What’s the Best Way to Grade Teachers?

After the Los Angeles Times made public a database of “value-added” rankings – which measure teacher performance by student test scores – the question of what makes a strong teacher became a controversial one. Below, ahead of Zocalo’s panel asking whether teacher rankings work, three education scholars tell us what single measure best captures a teacher’s performance.

No single measure, not even value-added

Aaron PallasI don’t think a question like this can be answered without understanding the different reasons for measuring teacher performance. A measure that might be useful for evaluating a teacher for tenure might not be particularly useful for some other purpose, such as identifying teachers who might benefit from professional development to address a particular concern in their teaching, or teachers who are to receive bonuses for excellent performance. Teacher evaluation systems need to be built from the ground up with a clear sense of the multiple purposes to which the evaluations will be applied.

Not surprisingly, if there are multiple purposes, it’s unlikely that a single measure of teacher performance will suffice. The next generation of teacher evaluations will include measures of teachers’ contributions to student learning and measures of whether teachers are engaging in good teaching practices. But we are still in the early stages of developing measures that we can count on to tell us what we want to know.

For example, value-added measures of teachers’ contributions to student learning, such as the ones reported by the Los Angeles Times in August of this year, might tell us something useful about teachers’ contributions to the kinds of skills measured by the California Standards Tests, but they won’t reveal teachers’ contributions to a student’s ability to solve problems, to think critically, or to become an engaged citizen willing to take responsibility for his community. If these goals are important, as most of the public believes, they should be reflected in how we measure teachers’ impacts on student learning.

Conversely, developing measures of good teaching practice requires first agreeing upon which practices are exemplary, and then training observers to identify these practices when they occur.  It’s a challenging agenda, because every classroom is different, and the practices which might be important in teaching beginning reading may not be the practices that matter most in teaching high school chemistry. We can make a great deal of progress in constructing measures of teachers’ contributions to student learning and of teachers’ engagement in exemplary practices, but we’re a long way from where we want to be.

Aaron Pallas is Professor of Sociology and Education at Teachers College, Columbia University in New York City.


First ask what they’re teaching

Morgan PolikoffOften overlooked in the conversation about teacher effectiveness, the content of the enacted curriculum is the most important measure for understanding teachers’ unique contributions to student learning. By “enacted curriculum,” I mean the particular topics and skills that the teacher covers in his or her instruction. There is simply no substitute for high quality curriculum; this is the key belief that underlies the last two decades of educational reform, and it based on the intuitive fact that students learn that which they are taught. There are several straightforward ways to measure the enacted curriculum, including daily logs and high-quality teacher surveys.

Unfortunately, comparative international studies at all levels indicate that U.S. curricula across subjects and grades are poorly structured relative to the curricula of higher-performing nations. Our curricula are broad (i.e., covering many topics), shallow (i.e., covering them superficially), and repetitive (i.e., covering them in grade after grade). The result is that our K-12 students know a little about a lot. This is not the kind of deep knowledge we want to be cultivating in the increasingly global economy.

In my view, it is not appropriate to have serious discussions about teacher “effectiveness” without considering the more important issue of what teachers are teaching. To understand which teachers are “good” at teaching fractions, we would first want to be sure that all teachers are actually teaching fractions. This means creating high quality standards (the newly developed Common Core State Standards are a start) and working seriously to develop outstanding curriculum materials and pacing guides to ensure that teachers are covering the key topics and skills at each grade. Once we are confident that teachers are enacting a high quality curriculum, it will enable us to target instructional quality more directly in our policies.

Morgan Polikoff is an Assistant Professor of K-12 Policy and Leadership in the Rossier School of Education, University of Southern California.


Value-added helps, but not enough

Peter GoldschmidtDeveloping an indicator of performance requires that we first have a notion of what information related to performance the indicator should convey – that is, what is the intended purpose of the indicator? Different stakeholders will have different conceptions of use: parents may want to know whether their student is learning to read, to do algebra, or is gaining social skills; principals may want to know how well students in particular grades are doing, or whether particular instructional choices are working; state education leaders may want to know whether students are college and workforce ready – whether the curriculum is adequate to prepare students to meet state standards. Given this myriad of uses for teacher performance information, it is clear that a single indicator cannot adequately supply the required information – multiple indicators are the only reasonable means of evaluating teachers’ performance.

Value added models (VAM) are currently receiving a substantial amount of attention because they potentially provide unbiased estimates of teacher performance based on student assessment results. VAMs rely on a set of assumptions that may or may not be tenable in practice (e.g. sorting, if any takes place, only depends on static student characteristics). Still, it is likely that value added estimates of teacher effectiveness is likely better than the ubiquitous teacher evaluations that rate virtually all teachers satisfactory. VAMs can provide information about the variability of teacher performance, but these estimates have issues with reliability and stability over time. Further, VAMs require at least two assessment results, which means that under current assessment systems approximately 80% of teachers would be excluded.

A better indicator of teacher performance might be based on measuring the opportunities to learn (OTL) – the materials teachers provide in their classrooms. There is considerable evidence that links OTL and student performance. The OTL indicator, based on direct student input, does not require dramatic changes to district assessment systems. Assessing OTL requires a carefully developed instrument that provides an unbiased, fair, and reliable score from which valid inferences about teachers’ facilitation of OTL can be made. Even this indicator will not meet the aforementioned need to use multiple measures, but it should certainly be included in any comprehensive teacher evaluation system.

Pete Goldschmidt is an Associate Professor at California State University Northridge and a Senior Research at the National Center for Research on Evaluation, Standards, and Student Testing.

*Photos courtesy respondents. Photo of blackboard courtesy pareeerica.