’enfants normaux et non r&&s pour un souwon queiconque de maltraitance avec des poupees anatomiques. Les auteurs ont observe 144 enfants ages de 3 i 8 ans dans une salle de jeux qui contenait des poupees de ce genre, (I) en presence d’un adulte, (2) sans adulte, et (3) lorsque les poupees dtaient deshabillies. Les observations faites montrent que ces enfants normaux n’ont pas trot& ces poupdes anatomiques plus intiressantes que d’autres jeux. On a observe peu d’agressivite ni d’activite explicitement sexuelle. Par comparaison avec I’observation clinique d’enfants ayant subi des &ices sexuels; le jeu avec les poupkes des enfants temoins, differe par l’absence d’agressivite ou de preoccupation sexuelle. Ainsi lorsque ce genre de comportement est present chez des enfants qui jouent avec de telles poup&es, il devrait itre pris au sirieux. Resumen-Muiiecas de trapo anatomicas (sexualmente explicitas) son usadas frecuentemente en entrevistas con menores de 10s cuales se sospecha que han sido sexualmente abusados. Se ha notado clinicamente que 10s niiios abusados son agresivos y a veces sexualmente precoces cuando juegan con Lasmuiinecas, y testimonio a tal efecto es f~cuentemente aceptado por el tribunal coma evidencia. Sin embargo hasta el presente no existen datos normativos con respect0 al juego con muiiecas sexualmente explicitas. Esta invest&a&t pilot0 provee info~aci~n empirica acerca del juego de niiios normales con muiiecas anatomicas. Ciento cuarenta y cuatro niiios de tres a echo adios de edad fueron observados en un cuarto de juego conteniendo muiiecas anatomicas ( 1) con un adult0 presente, (2) sin un adulto, (3) con las muiiecas desvestidas. Las observaciones mostraron que 10s niiios normales no encontraron estas muiiecas mis interesantes que 10s otros juguetes. No se observe mas que poca agresion y ninguna actividad sexual explicita. En contraste con las observaciones clinicas de niiios abusados el juego con las muiiecas de nirios not-males no tiende a ser caracterizado por la agresion o la sexualidad: por tanto este tipo de conducta debe ser tornado seriamente.
Nqkcr, Vol. IT!, pp. 295-304, Printed in the U.S.A. All rights reserved.
0145-2134/88$3.00+ .oO Copyrigtrt 0 1988 Pe~gamon Press plc
1988
r~E~~I~N OF NORMAL CHILDREN WITH ANATOMICAL DOLLS ABIGAIL B. SIVAN, PH.D., DAVID P. SCHOR, M.D., GINA K. KOEPPL, PH.D. AND LXNND. NOBLE, PH.D. Child Development Clinic, Division ofDevelopmental Disabilities, Department of Pediatrics, University of Iowa, Iowa City, IA
Abstr&ct-A~atorni~~ (sexually exphcit) rag dolls are frequently used in interviews of children suspected of being sexually abused. Abused children have been noted clinically to be aggressive and sometimes sexually precocious in their doll play, and testimony to that effect is often accepted in the caurtroom as pertinent evidence. However, to date, normative data on the play behaviors with the sexually explicit dolls have been unavailable. This pilot study provides empirical information on the play interaction of a relatively large number of normal (nonreferred) children with anatomical dolls. One hundred forty-four children (ages 3 to 8 years) were observed in a playroom containing these special dolls during three conditions: (I) with an adult present, (2) without an adult, and (3) with the dolls undressed. The observations showed that nanreferred children found these dolls no mom interesting than other toys, Little aggression and no explicit sexual activity were observed. In contrast to clinical observation of abused children, the doll play of nonreferred children is unlikely to be characterized by aggression or sexual concerns; thus these behaviors when observed in interaction with these dolls should be taken seriously.
