Forty-six of 97 available families (47%) participated in an investigation of mothers and their high-risk infants who had been discharged from a neonatal intensive care unit. Twenty-three infant demographic, infant medical, maternal demographic, and maternal medical variables were analyzed to determine if differences existed between participating and nonparticipating subjects. Subjects were categorized into three groups: (1) those who completed and returned their questionnaires, (2) those who agreed to participate but did not return their questionnaires, and (3) those who were contactable by letter only. Multivariate analyses of variance yielded no significant effect for Group Status for the Infant Medical, Maternal Demographic, and Maternal Medical variables. The multivariate analysis of the infant demographic variables yielded a significant Group effect with univariate analyses indicating that the groups differed on the Transport variable. Participating mothers more frequently had infants who had been transported to our facility than did nonparticipating mothers. The findings are discussed in relation to sample representativeness and suggestions are offered for reporting subject-loss data.