Why Won't Arne Duncan and Other True Believers in High-Stakes Testing Face Reality?

Democrats who support test-driven reform are showcasing a kinder, gentler soundbite to defend the indefensible. No longer do they take the "Sister Soldja" position, showing how macho they are by beating up on two of their most loyal constituencies - teachers and unions. Now, the rationale for attaching stakes to standardized tests isn't teachers being so all-fired awful; now, the voters supposedly demand it.

The current argument is that high-stakes testing may have caused severe damage to students - especially poor children of color - but we can't spend federal dollars without attaching test-driven accountability. Parents are thus being told that their kids are subject to the educational malpractice of test, sort, and punish because school reformers believe that they wouldn't invest in schools unless somebody has to give up a pound of flesh.

If this spin works, Secretary of Education Arne Duncan will soon hold his nose as he promotes the worst of No Child Left Behind, explaining to teachers, students, and parents that imposing his agenda hurts him more than it hurts us.

As high-stakes testing becomes even more unpopular, we can also anticipate more of his efforts to pretend that it isn't punitive. But the use of test scores to punish is, by definition, the use of tests to punish.

Let's engage in a "thought experiment" about what Arne Duncan would have proposed for renewing NCLB if he was focused on improving schools, not defeating adult opponents. Surely he knows that NCLB-type testing failed dramatically. Curriculum was narrowed and poor children of color were disproportionately subjected to drill and kill. Student performance growth on the reliable NAEP tests slowed as the punitive portions of NCLB tests took effect and it declined further after Duncan put bubble-in accountability on steroids.

So what would Duncan have proposed for accountability purposes if he was willing to grapple with the sad lessons of the last 15 years?

Let's start by facing the real-world significance of the seemingly miraculous increase in NCLB test scores at a time when NAEP test score growth mostly decreased or remained pretty stable. The "bubble" was not a victimless scam. At minimum, it meant that the energies of educators were diverted from school improvement to "juking the stats." Whether we refer to the process that jacked up primitive standardized test outcomes as massaging the data or flat-out fabricating it, the integrity of our public education system was compromised. Neither should we ignore the hard work of students and educators that was largely wasted in order to jack up test scores.

If Duncan and company were focused on honest data to highlight and rectify inequities, would they be doubling down on discredited bubble-in testing, or would they pattern their metrics on the highly-respected NAEP methodology? If Duncan really believed testing is a civil rights issue, would he still be stressing the quantity of test scores over their validity? Wouldn't he be proposing an assessment system that minimizes, or even eliminates, the distortions and the education malpractice encouraged by NCLB and his own high-stakes testing regime? Who would not trust results that are based on sampling or grade span testing more than those of high-stakes testing?

For argument's sake, let's think of how we would devise the best possible testing system. Let's just accept the premise that the more tests the better. We could then continue to test almost every student, every year, while using those results as a Consumers' Report, a body of diagnostic data.

To gain the benefits of testing while reducing its unintended harm, all we would have to do would be to create a "firewall" between individuals' test score result and the evaluation of individual teachers and administrators. States would be encouraged to use value-added estimates for decision-making, especially in the hands of objective school evaluators or an "Inspectorate." If state's sought to use test score growth to evaluate educators, they would have to do so on their own dime.

The obvious problem is that Duncan, with the assistance of the Gates Foundation, has already coerced states into changing their laws. By now, I bet, most states would love to toss value-added evaluations into the ash can of history. What lawmakers need is a fig leaf to allow them to undo a rash mistake without getting blamed for having leaped before they looked into the merits of using test score growth to evaluate educators.

The bigger problem, I suspect, is that it would be hard to create a fig leaf huge enough to provide cover for Arne Duncan and Bill Gates, the architects of the sham which is vams for teacher evaluations. Whether or not it was Gates behind the scenes pressuring him to do so, it was Duncan who coerced states into linking the individual teachers' test score outcomes and their evaluations. He's not likely to admit to the foolishness of this overreach.

But we Democrats should not simply lay the blame on Duncan and Gates. To a greater or lesser degree aren't we all guilty of trying to look tough so that Republicans can't paint us a liberal wimps?

And that leads to another thought experiment. What if we abandoned the blame game of the last 15 years? What if reformers who say that hate the use of tests to punish agreed to stop using tests to punish? What if we worked for a real civil rights movement of the 21st century? What if we dared to say out loud that Americans should invest in schools where all children get the education that they need and deserve?