A Technique OFTEN USED in the investigation of sexual abuse is an interview using anatomical (sexually explicit) rag dolls. Although testimony based on the use of these dolls is often accepted in the courtroom, little data exist which provide information about how normal (nonreferred) children freely interact with these dolls. This pilot project was developed to explore the free play interactions of nonreferred children with anatomical dolls and to compare these findings with clinical impressions of abused children. The data presented here were obtained as a part of a larger study. Many health professionals who work daily with children maintain that the play of abused children is substantially different from that of normal (nonreferred) children and testify to that effect in court. Guidelines have been developed for standardizing interviews with sexually abused children f I, 41 and recent research presented ~orn~~n~ the responses of abused and nonabused children f5]. However, the one published ~nves~~~on [3] which has explored the free play of normal children with these dolts is plagued by small sample sizes (n = 20) and broad units of analysis, e.g., sexual activity vs. no sexual activity. Many investigators (social workers and Iaw enforcement officers) have limited knowledge of This research was funded by the Iowa State Criminal and Juvenile Justice Agency as well as the Office of the Vice President for Educational Development and Research, University of Iowa. An earlier version of this paper was presented at the biennial meeting of the Society for Research in Child Develop ment, Baltimore, MD, April 25, 1987. Received for pubiication May 27,1987: final revision received October 15,1987; accepted February 19,198g. Reprint requests may be sent to A. Sivan, Ph.D., Division of Child Psychiitry, Rush-~~b~e~an-St. Center, 1720West Polk Street, Chicago, IL 6Q6i 2. 295
Luke’s Medical
296
A. B. Sivan, D. P. Schor, G. K. Koeppl and L. D. Noble Table 1. Mean (and Range) Age of the Subjects in Months Sex of Interviewer Sex of Child
Boy 71.29 (37-107)
Girl 72.78 (37-107)
Female
Male
51.08 (38-57)
48.75 (37-58)
69.58 (60-79)
70.92 (60-83)
94.08 (85-107)
93.33 (84-106)
47.92 (38-54)
48.25 (37-47)
72.58 (60-83)
74.08 (66-83)
95.42 (87-101)
98.42 (85-107)
71.78 (38-107)
72.79 (37-107)
normal child development and child behavior, and those child specialists who possess ample experience lack specific comparison data which describe the typical interactions (behavioral and verbal) of normal (nonreferred) children with the dolls. Without these comparison data, the use of doll play observations of investigated children may be open to challenge in court. For the above reasons, this pilot study was designed to explore some of commonly held but unresearched assumptions about the dolls. Specifically, it addressed the following questions: 1. In what ways do the children interact with the dolls (i.e., do they exhibit particular themes in their interactions such as caretaking or creative role playing)? 2. Are there differences in the children’s interactions which are systematically related to age or gender? 3. Are there differences in the children’s interactions with the dolls related to the presence or absence of an adult professional? 4. Does the gender of the adult present affect the child’s interactions? 5. For what amount of time do these dolls keep the child’s interest and is this related to age or gender? All these questions have clear clinical implications for professionals who are asked to relate observations of children’s play as a part of investigation and evaluation of the emotional and intellectual competence of the abused victim.
METHOD Subjects Subjects were 144 children (3-8 years old) whose parents agreed to their participation. jects were solicited through advertisements posted in local nursery schools and pediatric child clinics and through news releases. Subjects were stratified by age (3 and 4 years, 5 years, 7 and 8 years), gender of the child, and gender of interviewer into 12 groups children each. This stratification is depicted in Table 1,
Subwelland 6 of 12
Anatomical doll play Table 2. Scores on the PPVT-R months)
297
(Age equivalents
in
Sex of Interviewer Sex of Child
Female 57.08
ROY
83.53 (29-141)
Girl 14.86 1n :39)
Male 60.25
(30-74) 87.42 (66-l 12) 109.92 (89-141)
(29-86) 75.33 (51-98) 111.17 (78-136)
58.33 (39-88)
54.42 (40-77)
83.92
89.92 (68-128)
(62-121) (86-130)
I 17.00 (84- 139)
83.7 I (30-141)
84.68 (29-139)
105.58
Socioeconomic status was to be considered, but attempts to find a low status group in the greater Iowa City area met with failure. All the fathers reported at least a high school education; 63.2% had at least one college degree; and 72.1% held semiprofessional or professional jobs. Only two mothers did not complete high school; 60.5% had at least one college degree; and 43.8% held semiprofessional or professional jobs. Homemakers or students comprised 40.3% of the mothers. No information was gathered concerning marital status of the parents. Procedures
