[Susan notes: Here's a fine letter. Enjoy the basketball analogy.]
Submitted to but not published
The $4.35 billion Race to the Top Fund presents you with a unique opportunity. By using this program to reward boldness and creativity, you could support a wide range of projects to expand the knowledge base about teaching and learning, fostering valuable innovations in our nation's schools.
Unfortunately, your proposed priorities for Race to the Top would squander that opportunity by restricting federal funding to a set of preconceived notions about "reform," which may be ideologically fashionable but are largely unsupported by scientific research.
Our organization is especially concerned by your insistence that standardized test scores be used as part of teacher-compensation systems. In the absence of evidence that such a change would be beneficial, it would be irresponsible Ă˘€“ not to mention undemocratic Ă˘€“ to force states to bring their laws into conformance with your plan.
As evidence for this mandate, your proposal cites only a handful of economists, far removed from actual classrooms, who were unable to isolate the observable characteristics of effective teachers-- "effective" as determined by their studentsĂ˘€™ test scores. So, the logic goes, why not just evaluate and pay teachers on the basis of those scores rather than on their years of teaching experience or academic credentials?
Perhaps, lacking any background in education, the economists were "observing" in the wrong places and failed to consider the myriad of talents and skills that inspire children to learn. Or maybe their study designs were slanted, consciously or otherwise, to bolster a hypothesis that monetary incentives based on test data are key to improving teacher quality. (Several of the authors are members of the Future of American Education working group at the American Enterprise Institute, which is associated with that position.) Whatever the case, your proposal is based on research that is admittedly inconclusive and on a theory of teacher motivation that remains unproven.
The grant criteria would also place an undue reliance on standardized tests that offer, at best, a blurry snapshot of student progress. For English language learners (ELLs) in particular, such tests are rarely valid or reliable. Because these students cannot fully show what they have learned when assessed in a language they have yet to master, their scores typically lag far behind those of English-proficient peers. If teachers are to be penalized for an "achievement gap" over which they have no control, how many will want to teach ELLs? It is also well established that these children's progress in speaking, comprehending, reading, and writing English is never a straight-line trajectory. How could any "growth model" fairly accommodate that reality?
During his campaign, President Obama raised hopes that his administration would limit the uses (and abuses) of high-stakes testing. But paying teachers on the basis of test scores can only raise those stakes, at considerable cost to kids.
Surely, Mr. Duncan, you must be aware of the growing body of evidence about the perverse effects of high-stakes testing: narrowing the curriculum, teaching to the test, stressing basic skills over critical thinking, limiting bilingual instruction, unfairly labeling and sanctioning schools, demoralizing dedicated educators, fostering corrupt practices, encouraging educational triage, and Ă˘€“ worst of all Ă˘€“ creating incentives to push low-scoring students out of school before test day.
Or perhaps you, like the economists you cite, are unfamiliar with what takes place in actual classrooms after your ceremonial visits are over. So here's a basketball analogy that you and the President should be able to appreciate.
Suppose that NBA team owners woke up one day and decided they no longer trusted scouts and coaches to rate players. There were just too many unobservable traits that required human judgments to assess: motivation, leadership, flexibility, ability to work as a team, court smarts, and so forth. It wasnĂ˘€™t clear how those characteristics correlated with player effectiveness, as measured by objective performance data. How could the owners tell whether they were getting their money's worth? So they decided it would be simpler to pay the players based on a single measure: points scored per game.
You can imagine how that would work out. The long jump-shot would be highly valued, while skills like ball-handling, rebounding, and assists would be expendable. Nobody would pass the ball or worry about playing defense. In fact, the players would all be competing against their own teammates in an individual Ă˘€śrace to the top.Ă˘€ť Winning wouldnĂ˘€™t matter anymore Ă˘€“ only point totals. Basketball would be an entirely new game, drudgery to play or watch. But whoever said it had to be fun?
Can you now envision how schooling, a far more complex endeavor than basketball, might be harmed by a pay system that gives significant weight to one crude performance indicator? You yourself have complained about the quality of standardized tests. So how can you propose a central role for such tests in making major decisions about teachers, which, in turn, could have cascading, negative effects on their students?
We encourage you to rethink this approach and consider not only the potential waste of federal funds but, more importantly, the potential damage likely to be done by Race to the Top as presently conceived.
You might also consider the need for a kind of Hippocratic Oath among self-styled school reformers: First, do no harm. Or to put it another way: Until you have solid evidence to support your policies, don't try to impose them on our schools.
James Crawford is President, Institute for Language and Education Policy.