Times hits raw nerve with data on teachers
The Los Angeles Times has created a firestorm – and prompted a call for a boycott of the paper by the head of the teachers union – for evaluating the performance, based on standardized tests results, of 6,000 third- through fifth-grade teachers. The Times plans to publish the effectiveness ratings of individual teacher later this month, after giving them until Thursday to comment on their individual results.
Not surprisingly, the story in Sunday’s edition has generated a rush of comments, with Secretary of Education Arne Duncan, writer Diana Ravitch and some nationally prominent superintendents also weighing in on the Times article.
Duncan told the Times he endorsed the public release of the data, because parents have a right to know if their children’s teachers are effective. “What’s there to hide?” he asked. California Education Secretary Bonnie Reiss said the state would encourage districts to develop and release similar information. But Ravitch, who opposes using standardized test scores as a basis for firing teachers, called the article “disgraceful,” and California Teachers Association President David Sanchez said publishing the database of teachers’ effectiveness would be “irresponsible.”
Readers’ reactions were split. Some called for an end to tenure; others said the database would further erode the morale of teachers; one reader called for the Times to publish reporters’ reviews – so they can feel what it’s like to be under an endoscope.
The research raised two distinct questions: Should teachers be measured by the results of their students’ STAR scores in English language arts and math, and should the evaluations of those scores be published, with teachers’ names?
The former seems obvious: of course, if the limitations of the data are made clear and if the results are placed in the context of a teacher’s overall performance. The second question is tougher – and troubling – even if, as in the Times’ case, the methodology is sound. The public will inevitably inflate the importance of the ratings – even if the district and principals do not. Just as they push to get into schools with the highest API scores, ignoring other factors, parents will clamor for teachers that moved students the greatest tenth of a percentile. Many teachers will respond by fixating on STAR tests and the preparation for them even more than they do now. The numbers could become a straight jacket for their teaching, a distraction and source of tension for an increasingly joyless profession. For many teachers, the surprise publication of their scores will be an embarrassment.
At the same time, results of test scores should be an integral component of a teacher’s internal performance review. This is especially true at the margins – for the 10 percent of teachers whose students consistently regress and for the 10 percent whose students consistently outperform. The former need to improve, or, failing that, to exit the profession, while the latter should be recognized.
Methodology is critical. The Times used a technique known as a value-added analysis, which is favored by the Obama administration and many education reformers. It’s more rigorous than simply comparing raw test scores. As the Times explained, the value-added approach “rates teachers based on their students’ progress on standardized tests from year to year. Each student’s performance is compared with his or her own in past years, which largely controls for outside influences often blamed for academic failure: poverty, prior learning and other factors.”
A student who entered the year at the 40th percentile in English language arts or math would be expected to be at that level at year-end. If she falls to the 30th percentile or rises to 50th, the difference is credited to the teacher’s impact, though other influences could be at work, too.
The Times’ research was a huge undertaking. The newspaper hired a RAND Corporation researcher and economist, Richard Buddin, who did the type of data analysis that Superintendent Ramon Cortines acknowledged that the district should be doing — and will. “I think it’s the next step. It has to be done.”
Value-added analysis has its critics, who warn against using it in a high-stakes context, as the sole basis for rewarding or firing. It won’t account for the impact of larger class sizes, the bad-luck draw of disruptive students, parental divorce, a stomach ache on testing day. But the Times looked at seven years of student results – a length of time that can minimize the effect of data noise and yearly fluctuations. As the Times article fairly noted, “value-added analysis offers the closest thing available to an objective assessment of teachers.”
The Times’ macro findings were interesting and, to an extent, unexpected.
- The most effective teachers – those who raised students’ scores up to 25 percentage points in math and 17 percentage points in English — were scattered throughout the district, not concentrated in affluent neighborhoods. And the worst weren’t massed in poor neighborhoods. (Measuring value added actually presents special challenges in wealthy schools, because students may enter second grade, when tests are first given, already highly proficient, with little room for added improvement.)
- The teacher a student gets has a far greater impact than the school she attends. Some students in the study were assigned least effective teachers multiple years in a row, which can leave children falling farther behind their peers.
Credible advocates of value-added analysis don’t argue that test scores should be the sole or even the primary way to measure a teacher. But STAR program does measure the standards that teachers should be teaching, and there will be no retreating from that state commitment, regardless of what happens on a federal level to the No Child Left Behind law. The issue shouldn’t be whether value-added analysis should be done but how the data would be used.
Leaders of United Teachers Los Angeles, always 20 steps behind and 20 decibels too loud, cling to the position that STAR test results shouldn’t be part of a teacher’s performance review. As long as they do, parents and school critics will seize on the results as evidence that teachers are making excuses and covering for the weakest colleagues. They should be welcoming sophisticated analyses of student test results as tools for their improvement and as proof of their accomplishments.
Parents consider the effectiveness ratings important; teachers should assure parents that they do, too – and prove it by incorporating them into performance reviews, with consequences and rewards.
Publishing the database may finally force the UTLA to confront an issue that it has reflexively resisted.