Informed consent was obtained, and all subjects completed the Peabody Picture Vocabulary Test-Revised (PPVT-R) [2] to provide a screening of general intellectual level (IQ equivalent above 85). Table 2 presents data from the PPVT-R showing both the mean score and range of scores for each of the 12 groups. Data collection involved observing and recording the children’s interactions with the dolls under three conditions designed to help the children feel comfortable in the unfamiliar surroundings of a clinic playroom while at the same time broadening the base for observation. The conditions in the current study did not mirror those used in child abuse investigations. Because of time pressures and lack of training, child abuse investigations often involve asking many situation-specific questions after a very short period, if 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. A family of four dolls (Analeka Industries) was presented to the children: two dolls with adult features and clothes (one male, one female) and two smaller dolls with children’s features and clothes (again, one male and one female). The dolls are constructed of cloth and their body parts are proportional. They had removable clothing, their genitalia being visible only when the clothes were removed. Other toys were present in the playroom as well. The toys and furniture were arranged in a standardized manner before each subject entered the playroom. In Condition I(7 minutes), the children were allowed to play with the dolls and other toys
298
A. B. Sivan, D. P. Schor, G. K. Koeppl and L. D. Noble
in the presence of an unobtrusive adult who sat at the side of the room and occupied him/ herself with paperwork while the child explored the playroom. The adult did not interfere with the child’s free interaction and gave minimal answers to any comments addressed to him/her. In Condition II (7 minutes), the adult excused him/herself and left the playroom telling the child s/he would return shortly. This condition allowed the child to interact with the dolls and other toys without an adult present and provided an opportunity to examine whether an adult’s presence was an inhibiting factor in the development of the child’s play activity. Following Condition II, the same adult returned to the playroom to interview the child. S/he asked the child how these dolls differed from other dolls s/he had played with previously. The children were also asked for the names of specific body parts, starting first with neutral parts such as head, neck, arms, etc. and proceeding to those body parts involved in sexual activity. In the process, the dolls were undressed, and the child was asked about all four dolls. The lexicon obtained from this condition will be presented in a separate article. In Condition III (5 minutes), the adult again left the room. The difference between this condition and Condition II was that the dolls were left undressed, perhaps serving as a stimulus for a different kind of play. Parents of the subjects were encouraged to observe the play and the interview through a one-way mirror so that they could appropriately answer questions that might arise after the session. The child and parents were given the opportunity to discuss their observations and any questions they had with one of the investigators at the end of the session. Subjects were paid a sum of $10 for their time and effort. Sessions took place on weekends and evenings in order not to disturb the regular family routines. One of the principal investigators was on call during all sessions in the event that any difficult situations would arise, such as the disclosure of abuse. A beeper was used for contacting the on-call person. No cases of abuse were identified or reported during the study. Technical Considerations The entire session with the children and the dolls was recorded in the following manner: The children had a wireless FM microphone clipped to their clothing upon entering the playroom. Transmission of the child’s voice was received in an observation room next to the playroom by an FM receiver. The verbalizations of the child were recorded on one channel of a stereo tape recorder; simultaneously, a trained observer in the observation room gave a running account of the child’s behavior (interaction with the dolls or other toys, location in the room, etc.) on a second channel of the same stereo recorder. A mixer was added to the stereo set-up so that a second tape recorder could be used to play a signal tape marking oneminute intervals for the data analysis. The tapes were then transcribed by a secretary for later content analysis. Data Collection and Analysis A prepackaged observational scheme was not chosen for use because the researchers did not want to introduce prior expectations as to the types of interactions which might be observed. Instead, development of a tailor-made observational scheme was done during a prepilot phase in which the technical aspects of the study were tested. The scheme developed used descriptors in the form of phrases containing an actor, a verb, an object, and (if helpful) a body part used to accomplish the action. Descriptions of actions with the dolls were given priority over other actions. Variables such as location and intensity of action were initially considered but finally discarded as neither useful nor reliable. (A list of descriptors used by the two observers can be obtained from the senior author).
