The purpose of this study was to identify the factors in preschool educational television that engage and sustain children's attention. The method used to measure the children's attention was decided on in a pilot study of three measurement techniques, and consisted of an observer rating scale with the periodic introduction of a kaleidoscopic distractor. There were three groups of children observed: 2- and 3-year-old middle class children, 4- and 5-year-old middle class children, and 4- and 5-year-old Mexican-American children from lower income families. Five television programs for preschool children and one set of animated cartoons were viewed. To check for any possible correlation betwee n language development and attentiveness, the children were given the Peabody Picture Vocabulary Test. The results of the betwee n groups analyses revealed a very high degree of generality in terms of the type of program content that appealed to the age, sex, and social groups studied. Further, there was no correlation between Peabody scores and attention levels, once age was partialled out. Attention level was very low; analysis of the program content showed that children paid greater attention to animated cartoons, introduction to novel objects, and initiation of novel action by the teacher-performer. (MH)