Using the NCAST Feeding, Teaching, and HOME Scales, we tested 37 high-risk infants matched with 37 healthy infants on gender, race, and socioeconomic status. All infants were 8 months old. A one-to-one matched case-control design was used to determine whether increased risk of impaired mother—infant interaction was associated with case status. Conditional logistic regression was used to control for possible confounding and to evaluate interaction. Of the 37 high-risk infant—mother dyads, 25 had a low score on one or more scales while only 10 of the control dyads had a low score on one or more of the three scales. The Feeding scale was the only assessment in which the association found in the univariate analysis persisted after adjusting for other variables. Because of its low cost and high efficiency, the NCAST battery appears to be valuable for directing more specialized intervention services in a high-risk infant population.