Stakes and mistakes in assessing teacher effectiveness
Ms. Mathews and I are off in France, where I haven't been in 45 years and she has never been. I have left some blog posts behind that will pop up while I am away. That includes this guest column by Robert C. Pianta. He is dean, and Novartis U.S. Foundation Professor of Education, at the Curry School of Education at the University of Virginia.
He is one of America's smartest educational scholars, with a long history of observing kids and teachers in classrooms. His work focuses on the assessment of teacher quality, and policy and practice that enhance children’s outcomes, school readiness, and later achievement. Here, he addresses the question of the year: What is the best way to figure out how well a teacher is doing?
Teacher evaluation is emerging as the central flash point in education policy debates. The recent controversy in Los Angeles over publication of teachers’ student test score gains illustrates this. So does D.C. Mayor Adrian Fenty’s reelection loss following his school chancellor’s firing of 173 teachers who were rated “ineffective.”
Both incidents drew national attention because they exemplify an approach to teacher effectiveness aggressively promoted by President Obama and Secretary of Education Arne Duncan -- both rhetorically and in the Race to the Top and I-3 grant programs. Teacher evaluation was the main focus of NBC’s “Education Nation” coverage; one segment featured New Jersey Governor Chris Christie ranting over teacher unions’ defensive stance on evaluation.
Teacher evaluation is controversial because it combines two elements new to education professionals and the public – quantifiable measurement of performance, and stakes like firing or public exposure. Teachers matter. But the core problem in public education is not identifying effective teachers. It’s that our existing system does not produce effective teaching in sufficient scope, scale, regularity, or intensity.
Student test scores, and whether we post them, pay for them, fire teachers, or reward kids, are simply not in themselves a sustainable solution for the lack of effective teaching. Consider the Los Angeles example: How is posting a teacher’s student test scores supposed to lead to better teachers?
Using only test scores as a way to improve teaching is like telling an orthopedic surgeon performing a hip replacement, “Good job, the patient lived” or “bad job, the patient died.” Mortality is important to the surgeon, just as student achievement is important to a teacher. But it carries virtually no information on how to do hip replacement surgery any better the next time.
It’s the same with test scores and how to teach reading to 30 seven-year-olds. Our nation, and particularly its educators and children, are desperate for better information on the techniques teachers can use to foster achievement, problem-solving, getting along with other kids, and a host of other things we value. We even more desperately need ways to produce those behaviors in the 350,000 new teachers who enter the profession each year and the millions already in classrooms.
There’s hope on both fronts. Over the last ten years, at the University of Virginia, we have developed observational assessments of teachers’ behaviors in classrooms. These produce quantitative measures of teaching quality in the classroom. We have found direct links between the extent to which teachers implement certain types of behaviors that we can observe and the resulting student learning gains, motivation, engagement, and social development.
These observational methods are not just one person’s view of effective teaching. We approached observation of teachers’ behavior with the same scientific rigor used to develop measures of student achievement, so these quantitative measures provide robust estimates that regularly forecast student learning.
More importantly, we have shown that if you use a standardized observation of what teachers do that matters, it is then possible to engineer supports to target and improve those behaviors. Not surprisingly, we and others taking this approach have found that targeted systems of professional development that focus on teachers’ observed classroom behavior produce effective teaching and improvements in student learning.
Nor are these observational measures and associated supports too difficult or expensive to implement at scale. Unlike nearly every other method for observing teachers' behavior and interactions in the classroom, this approach has been reliably replicated in tens of thousands of classrooms. The work in these classrooms shows that scalable, standardized measures of teaching behaviors meet the dual challenge of both assessment and improvement.
So what’s the take-away for school districts or reformers? First, if you want to not only measure teacher performance, but also build a system that improves classroom teaching directly and systematically, observational measures are a good bet.
Second, don’t think that local, homegrown, observational “rubrics” or measures mashed-up from an assortment of approaches that look good to a committee will do the trick. If we are going to get traction on the core problem, then individual teachers, schools, and districts must all use quantifiable, non-subjective, proven-to-work measures. There is no shortcut. Local, bottom-up construction of such measures wastes time and is likely to fail.
Third, connect the observational metrics to both carrots and sticks. This includes sanctions for teachers who interact so poorly with students that they actually contribute to student disengagement and failure. But also link them to incentives, rewards, and most importantly, targeted supports to help teachers produce effective behaviors in the classroom.
Test scores alone won’t produce effective teaching, and neither will fuzzy observations and coaching. In the end there is no substitute for rigorous, standardized measures and proven-effective approaches to improve teachers’ instructional behaviors. We already rely too much on luck to produce an effective classroom teacher. Let’s not keep repeating that mistake in our efforts toward reform.
| October 24, 2010; 6:00 PM ET
Categories: Jay on the Web | Tags: Robert Pianto, University of Virginia, how best to assess teachers, research on teacher observation
Save & Share: Previous: Students reject their school's high rank
Posted by: mport84 | October 24, 2010 10:37 PM | Report abuse
Posted by: edlharris | October 25, 2010 12:41 AM | Report abuse
Posted by: fortune1002003 | October 25, 2010 8:49 AM | Report abuse
Posted by: dreyfuss1 | October 25, 2010 9:51 AM | Report abuse
Posted by: diana_prince | October 25, 2010 10:08 AM | Report abuse
Posted by: 12191946 | October 25, 2010 11:13 AM | Report abuse
Posted by: DrDemocracy | October 25, 2010 11:51 AM | Report abuse
Posted by: Title1SoccerMom | October 25, 2010 12:10 PM | Report abuse
Posted by: shadwell1 | October 25, 2010 12:44 PM | Report abuse
Posted by: Linda/RetiredTeacher | October 25, 2010 12:59 PM | Report abuse
Posted by: Linda/RetiredTeacher | October 25, 2010 1:38 PM | Report abuse
Posted by: jlp19 | October 25, 2010 4:59 PM | Report abuse
Posted by: axolotl | October 25, 2010 5:28 PM | Report abuse
Posted by: dcproud1 | October 25, 2010 7:18 PM | Report abuse
Posted by: axolotl | October 25, 2010 8:09 PM | Report abuse
Posted by: Linda/RetiredTeacher | October 25, 2010 8:47 PM | Report abuse
Posted by: axolotl | October 26, 2010 10:29 AM | Report abuse
Posted by: Linda/RetiredTeacher | October 26, 2010 12:16 PM | Report abuse
Posted by: efavorite | October 26, 2010 1:13 PM | Report abuse