Subject Areas



Too many ideas, too little understanding

Testing Teachers to Pay for Performance

We humans have a wonderful gift of imagination. We can envision ideas that seem very clear and valid, but are failures when subjected to real world conditions.  I have previously written to dispel the belief that schools and teachers are the cause of the country’s education troubles. Unfortunately, some critics continue to imagine that teacher incompetence must be the problem. 

Pay for Performance has been proposed as a method of improving teaching effectiveness; but as our elections for school board and congress demonstrate, it is very difficult to reliably identify superior performance. To make teacher assessment more objective, proponents of performance pay intend to rely on analysis of student test scores.

Businesses and governments have long attempted to use numerical ratings to measure relative quality among their workers. Numerical ratings have been used to justify pay and promotions, but  workers and supervisors eventually corrupt the process by “gaming” the system to influence results. The inherent difficulty of using numerical ratings to measure worker effectiveness is illustrated by the constant need to “refine” criteria and develop new systems.

Deriving teacher effectiveness from student test scores is highly problematical.  Due to the multiple variables involved, reaching valid conclusions would require a level of statistical analysis far beyond the competence of politicians or educators. Absent that expertise, identifying superior performance from student test results would be akin to reading tea leaves.

Pay for Performance may sound simple; but it isn't’t. While it is easy to compare salespeople selling identical products to similar buyers, when selling different products to dissimilar buyers, the comparison becomes mostly guesswork. Teachers are salespeople who sell learning. Teachers’ products differ greatly, depending on the subjects and grades taught. Their potential buyers (students) also vary: some students are eager to learn; a majority will want to learn if the information is well presented; and some students are a harder sell, but can be convinced by excellent teaching. However, some students simply  don’t believe that the learning product will be helpful to them, and are unwilling to listen to even the most ingenious sales pitch.

The premise of performance pay is that giving bonuses to a few teachers will modify group  behavior.  An irony is that we would be telling teachers —in effect— “We know you haven’t been doing your best, so we are going to pay a few of you some extra dollars to encourage the rest of you to work harder.” There is no reason to believe that teachers are not already doing their best. Furthermore, the hope that paying bonuses will raise effectiveness is not supported by experience.  Even if bonuses were successful in improving proficiency, there would  be insufficient funds to reward all who improved, and the extra payment would need to be continued forever.

A word of caution: Pay for Performance can made to appear to be working. Forms can be filled out, test data can be crunched and “adjusted;” and  awards can be made. However, the process would be very complex, unintelligible to observers, and would result in highly questionable  conclusions. Placing teachers under the additional stress of unnecessary competition would produce unfortunate consequences. Teachers would recognize that the plan is unfeasible, and the rewards not justified by impartial observation. The over reliance on test numbers would detract from traditional teaching recognitions, and lower motivation and morale.  Efforts will inevitably be made to further concentrate on teaching to the test, instead of broader educational goals. At best, the impact on academic effectiveness would be nil; at worst it would further distraction for our schools.

There are serious problems in public education; but they largely are not related to teacher performance. Pay for Performance hopes to use test scores—not necessarily indicators of teaching quality— to solve a nonexistent  teaching problem. In a time of job losses, furloughs, and pay cuts; it is counterproductive to pursue the fantasy that test scores can provide any meaningful guide to teacher effectiveness.

Analyzing student tests won’t measure teacher quality