Key points are not available for this paper at this time.
Abstract A total of 406 heterogeneously grouped students in Grades 3, 4, 6, 7, and 8 in three K through 8 Chicago public schools were assigned randomly to two conditions, ordinary standardized-test instructions (control) and special instructions, to do as well as possible for themselves, their parents, and their teachers (experimental). On average, students given special instructions did significantly better (p < .01) than the control students did on the criterion measure, the mathematics section of the commonly used Iowa Test of Basic Skills. The three schools differed significantly in achievement (p < .05), but girls and boys and grade levels did not differ measurably. The motivational effect was constant across grade levels and boys and girls, but differed significantly (p < .05) across schools. The average effect was moderately large, .303 standard deviations, which implies that the special instructions raise the typical student's scores from the 50th to the 62nd percentile.
Brown et al. (Fri,) studied this question.