One big advantage we have at CivicScience is our ability to capture and process lots of polling data. Our current database has over 120 Million poll results cataloged. Mining this database allows us to gather and infer lots of other characteristics about the people we poll, characteristics that are simply out of the reach of conventional pollsters. In the case of the recent numbers we crunched on the Allegheny County Executive race, we started with an initial group of 19,000 respondents, which allows for a lot more slicing and dicing than your standard polling sample.
Beyond all of the normal characteristics our algorithms process, like Age, Gender, Party ID, and likelihood-to-vote, we can find lots of other interesting (if sometimes benign) correlations. For example, people who use T-Mobile as their wireless carrier are 3-times more likely to support Rich Fitzgerald than people who use Verizon (N=3,173 for those of you keeping score at home). Curious, right?
Or, how about this one: Cat lovers are 2.5-times more likely to support Fitzgerald than Dog lovers (N=4,022). Oh, and another fun one: Fitzgerald leads handily among people whose favorite casual restaurants are Eat-N-Park, Panera, and Applebee’s. Raja leads among people who prefer King’s, Atria’s, and Denny’s (N=2,575). Unfortunately for Raja, way more people like Eat-N-Park and Panera.
OK. We admit that most of this is just useless trivia, though we could see how a candidate might choose to shake hands at one restaurant rather than another. Or, perhaps they’ll picture themselves in a commercial with their family cat instead of a family dog.
One not-so-useless observation, however, is the disparity among voters in the County Executive race along socio-economic lines. It comes as no surprise that wealthier voters are more likely to support a Republican candidate than a Democratic one, given their stated concerns over taxes, free enterprise, etc. But the extent to which the Fitzgerald-Raja race breaks across income brackets is particularly severe. Fitzgerald leads by a minimum of 15 points and as many as 34 points among every income bracket from $25,000 a year in household income to $125,000. But as soon as you move from voters making $100-$125k/year to those making over $125k/year, Fitzgerald’s lead goes from 20 points, to only 3. Normally you would expect a subtle shift from one bracket to another. These numbers are dramatic and pinpoint.
Maybe the current economic climate and “Occupy” movements are exacerbating the current divide between the rich and not-rich. Or maybe there are particular traits about Raja and Fitzgerald that engender such a clear disparity. We have a lot to learn about what CIVICSCIENCE data can tell us but as we gather and process more of it, the picture becomes clearer. If you geek out about this stuff like we do, it should be fun to watch.