Comment by vitaminj | Hacker News Reader

Comment by vitaminj | Hacker News Reader

The title of this story is a bit misleading. From the Urban Institute website, the Teach for America teachers were "high-achieving college graduates ... [who in North Carolina] tended to have graduated from more selective colleges and universities". So they're essentially comparing the cream of the crop college graduates with a bunch of experienced (but presumably average) teachers. For the claim in the title to be really tested, high achievers with teaching experience need to be evaluated against high achievers with no experience.

Notwithstanding, I suppose the interesting conclusion is that intelligence and talent beats out experience (on average) in teaching. Ultimately the ideal situation would be to have teachers with intelligence AND experience.

That said however, I personally feel that using test scores as the metric for evaluating teachers is a bit dubious.

I just used the title of the original blog post.

What other metric could be used for evaluating teachers?

Derivative of test scores.

A good teacher should make your scores improve. The absolute numbers are much less meaningful.

No, it really depends on the teacher. My precalculus teacher (probably the best math teacher I've ever had) gives progressively harder tests/quizzes as we learn more about each topic. He ignores the textbook and writes the tests himself.

(Incidentally, we've recently been covering derivatives.)

What do we want to measure in students? Skills, creativity, "intelligence"...?

Standardized metrics are themselves dubious when we are dealing with highly chaotic systems (see: combinatorial explosion, etc) with drastically different properties.

I don't know, nor do I think a good metric necessarily exists. Teaching doesn't fit in neatly with the typical production metrics we use in industry, eg. widgets per hour. Those outputs are directly measurable.

But what is the unit of output for a teacher? Test scores could only hope to measure teaching quality indirectly. It'd be like using record sales to assess the quality of a musician... and we'd all agree that Britney and Mariah aren't the best musicians going around.

Reply on news.ycombinator.com