Nate Silver has already changed the way people evaluate baseball players. During this election cycle, he also changed the conventional wisdom about whether that the outcome would be “too close to call.”
So I’m thinking it may be time for him to turn his talents to education.
Silver first came to national prominence with his statistical modeling in baseball. At a time when conventional wisdom suggested that players could only be evaluated on “heart and hustle,” he took a quantitative approach. He created his own statistical model, PECOTA (Player Empirical Comparison and Optimization Test Algorithm), which made predictions about player performance based on factors like their minor-league stats, weighted averages and regressions, which produced an estimate of a player’s true talent, and a career-path adjustment.
It was wildly unpopular. (Hall of Famer Joe Morgan called it “a joke.” It was also wildly accurate. (In 2007, he was criticized for suggesting that the White Sox, winners of the 2005 World Series, would have a record of 72-90. Until they turned in a record of . . . 72-90.)
Since 2008, Silver has turned his statistical modeling to election forecasts through his blog Fivethirtyeight.com. And he again demonstrated deadly accuracy (as of this writing, with Florida as yet uncalled, he is 49-for-49 in state predictions). In 2010, he predicted the Republicans would pick up 7 seats in the Senate. They picked up 6.
He’s again riling up the traditionalists, from David Brooks, who suggested he couldn’t predict human behavior with statistics, to Peggy Noonan who said . . . well, something about yard signs.
So now it may be time for Nate to take on the world of educational data. And just as he figured out a way to estimate how a baseball player’s minor league stats would predict major league performance, he can mine data sources to see what metrics most closely predict high performance in the classroom.
Acceptance of education data as a predictor is perhaps where baseball was in 2007. While Rick Hess wants to suggest that it was the Common Core that cost Tony Bennett his election, the woman who defeated him attributes her win to a different voter concern. “It really was a referendum on education policy, things like high-stakes testing,” she told the Huffington Post.
Today, no baseball manager would ever forego statistical information for a “heart and hustle” evaluation of a prospective ballplayer. And after Silver’s strong run, we’re more likely to see statistical evaluations of political races take on greater importance.
That we need data to make educational decisions should by now be a settled matter. The precise form that data should take is, obviously, still under development. But maybe Nate can help.
Photo Credits: Steve Garfield Tumbler and Craze Base