Seventy-five comments into a thread, Dan Meyer asks a really important question:

I got my 07-08 Geometry results back yesterday and they were not acceptable. Too many kids listing along at Basic levels, not enough kids rising to Proficiency. My question to so many commenters here: what would you have me do with that data?

As a principal who is both against standardized assessments and also very much measured by them, here’s what I’d do:

First, let’s work under the assumption that I’ve watched you teach, and I feel that you are a good teacher.

1. I would take the scores and compare them to grades. I believe that the multiple data points that go into making up a grade give us a richer sense a student’s learning. So the first question is this — Is there a correlation between student grades and scores? If there is — or if there isn’t — what does that tell us?
2. The next thing I’d do is ask you for your assessment: Most importantly, what is your assessment of how the students learned Geometry? How does that line up with what the scores suggest? What surprised you? What was what you expected?
3. If — as I would think — you were surprised by the scores and you honestly feel like there was deeper learning than the scores suggest, then the question is this: Is there a disconnect between the way you’re teaching the skills or the process or just the language and the way the state test assesses the learning? This raises several more questions:
   1. What are the assessments that you did in your classroom that would lead us to believe that the learning was more successful than the tests suggest?
   2. If we believe that your methods are successful, what do we do about the tests? Given that they are the coin of the realm, we cannot ignore them, so are there modifications we need to make? Can we make them without harming the learning you see going on?
   3. There is something going on in your class if your sense — based on the work you see every day — is that the scores really are not reflective of what they’ve learned. Is there something going on with the multiple opportunity style of assessment that you’re doing such that on a one-shot test the kids aren’t able to replicate their learning? Do you need to just take two weeks before the test to do some explicit teaching on how the way they’ve learned can translate to a test? Do you need to give them more opportunities during the year to take tests that mimic the structure of the state test?

Much of test taking is about the skill of making sure your knowledge and skills translate well on the test. The hard part, I really believe, is making sure that the learning you see every day in your class is measured on the tests, especially if you don’t teach in a pedagogical fashion that is in line with the state assessment. And I really do believe it’s important to tell us that the tests tell us something, but they don’t come close to telling us everything.

Anyway, that’s what I’d do.

[Update: Gary Stager takes a whack at answering Dan as well, and his answer is brilliant and paradigm-shifting and highlights why he’s one of the best thinkers we’ve got.]

Tags: danmeyer, high-stakes tests, assessment