COMPETITIVE PRESSURE A GOOD THING FOR FLA. SCHOOLS

Saturday, August 23, 2003

Florida's A+ program for school reform combines some of the most controversial proposals for improving performance -- high-stakes testing, grades for schools and vouchers for students in schools that get a failing grade.

That adds up to real competition, and, say Jay Greene and Marcus Winters, in a new working paper from the Manhattan Institute, competition works.

That comes as no surprise to the Manhattan Institute, of course, but it ought to be a revelation to those in the public school establishment who oppose all these proposals.

Florida issued its first school grades in 1998-99, based largely on the Florida Comprehensive Assessment Test. If a school receives two or more F grades within a four-year period, its students can apply for vouchers they can use at private or other public schools.

Greene and Marcus divided schools into categories ``based on the degree of threat each school faces from voucher competition.'' They controlled for such variables as poor and minority enrollment, school spending and the proportion of children with limited English skills. The result: the more imminent the threat, the greater the improvement in FCAT scores.

* Voucher-eligible schools, nine of them, received two Fs in the first four years, so their students have already been offered vouchers. They have the biggest incentive to improve, and indeed their scores went up the most -- roughly 10 points each on the math and reading parts of the FCAT, and five percentile points on the Stanford-9 exam. That's not a huge difference, but it is statistically significant.

* Voucher-threatened schools (50) received one F in the previous three years, so an F grade in 2002-2003 would put them in the voucher-eligible category. Their improvements were the second largest, and also statistically significant.

* Always D schools (63) -- nothing but D grades for the first four years -- have plenty of reason to worry about their performance, but they are at least two years away from actually losing students. Improvements slight, and not highly significant.

* Ever D schools (570) -- at least one D in those four years -- have less reason for concern. Their scores barely budged and what change there was, wasn't significant.

So far, these results look much like the ones from an earlier study, but by now the program has been running long enough to identify a fifth category:

* Formerly threatened schools received an F in the first year of grading, but not since. Thus, if they get another F, it will be the first in a new four-year period, and so not put them immediately on the voucher list. Their scores went down, and in most test categories the decline was statistically significant.

That is particularly interesting, the authors note, because it debunks the argument that F schools improve only because of the stigma conferred by the grade. ``It is implausible,'' the authors say, ``that the stigma effect only exists for three years and then suddenly disappears.'' A more likely explanation, they say, is that the effect suddenly disappears ``when the four year voucher threat period expires.''

With so much of a school's grade riding on FCAT scores, it could be that the improvements were more apparent than real, the results of strategies to improve test scores even though children weren't in fact learning more. But score increases on the Stanford-9 test were equally significant, and the schools have no comparable incentive to game those.

Greene and Winters also examine the possibility that the scores are regressing toward the mean, a statistical phenomenon more readily understood as, ``When you're at the bottom, there's no place to go but up.'' But that explanation doesn't work either. Because the school grades are based on more than just FCAT scores, there are schools with scores very similar to the voucher-eligible and voucher-threatened schools, but they did not show similar improvements.

It could also be that the worst schools got better because their lowest-performing students were the ones who took the vouchers and left.

If that were happening, though, it would undercut voucher opponents' argument that vouchers would allow private schools to ``cream'' the best students from public schools. Instead, they'd be ``dredging'' the weakest students. But the authors say there's no reason to believe it is happening, and if it did it wouldn't be a bad thing -- it would mean that the vouchers are serving precisely the students most in need.

Though most of the students eligible for vouchers won't actually use them, educators have to take the possibility seriously. One ``Triple-F'' elementary school in Florida closed because of low enrollment.

If just a little competition accomplishes so much, what would a lot of competition do?