According to the New York Times, the National Education Association has an official policy on student evaluation outcomes in teacher evaluations: Use them, but only if they're good tests. The article does a good job of sampling the union's position and laying out some of the big picture of what this means. (Full disclosure: I am a proud member of a local affiliate of the NEA. While I'm not a union shill, the casual observer would be forgiven for thinking I was.) From my perspective, the union's new policy feels like sense, and it resonates with my thinking on the matter.
We need data to drive effective instruction. We need good data about educational outcomes for students, and we need good data for educational outcomes for teachers. (Someday we'll get to have the conversation about data for educational outcomes for administrators and school boards. But in the meantime, teachers are the proxy for all those levels of the education machine. And that's to put entirely aside the effects of poverty, parent responsibility, and all the other smoke screens we teachers like to release when people try to look at us too hard.) This need for data means that we're going to have to include student testing in decision-making at some level; if data-driven decision making is going to be used to improve education, we need good data.
There are two key phrases: the more immediate concern is "good data." Neither I nor any of my colleagues I've talked to about this (a pitifully small sample size; even if you count my former colleagues at my last posting, the total number of professional educators I've worked with adds up to less than 50, and the ones I've talked to add up to less than 15) trust any of the current assessments. The disconnect between the assessments and educational reality is simply too great, for a lot of reasons, and it isn't necessarily because the ACT is "too hard." It's not. The reality is that assessments sample such a small number of learning goals, and do so in such a cursory manner, that drawing meaningful conclusions is tough. Equally importantly, the current assessments are not really designed to evaluate the skill of instruction. The good standardized tests assess really big ideas--critical thinking, drawing conclusions from data, things like that. But they don't do it in a way that means a great deal. Colleges have been de-emphasizing their focus on standardized test scores as an admission requirement, even as everybody else ramps it up. The union's director of teacher quality makes a valid point: the Common Core curriculum might be a starting place for getting better standardized tests.
The second, long-term consideration about data is longer-term political. I said, "if we're going to use data to improve education". That wasn't a rhetorical flourish. As trained professionals, many of us see the value in data (even if we lack the know-how, the resources, or the time to do anything about it). But I for one don't trust the motivations of many of the people pushing "testing testing testing" as the new standard for teacher (and by extension public education) efficacy. Many of them have a track record of being distinctly anti-public-education. These are all the usual boogeymen for this blog: the Mackinac Foundation, private charter school management companies, people who think that a lack of officially-sanctioned prayer in school is de facto a reason to be against public education. For many of these organizations (cue the "straw man" arguments, and I see your point), data-driven instruction may simply be another weapon to attack public education with.
Conversely, critics of the NEA (and, in this particular instance, I don't count myself among them) could argue a similar thing: In name, the union has accepted testing as part of teacher evaluations, while in practice rejecting any existing tests. It will be at least a decade before the kind of tests the union wants are available, by which time this argument will likely be moot. Good practice will show that high-stakes testing doesn't produce notably better education outcomes, and opponents of public education will move on to some other angle of attack.
So what's to be done? Well, I can't do a whole lot about public policy right now. Thanks to a recent vote in the Michigan Senate, my local union has less power than ever to do anything about teacher evaluations. What I can do is this: use the (pretty crappy) data we have to make decisions about what's good for my students. Make sure my learning goals are crystal clear in my head, and make sure my formative assessments are as good as they can possibly be. Get the data that I need myself. Make sure my class is so hard, when someone throws a standardized test at my students, they don't think twice about knocking it out of the park. Make sure my support structures are so strong, that all of them get it. Ray, I think you have the right idea: What do you want students to know? How are you going to help them learn it? How are you going to know if they got it? What are you going to do about it? The key questions are the only ones that matter.
(Edited to fix spelling error in the title. Thanks, Jamie. Who's the English teacher around here again?)