What’s wrong with this picture?

Print More

The current issue of The Atlantic features a story on political prejudice, and concludes that Jefferson County, New York (Watertown) is the least politically prejudiced county in the nation. Moreover, we New Yorkers are a remarkably open minded state. I didn’t attempt to cobble together at statewide average, but not another state comes close!

If you believe the data. The story is based on surveys designed and analyzed by PredictWise. The pollsters acknowledge sampling bias: “Like all modeled data, these results are imperfect; we did not survey every American.” But the problem here is much more fundamental than sampling bias, although the authors provide a link to a detailed methodology. Replete with the usual technical jargon, including a number of lengthy mathematical formulae, the results fail the most basic of “smell tests.”

Not only do I find the consistent open mindedness of New Yorkers to be implausible but I am skeptical that Floridians are just a prejudiced as New Yorkers are open minded (particularly knowing how many Floridians were once New Yorkers!).

Is it remotely plausible that residents of Dutchess County are in the 8th percentile of prejudice while their neighbors across the Hudson in Litchfield County, Conn. are in the 66th percentile? Or that Clinton County shares 8th percentile status while the adjacent county in Vermont (Chittenden) is in the 89th percentile? 

According to these data, Nassau County, Florida is a terribly prejudiced place—ranking at the 93rd percentile (thus only 7 percent of the nation’s counties display more bias). Perhaps Nassau residents should move a few miles north to Camden County, Georgia, ranking at the 20th percentile.

No fancy formulas or defensive language can defend PredictWise or The Atlantic. Here’s PredictWise again:

Yes, we do suspect some of the sharp state differences can be artifacts of how party data is collected at the state level. Counter argument? We only run into these differences in three out of 50 states, and the most blatant examples, South Carolina and Florida, are states in which we find partisans to be very insulated in their neighborhoods by age and partisan affiliation.

Nonsense. This story and the data upon which it is based lacks all creditability. 

I don’t want to take anything away from Watertown’s day in the sun, but I suspect that there’s something fundamentally flawed with the underlying data. Consider the helpful infographic provided with the story. 

4 thoughts on “What’s wrong with this picture?

  1. Psst. Jefferson County has a single economic driver (Fort Drum). Might that not influence population demographics and the apparent uniformity of political thought?

  2. The fundamental problem is that the poll asks people nationwide how they respond to various questions that are perceived to indicate political tolerance, then it applies the results by county based on how that county demographically matches the survey’s respondents (including “political isolation,” or how much people live near people of a different party). But there is no way to know if a New Yorker who demographically matches a Floridian has a similar attitude. And the fact that independents are eliminated from the study means that an important segment of the population for real political open-mindedness is missing.

    The Atlantic should be ashamed of itself for presenting such a half-baked analysis. They present no evidence that the answers of the 2000 people nationwide can be extrapolated to every county in the country based on the demographic characteristics used. As Disraeli reputedly said, “There are three kinds of lies: lies, damned lies, and statistics.”

Leave a Reply

Your email address will not be published. Required fields are marked *