The National Science Board has set new rules on how the National Science Foundation must spend money for science education, including a significant policy change that would eliminate programs that divvy up money based solely on the number of students in a state or school district.

On June 21, the science board said in a policy statement that the NSF must avoid general grant programs and concentrate its spending on three areas: writing new curricula, using computers and videotape machines as teachers, and extra training for science teachers.

The science board, the policy body that oversees the NSF, ordered NSF director Edward A. Knapp to produce a spending plan by October that is in line with its new policies.

Lewis Branscomb, chairman of the science board, wrote in the report of the board's meeting that the foundation "should seek to use its resources in ways that exert high leverage, . . . such as selective teacher training and model course content development in mathematics, science and technology."

In rebuilding its dismantled science programs, the board hopes to use its "best asset," scientists, to create materials and courses that will have a broad effect rather than spreading the money more thinly through grants to schools.

The amount of money available to carry out the program is expected to be more than double what it has been in recent years, though still not up to earlier levels.

Congress has approved an appropriation of about $75 million for fiscal 1984, compared with the $20 million and $30 million spent in 1982 and 1983 on science education. Those figures would have been almost zero but for previous commitments and graduate student fellowships that had to be continued.

Funds for 1984 will be divided up this way, following congressional suggestions and the new board policies:

About $19 million is expected to be spent on graduate fellowships. The number of grants is expected to rise from about 1,500 per year now to as many as 1,800 per year over the next three years--but still well below the historical high of 2,600 fellowships and another 5,600 trainee grants.

Another $34 million probably will be put into "merit-based materials development" or to pay for textbooks and course materials that will be put together by academics on some sort of competitive grant basis.

About $20 million will be spent to bolster training of the best science teachers in summer workshops and training sessions.

The science board's new policy also directs the NSF to concentrate its efforts on science education below the college level. FACTS ON SCIENTISTS . . . The Science Manpower Commission, a professional organization that provides statistical data on scientists to federal agencies and professional societies, produced a new report on science manpower this week that shows that the single greatest change in science employment is among women and minorities.

The number of science professionals has declined over the past two decades, but the number of women earning degrees in the sciences has increased by 15 percent.

In 1970, women earned 41.5 percent of the bachelor's degrees, 39.7 percent of the master's degrees and 13.3 percent of the doctorates. By 1981, the numbers had risen to 49.8 percent, 50.3 percent and 31.1 percent.