Bill Gates says that his $5 billion experiment in education has not failed. Gates also claims that he trusts in science. That prompted me to reread the National Academy of Sciences' analysis of the failure of test-driven accountability.

A funny thing happened as I tried to find a bifocal-friendly PDF. I accidentally found the Brookings Institute's "Volatility in School Test Scores." Its preface cited John Dryden's poetry: "Errors, like straws, upon the surface flow; He who would search for pearls must dive below."

The Brookings report explained the damage caused by those errors, concluding that test-driven accountability could make teachers behave like dogs who were trained by electrical shocks. When the tests changed, dogs just laid there and waited for the pain to stop. The report then warned of similar "perverse incentives that may actually harm rather than help."

These metaphors perfectly describe the Gates Foundation's "teacher quality" research. Gates is empowering districts to either dive deep and seek pearls, or skim the surface, and condition educators to cower and endure the indignities of test-driven schooling. As the toxicity dumped on teachers pours down on students, a generation or more of them might be robbed of a respectful education.

I then found the NAS's new blue ribbon panel's report, "Incentives and Test-Based Accountability in Education," and reread that masterpiece. It showed the failure of NCLB-type accountability to improve student performance, and explained why those "reforms" had damaged students. For instance, it showed that graduation examinations had decreased graduation rates by 2%, while not increasing student performance. It also summarized the huge body of literature on the unintended, destructive consequences of the accountability movement.

The problem is that for every effort to use testing as a part of a dive-for-pearls experiment in helping students, there is one (or more) effort(s) to grab at straws. The NAS argued,

Test-based incentives for students may cause some students to achieve more and others to drop out, even with extra support and remediation. Test-based incentives for teachers may cause some teachers to become more effective and others to leave the profession. Test-based incentives for schools may cause some to focus on the full curriculum and others to focus on test preparation.

The report also issued a not-so-ringing endorsement of the hypothesis that Gates-funded teacher quality research might pan out, "The extent to which value-added models can realize their promise has not yet been determined," because "it is not yet clear how fully these models can account for student differences to provide accurate measures of teacher effectiveness."

I am cautiously optimistic that the Duncan Administration might heed the experts' warning, but I am more hopeful that the report will change the course of Gates' research. After all, their work is a product of brilliant academics. Those scholars do not want to look in the mirror and see an alchemist peering back at them. Or worse, they do not want to land in the History of Science's Hall of Shame of scientists whose discoveries produced great harm. The jury is not in, completely, but surely it is time for social scientists like Paul Hill and Tom Kane to read the NAS report and ask whether they want to invest their considerable talents in a project which is likely to produce more destruction than good.

And another funny thing happened when I found a PDF of "Incentives and Test-Based Accountability in Education." Among the noted scholars who produced the NAS report were Paul Hill and Tom Kane! In fact, it also was Kane who had written "Volatility in School Test Scores," which cited Dryden to warn that schools grasping at straws were likely to produce teachers and students who learned that there was little they could do to avoid the abuses of test-driven accountability, and who thus tried to wait out the abuse. So maybe it is time for Kane, the head of Gates' "Measures of Effective Teacher Project," to heed his own words, and choose science over political spin.