Thursday, March 19, 2015

Yelp for Teachers?

At Yahoo News (yes, that's a thing) Matt Bai, who usually blogs about politics, has a piece arguing for more transparent metrics for teachers, a "Yelp for teachers.  Let's take a look.

He spins off from a suit by a Virginia parent to force the state to release "the ratings that public school teachers get based on the test scores of their students."

Bai considers the arguments, and does so in a fairly even-handed manner. In fact. he uses what I'd call a pretty good baseball analogy for why the test-based VAM scores are not exactly the best possible way to measure teachers.

Did Derek Jeter give up more runs, statistically speaking, than the average shortstop? Yes. Did you want Jeter as your shortstop anyway? Absolutely. A public school teacher who ignites the imagination (and most of us had someone like that) might just be worth more than the one whose students have demonstrated mastery of the Pythagorean theorem.

But Bai is not so much interested in techy test-based teacher tallies and how to best share them with the world. He's more interested in the matter of customer-based ratings. If and Yelp can use customers ratings in a transparent manner, why not do the same for teachers.

Ask the parents at any bus stop which teachers have lost their energy for the job or can’t control their tempers, and you’ll find out pretty quickly that they know better than anyone else. But there’s no mechanism for parents to pool that knowledge or to make the school system respond to it.

 There are a few issues with this. Some Bai recognizes; some, not so much.

First, the platform that he wants to see has been created, and more than once. For instance, check out, which has been around for over a decade. My school is there, with thirteen teachers rated (including one who doesn't actually teach here); I have six ratings myself, placed in the system between 2004 and 2007.

And there's your next problem-- how do you get enough people to use the platform to get meaningful results? Those six ratings (assuming of course that they were written by students that I actually taught) represent approximately .5% of my customer base. Is that enough data for my employer to make decisions about my job performance, or for me to make tweaks in my teaching style?

How would we get all the parents of my district to log in and evaluate teachers, and would we get data that was useful? Voluntary participation would insure that the few who left comments would really mean what they said, but only teachers who evoked particularly strong feelings would elicit comments-- twenty-five sets of parents from my class might say nothing because they think I do a fine regular old vanilla job, but the twenty-sixth parent, who's angry about how Little Chris flunked for never doing work, might blow me up. On the other hand, if I somehow encourage parents to just fill it out, go ahead, say anything, just do it-- they may well take to heart the "say anything" implied in that sort of set-up.

I'm actually a fan of feedback. I have course and teacher evaluation forms of my own that my students fill out anonymously at the end of the year, and on more than one occasion it has changed how I do things. But there's something missing in the techy on-line customer comments model of teacher feedback, and while I'm not sure what it is, the very fact that it doesn't exist is our biggest evidence that nobody yet knows what it should look like. After all, we've made it possible for farmers to find their soulmates. Heck-- if this kind of web-based eval was really desired by parents, it would already exist as a bunch of facebook pages.

There are other problems with the customer feedback model for education. For instance, the quality of the "product" is often not evidenced for years. Every high school teacher has stories of the kid who comes back years later to say, "Boy, at the time, I hated you and I hated your class. But I want to thank you because it turned out you were right to push me." Also, parents are not the only customers-- all taxpayers and business owners and community members are also customers of public ed.

Maybe this kind of transparent open-source feedback would be useful. After all, how often are comments and rating features of websites commandeered by trolls and cranks? Well, Bia's comment section is up to 1,500 entries, and I can tell you they're not all winners. Would they be a good evaluation tool for Bia's skill as a writer?

His last line is the final stake in this idea's chest:

Shouldn’t we teach to the parents at least as much as we teach to the test?

I understand his point (isn't parent feedback at least as useful as BS Test results soaked in VAM sauce), but the correct answer is "We should not teach to either. We should teach students."

1 comment:

  1. Very thoughtful, complete analysis. Good points. (And makes me realize good writing is often dependent on good thinking - which takes time, experience, and practice.)