Super Tuesday confirmed two trends that had been somewhat evident prior to yesterday. First, while there is an ideological dimension to this contest – with very conservative voters backing Rick Santorum or Newt Gingrich and less conservative voters backing Mitt Romney – there is also a regional dynamic overlaid atop this. Put another way, Romney does better or worse among voters in any given state the more conservative they are, but in different regions he will do systematically among all ideological groups.

To appreciate this, let’s break down the exit polls by region, and then by ideological groups. We’ll look at Romney’s margin of victory or defeat over his nearest opponent.

First, here is the Northeastern states that have voted so far:

As we can see, Romney does worse with the “very conservative,” best with the “somewhat conservative.” Meanwhile, the “moderate/liberal” vote breaks his way, but Ron Paul cuts into his margins here. That’s the same basic pattern we have seen all across the country, but notice that he wins every group in New England.

In the South, the story is the same, but also different:

Again, same pattern across ideological – best with somewhat conservative and moderate/liberal, worst with very conservative – but he performs worse among all groups relative to New England.

And unsurprisingly the Midwest basically has split the difference between the South and Northeast (which it has done for two centuries of presidential politics!).

Once again, Romney does best among the “somewhat conservative” and “moderate/liberal” while doing worst among “very conservative.” But his performance with each group is better than it was in the South, worse than in the Northeast.

Overall, then, ideological orientation is not sufficient to explain the results. Geography also matters. If you are a “very conservative” person living in New England, you are going to be much more inclined to vote for Romney than you would be if you lived in the South. Similarly, a moderate in Ohio would be much more inclined to vote for him than a moderate in Tennessee.

(This is actually typical of nomination battles going back to the earliest days – with regional factions backing one candidate over another for reasons that have little to do with ideology.)

The second important factor that also has little to do with ideology is the type of area that one lives in – urban, suburban, or rural. Michael Barone noted last week that Romney is doing well in the suburban areas, and indeed there is definitely a divide between rural communities and suburban/urban ones.

Unfortunately, we do not have the data to look at ideological groups within the urban/suburban/rural categories, but my gut tells me that, much like the regional divide, the “very conservative” in urban areas are more likely to vote for Romney than the “very conservative” in rural areas.

So overall, when people tell you that Romney’s problem is with conservatives, that is not precisely true. His problem has to do with “very conservatives” in part, but it also matters where one lives – North v. South v. Midwest and urban v. suburban v. rural.

Next Page