IN THE LATE 1960s, a new kind of national test began to be given to American schoolchildren. Called the National Assessment of Educational Progress, it differed from other tests not so much in what it sought to measure as in how it expressed the results. Instead of the usual percentiles that describe how groups compare, the assessment set out to describe what they actually knew. How many could answer this civics question, how many solve that problem in math, how many write a grammatical letter?

That seemingly innocent idea ran into fierce resistance. Chief state school officers especially said it could lead to a national curriculum in violation of the tradition of local control. Their critics said that what they really feared was exposure and accountability, but to a considerable extent the superintendents prevailed. The tests were authorized, but the testing organization was permitted to report the results only in regional terms, for Northeast, Southeast, central states and West. The usefulness of the test was thus blurred.

Now there is a move to sharpen the reporting. The number of children tested would be increased to create valid samples in, and allow results to be reported for, each state. The list of subjects tested would also be enlarged. The proposal is coming from strange bedfellows -- the likes of both William Bennett and Edward Kennedy, the Reagan Department of Education and the Senate Labor and Human Resources Committee. It is part of a national interest all across the political spectrum in performance standards for the schools. The state school officers, themselves now steeped in the vocabulary of accountability, are no longer resisting. The expansion will likely be written into the Senate version of the bill reauthorizing the main federal education programs.

It's a good idea. Yes, if it works as intended it will increase the pressure on a lot of schools. But what harm does that do? The testing program so far has not so much imposed a national curriculum as reflected the one that already implicitly exists. Most school systems have about the same basic goals. The idea here is to measure better how they live up to the