Romney ’08 vs. Romney ’12: Not the Same Voters

Jan 5 '12
This is a guest post from Tobin Grant:<br>

Most early analysis noted that Romney received around the same percentage of the caucus vote and number of votes as he did four years ago.

During the punditry last night, some even suggested that Romney attracted the same voters except those that died in the interim.

Polling results further suggested that the Republican electorate was very similar to four years ago and that Romney did about the same in the major demographic groups as he did in 2008 (see here and here for examples).

Entrance polls are only one way to slice the results, however. Another is to look at the geography of the contest. A comparison of the county results shows that counties that backed Romney in 2008 did so again in 2012, but the results from the two contests are correlated at only 0.69–positive but relatively low for aggregate data like this.

More puzzling: At the county level, Romney received only 55 percent of the vote percentages in ’12 that he did in ’08.

How, then, did he end up with the same statewide vote? The answer is that Romney’s support came much more from the more populous counties in the state than it did four years ago. Some smaller, rural counties dropped in their support for him, but he made it up by losing little or gaining in the more populous counties. For example, Romney improved in Des Moines, Iowa City, Ames, and Iowa City.

Heading into the homestretch, Romney spent most of his time and money in the areas where he did well in 2008. That strategy seems to have paid off.  Now it’s on to New Hampshire for another re-do of 2008.