Anatomical doll play
299
Two observers were trained to give a running account of the events in the playroom. When an activity was repeated, the phrase “again” could be used. Although many observations were made within the one-minute intervals, no particular time interval was used in making the observational statements. Had a more exact time interval been used, a different coding system using each observation might have been used. Instead, the transcripts were coded in oneminute intervals, and each interval was assigned three codes encompassing activities which occurred during that period of time. The observations were extremely repetitious. Interobserver reliabilities were calculated on duplicate observations in which the two observers both observed the same sessions (buffered by industrial strength earphones to prevent the observers from hearing each other speaking). Reliability coefficients ranged from .80 to .95. Interrater reliabilities were calculated by comparing the coding of randomly duplicated transcripts included in the pool of transcripts. These coefficients ranged from .83 to 1.OOwith a mean of 90.7 and a median of 90.5. Coding of the transcripts was done with the implication that the child was always the actor/ subject of coded phrase or action. Three coding categories were used: 1. Activities/Verbs-( 11) exploration/imagination (touch, hold, moves, inspects); ( 12) imagination (caretaking, feeding, moving cars); ( 13) task orientation (playing by rules, reading); ( 14) dressing/undressing; and ( 15) throws/kicks. 2. Objects-(21) dolls; (22) toys; (23) people (self, others); (24) environment; (25) clothes; and (26) microphone. 3. Other (97) talking (when initiated by child) (98) wandering, sitting or doing nothing; (99) withdrawal (opening door, fear responses). Doll events were easily coded. Most difficulty was encountered in trying to determine when the microphone was interfering, when the child was wandering about the room, as opposed to exploring some aspect of the room itself such as the blackboard. There were some technical problems with the child’s microphone, and at times the child was inaudible. For this reason, no specific analysis was done on the talking children did during the session. However, there were times when the child’s speech made a difference in the coding. For example, if a child was reading a book aloud and the speech was clear, it allowed the coder to consider the activity as “task-oriented” play as opposed to mere “exploration” of the book. Much of the child’s talking was directed to the dolls as characters in role play activities, e.g., “Have some more soup, ” “That’s a good girl,” etc., and many verbalizations were singing or nonsense talk. A second approach to the analysis of the data was an analysis of sequences of behavior. Specifically, the sequence of events in the coded observation protocols was noted to allow examination of probable and nonprobable sequences. These analyses showed that children tended to repeat an activity many times in the course of an observational session, suggesting that the observations of the behavior of the children were not random. Moreover, the fact that observed events tended to repeat themselves many times during a given condition indicated that the observational system itself had strong internal consistency.
RESULTS The coded observations were analyzed using a number of chi-square tests (SAS package) to address the questions listed earlier. Significant differences were found on all but a few tests. The large number of significant differences may be, in part, an artifact of the large (7,302) number of observations considered. Similarly, some of the chi-square values were very large. For example, comparing doll vs. non-doll events for the three conditions yielded a chi-square of 2,657.9, d’ = 2. For this reason, the findings are presented below in terms of trends
A. B. Sivan, D. P. Schor, G. K. Koeppl and L. D. Noble
300
Table 3. Percentage
of Doll VS. Non-Doll
Events for 144 Children Sex of Child
Boys Srx cflnterviewer.
Aee 3-4 years 5-6 years 7-8 years
Girls
Female
Female
Male
Male
Doll Events
NonDoll Events
Doll Events
NonDoll Events
Doll Events
NonDoll Events
Doll Events
NonDoll Events
I II 111
7.98 11.23 23.24
92.02 88.77 76.76
4.81 6.71 2.88
95.19 93.29 97.12
21.30 31.58 38.19
73.70 68.48 61.81
10.26 12.00 34.57
89.74 88.00 65.43
I
17.01 II.17 16.1 I
82.99 88.83 83.89
14.63 6.52 I.52
85.37 93.48 98.47
35.37 38.68 39.02
64.63 61.32 60.98
13.61 19.35 37.90
86.89 80.65 62.10
7.66 21.10 5.13
92.34 78.90 94.87
10.47 6.49 16.57
89.53 93.5 I 83.43
36.23 28.94 39.66
63.77 7 1.06 60.34
34.05 28.92 44.7 I
65.95 71.08 55.29
Conditions
II III
I II III
