A NATIONAL SURVEY by a school reform group has raised some serious questions about the significance of rising scores on standardized achievement tests. The Friends for Education, a West Virginia-based research group, says that the success often indicated by improving test scores is greatly exaggerated or -- in some instances -- practically meaningless. In one sense, however, the survey points out something that should be obvious: school districts must rely on a number of yardsticks to prove that progress is being made in the classrooms.

Standardized achievement tests have always used the term "national norm" as the median or the 50th percentile. But that standard, according to the survey, has not been based on how the country's students have performed each year. The "norms" came from a sample of students tested before the exams were in widespread use -- and have not been adjusted upward as the tests have become more familiar. Performance improves and scores rise. That, it turns out, is why such gratifyingly high percentages of children in every school system are reported to be above the national norm. In Virginia, for example, 112 of 134 school divisions scored above the national norm.

Does this mean that the appearance of progress in recent years is fake? Not necessarily. Other standards also offer clear evidence of academic progress, stagnation or regression. The District's public schools are a good example. School attendance rates are up from 87 percent in 1979 to 91 percent. The percentage of students who go on to college has risen from 40 percent in 1980 to 57 percent. The D.C. schools also implemented a student progress plan in 1981 that required the mastery of certain skills in order for a student to be promoted. In that year only 68 percent of the students were promoted. Now, the rate has surpassed 90 percent. The District's schools really have improved.

To be useful and credible, those who develop and score these tests will have to be a bit more candid and careful. Test norms will have to be updated sooner, or different tests will be needed more often. School districts should not use the same test for more than three or four years. Standardized achievement tests can still be a valuable tool, but not when they reflect inflated levels of academic progress.