College Sports Attendance & Gender – Final Update

First off, I took another approach to look at the data set. I used the controlled ratio I discussed in my previous post and compared the average value for basketball conferences. I took the average ratio for each conferences and then grouped the larger, well-known conferences known as the Power 6 (Big Ten, Big 12, SEC, ACC, Big East, Pac-12) and compared them to the smaller mid-major conferences. I would have done this for the other sports but basketball was the only one where almost all the schools had their men’s and women’s teams in the same conference. I’ve posted the graph below (found out how to make high-res pictures in RStudio, so now they’re a little more readable readable; any remaining issues I’m blaming on this site).


So the population sizes are too small to really do any statistical test, but it is clear that the Power 6 schools have a much lower controlled ratio then mid-major teams (indeed, the average for Power 6 was .315, and the average for mid-majors was .528). It appears that the schools that have the fame and power to really alter the conversation on women’s sports might not be doing all that much.

I’d like to spend the rest of this post going into some of the (many) flaws in this analysis and why I refrained from making the kind of conclusions I just made above about blaming any differences on schools or fans. First off, the reason I created the controlled ratio in the first place was because I did not have the statistical background to do the regression I really should have done. While it worked fine for the most part, the nature of the win percentage value really created some skewed ratios (in theory a school that won no games would have a almost infinite controlled ratio, which makes no sense). Also my ratio doesn’t describe an attendance rate, but rather a kind of elasticity of attendance based on win percentage. This does not really show that a men’s program has higher attendance than a women’s program, but that the men’s program is rewarded with more attendees than a women’s program for the same increase in win percentage. Also comparing in schools controlled for lots of crucial factors, it would be much more difficult to parse out things like day of the week the game occurred on, or popular players on the team. I also did not include the price for the game, and assumed that many schools are like WM and students can get into games for almost free (this is not the case for big schools).

Because of the reasons above, I did not feel comfortable making strong conclusions about the results of my analyses. Similarly, I chose to abandon the idea of tying discrimination into my paper as I had originally planned, because even if I had established a significant differences without question, I had no clear way of blaming this difference on the preferences of the consumer or the schools and the NCAA. Who would be responsible for causing the discrimination would play a large factor in determining what policies would be best to fix this gap. I went into this research with the opinion that discrimination did exist in college sports, and nothing I found seem to contradict that opinion; nonetheless, I think it would be dishonest to claim I was able to prove anything more than I actually did.

Overall, I enjoyed working on this project. Not only did I develop some more data analysis skills, I also improved on the other parts of doing research that I don’t find as interesting but are as vital to having a successful project. I don’t think I will be building on this project in future years like I might’ve originally hoped, but the experience will be valuable in the future.