and descriptions rather than significance levels. The few nonsignificant findings will be emphasized. During the coding of the data, it was found that seven pre-assigned observational categories were not used. These were imaginative play with people, imaginative play with environment, imaginative play with doll clothes, imaginative play with the’microphone, throwing or kicking people (self or others), throwing or kicking the environment, and throwing or kicking the microphone. Of note was the fact that kicking or throwing the dolls was never observed when the dolls were undressed, and that kicking or throwing the doll clothes was only observed under this condition, i.e., when the dolls were undressed. Throughout this discussion, “conditions” will be used to delineate the three settings in which the observational data was collected. Condition I had an adult present; Condition II had no adult in the room while the child was playing; Condition III was after the interview when the dolls had been undressed and the adult left the playroom. Of most significance is the finding that the dolls occupied little of the children’s interest. Table 3 depicts the percentages of doll as opposed to non-doll events for all 144 children across the three conditions. These percentages are broken down by age and sex of the child as well as gender of the interviewer. Across all the conditions, ages, and genders of both the children and the interviewers, the highest percentage of time the children were occupied with the dolls was 25.6%. These data are presented in Table 4. Only four categories of doll activities were observed in the children’s interactions with the dolls. These were exploration (inspecting, touching, holding, moving, looking); role play
Table 4. Percentage of Doll vs. Non-Doll Events for 144 Children (Collapsed over age and sex of child as well as gender of interviewer) Conditions Events
I
II
III
Doll Non-Doll
17.81 82.19
19.65 80.35
25.60 74.40
Anatomical doll play
301
Table 5. Percentages of Doll and Non-Doll Activities for Boys and Girls Conditions I
III
II
Doll Activity
Boys
Girls
Boys
Girls
Girls
Exploration Role taking Dressing Aggression
6.29 I .64 2.16 .20
10.10 4.23 10.35 .38
5.10 3.15 2.30 .43
10.02 6.17 11.18 .I5
6.53 1.05 3.97
8.00 2.09 29.47
Totals Doll Non-Doll
10.29 89.7 1
25.06 14.96
10.97 89.03
27.53 72.47
11.55 88.45
39.56 60.44
Boys
(feeding, caretaking); dressing; and kicking or throwing (aggression). Exploration was the most often observed category of activity under Conditions I and II, whereas dressing was the most frequently observed category or activity under Condition III when the dolls were left undressed. Under all the observational conditions, aggression (kicking or throwing) toward the dolls had a frequency of less than 1% of the observed interactions. In fact, the frequencies were low enough as to be notable. Under Conditions I and II combined, 16 aggressive events towards the dolls were observed in a total of 5,583 events. As noted earlier, no aggression was observed in Condition III. Feeding was the most popular role taking activity with the dolls. Table 5 depicts the percentages of different types of doll activity as well as the total non-doll activities under the three observational conditions for boys and girls. As is obvious from Table 5, girls played with the dolls more than boys under all the conditions (Condition I, chi-square = 115.9, df= 1, p < .OOl; Condition II, chi-square = 107.1, df = 1, p < .OO1; Condition III, chi-square = 177, df = 1, p < .OO1). Girls were also more influenced by the changing circumstances of Condition III and engaged in more dressing behaviors with the dolls when the dolls were presented undressed. The gender of the interviewer had an effect on the children’s interactions with the dolls. In the first two conditions, children played with the dolls more when a female interviewer had been present in Condition I. Table 6 depicts this finding (Condition 1, chi-square = 20.8, df = 1, p < .OOl; Condition II, chi-square = 43.0, df = 1, p c.001; Condition III, chi-square = 3.1, ns). In contrast to the gender variables, the age ofthe child had a differential effect on the interactions between the children and the dolls. Younger children interacted significantly less with the dolls than did older children under the first two conditions, and in Condition III, no significant differences were found among the three age groups: 3-4 year olds, 5-6 year olds and
Table 6. Percentage of Doll vs. Non-Doll
Events for Female and Male Interviewers
Conditions I
II
III
Events
Female
Male
Female
Male
Female
Male
Doll Non-Doll
20.97 79.03
14.71 85.29
24.63 75.37
14.14 85.86
27.30 72.10
23.57 76.43
A. B. Sivan, D. P. Schor, G. K. Koeppl
302
and L. D. Noble
Table 7. Percentage of Doll vs. Non-Doll Events for Three Age Groups Conditions I
II
III
Ages
Doll Events
Non-Doll Events
Doll Events
Non-Doll Events
Doll Events
Non-Doll Events
3-4 5-6 7-8
11.17 20.06 22.57
88.83 79.94 77.43
16.39 20.34 21.53
83.61 79.65 78.47
25.27 24.12 27.06
74.73 75.88 72.94
