Total Pageviews

Tuesday, March 10, 2009

Teaching and Baseball (Statistics)

In baseball there is an epic battle between new wave statisticians and old style "heart of the game" types. Of course by epic I mean a lot of guys are blogging about it. There are a core group of baseball researchers that believe that through the study of the immense amounts of data surrounding baseball that we can better predict and evaluate a player's performance. On the other side is a bunch of guys saying, "Nuh-uh."

I find myself siding with the stats mainly because the other guys always want to talk about grit and determination. They also refer to people as gamers. However, in the classroom I can't seem to get behind the data as firmly as I do on the diamond.

Data driven decisions in the classroom are a lot trickier, kind of like defensive statistics in baseball. It obviously works on things like attendance and graduation rates, but in other areas I feel that there are too many variables.

First of all we need to take into account park effect. Baseball stats are often adjusted for the players home park. For example pitching stats are often inflated in Colorado and the Green Monster obviously offers and advantage to right handed hitters in Fenway. I propose that when we get testing data that it should be adjusted to reflect the home effect and the school effect.

For instance if a student scores lower on a test but they come from a home with parents that just barely graduated high school then those scores should be adjusted to reflect that. By the same token a student that have their own little study oasis at home and private tutors to help them along should also have their test scores normalized as well.

Secondly, much of the data that we use is subjective and like the defensive statistic that I mention earlier, there really is a way to objectify them. Every teacher will grade differently no matter how many PD sessions we have. Every student will perform differently depending on how much sleep they got the night before and whether they ate oatmeal or Fruit Loops for breakfast. Furthermore we are often compare players from the low minors with All Stars. Hitting in the Carolina league is nothing like facing a Mariano Rivera in Yankee Stadium and a player that makes that jump will look foolish every time.

Finally, if we are going to use data to drive our decisions we should realize that a .200 hitter will never win a batting title and no amount of steroids will turn a pitcher with 3.4 K/9 into Nolan Ryan.

Each district should hire a statistician to analyze and compile the data just like every major league club relies on at least one come draft day.