Home Field Advantage in College Football

Just read some of JR Miller's book about beating the Pro Football Pointspread. Football is football is football, and it got me thinking about college football and its relevance to today's topic: home field advantage.

Some JR Miller quotes:
"There are actually wide differences in home field advantages from team to team for extended periods of time" (2-4).

"There is a general tendency for teams to play better on home turf" (2-4).

"For a three year period ending in the early '90s, the Washington Redskins...actually performed 3 points worse at home than on the road" (2-4).

In college football I'm sure the effect is magnified. The players are younger, more emotional, less confident, and less professional (on average) than their professional counterparts.

So how could this study be done and its effects measured?

Well one could look at the difference in average score between home and road, difference in average margin of victory or defeat, and also look at how visitors do at a certain stadium compared to the road and overall average.

The problem of course is not all college teams are made equal. Ohio State's stadium will have given up fewer points than most teams because they traditionally emphasize defense and ball control.

Over the years that will have given them a skewed measurement of how much effect their stadium has because they also have above average defenses. So one would have to look at how opponents do in a given season against common opponents.

This is actually a very tricky question because teams have certain identities over the years. One could look at the difference between consecutive years playing Ohio State on the road or at home to get an idea of the difference that stadiums and home field advantage have.

Or if teams play twice in one season, looking at the difference in scores. Unfortunately I'm dealing with something that has high variability over the course of a season(injuries) and between seasons (player turnover and graduation) and very few instances.

I could look at various statistical measures such as a t-value (for average variability around an unknown value). But that would be getting very complicated very fast.

Lastly, the effect could be measured for young teams (freshman, sophomores) versus teams that are experienced/older (juniors, seniors). Also comparing seasons started per team. I'd think that more experienced teams would perform better.

Also, a three year starter would have played at Ohio State, then hosted them his junior year, and then go back again his senior year to the horseshoe. So that would be a real advantage because he'd already have played their once and seen the kind of debacle he was walking into.