7-8 year olds (Condition I, chi-square = 52.2, df = 2, p < .OOl; Condition 11,chi-square = 7.2, & = 2, p < .05; Condition III, chi-square = 1.4, df = 2, ns). These findings are depicted in Table 7. Examination of individu~ categories under the various conditions yields several trends. Under the first two conditions (i.e., before the dolls were undressed), the most frequent activity, it-regardless of age or sex of the child or gender of the adult, was exploration of the other toys in the playroom. This remained the most popular activity for boys in the third condition as well. Girls, however, switched their focus to dressing the dolls under the third condition. In addition, a notable increase in wandering around the room was noted for both boys and girls in the third condition. A notable trend was seen in the youngest age group (3-4 years old); these children were observed to refuse to stay alone in the playroom during the third condition. These responses comprised 5.1% of their total responses during the third condition. In addition, an increase in wandering and exploration of the microphone by the older children was also noted in the third condition. A sequential analysis of the data was done by calculating the probability that one event would follow another specified event. Again, the data were studied by examining the differences among the conditions, the age and gender of the child, and the gender of the adult interviewer. In all conditions, those categories which included over 5% of the observations were found to be preceded by themselves; in other words, the event most likely to precede a given event was the same event itself. Doll activity was preceded by doll activity; role playing was preceded by role playing. This finding indicated that the observation system was not random; children tended to repeat an activity many times in the course of an observational session. Moreover, these probabilities were not insignificant; for all children, the probabilities for events comprising over 5% of the total observations ranged from 69.43 to 9 1.7 1. Similar values were obtained when the sample was broken down in the twelve groups.
DISCUSSION Anatomical (sexually explicit) dolls are not particularly interesting toys for normal children. In fact, given the choice between what appeared to be many ordinary toys, these novel dolls were found to have marginal attractiveness. Moreover, although the sexual body parts of these dolls were inspected by many children, role playing of explicit sexual behavior was not observed. Contrary to the assertion sometimes made about the dolls in judicial proceedings, these dolls did not appear to stimulate discussions or activities of a sexual nature. Aggression, in general and toward the dolls in particular, was a rare event. Less than 1% of the observed interactions were aggressive. Of particular interest was the observation that when the dolls were left undressed, no incidents of aggression towards the dolls were observed.
Anatomical doll play
303
As might be expected, girls found the dolls more attractive and played with them more than did boys. However, the finding that the presence of a female interviewer stimulated children’s play with the dolls was unexpected. It is interesting to speculate whether children see females as more nurturant and therefore identify with the role or whether in some way the female interviewers themselves subtly suggest doll play more than do the male interviewers. Should this finding be replicated, investigating agencies might want to consider using only females in those investigations in which doll interviews or doll play are utilized. The youngest children (3-4 year olds) were observed to play less with the dolls when the dolls were dressed than did other children. During the study, these young children sought to communicate with the adult when s/he was in the room with the child. Once the adult left the room, these children behaved in a manner more similar to that of the older children. The youngest children also had a tendency to leave the room or to refuse to stay alone after the doll interview. In general, they appeared to be less comfortable and more unsure of the expectations for their behavior under the different conditions. It was hoped that a sequential analysis of the data would provide some insight into the antecedents of particular behaviors such as aggression, wandering, or leaving the room. Instead, these analyses provided validity information on the observation system itself. Since this pilot study involved only middle class subjects, the degree to which its findings apply to other socioeconomic and cultural groups needs to be determined. Replication of the study is suggested for other groups of subjects and might best be done with only minor changes in methodology. Likewise, a larger, more systematic study exploring the differences between the doll play of nonabused and abused children might follow; data from such a study could provide information which would be invaluable in the training of child health workers involved in the investigation of abuse incidents. The authors are now considering the possibility of developing a projective evaluative technique based on free play/observation of interactions of children with the dolls. The noninvasive nature of such a technique would provide many advantages over the often criticized interview methodologies. The results of the present study suggest that in most ways children’s reactions to anatomical dolls are similar to their reactions to other toys. Thus, unusual behavior observed in interactions with these dolls should be taken seriously. In direct contrast to clinical observations of the play of abused children, observations of these 144 nonreferred children suggest that the doll play of nonabused children is, for the most part, free of aggression and sexual concerns.
