Factors That Contribute to More Detailed Disclosures of Child Sexual Abuse During Forensic Interviews

Background and Purpose:  Research shows that children’s capacities to disclose sexual abuse appear to be related to one or several factors including age (Lamb et. al, 2003), gender (Lamb & Garretson, 2003), and relationship to the perpetrator (Alaggia, 2004).  While demographic and contextual factors have been studied in relation to children’s ability to disclose sexual abuse, many of these aspects have yet to be explored within the context of forensic interviews.  Establishing rapport at the beginning of the interview through the use of open ended questions, a technique called narrative practice, and supportive statements from interviewers can result in children providing more detailed disclosures (Davies, Wescott, & Horan, 2000) as compared to non-narrative practice which seeks to establish rapport, but uses fewer open ended questions. As part of a program evaluation, administrators at a children’s advocacy center in the Upper Midwest wanted to know whether a revised version of their interview that incorporated narrative practice would elicit more detailed disclosures than the interview they had used previously.  We hypothesized that 1) the revised, narrative-focused interview protocol would result in children’s disclosure of more details than the non-narrative protocol, and 2) child characteristics and contextual factors related to the abuse would result in significant differences in the number of details that children provided.   

Methods:  This study compared the previous and revised interviews through content analysis of 100 videotaped forensic interviews and case files where children disclosed sexual abuse.  We developed a coding scheme and tested it for inter-rater reliability (n=15) with 90% agreement between coders.  Coded data included demographic information, relationship of child to the alleged perpetrator, type of abuse allegation, and number of details provided by the child during the interview after disclosure. The outcome variable was number of details and included language related to time, place, people, events, and descriptive words used by the children regarding the abuse.  Data analysis included descriptive statistics, independent sample t-tests, and linear regression.

Results:  Comparison of the non-narrative interview group to the narrative group showed a statistically significant difference in the number of details provided (t = 1.7052, df = 58.006, p = 0.04676).  Linear regression showed that older children (F=53.18, df=98, p<0.001) and an increase in supportive interviewer statements (F=13.07, df=98, p<0.001) significantly predicted more details.  Girls provided significantly more details than boys (t = 3.57, df = 58, p<0.001).  Children related to the alleged perpetrator provided significantly fewer details than those who were unrelated to the alleged perpetrator (t = -2.45, df = 68.56, p= 0.017). 

Implications:  Our results show that the revised, narrative-supportive protocol elicits more detail than the former, non-narrative interview protocol and that age, gender, and relationship to perpetrators are significant in predicting the number of details children provide. Of note in the present research is that children require considerable support during forensic interviews when they have family relationships with alleged perpetrators. This has implications for further revisions of the interview. The results also suggest the need for more effective ways to interview boys and younger children.