THE SCHOOL reforms of the last 15 years have been carried out at an unexpectedly high cost. Important and morally admirable as the reforms were, they diverted attention from the ablest children and the technical subjects. Worse, the reform movement in its fervor derided high standards as elitism, a term that became one of the great pejoratives of the last decade. A romantic bias against technology inflicted real damage on the schools and their students.

The great contribution of the reforms was to bring important kinds of help to the children traditionally most neglected in American schools -- those who learn least easily, those whose parents are poor and illiterate, those who had been discriminated against over the years. In the 1970s, there were heartening improvements in the education of these children, particularly in reading. It is demonstrated in the comparisons recently by the National Assessment of Educational Progress, a research organization set up by Congress to keep track of achievement in the schools.

But the Assessment's surveys also showed other and deeply disquieting trends in the typical high school. Among the best students, capability in math and science deteriorated significantly in the four or five years between the early and late 1970s. That is as true for the blacks among the high achievers as for the whites.

The Assessment's report will only confirm the view to which most people came some years ago. In the Washington area's schools, programs for gifted children and advanced placement courses in the academic subjects are a response of the most useful sort.

But it's going to take more than that to build satisfactory standards for math and science in the schools. Stopping the decline is hardly adequate. Better teaching for more children, in these demanding subjects, is not consistent with the politics of repeated tax cuts. That is true equally at the local and federal levels.

President Reagan, in his budget message two weeks ago, promised aid to the states in meeting the increasingly severe shortage of math and science teachers. But he has in mind the very modest sum of $50 million a year, to begin a couple of years from now. It would be a great deal more productive, although more expensive to consider serious incentives like forgiving the student loans of science and math graduates who go into high school teaching.

"The challenge now," says the Assessment in its report, "is to give attention to science and mathematics, while not losing ground in reading...." This is right. Or, to put it less gently, a degree of success in teaching reading to fourth graders is no excuse for failure to teach calculus and chemistry to twelfth graders. These subjects are now central to this country's culture, not to mention the way it earns its living. The costs of encouraging them are very small, compared with the costs of neglecting them.