We’ve often referred to Jacob Vigdor’s work on the implications of creating an evidence-based salary schedule for teachers:
Relative to a teacher just beginning in the profession, teachers with one or two years of experience raise test scores by an extra 5 percent of a standard deviation. They are paid, on average, 2 percent more than starting teachers. If the standard were to pay teachers an extra 1 percent of salary when they raise test scores by 2.5 percent of a standard deviation, then highly experienced teachers who post a 25 percent test-score advantage over rookies should be paid a 10 percent premium. Instead, their premium approaches 70 percent. Visually, the darker bars rise quickly at first, moving from left to right, but largely level off once a teacher has six years of experience. The salary schedule marches right along, providing continuously increasing rewards to teachers as they progress from 6 to 27 years of experience, even though their classroom effectiveness has barely improved.
The existing salary schedule rewards teachers too little for the substantial improvements they post in the first few years on the job, and too much for the later years of their career, when they show only incremental advances. An evidence-based salary schedule would alter this arrangement, focusing the rewards on the early rungs of the experience ladder.
An evidence-based salary schedule would represent a marked improvement over the status quo. But Vigdor begins his meditation on the salary schedule for teachers by asking a question that is not generally raised:
On what basis should we distribute rewards to salespeople?
The idea is that we accept that different salespeople should be compensated differently, according to a decentralized trial-and-error discovery process in which different firms will embrace different approach. This came to mind as I read Gene Hickok’s thoughts on shifting from collective bargaining to individual teacher contracts:
Why not borrow a practice from higher education where faculty sign individual contracts with their institutions? This would give school district and building officials the flexibility to negotiate salary and employment benefits based upon the individual teacher’s experience and subject expertise. It would focus more attention on the talents and skills of the teacher, the needs of the students, and the efficient allocation of revenues by the district/school. A qualified teacher in a subject that is difficult to staff might be able to negotiate a contract reflecting that fact. A low performing school might be able to attract talented teachers by offering more lucrative contracts or signing bonuses. As it currently stands, in most places, seniority is what drives differences in teacher compensation and more often than not the more senior and experienced educators are not assigned to schools that need them most. This makes very little educational sense.
Think about how such an approach might transform the teaching profession. The focus would be on the individual educator and his or her relationship to the school, the classroom and the students. The teacher would no longer be a “faceless member of a bargaining unit.” Every few years, the teacher would negotiate another contract taking into account such things as performance, professional development, and service to the school. It might be possible for a teacher to “earn” tenure after a number of years of superlative performance, just as faculty do in higher education. Currently, tenure in American K-12 education is awarded after only a few years of teaching, has almost no relationship to job performance and too often provides employment protections for those who need to find another profession. Indeed, individual contract negotiations might help to make teaching the profession it should be.
The chief virtue of this approach is that unlike most proposals for merit pay, it allows managers to take into account variables that models operating at a system-wide level would inevitably fail to capture. This brings to mind Amar Bhide’s critique of “robotic finance,” which Yves Smith referenced two years ago:
Because natural laws and mathematical inferences cannot predict behavior, algorithms are built upon statistical models. But for all their econometric sophistication, statistical models are ultimately a simplified form of history, a terse numerical narrative of what happened in the past. (The simplifying assumptions of most statistical models are in fact so great that they can almost never be used successfully to reconstruct the very historical data used to construct the models.) They reveal broad tendencies and recurring patterns, but in a dynamic society shot through with willful and imaginative people making conscious choices, they cannot make reliable predictions….
This doesn’t mean statistical controls and data-mining programs are useless in human affairs. They can debunk false assumptions and stereotypes or suggest new rules of thumb. Faced with a large number of choices (as when thousands apply for one job), they can provide a quick, objective first-cut screen. But predictions of human activity based on statistical patterns are dangerous when used as a substitute for careful case-by-case judgment. They nonetheless continue to gain ascendency. Nowhere has this been more apparent—or more dangerous—than in the financial industry. [Emphasis added]
It is easy to see why teachers might chafe at mechanisti merit pay formulas. Hickok’s approach is essentially a way to bring case-by-case judgment back into play.