In order to prevent children with only one participating parent from being excluded from the analyses, missing on the parental alcohol and mental distress variables was coded into separate categories, thus providing results for children of non-responding parents as well. In addition, to keep children with unidentified or dead fathers (N = 100) and mothers (N = 9) in the analyses, the missing parents' age was estimated from the age of the co-parent. This treatment of the data permitted all adolescents with valid school data to be included in the final sample.
