One of the things that I find incredibly frustrating about the current discussion on teacher evaluation is that it’s almost entirely focused on adults, rather than kids. Obviously, “putting adult interests ahead of kids” is a complaint you hear a lot in education reform conversations, and there’s clearly an element to that here: In the debate over teacher evaluation systems, there’s a tremendous emphasis on whether these systems are fair to teachers: Are student learning gain measures are accurate and valid reflections of teachers’ impact? Are observers unbiased and sufficiently trained? Are teachers identified as ineffective are given sufficient opportunity to improve? Is the system as a whole is sufficiently valid and reliable to serve as a grounds for dismissal decisions? And these are important questions--after all, we are talking about people’s livelihoods here. But in debating what’s fair to teachers, we shouldn’t ignore what’s fair to kids. Kids’ and teachers’ interests are very often aligned (unfairly dismissing good teachers would also be unfair to kids), but there are also trade-offs. The higher we set the bar for identifying ineffective teachers or taking corrective actions towards them, the more kids will be unfairly subjected to ineffective teaching.
But there’s a second level at which we focus on adults rather than kids here. Nearly all of our conversations around teacher evaluations focus on teachers as the unit of analysis--for example, many states have provisions designed to dismiss teachers who are rated ineffective for two years. What if we flip that and look at kids as the unit of analysis, using teacher evaluation data to track the quality of instruction to which children are exposed over time. We could say, for example, that no child should be assigned to two ineffective teachers in consecutive years. We could say that districts should make every effort to assign students who had a teacher rated “ineffective” or “needs improvement” in the past year to a teacher rated “effective” or “highly effective” this year. We could say that kids who are particularly struggling in math should be assigned to teachers who perform well on student learning measures in math. And so forth. From the current debate, you’d think that value-added data is only good for evaluating teachers, but, as Craig Jerald has written, it can provide a wealth of information that can be used to inform both instruction and smarter student assignment decisions for the individual child--and should be. Such conversations have the potential to dramatically shift the way we currently distribute and prioritize teaching talent to students in schools. Given the body of evidence that teachers impacts on student learning are cumulative--students with three effective teachers in a row will learn substantially more than those with three ineffective teachers--it seems insane not to look at how we use teacher evaluation data from the lens of the child, not just the teacher.