Tuesday, January 20, 2015

Lessons on Using Data for Assessment From A Football Game

via NY Post
The last couple of days, I've been a little bit down, emotionally. I'm a Packer fan, and an "owner," and I was coming to grips with the fact that the team had thrown away a great chance to advance to the Super Bowl in one of the most epic collapses in NFL history

As I perused some of the articles written about the game, I was startled to see the statistics from the two quarterbacks in the game, the Packers Aaron Rodgers, and Seattle's Russell Wilson.

QB Ratings for NFC Championship
via ESPN

One of the ways that quarterbacks in the NFL are evaluated is by Quarterback Rating. This rating is based on a complicated formula involving Completion Percentage, Yards Per Attempt, Touchdowns and Interceptions. While some believe that the formula is flawed, it has been used for many years to evaluate a quarterback's performance.

Based on this generally accepted formula, neither quarterback had a very effective game. In fact, Wilson had a 0 for a rating through most of the game. Still if you used this formula, it would appear that Rodgers played close to 4 times better than Wilson based on the statistics. Despite this, Wilson's last three passes resulted in a 2 pt. conversion, a key 3rd down reception, and a touchdown that won the game for his team. 

This got me to thinking about how students and teachers get evaluated. Some teachers complain about the student who coasts all semester, only to turn it on in the last few weeks to pass the class. They want the grade to reflect that "lack of effort."
Others want to evaluate teachers based on standardized test scores of their students, despite research that questions the correlation

Based only on the statistics, it appears Rodgers had the better game. But I think he'd take Wilson's numbers if it meant he'd be playing in the Super Bowl in a few days. 

There are some flaws to this analogy, but I hope it encourages people to think about the data they use to assess students and teachers, and how sometimes, numbers don't tell the whole story. 
Post a Comment