Friday, July 14, 2006

EDITORIAL >> Look at tests cautiously

Tuesday was a grand day for celebration because the state education commissioner announced that Arkan-sas children performed beautifully this year on both the standardized test that the state devised especially for Arkansas students and on the nationally normed Iowa Test of Basic Skills. We should congratulate ourselves, he said, on the steadily improving job that we are doing educating our youngsters.

Maybe we should. But why do we have nagging doubts? Why not clutch the good news and relax for a change?

One reason is that we have experienced these days before, many times. Just look at the results from the Pulaski County Special School District. See the disturbing article on the front page.

Since Arkansas finally began to enlist in the school accountability drive a quarter-century ago, the yearly release of Arkansas scores on standardized tests has given us comfort and concern. We are either near the national average or above it, and we are gaining on the other states. Later, some objective tabulation of nationwide results on one test or another — often the standardized college-entrance exams — brings us down to earth.

Another source of skepticism is the discovery that all across the land other state education agencies are announcing similar good results and gains against the national average. Everyone, of course, cannot be above the national average. It is the Lake Wobegon syndrome, after Garrison Keillor’s fictional Minnesota community “where all the children are above average.”

When states announce that their children are performing above the national average in, say, seventh-grade mathematics that is not necessarily so. Arkansas children are measured against a national group that took the test, not their contemporary peers across the country. So, indeed, every state may announce that its kids are performing above the national average.

Despite the No Child Left Behind Act and state accountability reforms, education statistics still have a way of lending themselves to universal optimism.

Education Commissioner Ken James said the number of Arkansas children who earned high marks on Arkansas’s Benchmark exam — that is the test devised specifically to measure how much they have gained of the knowledge that we want Arkansas students to have at that age— and the Iowa Test of Basic Skills improved demonstrably from 2005. He called it “fantastic news.” And the figures across the grades were almost uniformly heartening.

But let us take one statistic cited by Mr. James. Fifty-seven percent of Arkansas sixth graders who took the Benchmark math exam scored proficient or above, compared with 43 percent a year ago. What would account for such a stunning improvement in a single year across the whole state? Since the first Benchmark exams were devised and administered in the late ’90s, we have seen such dramatic year-to-year increases — indeed, far more dramatic ones in individual school districts.

Could sixth-grade math teachers, or fifth-grade teachers, have so sharply improved their methods in 12 months? Did teachers convert to a new teaching doctrine? Were dramatic changes made in instruction at the lower-grade level the year before?

Did more and more teachers in 2005-06 decide to spend more time teaching the tests, or did school administrators apply the additional pressure? There is anecdotal evidence of that, but it is only anecdotal. But this much is incontrovertible: The pressure of rewards and penalties for good standardized test scores is practically insurmountable now. Schools can lose their autonomy by failing to raise low test scores.

The movement to base teachers’ salaries on standardized scores is gaining steam. Soon, any teacher who does not devote a good deal of a semester drilling students on the Benchmark and other standardized tests will be insubordinate and looking for a job. A better picture is how students do on the National Assessment of Educational Progress, which compares children nationally on relatively objective measurements.

Last week, the U. S. Education Department cited Arkansas among many states that employed tests of questionable validity. Arkansas is going to have to persuade the feds that its Benchmark exams are not designed to make kids (and, perforce, the state’s political leaders) seem better than they are. Some states are guilty of continually dumbing down their tests so that they appear to be making the progress that the No Child Left Behind Act requires. Arkansas has revised its tests and students perform better afterward. But state leaders say it was simply making mid-course corrections and the purpose was not to improve appearances. The state will now have a chance to prove that it is so.

Meantime, let us be encouraged by the good numbers but not surrender our skepticism. Neither should we abandon our zeal to hold education and political leaders accountable for improving education, starting with the candidates for governor and for the General Assembly.