Good teachers, good students

The role of test scores in evaluating teachers is a prickly and complicated issue, which is why California has been avoiding the conversation for so long. Fortunately, that procrastination is no longer possible after The Times took the bold step of analyzing standardized test scores in the Los Angeles Unified School District to see whether individual teachers appeared to be successful at raising their students’ scores.

Given the current nationwide push to include test data in teacher evaluations, it was time to strip away the mystery about test scores and take a close look at what they are, what they show and don’t show, and what teachers, administrators and the rest of us might learn from them. The Times’ articles and online database rating nearly 6,000 elementary school teachers allow the examination to begin.

Critics, and especially union leaders, railed against The Times’ decision to release this information, saying that it unfairly casts teachers as good or bad. They predicted that parents would throng local schools demanding that their children be assigned to one teacher or another based on scores that are at best an incomplete measure of a teacher’s effectiveness.

Test scores are indeed just one indicator of a teacher’s performance. It’s too early to determine the long-term response, but so far, parents seem to understand that; L.A. Unified reports that there has been no headlong rush to principals’ offices for a change of teachers before classes start later this month.

But it’s revealing, and disturbing, to read the comments of some teachers who don’t seem to care whether their students’ scores slide. They argue that they’re focused on more important things than the tests measure. That’s unpersuasive. The state has carefully constructed some of the best curriculum standards in the nation, which are about to become better with the adoption of new English and math standards. These represent widespread agreement among educational experts on what students should learn by certain grades. We’re far past the point of allowing individual teachers to decide how much of the curriculum they want to impart, or sitting by while low-income students enter high school illiterate and without a basic grasp of multiplication.

There are better ways to respond to the newly published information. Instead of defending lackluster ratings, some teachers talked about learning how to improve — not by “teaching to the test” but by adding new skills to their pedagogical repertoire. Times reporters uncovered ample anecdotal evidence that teachers whose student scores improve the most are not stressing testing skills but rather engaging students in challenging, often lively, academic work.

There are teachers who, year after year, show marked improvements in their students’ scores. We’d like to think other teachers would flock to their classrooms, to observe and learn their secrets, and that administrators would be looking for ways to reward them for their work.

This page has never believed that test scores should count for all of a teacher’s evaluation — or even be the most important factor. But they should be a part of it. There is more to good instruction than preparing students for the tests, but covering the required material effectively is a necessary part of good teaching.

To be sure, there are weaknesses in California’s tests, which are almost solely multiple choice because open-answer questions are too expensive to grade. The only exceptions are in fourth and seventh grades, when students must write an essay. How can we judge how well high school English teachers impart sophisticated writing skills — one of the most important abilities needed for college — if students don’t have to show after seventh grade that they can string a few sentences together? How can we judge the teaching of history when the tests ask rote questions instead of measuring critical thinking skills? Parents should be informed about whether teachers have successfully raised student achievement on tests, but they should also be informed about the tests themselves; sample questions are on the California Department of Education website.

Right now, the “value-added” scores The Times has been reporting are more useful for evaluating schools than teachers. Many factors can throw off the data at the classroom level. The scores are better at telling us about the performance of the best and worst teachers than the majority in the middle, and several years of data are better than a couple. In many elementary schools, students have multiple teachers, but only one receives credit or blame for a certain group’s test performance. Students have no stake in doing well on the tests, which do not affect their grades or promotion. The list goes on.

That’s why we think the Obama administration has been too hasty to push states into linking test scores to teacher evaluations and to reward states that overemphasize the scores, making them count for half or more of a teacher’s worth. The administration’s first priorities should have been developing better tests, which it’s working on now — if we’re going to judge teachers in part by these scores, it’s unacceptable to say that top-notch tests are too expensive — and statistical models that minimize random factors and make the scores a better evaluation tool.

Current teacher evaluation practices are ripe for overhaul. Performance reviews should include, at minimum, classroom observations, portfolios of student work over the academic year and, yes, objective test data. Most important, they should be used to help calculate teacher compensation. The Times’ investigation found that experience had little to do with which teachers were the most effective at raising test scores, yet teacher pay scales and job security rely mostly on seniority. We must reward great teachers, train less-than-great ones, and overcome ridiculous tenure rules that block schools from firing those who don’t improve.