a b s t r a c tObjective: The study focused on children's nonverbal behavior in investigative interviews exploring suspicions of child abuse. The key aims were to determine whether non-verbal behavior in the pre-substantive phases of the interview predicted whether or not children would disclose the alleged abuse later in the interview and to identify differences in the nonverbal behaviors of disclosing and non-disclosing children. Method: We studied DVD-recorded interviews of 40 alleged victims of child abuse. In all cases, there was external evidence strongly suggesting that abuse had occurred. However, half of the children disclosed abuse when interviewed using the NICHD Investigative Interview Protocol, whereas the other half did not. Two raters, unaware whether or not the children disclosed, independently coded the videotapes for nonverbal indices of positive and negative emotions, stress, and physical disengagement in each 15-second unit of the introductory, rapport building, and substantive interview phases. Results: Indicators of stress and physical disengagement increased as the interviews progressed while indices of positive emotions decreased. Non-disclosers showed proportionately more physical disengagement than disclosers in both the introductory and substantive phases. Conclusions: Awareness of non-verbal behavior may help investigators identify reluctant children early in forensic interviews. Practice implications: There is substantial evidence that, when questioned by investigators, many children do not disclose that they have been abused. The early detection of reluctance to disclose may allow interviewers to alter their behavior, helping the children overcome their reluctance by providing non-suggestive support before the possibility of abuse is discussed. Of course, nonverbal behavior alone should not be used to assess children in investigative interviews. However, nonverbal cues may nonetheless provide additional information to interviewers and assist them in identifying reluctant children.