Counterparties

By Felix Salmon
March 8, 2011
WSJ

" data-share-img="" data-share="twitter,facebook,linkedin,reddit,google" data-share-count="true">

Measuring teacher performance is really hard. But who cares about getting tenure if you’re thinking of leaving anyway? — NYT

With Sir Michael Gambon, as Blenheim Palace, and a brief but scene-stealing turn from Dame Judi Dench, as a wingback chair with cabriole legs — TNY

Kinsley on movie math — LAT, see also me in 2008

Open City, Closed: Acclaimed Literary Journal Says Goodbye — NYO

The TechCrunch experience with Facebook Comments — TC

The best part of Newsweek’s interview with Larry Summers is the photo caption — Newsweek

SCOTUS eliminates a FOIA exemption — HuffPo

Emily Badger’s great article on Slugging, the People’s Transit in DC — Miller-McCune

Since when does the WSJ consider a 4-digit sum a “Big Payday”? When it’s “for Some Hill Staffers” of course — WSJ

More From Felix Salmon
Post Felix
The Piketty pessimist
The most expensive lottery ticket in the world
The problems of HFT, Joe Stiglitz edition
Private equity math, Nuveen edition
Five explanations for Greece’s bond yield
Comments
6 comments so far

Anybody can construct a complicated model, but how do you demonstrate that model reflects teacher quality? Even if you show a strong correlation in the year-to-year results (and the article did not say that had ever been done), that correlation could equally be attributed to the courses or the students involved.

I know there are annual job evaluations in business as well, but in my experience they NEVER result in talented, hard-working individuals being fired. If you construct a teacher evaluation model that is flawed, you will strongly encourage your best teachers to leave for business — not so they can earn twice as much but so they can have job security.

Posted by TFF | Report as abusive

Do they control for the average hormonal imbalance of each student in her class? Average divorce rate?

Trying to not only model an average student, but create a predictor of the student’s school performance sounds like a tough stat problem.

Plus, with only roughly 30 sample sizes to draw from, you’re going get large errors, like the article mentioned.

It seems like the best approach is to simply collect quarterly evaluations from every source possible – Principal, Students, Student’s Parents – and use that as a training tool at the very least, not a pure evaluation tool. That type of feedback seems few and unformalized based on teacher’s I know.

Performance evals in my line of business have all sorts of data around profitability and growth, yet borderline on useless as well.

Posted by djiddish98 | Report as abusive

I know there are annual job evaluations in business as well, but in my experience they NEVER result in talented, hard-working individuals being fired.

In my experience they routinely do, and for good reason. Just because you are talented and hard-working doesn’t mean you are a good fit for a particular job or organization.

Posted by right | Report as abusive

Using statistical models to determine if a teacher is any good is like using economic models to determine how people will react…

I really thought that the equation was a joke… silly me, but in this world where statisticians and Quants are revered (even though many chose those jobs because they lack empathy and humanity) we can expect no less. If we can’t understand the equation and make it impossibly complicated, how can we trust the results?

If that is how they evaluate teachers then they are devaluing teachers! That such a model could ever be used to determine the quality of a teacher means no one had the guts prior to this fiasco to say the emperor has no clothes. Shame on the system!

Posted by hsvkitty | Report as abusive

“Just because you are talented and hard-working doesn’t mean you are a good fit for a particular job or organization.”

That is true… I was thinking more of individuals who ARE doing a good job, and keeping their nose clean. Not misfit situations.

Posted by TFF | Report as abusive

hsvkitty, most politics these days revolves around “sound bites”. People believe what they want to believe and don’t bother to see whether or not the emperor has any clothes.

This is the biggest reason to depoliticize education funding to the greatest extent possible.

Posted by TFF | Report as abusive
Post Your Comment

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see http://blogs.reuters.com/fulldisclosure/2010/09/27/toward-a-more-thoughtful-conversation-on-stories/