Teacher accountability is a thorny issue that has increasingly come under discussion as student performance has declined relative both to the past and to the performance of students in other nations. It is opposed, understandably, by teacher unions, though not always by teachers themselves. Even among those who approve of accountability, the method for achieving it is not by any means clear.
Teacher performance is widely recognized as the single most important factor in student achievement, more important than class size, quality and availability of materials and resources, and physical plant. Yet standards of teacher performance are thought to be elusive or even nonexistent.
Scores on standardized tests are the only readily available means of assessing the effects of teaching on students, but these are subject to many variables outside the teacher's sphere of influence.
Teachers are in an unenviable position. They have large workloads with low pay. They must constantly juggle the competing demands of students, parents, and administration. They face an unending, but somehow always increasing, load of paperwork and extra duties.
It seems almost churlish to demand that their pay, or even their jobs, be made to depend on the performance of other people.
The fact is, though, that standardized achievement tests do provide a valid, if limited and imperfect, measure of student progress and, over time, of teacher effectiveness. Average teachers generally teach about a year's worth of academic progress in a school year. Really good teachers effect a year and a half or more of progress in their students. Poor teachers, on the other hand, teach only half a year or less of academic material.
These results can be demonstrated consistently over a three- or four-year period.
It is clear that the difference between excellent and poor teachers is enormous. It has been estimated that student performance could be enhanced significantly by getting rid of the poorest seven to ten percent of teachers, even if they could be replaced by merely mediocre ones. This could be done by measuring yearly student progress over a three-year period for every teacher.
This method would obviate the variability between different schools, classes, and grade levels, since students' progress would be measured only against their own yearly performance, not against that of other students. Measuring progress over several years would eliminate the possibility that poor student progress in a single year might be due to specific factors outside the students' or the teacher's control.
Most schools already have the data needed to perform these analyses. Those that do not could fairly easily implement procedures for obtaining this information. The results could be used to identify superior teachers and to provide feedback to all teachers to use to improve their performance; even poor teachers might be motivated to improve, especially knowing that their jobs might be on the line in a couple of years.
The question of "teaching the test" does arise. Since average teachers would not be affected by their students' test results, they would have no motivation to do so. Poor teachers might be inclined to do this, but being able to teach their classes to show a year's worth of progress on a standardized test would represent a significant improvement in their teaching! Teaching the test is at least teaching something.
Measuring student progress on testing can be an effective way to identify and remove poor teachers.
Use of this information should be strictly criterion-referenced; that is, it should require a specific level of performance, not a ranking of teachers. A school might seek to remove teachers whose students demonstrate a half a year or less of academic progress over a three-year period, not to remove the lowest performing seven to ten percent. The goal would be have all teachers achieving a year or more of academic progress, at which point there would be no need to remove any.
Though consistent academic progress can help to identify superior teachers, using the information to award bonus pay might be problematic. Teaching to the test might then become an issue where it does not represent a real improvement in teacher effectiveness. Being able to identify the best teachers could, however, enable a school to develop master teacher and mentoring programs, to which additional pay might be applied for additional or higher-level work.
Holding teachers accountable for their students' academic progress is not unreasonable, but it does need to be done fairly, using valid and reliable standards over time. The method described above could accomplish this goal without placing undue pressure on the vast majority of teachers who do a good job and deserve our support.