Evaluation results of a program (Infant School Program at New York University Medical Center) serving 40 multihandicapped preschool children over a 3-year period are given. Originally presented with accompanying slides, the paper lists goals of early childhood education and describes methods of data collection and program procedures. Ss were evaluated at admission and prior to discharge in the following areas: interaction with materials, awareness of the environment, social responsiveness, communication, affect, gross motor, and fine motor. It was concluded that analysis of the data on the entire group was not meaningful, but that various subgroups, based on levels of functioning, did exist which reflected various patterns of change; and that data (particular teacher observational data) on the individual child was the most useful way to plan a specific educational program for each child. (DB)