So, I see this article show up on Bleacher Report (here) written by some dude who attended the 'University' of Arizona.  In it, the 'author' purports to rank the Pac-10 fair weather fans.  Knowing what I know of LA football, I decided to indulge the author and get a good laugh at USC's expense.  However, as I clicked through the slides and noticed USC ranked as having less fair weather fans than UCLA, I realized that it was garbage.   Other than the low hanging fruit, like Oregon & Utah selling out their stadiums, the rest of the slides were pure conjecture and opinion.  

As a result, I did a quick search for UCLA and USC attendance figures to do a little more in-depth analysis.  I found the results on the NCAA website, but only as far back as 1998.  However, I thought this a relatively appropriate time period upon which to run a simple statistical analysis.  For the period 1998-2001, SUC amassed a pretty mediocre 25-24 record, while UCLA amassed a 27-19 record.  Meanwhile, for the period 2002-2009, USC amassed an impressive 91-13 record, while UCLA amassed a pretty horrid 53-47 record.  Subsequently, I was able to find UCLA's attendance figures back to 1994 thanks to an LA Times article, but was not able to do so with USC.

Admittedly, the data is sparse with only 16 years of data for UCLA and 12 years for USC.  However, that is the nature of the beast and, I believe, captures enough down time for both programs, as well as capturing the recent pinnacles of each.  I also decided to list the wins in both the year of analysis, as well as the previous year's win count.  I then determined the standard deviation and R-square, Pearson & correlation of attendance to current year wins and previous year wins.  Results are below:

0001K7 (via Scotucla)

To begin with, you'll notice UCLA has a much lower standard deviation than USC.  On the face of it, it would appear that UCLA's fan base is much more steady, attracting a more consistent fan base in good times and bad (i.e. is less volatile).  Meanwhile, the standard deviation for USC is noticeably greater than UCLA (14,154 vs. 8,474), which would seemingly imply that their attendance fluctuates more wildly.

However, what I find most interesting is the correlation/R-squared/Pearson analyses (which are all the same for each team due to the small sample size, I presume).  The results indicate there is very little correlation between attendance and current or previous year wins at 15.47% and 4.92%.  It would appear the previous year wins correlation is depressed since attendance fell so dramatically after our 1998 season (any ideas as to why, anyone?).

Meanwhile, there appears to be a VERY high correlation (remember, a 1.0 implies perfect correlation) between both USC current and previous year wins at 83.24% AND 92.51%(!).

So, while not a statistician by any means, who would appear to have the more fair weather fan base (i.e band wagoners!).

And I welcome critiques and suggestions to improve the analyses by those more knowledgeable in such areas than myself!