As promised… a weekend follow-up to my Friday post about education reform. 

The real centerpiece of reform efforts is the notion of “data-driven education,” i.e., that teachers can be objectively measured according to certain benchmarks.  While perhaps a reasonable principle, what this has meant in practice is an increased reliance on standardized testing.  Standardized testing in education is an abominable failure as a scientific method of data collection: different populations are measured against each other without even establishing a baseline level of performance for any of them.  Any college freshman suggesting this kind of experimental design would fail a research methods class in a heartbeat. 

If data-driven education truly wishes to succeed, then data must be collected objectively.  The proper way to do this in an educational setting would be to test students at the beginning of each school year, to establish a baseline for their performance, and then test them again at the year’s end, to discover how much they have learned.  This process must be repeated each year, so that if many students, over several years, learn significantly more in their fifth-grade classroom than in their fourth-grade classroom, we might then reasonably ascertain that the fifth-grade teacher is more effective than the fourth-grade teacher (alternatively, if it was discovered that students of all different backgrounds across the country learned more in fifth grade than in fourth, we might conclude that, developmentally, the fifth grade is an important year of significant cognitive growth, more so than the fourth grade).  Such a metric would involve more time and probably much more attention to quantitative research methods amongst educational administration instruction and practice, but it would have the benefit of actually telling us something, instead of the current, relatively useless system of expecting teachers in core subject areas and key grade levels to be wholly responsible for their students passing state assessment tests without any sense of what those students’ capabilities were upon entering the grade. 

That, then, is problem number one with current reform efforts: remarkably shitty experimental design.  Problem number two is how this poor design translates into poor incentives for teachers.

The best example of these skewed examples is the recent rash of cheating scandals, in which teachers in a number of districts — in struggling schools, who felt that their jobs (or bonuses) were entirely dependent upon student standardized test scores — were found to have falsified student tests to increase their average scores.  Instead of the multi-year assessment and ongoing quantitative observation of classroom learning that would not only provide greater insight into teacher performance but also offer more of a window of time for improvement, current education policy often ties teachers’ salaries and bonuses — or their employment altogether — to standardized test performance.  Although known as “merit pay”, there is very little proven correlation between student standardized test scores (taken as independent variables) and the merit of their teachers.  By reducing the entire year of teaching to a single variable — test scores — reformers, administrators, and policymakers have essentially invited the unscrupulous to manipulate that variable.  And, because of the one-to-one correlation between test scores and pay, people are doing just that.

Now, another cherished effort of contemporary reform is the development of alternative teaching programs: charter schools, Teach for America, the New Teacher Project, etc.  These projects are celebrated for helping to attract “the best and the brightest” back into teaching, and although it is questionable how many of their teachers remain career educators, they have nonetheless attracted a great deal of attention, funding, and enthusiasm.  So far, they seem to produce marginally higher scores than their traditional public counterparts, though nothing that is remarkable.

What is striking about these programs, however, is how they exist at the exact opposite motivational pole as “merit pay.”  Merit pay relies on the capitalist market norm that performance and financial reward should be highly correlated, and that the promise of greater financial reward will, in turn, inspire greater performance.  While participants in Teach for America and prominent charter programs like KIPP are paid at typical district salaries according to their experience, those teachers more often cite other things as the reason that they became teachers: giving back to the community, effecting positive change, etc.  I’d venture to guess that a high percentage of entry-level teachers outside of these programs feel similarly, but very few public school districts make a point of emphasizing that element of teaching, or of making their teachers feel like members of a community with a clear social mission.  The notion of teaching as a “noble profession” is in direct contradiction to the reformist efforts towards increasing monetization.

Why is this important?  As Dan Pink explains so eloquently in this TEDTalk, recent psychological explorations demonstrate that tying performance to fiscal reward is a very, very poor way to produce strong results — a counter-intuitive proposition, perhaps, but one that manifests again and again.  Of course teachers should be paid well enough to live comfortably, but the idea that the greater reward must be ideological and reputational, so shocking to our free market perspective, is falling away just when it needs to be resuscitated most.  The ascendance of merit pay, particularly as its tied to horribly flawed metrics of performance, is truly troubling.  If you disagree, here’s a thought experiment: recent data pegs the true value of a good kindergarten teacher at about $350,000.  Now, if school systems and policymakers decided to observe the market and pay teachers according to their true social value, we would pay good kindergarten teachers that amount.  However, imagine that we have only our current system of measurement for teacher performance.  A kindergarten teacher can only earn that pay if his or her students perform well on state-mandated reading and math assessments at the year’s end.

What sort of kindergarten teachers is such a system likely to attract?  Passionate, curious men and women who genuinely care about the well-being of children, who are not interested in giant payday so much as improving kids’ lives?  Probably there would still be a few, and they would be less likely to cheat on standardized tests.  But such a system would probably also attract a lot of people who would otherwise go into fields like finance, seeking an easy paycheck.  After all, all they have to do is get some decent test scores!  For folks who have no compunction about engaging in “creative accounting” over years to shareholders and regulators, once-annual deception of the state department of education would seem like a cakewalk. Does anyone want the sort of people who toppled the entire global system of finance to be in classrooms, teaching children?

This is, of course, a worst-case scenario.  But it is also where current reform efforts — with an increased emphasis on merit pay, based only on a single, terrifically poor measurement of teacher performance — ultimately point.  As data about teachers becomes more easily available (even published in major newspapers), and as the teaching profession is commoditized more and more, we must think critically of reform efforts and be sure that the measurements and incentives involved operate together to attract, and maintain, a high caliber of educators for all students.