Thursday, November 06, 2008

A better way to debate

by Andrew Gillen

I recently stumbled across a really fascinating argument occurring between the Boston and Minnesota Federal Reserve Banks illustrating the potential power of decentralization. Researchers at the Minnesota Fed came out with a piece showing that the credit crisis was being overblown and was largely confined to the financial sector. They showed that typical businesses were still capable of finding financing.

But then researchers at the Boston Fed came out with a piece arguing that the Minnesota Fed was wrong, the credit crunch was a big problem and the Minnesota folks were missing the real story because they focused on the wrong things.

I have absolutely no idea who is right, but I love the fact that they are debating each other (with evidence no less!). What is fascinating to me is that they are both part of the same organization, yet there is no effort to suppress either view.

This got me thinking… why couldn’t the Department of Education be more like the Fed? There seems to be quite a bit of disagreement on many aspects of education, so why not let each school of thought run their own experiment and then compare the results? This might actually provide evidence with which to argue, rather than base everything on untested theories and faith.

To take but one example, those who believe
there is nothing wrong in hypothesizing that data-driven accountability could add more than it subtracts to education. But we have not had a real debate on that hypothesis

are right in one respect. There hasn’t been a real debate, mostly because there isn’t enough variation and data to conclusively prove one side or the other.

But they are wrong (in my view) that the benefits of data driven accountability are outweighed by the costs. There are those of us who've noticed that
When [schools] start to be evaluated based on student outcomes, student outcomes tend to improve.

It’d be great if those who think such efforts are a waste could go on doing whatever it was they hypothesize will educate children, while those who think that accountability will help could implement such programs.

While such decentralized experiments, with each system based on the assumptions and theories of those running the schools, would put some kids in inferior systems, at least in time we would be able to figure out which systems, theories and assumptions are truly inferior and abandon them. Such an outcome is vastly superior to the one where we constantly make everyone jump from one unconfirmed theory to another, never having evidence to determine if it helps or not.


capeman said...

One way for those who believe that

"When [schools] start to be evaluated based on student outcomes, student outcomes tend to improve"

to be taken seriously in the academic world is to publish their papers in peer-reviewed journals, rather than as glossy output from the agenda-driven think tanks. Maybe the gloss will influence some policy makers, maybe not. My guess is the conservative think tanks are going to have a much harder time, at least in Washington D.C.

~ said...

Actually, this view is not confined to conservative agenda-driven think tanks. The quote is pulled from a report, College Rankings Reformed, by Education Sector, which leans left. While in the past, this issue was not a cause of left, almost getting Andrew Rotherham expelled from polite circles after he wrote
Toward Performance-Based Federal Education Funding, for the left's Progressive Policy Institute, I would say the idea now spans ideological divides in the sense that you can find lots of people on all sides who think it is a good idea.