The differences in learning potential among educable adolescents is discussed in terms of high scorers, those who learned to solve problems on the Kohs test while doing the first trial, gainers who experienced difficulty but showed marked improvement following tutoring, and nongainers who did not improve after tutoring. The reliability of the Kohs block designs is indicated and the relation of learning potential status to social, demographic, school, and test variables is discussed. Evidence is given to support the suggestion that severe social deprivation during or prior to school age tends to be associated with a higher proportion of nongainers, while less severe disadvantage results in higher proportions of gainers and high scorers. The latter groups are portrayed as more educationally rather than intrinsically retarded. (RJ)