Reactions of children to interviews using anatomically correct dolls.

Abstract

This study tested an underlying assumption of professionals who interview young children with anatomically correct dolls: children who have been sexually abused react differently to interviews with the dolls than children who have not been sexually abused. The behavior of a group of children who were referred to a mental health clinic in the southwestern United States because of suspected sexual abuse was compared to the behavior of a group of children referred to the same clinic for other reasons while the children were interviewed by clinicians using anatomically correct dolls. The study examined four categories of behavior which consisted of indicators of child sexual abuse from the literature. The four categories were: (1) sexual behavior; (2) anger/aggression; (3) anxiety/regression; and (4) avoidant behavior. Clinicians at the mental health clinic identified potential subjects for the study from the outpatient population. Parents were given written and verbal descriptions of the study and asked to contact the researcher if they were interested in allowing their child to participate. The voluntary nature of participation in the study was stressed. Eleven children who were referred because of suspected sexual abuse and eleven children referred for other reasons were successfully recruited. Groups were matched as closely as possible as to sex, age, racial/ethnic group and developmental level. Two dependent measures were employed: the Behavioral Checklist and the Likelihood of Victimization Scale. Both instruments were designed for the research study. The Behavioral Checklist was completed by two observers who watched each interview from behind a one-way mirror. The Likelihood of Victimization Scale was completed by the clinicians who interviewed the children. Observers as well as interviewers were blind to the referral status of the children. Statistically significant differences were obtained for two of the four categories of the Behavioral Checklist. The same two categories correlated significantly with the Likelihood of Victimization Scale. The results provided support for the assumption tested.