In recent years, reformers have sought to improve our failing public education system by tightening and standardizing the measures we use to judge performance. From the numerical requirements of the No Child Left Behind Act to California's increased focus on assessment and accountability, there's been a conscious attempt to use hard data to measure success at every level of the education system.
But one group does not have its performance measured this way: teachers. Determining the effectiveness of individual teachers -- are they helping our kids learn or not? -- remains a mostly subjective judgment. Yet there's no reason why teachers shouldn't also be evaluated against objective measures of student performance just as are schools, districts and states.
Teacher evaluations focus on what they do in the classroom -- the input of the learning process. In most school districts, principals show up at prearranged times to observe teachers' work, and then write their observations. In doing this, they typically use a checklist to guide their assessments. Evaluations usually consist of one or two written observations.
This superficial and largely subjective approach to evaluating teachers is something of a farce. In many instances, principals can only rate teachers "satisfactory" or "unsatisfactory." Multiple unsatisfactory evaluations can lead to dismissal. But faced with the prospect of battling the local teachers union to prove that a teacher's unsatisfactory evaluation is valid, most principals capitulate and rate virtually all teachers as satisfactory.
This rubber-stamp routine may make things easier for administrators, but not for the kids. Several researchers, among them Eric Hanushek of the Hoover Institution of Stanford University and Jonah Rockoff of Columbia University, have shown that teachers are not interchangeable when it comes to student learning. Given a year with an effective teacher -- one whose pupils previously showed test-score gains -- students can advance their learning by a grade level or more, according to research done by William L. Sanders while he was at the University of Tennessee. He also found that under a weak teacher, kids' progress can stall, and they can fall behind.
So why not include student test scores -- the output of the learning process -- in teachers' evaluations? Besides giving the evaluation process a much-needed shot of objectivity and rigor, this change could help administrators target assistance for struggling teachers and recognize those who are most effective in the classroom.
In its report this month, Gov. Arnold Schwarzenegger's nonpartisan committee of education experts agreed. Among other things, it recommended that teacher evaluations should be based in part on student achievement.
Teachers unions object to using student test scores to evaluate teachers. They argue that these scores are influenced by many factors beyond a teacher's control -- students' home environments, language abilities, whether they ate breakfast on the morning of a test. True enough, but this is not a reason to ignore student achievement altogether.
Of course, student test-score data should not be the sole measure of a teacher's performance. It should be combined with other factors to produce a well-rounded assessment, including more rigorous and more frequent classroom observations by principals, announced and unannounced, as well as reviews of teachers' lesson plans and homework assignments by principals or peers.
And incorporating student test data into teachers' evaluations should be done in a way that ensures fairness. For starters, not just absolute student test performance should be taken into account, but also how much students grow over the course of a year. For instance, a teacher could make phenomenal progress with struggling students but still not get them to a high achievement. In this case, the teacher should be rewarded, not penalized. This approach would prevent teachers from fleeing low-performing schools or classes.
Second, evaluation must consider extenuating circumstances. For instance, if a first-year English teacher is assigned to teach chemistry, he shouldn't be blamed for less-than-stellar test scores.
Finally, any attempt to use test scores to help evaluate teachers should not be done on the cheap. Policymakers may be tempted to co-opt existing assessments like California's STAR tests for the purposes of teacher evaluation. But these standardized tests are designed to give information about how a school, district or state is performing, and they don't cover all subject areas. To build a better system of evaluating teachers, it is worth the investment to design tests that measure how much individual students learn over the course of a year on the material the teacher is expected to teach.
There's no question that teachers have tough jobs. But the old evaluation system that ignores student achievement and finds virtually all teachers "satisfactory" simply sets the bar too low, lacks objectivity and does not address whether students are actually learning. If we want to give students the best chance at success, we need to do a better job of determining whether their teachers are helping them. Evaluating teachers with no hard evidence about their primary responsibility is just plain irresponsible.