A reader emails: “I keep checking eduwonk to get your take on the LA dustup (that I think marks a real turning point in teacher accountability), but all I see is a dead fish!” Indeed. Just getting back.
Big thanks to Jim Ryan, Sarah Usdin & Neerav Kingsland, Becca Bracey, Hae-Sin Thomas, Tim Daly, Victor Reinoso, Paul Herdman, and Terry Ryan for keeping it interesting. If you haven’t taken a look check out their stuff below.
If you left a comment that didn’t appear on the blog, more than a few got caught up in the spam filter “awaiting moderation” while I was away, they are all up now.
I, too, think that looks like a trigger fish.
LA Story
While important, I found the LA Times teacher value-added story itself less interesting than the reaction to it — on all sides. Everyone seemed to rush to position themselves relative to the story (‘Yeah, tough love at last! Get ’em!’ ‘I’m thoughtful and measured so though you’d think I would like this here’s why I don’t!’ ‘The LA Times is in bed with the billionaires, to the barricades!’) The big takeaway here, it seems to me and as the reader above notes, is that the teacher data genie is out of the bottle and though there will be a lot of debate, contention, and politics, it’s not going back in.
My only disagreement with what the paper did is the part about naming names of specific teachers. I think that crosses a line and I don’t see how the public interest would be less served by simply describing the teachers in more general but not personally identifiable terms. The same pressure would be brought to bear on the district to address the issues absent the public humiliation, which I’m not sure serves any point at all. Otherwise, this is fair game. Teachers do work in the public sector. And for their shortcomings value-added is one of the best tools at hand right now in a field starved for performance tools.
But, before everyone puts this all on the union, the district deserves a lot of blame for this situation, too.
More generally, it seems the common trend in our field is that until data are shown in painstaking detail about controversial issues nothing changes.* That’s a lack of leadership. Everyone knew that the money flows within school districts to Title I schools were unfair to poor kids. But there was no movement (and still not enough) until this was consistently shown in great detail. You’re going to see the same thing on the teacher quality debate. So, if the teachers’ unions are so concerned about stories like this they could do more to get off the dime on the teacher effectiveness issue more generally rather than waiting to be wrestled to the ground.
On the substance, the article didn’t find a lot that counters the general knowledge base among those who closely follow the issue. The one finding that caught my eye is the paper’s claim that low-performing teachers are not more concentrated in high-poverty schools in LA. I hope they unpack that more in future stories because it does run counter to other data. For instance in an analysis of value-added results in Title I schools by Jane Hannaway and some colleagues found that those schools disproportionately had less effective teachers as measured by value-add results. Is that a function of something unique to LA and how teachers are assigned there, the geography of the district, the overall composition of the schools? I hope they tell us!
As for the teachers’ union boycott, this tactic hasn’t worked well for them in the recent past. The bark is worse than the bite. Ask Wal-Mart. Seems ill-considered as a response. But then again, the LAT isn’t Wal-Mart! Stay tuned.
A few great edujobs
The Education Equality Project is seeking an executive director. This is a tremendous opportunity if you want to be in the heart of national reform advocacy in a leadership role. And the Boston Teacher Residency is seeking people for a few roles, in particular a Director of Student Learning. As you might infer, those roles are based in Boston.
And Bellwether’s full-time team of seven is expanding in response to demand and growth. A few opportunities available. If you’re committed to dramatic educational change on behalf of currently under-served students and thrive in a fast-paced entrepreneurial environment where you’re supported but also encouraged to act independently and grow professionally and personally, please check us out. And if your life circumstances mean you’re more interested in part-time work we have some need for that, as well. Email us with your information.
And NACSA, on whose policy advisory board I sit, has some great grant competitions going on for charter school authorizers right now. Resources to support authorizing work.
*See, for instance, this new CRPE report on math/science teachers and salary in Washington State.
The VAM (value added model) scores are just that: estimates from one statistical model. Different models fit the data differently. But if some measure of overall model fit from a pair of models is similar, the distribution of errors across data points can be different. The estimates of value added for particular teachers will be different then, depending on the model chosen.
A simple example. A cloud of X-Y points through which two lines, one a mild parabola, and one straight line are fit. Say there is no difference in the fit: the sums of squares or any other chosen loss function of difference between predicted vs real score y given x is similar, even identical. But, plainly the predicted y is different under the two models, and teachers will be judged differently under the two models.
Don’t be fooled, as other quants have been: over all classrooms, the variation in student outcomes –and typically only one or two short term outcomes are measured –is hardly do to teachers, most of the variation occuring within classrooms.
One more nasty thing: For many of the VAM efforts, there is no way to differentiate the “effect” of a teacher on classroom performance from the “effect” of particular students or constallations of students. It is just because peer student effects can be so large — usually seen negatively so — that schools increase overall happiness and learning by removing selected students from classrooms, and even expelling them from schools. Unless those students are their own kids, economists estimating the VAMs wouldn’t know about them.
If you and other reformers think that naming names crossed the line, please communicate with Duncan on this. We need to fire bad teachers (although I see VAMs as too dangeous and unnecessary for that task) but the way to get there requires us to respect some lines of common decency that should never be crossed.
Like “incredulous” (above), I agree that VAM’s should be treated with caution. As incredulous noted, there is no good way to isolate the effect of a teacher vs the effect of other variables on student performance (for example, the effect of their peers). As a former teacher, I believe that one or two very disruptive students will make it much more difficult for a teacher to teach and, presumably, such students affect the learning (and test scores) of their classmates. All I can say is that I’m glad I decided to leave teaching before this new “blame the teacher” craze took off…
The Los Angeles Times article saved me some money. First, I cancelled my subscription and vowed to read various newspapers at the nearby library. Next, when someone from the Democratic Party called (again) for money, I was able to give an emphatic No without my usual sense of guilt. I will not give another dime until this administration supports significant changes for our poorest children. For me this means medical care, infant and toddler monitoring, parent education and high-quality preschool for our youngest children. This is where the achievement gap starts and that’s where we must begin to address it.
I do agree with you that the teacher data genie is out of the bottle so teachers need to insist that these tests (1) are designed to measure teacher effectiveness and (2) are administered with the same security as other high-stakes tests, such as the SAT.
I agree with all of the commenters but I just want to add that an apology from Duncan would be huge for me. Perhaps I’m being naive, but I’ll twist myself into a pretzel to keep supporting Obama. They have given us a lot of money, but these blame the teacher insults are going to cost Dems in the fall. Perhaps I’m naive, but I’m hoping that this latest toxic stunt will show that the scapegoating has gone to far.
We all seem to agree today!
Insulting our teachers and subjecting them to public humiliation will not help children. That’s just common sense.
Another example of some really poor value-added work. And this had high-stakes attached to it! Wouldn’t you think someone at the state dept of ed, testing agency, or the people doing the value-added work should have asked about the equating of the tests across grades BEFORE implementing value-added?