If you have read my previous article about the height of Division I collegiate volleyball players, you might have noticed something a little odd. Let’s look at that graph again.
The most common height of a collegiate Division I volleyball player over the last decade is 6’0”. The next most common height is 6’1”, followed by 5’10”. In case you are not familiar with height distribution data, they usually reflect a “normal” distribution (sometimes called a bell curve) in which the numbers increase as you get closer to the most common height in the group, and then gradually start decreasing after the most common height (thus the bell shape). If you look closely at the data, however, you will see that there is a dip at 5’11”, when all signs should point to it increasing (if 6’0” is the most common height, then there should be more student athlete years at 5’11” than at 5’10”). Why do we think that is? Here are my theories:
When looking at the graph, the obvious thing to deduce is that there are simply fewer 5’11” volleyball players. That is objectively what the graph shows. While some people might be okay with believing that, I am not buying it.
Another possibility is that 5’11” players are disproportionately playing in other divisions. But why the 5’11” players more than the other athletes? Overall, average player heights in Division II and Division III are lower than in Division I, and at least without more of a story, this doesn’t seem to be a good reason that there are disproportionately fewer 5’11” players in Division 1 volleyball.
My final hypothesis is that the 5’11” players have not gone anywhere. They are right under our noses, hiding in plain sight. Lying about their height. You heard me, they have become pretend six footers. Why say you are 5’11” when you can hit that nice, round milestone of 6’0”? While this is just a theory, this seems both in-keeping with my own personal experiences and patterns we see in other sports.
Data Notes:
- So, anomalous patterns in data can sometimes be driven by how people systematically misreport data–often to make themselves look better. Of course anomalous patterns can be driven by other things as well, like data reporting or data entry errors. In cleaning the NCAA data used for these posts, there were some interesting data points that were (almost) certainly data reporting or transcription errors.
- If you go the NCAA D1 volleyball stats archive, you’ll see for example, that a decade ago there were dozens of athletes listed at exactly 4’0”. I’m not sure what led to people being listed at 4’0”, but in cleaning the data I chose to treat those athletes’ heights as ‘missing’. Overall, we have the heights of more than 98% of the athlete-years in this dataset. We can fill in the height data for some of the athlete’s that are missing in this dataset from team websites that I scraped, but I’ll need to write about the (small) differences in the height data on team websites and the NCAA data before doing that.
- Generally, the NCAA Stats data coverage for the things I’m reporting here is really quite good, although with slightly more missingness and apparent data entry errors for 2013 (and even more in the years before that, which I’m not analyzing here) than in more recent years. I’ll write more extensively about data quality and coverage when I post on topics that require using data with worse data coverage.