Teacher evaluation, aka accountability, continues to be a topic of wideranging debate. On the one hand, we have lots of folks who call for teacher "accountability." On the other hand, Race to the Top and the state waiver programs gave us systems of teacher evaluation that are spectacularly dysfunctional and conceptually stupid (how well some eighth graders do on a single, bad reading test should determine how good the shop teacher is?). And on the third hand, the new education law (ESSA) gives each state a chance to come up with new ways to make a hash of the whole business.
Critics (and I'm one of them) have said repeatedly that value-added measures and test-based ratings and a few other stupid things that have been tried are, in fact, stupid, destructive and bad for everybody. Supporters of the accountability movement have replied, "Fine then. What do you want to do instead?"
Okay, then. How do we get on the path to a useful method of teacher evaluation? Step one is to figure out what purpose the evaluation will serve. This may seem obvious. It isn't. Here are some of the goals that a teacher evaluation system might try to meet.
To find bad teachers. For a while, this was the focus-- we would find all the Bad Teachers and fire them, and then life would be swell. This remains the focus of attacks on tenure and seniority; we plan to cut your budget to the point that you have to fire people, and we want to be sure you fire the right ones. The bad ones.
To find good teachers. Let's locate all the super-duper teachers so we can move them to where they're needed, or reward them with generous quality bonuses.
To guide and support teachers. A new favorite. Let's figure out teachers' strengths and weaknesses so that we can help them grow and develop.
To compare teachers. We want an evaluation system that helps us stack rank teachers so that we can do comparisons across a building, district, state, or even nationally. We need measures that let us compare every teacher to every other teachers.
To let the taxpayers know whether or not they're getting their money's worth. Taxpayers spend a lot of money paying teachers; they are entitled to know if they are getting a good deal or if they would be better off hiring a minimum wage worker to push "play" on Khan Academy videos.
To give teachers a clear set of expectations. This is not discussed nearly enough. While some jobs come with clearly delineated descriptions of duties and responsibilities, teacher are more likely to get directives along the lines of, "Go in there and do teachy things, and do them real good." An evaluation system makes job expectations explicit, whether you want it to or not. That's one reason the test-linked systems suck-- they send the message that a teacher's job description is "Get students to score well on the Big Standardized Test."
To make the complex look simple. Trying to understand and make decisions about complex human behavior is hard. If we could just boil it all down to some simple digits and data points, it would be easy, and then we wouldn't have to do hard thinking about hard things. But if we could just take human judgment out of the whole hard business and simplify it to an algorithm, boy it would just be the berries. Which is both the most popular and the dumbest idea on the list. Fans of this idea are welcome to sign up at match.com and let the dating site pick their spouses for them. If the marriage results in children, they can also go ahead and use data to decide which child is the best one. Human beings are complex. Human relationships and interactions are complex. Lord knows we all have days when we wish it weren't so, but hey-- I have days when I wish I weren't so bald or that Donald Trump weren't a Presidential candidate. You can't build a strong system on a foundation of denial.
And none of this addresses the most important element of all, which is hiring practices. W. Edwards Deming used to refer to an insight from his own mentor--
if you have dead wood in your organization, either you hired dead wood,
or you hired live wood and killed it. Either way, you have a management
problem that directly impacts the quality of employee work.
Now, you may say, "Well, those all seem like swell ideas." You'd be wrong in some cases, but still-- you have to pick. You cannot have one system that does them all. It's not possible. Evaluating for the top and bottom qualities in teaching-- that's two entirely different searches. In fact, it's two entirely different types of searches, the difference between looking for something to include, and looking for something to exclude. And both of them are different from a system that would help a teacher develop by examining personal strengths and weaknesses from the standpoint of improving their craft. It's the difference between walking into a restaurant and saying "Don't bring me anything green," or "My dinner has to include a steak" and "Let's look at this whole meal and see what we could do to bring its elements more into balance."
Letting taxpayers know they got what they paid for is a laudable goal, but it leads us directly away from any sort of standardized measure that lets us compare teachers across the nation. Instead, being accountable to local taxpayers would start with asking them what they want for their money, and that is likely to vary from community to community. I think, however, it's the foundation of a real evaluation system.
I do think teacher evaluation is possible. I also think almost everything we're currently doing in the name of accountability is dead wrong. It can't be fixed-- what's needed is to start over, and to start by talking about the question of purpose. We can't get anything done until we do that.