Sportive Thoughts

The rules for Duncan's "Race to the Top" (RttT) are out today. RttT is a series of competitive awards of Federal money for education, totaling some $4 billion. The top criteria for an award are:

• persuasive reform agenda
• significant district contributions
• teacher and administrative performance evaluations

Sounds good; yet there are the usual naysayers, whose arguments don't hold that much water:

One organization unimpressed with the final rules is FairTest, a nonprofit that is critical of standardized testing. Monty Neil, the group’s co-executive director, said the final rules encourage the use of test data to help teachers improve instruction, rather than merely to hold schools accountable, which he called a minor improvement. “But most of the points will be awarded for work on new standardized tests,” Mr. Neil said.

Neil basically agrees with the new rules, yet is a bit marshy over the fact that standardized tests are needed to make reform work, and new tests must be developed. I suppose if your platform is flatly negative with regards to standardized testing, you won't find much cause for celebration.

The other naysayers are, of course, the union heads:

Both national teachers’ unions criticized the draft rules in terms that were often scathing. Randi Weingarten, president of the American Federation of Teachers, said in an August letter that the program had the “potential to spark innovation,” but argued that the administration had overstepped its authority with many elements of the competition. Ms. Weingarten reacted more positively to the final rules.

“The administration worked very hard to find the right balance,” she said. “This is not a ‘Kumbaya’ moment, but these rules suggest they won’t dictate from on high but will work together with teachers and their unions.”

Dennis Van Roekel, president of the National Education Association, said the draft rules had seemed to encourage states to evaluate teachers and principals largely “based on a single standardized test score,” which he likened to judging a basketball player based on the number of baskets in one game.

The final rules, he said, “put more emphasis on student growth, teacher practice and improving instruction. So I’m really pleased that they listened.”

Mr Van Roekel's analogy is a bit off the mark; a good standardized test operates much like an entire statistic table, like the following:

In the table above (Peyton Manning's career passing statistics), we have several categories: seven raw sums, two simple averages, and one weighted calculation.

A proprietary standardized test is similarly configured with raw scores in each type of question, simple averages in each category of question types, and a weighted calculation of overall performance. Students have math, science, reading and writing scores; and when complied according to some complicated algorithm, we see roughly how individual students, specific classrooms, schools and entire districts are performing. What's wrong with this?

To be fair, part of Van Roekel's complaint is that the test scores are restricted to one performance. But wouldn't the solution to that be more testing?