January 9, 2007

BCS and NCLB

Right now, the teachers who work at Boise Senior High School have to be some of most anti-formula individuals in the whole country. First, they're dinged on AYP because the school missed Idaho's targets for one population group (students with disabilities). Now, their beloved local Boise State Broncos are the only undefeated Division I-A football team in the country, but they're not the recognized national champion because the BCS formula gave the University of Florida (and several other BCS-conference teams) a statistical advantage, leaving the University of Florida as Ohio State's matchup in the national championship game last night. Florida trounced Ohio State, and Boise State ended up ranked 5th or 6th (depending on the poll).

So what's the connection with NCLB?


Here's the money quote from Dan Wetzel:
There is no way, no formula, no mix of opinion polls and computers that ever consistently can select the top two teams... No matter how hard it tries, this championship system turns up paper contenders as often as not (Oklahoma 2004, 2003; Nebraska 2001).

As every college football fan from 5 to 95 knows, the BCS formula is not only rigged for certain conferences but inherently arbitrary. The defenders of the system pretend that having a statistical formula means that the system is objective, and it certainly puts out a number. But the existence of a set of numbers does not mean that the formula is the best way to resolve a national championship in a fair, independent manner.

That fact is why most sane individuals—those who aren't tied to the most powerful conferences—favor a playoff, because in sports the way you have an independent judgment of which team is better is to play the game. Only in chess tournaments and Division I-A football do you decide who is the best by asking your TI-93.

One of the problems with formula-based accountability systems is the conflation of statistics with independence from conflict of interest. There is a tremendous need for mechanisms to hold schools accountable in ways that systems would not be unless someone is looking over their shoulders. That's a requirement for accountability independent of the political and other interests of school systems as organizations and collections of people.

Currently, high-stakes accountability addresses the need for independence through formulae: insert numbers, retrieve judgment. The argument in favor of such an approach is that it removes the inherent conflict of interest in having educators judge their own work.

The problem with this argument is that it assumes that statistics are the only way to fashion an independent judgment of school effectiveness. In other walks of life, though, we don't require statistics to remove the conflict of interest from judgments. In sports, you play a game, and the referee or umpire is the neutral dispute-resolution mechanism. (Unless you're into Fantasy Baseball, but I'm not talking about cults here.) In law, you go to court, and a hearing officer, arbitrator, or judge makes a judgment.

The key word here is judgment, that sometimes ineffable quality that allows humans to synthesize information and make a decision. In sports and courts, statistics are tools but not trump cards. Baseball managers have a wealth of statistics at their command, but I will take the late Billy Martin's judgment over any crude number-cruncher today. Judges will hear testimony from experts wielding tables and graphs, but the decisions tumble from their computers as a stream of words, not charts.

I have not only a history degree but a masters in demography, and I am not denigrating the value of well-crafted measures (not that most high-stakes test statistics deserve that description). But statistics cannot replace thought, and I am afraid that we have seen that in school accountability policies.

At least football is just a game. But No Child Left Behind's adequate-yearly-progress standard, a brain-dead mechanism that analysts knew was a problem in 2001? That's federal law.


What is it about the beginning of my semester that a Kappan article by Rick "Tough Love" Hess and Andy "Eduwonk" Rotherham is dangling in front of me while I have a gazillion things to do? World, will you stop putting these temptations in my path?

Listen to this article
Posted in Accountability Frankenstein on January 9, 2007 5:39 PM |