By Kristin
Like many other Washington parents, in the fall I checked my daughter's MAP scores. The test, the Measurement of Academic Progress, is Washington's best bet for measuring teacher impact because it's given in the fall, the winter, and the spring.
At least, a few weeks ago I thought it was the best tool. Now I'm not so sure.
In math, my daughter scored in the 68th percentile in the fall. I got a little nervous. My own elementary school spent more time on bible verses than numbers, and I entered junior high far behind my peers. I never caught up, and while I've never said "I have a math phobia" or "I'm not a math person," I'm not very good at math and I've always felt I was missing the party. I didn't want that for my daughter.
So, I bought little workbooks. We do her homework every night, plus workbooks. My husband is an elementary teacher and he sits with her and coaches her through math when she gets frustrated. When we're in the car, we play a game I call, "Geography, spelling, or math?" and the girls often pick math. Luckily, since they're only four and seven, I can stay ahead of them. I actually started that game because I was so sick of them trying to kill each other in the back seat, but it fit nicely in my "I don't want you to struggle in math!" campaign.
But it looks like it won't be long before my seven year old will be quizzing me on math, because her score jumped from the 68th percentile to the 98th percentile in three months. Did she just have a bad day in September, and all our work has been unnecessary? Did her parents improve her skills? Did her teacher, who is practically perfect in every way? Did the classroom aide, the well-loved Ms. Kathryn, who pulls the first graders out to do math? Who is responsible for that leap, and can we use that leap to say my daughter's teacher is better than another teacher, whose students went from 68% to 69%, or backwards?
A teacher's evaluation can't be based solely on test scores – classroom observations are essential. And you don't need privileged students to see leaps in test scores (although, obviously, that helps) - I've seen growth in my students who slept on the floor of any house they could find. There's lots of data to support that good teaching results in big leaps between September and June – we don't need to be afraid of teaching disadvantaged kids, we just have to teach really well.
If we keep pushing for a teacher's evaluation to be based mostly on observations, if we keep pushing for test data to be triangulated with demographics and attendance, and if we keep pushing for September-June assessments, I don't think we need to worry about the muddy area a privileged child's 20 percent jump presents.
If we stonewall, and refuse to be proactive in developing and supporting a powerful classroom instruction evaluation tool, if we stop pushing to get administrators some help in that huge task, we may see test scores become a bigger part of the equation because they are easy and cheap to use. Then, we're in trouble, and we know what will happen – a lot of good teachers will want to teach kids like my daughter, who has every advantage. And that would be a shame, because it's not kids like my daughter who need the best teachers.
In my opinion, the best way to use student test data as a tool for teacher evaluation is to look at how teachers actually USE the data.
That, of course, would mean a stronger emphasis on the creation and use of formative assessment, and the process by which a teacher uses that data to inform subsequent instruction.
My dad used to quiz me on math facts in the car. He was an electrical engineer. I turned out to be pretty good at math, and I think it was partly from practice, partly from genetics, partly from parental expectations, and partly from teachers, not all of whom were good. But my students are expected to be good because of me, period.
Kristin, you’re right: I get the credit when a students excels for all of the reasons above (and parents like you). And now I get the blame when my students have none of the above going for them.
I agree that we need “a powerful classroom instruction evaluation tool”, but I reject, adamantly, the notion that we can put a meaningful number to it.
Thanks David, I agree.
I think test data can show trends and patterns, but it shouldn’t be used as a golden ticket that launches one teacher into the top rating, nor should it be used to fire a teacher.
Now that there’s such a push to truly evaluate a teacher’s effectiveness – a push I support – we have to make sure there are accurate ways to do that, fairly and effectively. I wouldn’t say ignore test data, but even September-January-June data is impacted by the student and the parent as much as the teacher.
Right on, Kristin! Great example of why we need to hear more stories from school, and home, (and the car!). Barnett Berry just posted a blog about a Harvard study of VAM in math instruction, and the data included careful observations and additional data gathered about the teachers. Turns out that those with high VAM ratings sometimes were low in other ratings. The blog suggests that we should therefore proceed with caution re: VAM, while from my point of view, it undercuts VAM. If the teacher who makes more mistakes and uses weaker strategies has the higher VAM, why do we bother with it all? That study used math teachers – and I think all of this makes even LESS sense when we talk about reading and language arts.