Tuesday, March 17, 2009

Teachers' Unions

In the comments on that previous post about Teachers' Unions, Katya raised a point I'd like to give further justice than just a reply in the comments:
Do you think there's anything about teachers' unions that makes them particularly bad? Maybe the quality of a teacher is harder to measure . . . or it's easier to blame shortcomings on the students.

There are, of course, some unique challenges with determining the quality of a teacher. Their success will have some correlation with the success of the students they teach, however exactly what, or how strong, that correlation is would be difficult to determine. What bugs me is the Union's stance against merit based incentives of any kind. Their opinion seems to be "There's no perfect way of doing it, so we won't allow any merit based incentives". If you buy into this type of reasoning, then there's really no reason to ever do anything. You'll rarely, if ever, be able to do something perfectly. This is true of some very important things. Like voting: The Gibbard-Satterthwaite Theorem (in association with Arrow's Impossibility Theorem) proves that no voting system is fair. So, by the logic of the Union, we shouldn't bother with voting either, because we can't do it perfectly.

Obviously, such an approach to life is rather limiting. In pretty much every aspect of our lives we accept solutions which are less than optimal simply because perfect solutions are either impossible to find or the necessary effort required to find them outweighs the usefulness of having them.

But, how do you measure the quality of a teacher? Any such measurement will need to include a subjective evaluation from students, other teachers, and administrators. So you'll need a way of dealing with the subjectivity of these evaluations and prevent them from turning into political games and popularity contests. I'm not exactly sure how you go about doing this. I'm sure there are lots of people that work in subjective survey taking that have some good ideas on the matter though. All such evaluations would be anonymous, of course, to prevent retaliation and gaming the system.

Part of this, I think, could be to include "down-the-road" evaluations. Have students fill out a very simple evaluation of how good they think the teacher was a year after having had them. (In very controlled systems like public education I wouldn't think this would be too hard to organize). Then you can get some larger picture perspective from students, who otherwise might be bitter about a final grade or something.

You can't grade performance simply on students' grades. That would further increase the amount of GPA bloat already existing. Creating an outside exam creates the same problem we already have with standardized testing: teaching to the test. But you can include that as some part of the overall metric.

The most important thing, in my opinion, is to be sure that merit based incentives take a longer window than 1 year into account. A teacher may have a bad year, a bad group of students, or some other issue that comes up. However, if you take a 5 year span and see that student/peer/supervisor evaluations are consistently low, and student performance is consistently low, then you can probably start making a pretty good guess about the overall quality of that teacher. During those 5 years they can be encouraged to improve their teaching. If, however, they consistently show, during a 5 year span, that they just aren't cutting it, it's time to cut them loose.

This gives teachers plenty of time to get their act together, have a bad year, deal with a bad group of students, whatever, while still allowing school systems to reward teachers who consistently receive good evaluations.

If we really want to improve public education in this country the best thing we can do is remove the cruft from the teaching staff. I think we can all look back on our public education and remember teachers who were excellent as well as teachers that just sucked. Knowing that those excellent teachers will not be rewarded for their quality is frustrating and saddening.

3 comments:

Katya said...

Wow, I got a whole blog post for me! You might also be interested in Malcolm Gladwell's article here on one potential model of merit-based pay. Interesting stuff, and your first post reminded me of it.

Unknown said...

That's a pretty interesting article, I like Malcolm Gladwell. Reminds me of part of Blink, the study that showed that something like when students watched a 30 second clip without sound of a teacher and then filled out an evaluation their evaluations were, overall, the same as student evaluations for that teacher after an entire semester.

Katya said...

I need to read Blink. I keep meaning to check it out but it's always on hold and I'm too cheap to actually buy it. ;)