The rationale for this is that the actual record alone doesn't tell the whole story. If a conference is having a down year and/or gets some matchups that aren't favorable, you'd expect a poor record so even a 0.500 record may be an indication the conference did well or at least better than expected. Thus comparing the actual with expected record is useful, as is looking at how a conference does against the spread, as that tells you if the teams did better/worse than Vegas and the public bettors expected.
Here is the up to date table through the Sugar Bowl:
|Conference||Vegas||Computer||Most Likely||Thru 1/2||ATS|
From this we can see that the SEC and Big-12 were supposed to do the best, and they have, but the SEC is actually behind where they should be. Vegas and my computer had them favored in every game with my computer projecting a most likely record of 5-2, and with one game left to play they need to win it to just get to that 5-2 mark. Note that I'm not including the BCS championship game as two SEC teams are facing each other. The SEC is also just 3-3 against the spread, so performing about where the public betting said they would.
The Big-12 has arguably done the best, on track for the expected 6-2 or 7-1 but a great 6-1 against the spread.
A couple big conferences that didn't have a good outlook were the Big Ten and Pac-12. The Big Ten has another game to play and will finish right around the expecte 3-7 or 4-6, but they haven't done well against the spread going only 3-6. The Pac-12 similarly finished right between the expected 1-6 or 3-4 at 2-5, but also did poorly against the spread going 2-5.
The bowl games are often a chance for the smaller conferences to make a statement, but most haven't, the exception being Conference USA which is 3-1 overall, but a healthy 3-0-1 against the spread. The Mountain West is only 2-3 in each category and the WAC 0-3 and 1-2. The MAC has done decent but my computer expected them to.