Sunday, August 23, 2009

A discourse on basic stats

I make it a habit to regularly check the weather, usually via Weather.com. I've been monitoring the weather up in Hillsdale to gauge what it's normally like around this time of year, so sometimes I'll check the "monthly view", which gives you average and record temperatures for each day of the month. This "average" temperature thing puzzles me a bit...

People generally think that averages give a good idea of the normal numerical trends in things like temperature, class sizes, et cetera. The thing is that this isn't necessarily true. A college's typical class size isn't necessarily the average of the college's class sizes.

For example, COD's is 19. I laughed when I saw this figure because of course they're trying to pass this off as the 'typical' class size, though this is far from the truth either way it's measured. The maximum capacity for most COD classes is 35, and the more common gen ed classes tend to fill (retention is a different matter, but either way the typical enrollment is not 19). However, COD is a community college. They also have numerous specialty classes geared toward their myriad of applied degrees. Plenty of these don't see enrollment anywhere near maximum capacity, but they happen regardless: they're necessary for people in those applied programs to graduate. My sister once took a library science class that started with 6 and ended with 2. Enough classes like those will mess with COD's class averages, and they do.

In some cases, the median or mode of class enrollments might be a better indicator of what is 'typical'. But there exist enough cases where these aren't reliable indicators either. Something as mathematically simple as the average, median, or mode will in turn can give you simplistic answers, which may or may not give you what you're looking for.

To end: As to the typical weather in Hillsdale around this time, I'm not sure whether to believe the averages or not. I'd be much more comfortable if I could see the trends from the past few years, but I'm not really about to go digging that up.

No comments: