Objective: The present study extends field research on interviews with young children suspected of having been abused by examining multiple assessment interviews designed to be inquisitory and exploratory, rather than formal evidential or forensic interviews. Methods: Sixty-six interviews with 24 children between the ages of 3 and 6 years who were undergoing an assessment for suspected child abuse were examined. Each child was interviewed 2, 3, or 4 times. The interviewer's questions were categorized in terms of openness (open, closed or choice), in terms of the degree of interviewer input (free recall, direct, leading, suggestive), and for topic (whether the question was abuse-specific or nonabuse-related). Children's on-task responses were coded for amount of information (number of clauses) reported in relation to each question type and topic, and off-task responses were categorized as either ignoring the question or a diverted response. Results: Children provided a response to most questions, independent of question type or topic and typically responded with one or two simple clauses. Some children disclosed abuse in response to open-ended questions; generally, however, failure to respond to a question was more likely for abuse-specific than for nonabuse-related questions. Conclusion: The findings are discussed in terms of the growing literature on interviewing children about suspected abuse, particularly in interviews conducted over multiple sessions. Practice implications: Assessment of suspected child abuse may involve more than a single investigative interview. Research examining children's responses to questioning over multiple interviews (or single interviews conducted over multiple sessions) is necessary for the development of best practise guidelines for the assessment of abuse. (Contains 1 figure and 5 tables.)