Bad timing for test cheaters

Maybe I just missed it. But has anyone else seen warning signs for Gov. Sonny Perdue's pay-for-performance teacher salary proposal embedded in the budding scandal over standardized test cheating?

//

Perdue, who has single-handedly made Georgia educators secretly wish for the days of Roy Barnes, hatched a plan to allow teachers to be paid based on some measures of their classroom performance.

The idea has "awful" written all over it, for a whole lot of reasons. Chief among them is the fear that teachers under such a plan would be even more encouraged to teach children to take tests rather than to think and reason independently.

That's what makes this testing scandal such a huge red flag.

As Morris News Service Atlanta Bureau Chief Walter Jones wrote the other day, the testing problems were unearthed when a state official had the company that scores the Criterion Reference Competency Test to check the score sheets for unusual numbers of answers that were erased and corrected

Jones points out that, statistically, for Georgia's number of tests about 230 answer sheets should have had more erasures than the norm.

The analysis, however, showed more than 6,000 such overly-erased answer sheets.

There have been all sorts of speculative excuses while the school systems fingered by the analysis are being investigated, but the main assumption is that some teachers in some Georgia schools erased some children's answers to improve their test scores.

Now, why would they do such a thing? That's pretty obvious. If their students perform better, the teachers look better, don't they?

There might be other puzzling reasons, too, of course; for example, all but one of the school systems with the most erasures is majority black, while all of the school systems showing zero erasures are majority white. Why?

Perhaps some of those presumed-cheating systems feel more pressure to meet specific standards and sought a shortcut to get them there.

Just .63 percent of Columbia County's tests were "flagged," by the way, a number in the "cleared" area below 5.4 percent. Richmond County had 6.11 percent flagged, considered "minimal concern." Nearby Hancock and Taliaferro counties hit the "moderate concern" threshold with 13.01 and 11.11 percent flagged, respectively.

In any event, the obvious red flag is that standardized tests are assumed to be one of the ways teachers would be evaluated to decide if they are due "performance" pay under a plan such as Perdue's. If tests are manipulated now with no bonuses at stake, how much greater incentive would there be to monkey with scores if there was higher pay on the line?

Set that question aside for a moment, though.

After the governor unveiled his proposal, he also released the results of what he said was a poll of 20,000 teachers in the state. Perdue claims the results show most teachers like the idea of pay based on performance.

Baloney, says everyone with a firing synapse. Just in case, state Sen. Bill Jackson - a former school board member - asked Columbia County School Superintendent Charles Nagle to poll the county's teachers.

The responses were as predictable as they were swift: While just 56 favored the idea and 87 were undecided, a whopping 1,010 teachers oppose the concept.

Now, I fully believe an even more overwhelming majority of those teachers would never admit to even remotely entertaining the idea of fudging their students' tests, even if there were some profit motive involved.

But the CRCT investigation should make it pretty obvious that Perdue's well-intentioned plan (at least, I'm sure he think's it's well-intentioned) could suffer from the same unintended consequences that seem to doom most top-down public education reforms.

It won't stop him from trying, of course. He'll be out of office in a few months anyway, and it'll be someone else's problem then. Maybe even Roy Barnes'.

(Barry L. Paschal is publisher of The Columbia County News-Times. E-mail barry.paschal@newstimesonline.com. Follow at twitter.com/barrypaschal.)

* Regarding the headline for this column: No, there is no "good" timing for test cheaters.