Cheating on high stakes tests? Who woulda guessed

by David Safier

When test scores are make-or-break for teachers and administrators, it causes some people to bend the rules to save their bacon and/or boost their standing. Stories have emerged of outright cheating, like teachers and/or administrators erasing students' wrong answers and bubbling in the right ones. Less blatant ways to "help" students before and during the tests are even more typical.

And now this. A Georgia principal un-enrolled 13 third and fifth grade students right before the high stakes test in May, then reenrolled them after the test. She has since resigned.

Sounds terrible, right? Yes, but not unusual. Texas has been doing this kind of thing on a large scale for years to boost their tenth grade state test numbers. They figure out ways to drop low ranking students during the ninth grade, or they hold the students back a year, then leapfrog them from the ninth to the eleventh grade. This has been standard operating procedure all over the state, and so far as I know, it's still going on.

I suspect some similar manipulation in Florida, possibly in the eighth and tenth grades, but at this point, it's only a suspicion. We already know how the Florida NAEP scores are boosted in the fourth grade by their high third grade retention rate. Here's some of the most recent proof.

To the extent high stakes tests are reliable indicators of student achievement, they have to be given to a representative sample of students, and the students have to take them without any advanced preparation which is geared to the specific test. Otherwise, the results have little relationship to the attributes they're designed to test.