An archaeologist applies stats to reported instances of college football concussion and uncovers conspicuous instances of under(non)-reporting. Nice job, Shane.
Over the course of the semester I’ve been using data from college football as examples in my Quantitative Methods course.
One data set that I found particularly interesting is this map outlining recorded concussions in college football last year. The map is kind of cool, but what’s really interesting is all of the other data they gathered along with it.
So…I downloaded the .kmz, pulled it into ArcMap, converted it into a shapefile, and then pulled out the data table.
I was particularly interested in determining if some positions were more prone to concussions than others.
Here we go…the most straightforward breakdown. With chi-square tests you compare your observed frequencies to an expected frequency distribution – in this case, if spread out all 166 concussion in their sample evenly across all of the positions. To test whether any differences are statistically significant, we run a Pearson’s Goodness of Fit…
View original post 488 more words