In a recent guest editorial in the Seattle Times, Washington State School Superintendent Randy Dorn spoke in favor of using student achievement tests on teacher evaluations. Basically his rationale boils down to two reasons:
1. The state’s NCLB waiver is at risk. The Department of Education granted us a waiver from the onerous requirements of NCLB, but takes a dim view of our teacher evaluation system’s provision that student test scores can be used for evaluative purposes, instead of mandating that they must.
2. Using student test scores will make teacher evaluations more consistent, since these are tests all students must take, as opposed to district-based tests, which vary from district to district.
Let me respond to his second reason first, since it’s the weakest. As we’ve reported time and again on this blog, a main argument against using student test scores is that they aren’t consistent. The fact is, only a small minority of teachers teach in “tested” grades or subjects. Consider my school, which has 34 certificated employees. These include four music teachers, one PE teacher, one librarian, six special education teachers and one counselor. We also have three kindergarten teachers, four first grade teachers, three second grade teachers, and three third grade teachers.
None of these people teach grades or subjects for which state achievement tests could be used for their evaluations.
We also have three fourth grade teachers, three fifth grade teachers and two sixth grade teachers. That’s only eight teachers. Eight out of 34 teachers – less than 24% – for whom state tests could be used. The rest of our faculty would have to use district or classroom based tests. Yet Mr. Dorn argues that using state tests would be more consistent? How?
On the other hand, his other argument – the risk of losing the waiver – does make sense. I have to assume that Randy Dorn, Governor Inslee, or both of them have asked Washington’s congressional delegation to press Department of Education officials about the true risk to Washington’s waiver. And the fact that Mr. Dorn is still arguing in favor of capitulating to the DOE’s demands means he doesn’t think they’re bluffing. Either that or he just doesn’t want to take the chance that they aren’t.
And that’s where he and I agree. Like Dorn, I’m not willing to gamble that much money ($38 million) for the sake of fairer evaluations for teachers like me. Put another way, I’m willing to use state achievement tests instead of more meaningful district or classroom based tests as part of my evaluation if it means ensuring our NCLB waiver.
Good question, Todd, and lawmakers gloss over this all the time. How can a test be used with consistency for teacher evaluations, when only 17% of us have access to it?
But like I said, the issue is moot since the Legislature voted it down.
Title I funds amount to 0.5% of the state’s overall education budget.
It looks like the issue is moot:
http://seattletimes.com/html/localnews/2022939390_teacherevalxml.html
Apparently the legislature is more willing to to a gamble than I am! We’ll have to see how this turns out; could be interesting!
One other thing, your own school’s percentage of 24% of teachers who test in tested subjects or grades is much higher than the state’s average overall which is about 17%. Leaving 83% to do what?
Great post.
Todd Miller
My bigger fear: they will consider student growth percentiles in teacher evaluation. These are horribly flawed. Besides the margin of error of +/- 20%, and the fact that the statisticians who designed the percentiles deliberately state that the data is not intended to imply causation (teacher-causes-growth), there is the fact that it is akin to a Vegas sleight of hand show using numbers.
I feel like that $38 million ought to be a gap our state budget could compensate for somehow. What percentage of the state’s overall funding is that? On one hand I strongly oppose the use of state tests, but on the other hand the law requires multiple measures and two points in time… The state tests cannot achieve the latter. That means that state tests would be one of multiple data points, the rest of which would ultimately need to be assessments similar to what the other 84% of teachers not in tested levels or disciplines are using. From that perspective, using state test data just becomes a time wasting hoop that doesn’t impact teachers OR kids. Even if state data “must” be used, how could it have any valid and defensible effect if it is just one data point (and probably the least valid, at that)?
You mean the state achievement tests that are still being developed and haven’t been validated?