When it comes to standardized tests, there are two kinds of idiots: those who think the tests can measure nothing worth measuring and therefore want them abolished, and those who think the tests can measure anything, and therefore want them enshrined.

The first category includes the researcher who, in the course of arguing against reliance on standardized tests as a "single criterion," told me a couple of weeks ago that flunking a general math test should not be accepted as evidence of unfitness to teach math.

The second surely must include the people who adopted Maryland's statewide writing proficiency test. Officials there, dismayed that almost half the 11th graders who have taken the test have failed it, are now looking at a $183,000 packet of changes in the testing program. The chances are it will be wasted money.

As a professional writer, I believe it shouldn't be necessary to point out that I am not opposed to the teaching of writing skills, which I consider not just possible but vital. I have argued for years that children ought to be required to do far more writing than they now do: how-to writing, expository writing, descriptive writing, factual writing, argument, summary -- all sorts of writing.

I have threatened to organize sledgehammer-wielding posses to raid the schools for the salutary purpose of busting up those destroyers of writing: the ditto machines. The best way for children to learn writing, I have argued in my simple-minded way, is to let them write. The surest way to stunt whatever writing ability they may possess is to let them get by with filling in blanks on ditto sheets. ("Columbus's three ships were the (blank), the (blank) and the (blankety blank).)

But if teaching writing is vital, the standardized testing of writing is virtually impossible, like trying to devise standardized tests for painting, or singing or public speaking.

Maryland, to its credit, didn't fall into the trap that ensnared some other school systems a few years back: the attempt to test writing ability by penciling in the appropriate squares on machine-gradable answer sheets. Instead, it chose to rely on a panel of experts to assign numerical grades (based on necessarily subjective criteria) to writing samples. Still, they were embarrassed a few weeks back when newspapers published what seemed to be a quite decent sample written by an honor student who, nonetheless, flunked the test.

Teachers of writing can grade their pupils' writing ability against principles and techniques taught in class. But teachers -- at least the good ones -- understand that there are no absolute criteria, comparable to the correct answers in math and social studies.

They know better than to put their faith in standardized measures of good writing, including the absurd "fog index" somebody came up with several years ago, which deducted points for words of more than three syllables. (Overuse of big words can result in foggy writing, but why should the use of an ordinary word such as "impossibility" be penalized more than the use of the shorter "adumbrative"?) What the combatants on both sides of the testing debates ought to recognize is that there are some things that standardized tests can measure (mathematical computation, facts, spelling, reading comprehension) and some things that are best measured by those who teach them (artistic expression, teaching technique, writing).

Some who oppose virtually all standardized testing really are opposed (it seems to me) to the unpalatable truths that such tests can reveal, including the fact that low-income and minority students tend to do less well on the tests than middle-class whites. Outlawing tests in an effort to mask these unpleasant facts makes about as much sense as outlawing thermometers and electrocardiagraph machines to mask the fact that minorities tend to suffer poorer health.

But it is important to recognize that there are some things that standardized tests cannot do sufficiently well to make them worthwhile. Measuring writing skills is one of them.