"A child's learning is the function more of the characteristics of his classmates than those of the teacher." James Coleman, 1972

Tuesday, September 04, 2007

No Merit to Merit Pay

From Dave Riegel at Huffington Post:
I was surprised to hear that Barack Obama was sticking his big toe into the merit pay waters at the NEA convention and again at the most recent Democratic presidential debate. While Obama has not to my knowledge advocated "merit pay" by name or outlined a specific proposal, his apparent openness to the concept has excited advocates of pay for perfomance who are anxious to see a major figure on the left like Obama defy the prevailing Democratic wisdom and counter the NEA's opposition to the concept.

Marc Lampkin of the Strong American Schools campaign, nobly promoting the idea that education should be at the heart of the presidential discussion, took the NEA to task for suggesting that none of the Democratic candidates in Iowa for ABC's debate supported the concept of pay for performance. However, the candidate Lampkin points to -- Obama -- was rather circumspect in his support. In saying that pay shouldn't be tied to "standardized tests that don't take into account whether children are prepared before they get to school or not," Obama is trying to have it both ways, giving the appearance of supporting some vague pay for performance standard, but also insisting it not be tied to test scores. There's the rub: a pay system not tied to test scores isn't really a merit pay system at all.

Other kinds of financial incentives, such as paying teachers extra to work in high poverty districts or scarce fields like math or science, can't really be considered "merit pay" systems in the common parlance. Those are incentives to attract people to certain districts or fields. Pay for performance means an adherence to some type of evaluative standard, whether it be test scores or supervisors' evaluations (which are bound to be tied to test scores). And that's the problem.

The use of test scores for evaluation of teachers is fraught with difficulties that should be obvious to any outside observer. First among them, you can't pick your students upon whom your salary might depend. Those in favor of merit pay often use the private sector as a comparison point, saying essentially that most people are paid by how hard they work or how many cases they win or how much they sell. And all that's true. But a salesman isn't forced to spend his time on customers who clearly don't want to buy his products. Lawyers don't typically take cases they can't win. But the logic of paying teachers based on performance is similar to saying to a car salesman, "here are 30 adults chosen at random. Your salary depends on being able to sell all of them cars -- a standard car, at that -- regardless of their needs, desires, or ability to pay." Or to tell a lawyer, "you must win the next 30 cases that walk through your door, using limited resources, regardless of the merit of their suits, or the expense required to prosecute their cases."

Teachers don't get to choose who walks in their doors, like the hapless lawyer or car salesman in the examples above. It's the luck of the draw. Teachers (good ones) certainly believe all children can learn, and want them to. But success in terms of test scores depends on many factors, mostly too obvious to mention, outside the teachers' control. Not the least among these, and perhaps less obvious to outside observers, is the support of fellow practitioners. In many cases, a child's learning requires the support of others besides just the classroom teacher. It depends on an administrator who can effectively create an climate for learning in the school. It may depend on reading specialists who can help students comprehend their textbooks. It may depend on intervention specialists who help devise strategies for learning disabled students to make more effective gains. It even depends on successful foundations provided by teachers in previous grade levels. How do merit pay advocates propose to disaggregate the work of a classroom teacher from the support staff around her? For that matter, how would art, music, physical education or special education teachers be judged under a pay for performance system? Would we need to implement standardized tests in those areas?

I could go on and on about practical and logistical difficulties associated with merit pay. But the strongest arguments against it are philosophical. At a time when many progressives are questioning the effectiveness of high stakes testing mandated by NCLB, should we really be talking about entrenching that drill and test regime taking over education today by connecting it to teacher compensation? The real debate today should be about whether the schools created under they tyranny of NCLB are the kinds of schools we want to have. Do we really want high stakes tests driving our definition of education? And driving our definition of quality teaching?

I am always suspicious of merit pay arguments because they seem to insinuate that a teacher's effort is dependent upon his or her level of compensation. Instead of rewarding teachers for maximizing student achievement -- as most would insist they are trying to do anyway -- the right approach would be to reward activities that help teachers become better trained and more competent. For example, most local salary structures reward teachers for attaining a higher level of education -- teachers who earn a Master's degree earn more than teachers with similar experience who do not. Likewise many states offer annual stipends to teachers who achieve National Board Certification, a rigorous process which requires teachers to demonstrate and reflect upon their classroom practices. These sorts of rewards make sense to teachers: they understand the connection between professional development and effective instruction.

I find that merit pay advocates also hope that a compensation structure will do that job of evaluating teachers that should properly be done by effective building administrators. We shouldn't simply withhold monetary rewards from teachers who are ineffective: we should help them improve or evaluate them out of the profession. The canard that teachers' unions protect bad teachers from dismissal is not true: bad administrators protect bad teachers from dismissal or non-renewal. But teacher evaluation is more complicated than simply looking at test scores. It requires careful examination of specific teacher behaviors in the classroom, of how a teacher relates to students, and his or her command of the subject matter they are teaching. This cannot be judged simply by looking at test scores, which may be high in some cases in spite of uninspiring instruction: it requires an effective and highly skilled administrator who knows what she is looking for when she observes a teacher interacting with her students, and who is skilled at helping teachers improve. In short, pay for performance provides an easy way out when quality supervision of instruction is what should really be taking place.

Finally, the discussion of merit pay in the context of a presidential campaign continues a disturbing trend of increasing federal involvement in local decision making. Teacher salary structures and evaluation practices are negotiated locally between a board of education and a bargaining unit under the broad general guidelines of state law. If Denver teachers agree to a merit-based system, then good for them. They've decided in agreement with their board on a system that makes sense for them and their community. These kinds of contractual decisions are and should remain local, not the subject of federal intervention. An important reason why the NEA objects to merit pay proposals is precisely this -- that it takes away control from a local bargaining unit to decide their own fate. If Barack Obama truly believes that education proposals need the support of teachers, then those proposals should continue to be locally decided, not a subject of debate in a national election, unless it is clear that the debate is purely philosophical, and not bearing on any public policy he would enact as president. The federal government certainly has in important role in education. It establishes policies and guidelines that protect the education of handicapped children, for example, and provides funding to support that education. The federal government supports research in education and provides grants to support high poverty schools. But dictating the terms of local teaching contracts should not be a function of federal policy.

The debate about merit pay isn't the debate we need to be having right now. With the demands for charters and vouchers from the right, and the ongoing problems facing education in high poverty districts, the very existence of public education is being threatened. We need to be talking about why public education still matters, and what it should look like in the 21st century. Gimmicks like pay for performance are only getting us off track.


  1. Even worse - the money being held to reward the teachers of the higher achievers might otherwise have been budgeted for other needs: professional development, cost-of-living increases, etc.

  2. Anonymous10:09 PM


    In Texas, as in other states, we can't bargain with local school boards on salary. We take what we're given. Often, salary decisions aren't made until the school year is already started, when it's too late for teachers to change jobs.

    Great post, otherwise. It's amazing how accountability advocates forget the role of assistant principals, principals, and other administrators in student achievement (however it's measured).

    Q: What does it take to become a school principal in Texas?
    A: A master's degree and two losing seasons.