The purpose of this research was to study expectations of elementary school children in two ways: experimentally and observationally. Expectations may be roughly defined as a child's ideas of his own ability at a particular task. From the data it appears that childrens' expectations could be raised experimentally by a suitable adult and high expectations in one area generalize into other unrelated areas. The experiments are summarized in a number of published articles reproduced herein and listed in the bibliography. The observational data focus on children in first and second grades in a white middle-class school and in an integrated lower class school. From the time they enter school individual children are followed to see how their expectations for their own performance in reading, arithmetic, and conduct develop. Their expectations, and their parents' expectations for them, are repeatedly measured. Children in both places have, on the average, very high expectations for themselves before they get a report card, higher than their parents. These expectations do not diminish much when marks are lower than expected; in fact for the majority, expectations are maintained over first grade. Children whose marks improve are likely to be those whose expectations exceeded marks. (Author)