Thursday, October 25, 2007

Responding to the threat of vouchers 

In economics, you look often for natural experiments to test a hypothesis. In Florida, schools are graded and vouchers are given to students in schools who "fail" the test twice. Thus one failure doesn't invoke a penalty but increases incentives to improve performance.

Rajashri Chakrabarti writes about these incentives. (h/t: WSJ Real Time Economics). Because the Florida program only requires a school to get to minimum standards in one area of the three 'R's, Chakrabarti tests whether the schools focused on one subject over the other two. They do: the focus is on writing rather than reading and arithmetic "because it was easiest to improve in."
Case studies reported in Goldhaber and Hannaway (2004) are very much consistent with this picture: �Writing came first �because this is the easiest to pass�...�With writing there�s a script; it�s pretty much first we did this, then we did this, and finally we did that, and using that simple sequencing in your writing you would get a passing grade.��

Telephone interviews conducted by me with school administrators in several F schools in different Florida districts also show a similar picture. They reveal widespread beliefs among school administrators that writing scores were much easier to improve in than reading and math scores. They say that they focused on writing in various ways after the program. They established a �team approach in writing� which introduced writing across the curriculum. This approach incorporated writing components in other subject areas also such as history, geography, etc. to increase the students� practice in writing. They also introduced school wide projects in writing, longer time blocks in writing, and writing components in lower grades.
That is an interesting result, reasons for which I do not really know but can speculate. If you are grading writing, it has to be by a template -- "three points for this; five for that"; the template is something teachers will know. As a result, they can teach to the template. That and repetition, it appears, did the trick. Is this learning? To some extent yes, but it would be useful to know if there were cross-effects to other student subject areas. It doesn't appear to be so, but the results are not clear on that question.

The study also finds that schools focused on the low-performing students, but that the higher-performing students did not suffer a decrease in scores as a result. If anything, the higher-performing student results went up with the lower-performing students, just not as much.

The important point is that the threat of vouchers appears to motivate some behavioral change in teaching strategies that improve learning. It is another example of why the voucher program is worth further investigation and study, and supports the claims of its advocates.