By this point in the primary season, everyone has his or her preferred metric for predicting the outcome of a given contest. We’ve all noticed that Mitt Romney tends to perform well in states with small evangelical and rural populations, and in states with large numbers of college grads and affluent voters, while the opposite is generally true for Rick Santorum.
Since Wisconsin’s smallish evangelical population cuts the opposite way of its large-ish rural population, let’s keep things simple and take the final two demographic factors. According to the U.S. Census, 26 percent of Wisconsinites over age 25 have a four-year college degree, and the median household income in the state is $51,600. In Illinois, a neighboring state that Romney carried by 12 points, the numbers were 30 percent and $55,700 respectively. In Ohio, a nearby state where Romney eked out a 1-point victory, the numbers were 24 percent and $47,400. Which is to say, when it comes to socioeconomics, Wisconsin is almost smack in the middle of the two states, and so the consensus estimates of a 5-to-10 point Romney win sound about right.
The only twist here is the Milwaukee suburbs, which tend to be both very Republican and to boast high turnout. If you listen to people who know Wisconsin, you get hints that these areas—Waukesha County is the most prominent of them—aren’t just Republican hotbeds, but more conservative than you would expect based on sheer demographics. Which is to say, an affluent, college-educated Republican in Waukesha is more likely to be conservative than an affluent, college-educated Republican in an otherwise similar state. If that’s the case, then maybe Santorum has a shot of coming close in these areas even though they look more sympathetic to Romney on paper. That, in turn, could help him give Romney a scare statewide.
So is it true? Is a place like Waukesha County more conservative than you’d expect just from eyeballing the numbers? I think it might be. For example, based on sheer demographics, Waukesha looks like it should probably be more liberal than a suburban/exurban county in Illinois called Kendall (which Romney carried by 9 points a few weeks back). Thirty-nine percent of adults in Waukesha have a bachelor’s degree, versus only 32 percent in Kendall, and the median income of the two counties is pretty close ($75,000 in Waukesha versus $80,000 in Kendall). But, at least judging from the results of the 2004 election, Waukesha is significantly more conservative. George W. Bush carried it by 35 points over John Kerry, whereas he only carried Kendall by 22.
Don’t get me wrong. I’m not predicting an outright Santorum victory tonight. But there does appear to be something to the idea that Wisconsin’s most reliable Republican voters are more conservative than their demographic counterparts elsewhere in the Midwest. So the margin could be a bit closer than most of us are expecting. And if Romney only manages, say, a 3-point win when people are expecting 7 or 8, does it really count as a win? I'm sure we'll be litigating the rest of the week if it comes to pass.
P.S. One reasonable question is why the polling wouldn’t have picked this up? My guess is that it’s all about turnout—if my hypothesis is correct, then pollsters are probably under-sampling conservatives in counties like Waukesha.
Follow me on twitter: @noamscheiber