By Richard D. Bingham, Ph.D. and John S. Heywood, Ph.D.
The growing concern about the quality of education being provided by our nation’s public schools is being shared by citizens and public officials alike. Increasing evidence shows not only that American students are learning much less than their counterparts in other industrialized nations, but also that today’s American students, on average, are learning considerably less than their counterparts of twenty-five years ago. Hard questions are being asked about who is responsible for these sorry conditions and what can be done about them.
Unfortunately, we do not have irrefutable answers to either question. But prominent in proffered answers to each are teachers and improved teaching. It is teachers who spend large amounts of time with students and who many believe should have the greatest impact on student performance. But how can teachers be held accountable for student performance when to date we have no accurate way of measuring their contribution to student learning? And how can we improve teaching when we cannot prove what are the superior ways to teach?
This report describes a newly developed method which attempts to answer these questions. It outlines a systematic means of identifying teachers whose students annually have learned more than they were expected to learn, based on their previous performance. Superior teachers are identified through a procedure which involves statistically predicting how each student in a classroom should do in a given year based on previous performance and comparing the prediction with actual performance. The residuals (differences) between the two for each student are then aggregated across all students in the classroom to determine whether students as a whole did better, a positive sum, or worse, a negative sum, than predicted. Superior teachers are identified when students did much better than predicted and thus have a large positive residual. Inferior teachers are identified when students did much worse, as shown by a large negative residual for the classroom. A similar procedure is used to identify superior schools.
In the pilot project in which this study was undertaken, some extreme boundaries were set for identifying superior and inferior teachers and schools. The boundaries were arbitrarily set to prove that there is a continuum of results that would identify superior to inferior schools and teachers. Out of the almost 100 schools in the sample, seven schools had unusually large, positive results, and ten had unusually large, negative residuals. At the classroom level eight teachers were associated with unusually positive and twenty-four with unusually negative differences.
The most important conclusion is that one can begin to objectively evaluate teachers using this method. The method is flexible enough that it can be customized for any school district. It provides an identification of the superior and inferior teachers and holds teachers accountable for the performance of their students, taking into account other factors affecting performance.
Once the superior teachers have been identified, perhaps further answers as to how to improve teaching can be gleaned. The mere fact that teachers can be held accountable may help. But by identifying the superior teachers and then monitoring them, precisely what it is that makes them successful can be learned and passed on to others. Second, those superior teachers who desire it might be elevated to master teacher status so that they can directly convey their methods of teaching to others. Third, the successful teachers should be financially rewarded for achievement, creating incentives for them and others. This pilot method allows merit pay plans to be implemented because it identifies those teachers who are most deserving.
On the negative side, further testing and/or implementation of these procedures does impose certain costs. Complete and accurate records on each student must be maintained in a centralized, computerized file. Without such a record-keeping system the procedure should not be attempted.
It can be done
By Richard Bingham, Ph.D. and John Heywood, Ph.D.
December 1989
