Previous experiments with grade school children, designed to raise their expectations for their own performance, were repeated with pupil subgroups formed according to race, sex, and age. Positive results of the earlier studies were reproduced, and further analyses are presented to assess the effect of some status factors in the situation. The experimental procedure was effective across the subgroups studied. The relation of this work to "teacher expectancy" studies is discussed, and some implications of the results, in terms of other studies on expectations and the structure of competition in schools, are drawn.