REFERENCES 1. BOAT, B. W. and EVERSON, M. D. Using anatomical dolls: Guidelines for interviewing young children in sexual abuse investigations. Department of Psychiatry, University of North Carolina, Chapel Hill, NC (1986). 2. DUNN, L. M. and DUNN, L. M. Peabody Picture Vocabulury Test-Revised. American Guidance Service, Circle Pines, MN (1981). 3. JAMPOLE, L. and WEBER, M. K. An assessment of the behavior of sexually abused and nonsexually abused children with anatomically correct dolls. ChildAbuse& Neglect 11:187-192 (1986). 4. WHITE, S., STROM, G. A. and SANTILLI, G. Clinical protocol for interviewing preschoolers with sexually anatomically correct dolls. School of Medicine, Case Western Reserve University, Cleveland, OH (mime0 l/11/ 86). 5. WHITE, S., STROM, G. A., SANTILLI, G. and HALPIN, B. M. Interviewing young sexual abuse victims with anatomically correct dolls. ChildAbuse & Neglect IO:5 19-529 (1986). R&sum&-Des poupkes en Ctoffe, anatomiques (explicites du point de vue sexuel) sont souvent utiliies lors d’interviews d’enfants sou~onnk d’avoir &z maltraitb sexuellement. On a observk que ces enfants maltraitks ont tendance cliniquement B se montrer agressifs et parfois prk&ment sexuels dans leurs jeux avec la poupke; souvent dans les tribunaux un tkmoignage de cette sorte est accepti comme une preuve suffisante de la mat&alit& du d&lit. Cependant,
304
A. B. Sivan, D. P. Schor. G. K. Koeppl and L. D. Noble
il manque jusqu’i ce jour des don&es fondamentales SW le compo~em~nt de jeu dimontre par des enfants normaux lorsqu’ils jouent avec des poupkes sexuellement explicites. La presente etude apporte une information empirique sur le jeu interactif d’un relativement grand groupe temoin d’enfants normaux et non r&&s pour un souwon queiconque de maltraitance avec des poupees anatomiques. Les auteurs ont observe 144 enfants ages de 3 i 8 ans dans une salle de jeux qui contenait des poupees de ce genre, (I) en presence d’un adulte, (2) sans adulte, et (3) lorsque les poupees dtaient deshabillies. Les observations faites montrent que ces enfants normaux n’ont pas trot& ces poupdes anatomiques plus intiressantes que d’autres jeux. On a observe peu d’agressivite ni d’activite explicitement sexuelle. Par comparaison avec I’observation clinique d’enfants ayant subi des &ices sexuels; le jeu avec les poupkes des enfants temoins, differe par l’absence d’agressivite ou de preoccupation sexuelle. Ainsi lorsque ce genre de comportement est present chez des enfants qui jouent avec de telles poup&es, il devrait itre pris au sirieux. Resumen-Muiiecas de trapo anatomicas (sexualmente explicitas) son usadas frecuentemente en entrevistas con menores de 10s cuales se sospecha que han sido sexualmente abusados. Se ha notado clinicamente que 10s niiios abusados son agresivos y a veces sexualmente precoces cuando juegan con Lasmuiinecas, y testimonio a tal efecto es f~cuentemente aceptado por el tribunal coma evidencia. Sin embargo hasta el presente no existen datos normativos con respect0 al juego con muiiecas sexualmente explicitas. Esta invest&a&t pilot0 provee info~aci~n empirica acerca del juego de niiios normales con muiiecas anatomicas. Ciento cuarenta y cuatro niiios de tres a echo adios de edad fueron observados en un cuarto de juego conteniendo muiiecas anatomicas ( 1) con un adult0 presente, (2) sin un adulto, (3) con las muiiecas desvestidas. Las observaciones mostraron que 10s niiios normales no encontraron estas muiiecas mis interesantes que 10s otros juguetes. No se observe mas que poca agresion y ninguna actividad sexual explicita. En contraste con las observaciones clinicas de niiios abusados el juego con las muiiecas de nirios not-males no tiende a ser caracterizado por la agresion o la sexualidad: por tanto este tipo de conducta debe ser tornado seriamente.