Data from the National Incidence Studies of Missing, Abducted, Runaway and Thrownaway Children (NISMART), because they focused on a broad continuum of abductions varying in their degree of seriousness, were well suited to examining these issues. In NISMART's national probability sample, telephone contact was made with 10,544 households, where primary caretakers were asked about the experiences of 20,505 children aged 17 or younger. Two types of situations were included in the definition of "family abduction" used by NISMART: situations in which a family member took a child in violation of a custody agreement or decree or situations in which a family member (in violation of a custody agreement of decree) failed to return a child at the end of a legal or agreed upon visitation period, with the child being away at least one additional night. NISMART identified 104 family abduction incidents that involved 142 children. The study found that the length of time it takes to resolve a family abduction and the amount of harm the abducted child experienced are related, both logically and statistically. Greater duration of abductions was associated with an increased risk that a child would be harmed. Moreover, both are indicators of more serious events that should receive special attention from police and court officers. One common factor is apparently related to both duration and likelihood of harm. Perpetrator behaviors that are threatening, including the threat to hold the child permanently, have a positive relationship to both duration of episodes and the likelihood of harm to a child. 4 tables, 13 references, and 8 notes