The study is based on data for 6600 fourth graders and 7100 eighth graders who took part in the National Assessment of Educational Progress, which in 1996 for the first time included detailed surveys of computer use in classrooms nationwide. For fourth graders, the study found, using computers for math drills and practice--such as figuring out answers to simple arithmetic problems--had no effect on grades. When computers are used just for this purpose, "students would be better off not bothering at all" with them, says the study's author, ETS's Harold Wenglinsky.

CREDIT: "DOES IT COMPUTE? THE RELATIONSHIP BETWEEN EDUCATION TECHNOLOGY AND STUDENT ACHIEVEMENT IN MATHEMATICS"

By eighth grade, however, computers in many classrooms are employed to more sophisticated ends, such as simulations and applications of math concepts to real-life problems. For these students, the report found, computers had a "substantial" effect overall, jacking up performance by two-fifths of a grade. But minority students did not benefit as much--in eighth grade, for example, more than half of black computer users were still mainly in the drill-and-practice mode, which actually appears to depress grades at that age, compared with 30% of whites.

Douglas H. Clements, professor of education at the State University of New York, Stony Brook, says the study left undefined "too many variables, including the specific type and quality of the software and the teaching that incorporated it." Nonetheless, he agrees there's a message here: that "it's how you use [educational technology], not how much it is used," that counts.