About thirty years ago, a revolution in baseball analysis began. intelligent people, namely Bill James, began to look beyond the traditional numbers that had guided conventional thought in the game. The most important revolution (or at least the one that made Brad Pitt want to make a movie about) was in the recognition that the statistics that everyone thought were most important (stats like batting average and RBI) were much less significant than they had been considered previously. Of course, many of these analysts were ridiculed by traditionalists (as parodied in that same Brad Pitt movie). For much of the last 10 years, the narrative has been the “old school real baseball guys” (who believe in the traditional stats) versus the “new school eggheads” (who are looking at more advanced and/or peripheral stats). To put it in simple, sitcom-esque terms, it was the jocks versus the nerds.
However, what made this movement truly revolutionary was the emphasis on not only WHAT statistics mattered, but HOW analysts should go about gathering that information. Common statistical terms like sample size infiltrated general conversation about the game. In 2013, virtually every major league baseball team operates with a distinct eye towards not only “using” meaningful statistics, but operating with a defensible methodology as well.
Unfortunately, education is not there yet. And worse, the same logic that guided the “old school baseball guys” is winning the narrative war.
In the last 7 years or so, politicians and “education reformers” have taken up the cause of “data-driven” education. Their concepts are built around their professed desire for a more quantitative look at education and the development of real statistics that act as evidence of learning. Testing is the answer for these folks, as it provides data points to be tracked and analyzed for the purpose of determining what is being learned in a classroom and whether or not a teacher is doing his or her job. These are the statistics of education reformers.
And they’re the wrong statistics.
A standardized test IS a data point; however, it is a data point that, for the purpose of understanding learning, is woefully inadequate. First, it is a data point seeking an analyst–it can’t possibly prove what ed reformers claim it proves. Compare it with baseball: baseball is a series of binary events that can be evaluated. A pitcher throws to a hitter, who is judged by whether or not he can successful either (a) put that ball into play successfully or (b) make it to first base through other means. What is being evaluated is right there on display. We don’t use a hitter’s batting average to categorically state that he is able to “think better” than other hitters; a statistic is used to support the observable. Unfortunately, it is much more difficult to make learning an observable activity; given the numerous ways in which it can be demonstrated, it is more difficult to quantify. Furthermore, learning may not manifest itself on a single test; if a hitter in baseball strikes out three times, and in each at bat learns more about the way he’s being pitched, then finally gets a hit in his fourth at bat, he’s still a .250 hitter. He may improve over time, but if he were graded on only that game, he’d be considered below average.
And that raises the second problem with standardized testing: the issue of sample size. To continue with the baseball analogy, a major league player usually gets upwards of 600 at bats in a single season. There are six hundred data points from which to build an understanding of his success or failure. Meanwhile, even students in the most test-happy districts in the country take maybe 10 tests a year. It’s the equivalent of determining that a player is great or terrible based upon the first three games of a season. Students and their teachers are evaluated, hired, fired, held back, and put into special education classes on what we would consider to be less than 1% of adequate sample size in baseball.
So, to answer the question posed by the title of this article: What can baseball teach education? First of all, that we need to look beyond the terrible data points we’re currently using to evaluate students and teachers. And more importantly, that if factions of the education field are going to take up the mantle of “reformers,” they should think harder about the level of analysis required to really reform the education system and build that reform on justifiable data, not just quick numbers that look good to the public.