November-December 2006 > Right Now

Pedagogical Performance

Grading Teachers

by Ashley Pettus

Education researchers and policymakers, like the rest of us, have long known that a good teacher can make all the difference to a childs education. The federal No Child Left Behind law requires that school districts hire only highly qualified teachersthose with a bachelors degree, full state certification, and demonstrated competence in the subjects they teach (i.e. a college major or equivalent course credits, passage of a state examination, or a graduate degree). But do these credentials actually add up to efficacy in the classroom?

Thomas Kane, professor of education and economics at Harvard Graduate School of Education, has sought to measure the effectiveness of teachers based on their accreditation status. In a New York City study conducted with Douglas Staiger of Dartmouth College and Columbia Business School professor Jonah Rockoff, Ph.D. 04, Kane looked at students test scores in grades four through eight to estimate the value added by their teachers, whom the researchers divided into three groups: traditionally certified, alternatively certified (by the New York City Teaching Fellows Program or Teach for America), and uncertified at the time of hire. The researchers controlled for students prior-year test scores, and for classroom, school, and grade-related factors, as well as for the teachers experience levels.

They found only small differences among the three groups, but large differences in ability among individual teachers. To put it simply, the study authors conclude, teachers vary considerably in the extent to which they promote student learning, but whether a teacher is certified or not is largely irrelevant to predicting their effectiveness. Other academic credentials, such as selectivity of undergraduate institution or grade-point average, also proved to be poor predictors of a teachers classroom impact. (The study is available at www.nber.org/papers/w12155.)

As a result, Kane argues that school districts would do better to use performance on the job, rather than pre-hire criteria, as the basis for long-term teacher selection. The traditional approach has been to create a bunch of hurdles, like certifications and test scores, on the front end. But then once people are hired, theres little subsequent selection, he says. Fewer than 2 percent of teachers who left their jobs during the first three years of teaching reported having been involuntarily dismissed or transferred, according to a 2004 survey. After three years on the job, even the lowest-performing teachers are protected by tenure under a system that rewards longevity rather than results.

Kane and his coauthors found that schools can gain a fairly clear picture of a teachers effectiveness in the first two or three years. Beyond that point, the returns on experienceas teachers improve with practicebegin to diminish. Youd expect the distribution of teacher effectiveness to get tighter as the years pass, Kane says. But we see the opposite. Theres a tendency for the more effective teachers to get better faster, while the weaker teachers dont catch up. In order to hire the highest-quality teachers, he argues, school districts need to open the doors to a wide pool of candidates, certified or not, and then assess each teachers value over a three-year trial periodusing not only student test scores, but also classroom observations, reviews of student work, and parent evaluations. If were going to be selective, Kane asks, why dont we be selective at the point where we actually have some information?

Kane and Staiger also present their case in a second 2006 policy paper, written for the Brookings Institutions Hamilton Project, that was coauthored with Robert Gordon 93, a senior fellow at the Center for American Progress, another Washington, D.C., think tank. Drawing on student achievement data they collected from a sample of Los Angeles teachers observed during their first, second, and third years of teaching, the researchers ranked teachers according to their levels of effectiveness and quantified their estimated impact on student learning. They found that an average student with a bottom quartile teacher lost 5 percentile points relative to comparable students in terms of baseline test scores and demographics, while the average student assigned to a top quartile teacher gained an average of 5 percentile points. This 10-point spread suggests that the effects of having a top-quartile teacher for four years in a row would be enough to close the national black/white test score gap, now estimated at 34 percentile points.

Based on these findings, the authors propose awarding significant pay bonuses to highly effective teachers who are willing to teach in high-poverty schools. Current data from Los Angeles reveal that students in the poorest schools (where at least 90 percent of families qualify for free or reduced-cost lunch) are more than two and a half times as likely to have bottom-quartile teachers than are students in the wealthiest schools. The existing systemuniform pay based on experience and educational attainmentpromotes an inequitable distribution of teachers, be- cause principals in low-achieving, low-income schools have no way to compensate teachers for the additional challenges of working in often-blighted schools, whereas wealthier schools offering better conditions tend to attract better teachers. The authors argue that targeting salary increases to high-performing teachers in underserved schools could help counter these effects and even, potentially, bring more talent to the teaching profession.

Under Kanes proposal, performance-based measures would apply only to incoming teachers and would not affect existing teacher contracts. Nor would this two-tiered approach hinder the plans impact, because a majority of baby-boomer teachers (who entered the profession in the late 1960s) are approaching retirement. Schools are expected to face a severe teacher shortage in the coming years. A policy that minimizes traditional barriers to entry, while implementing more effective quality controls, could prove extremely timely.

Thomas Kane e-mail address: kanetogse [dot] harvard [dot] edu (kaneto@gse.harvard.edu)