Religion, not race, is the best single predictor of voting preferences
America’s founding fathers envisioned a republic in which free-thinking voters would carefully consider the proposals of office-seekers. Today, however, demography seems to govern voters’ choices. Since April 2017 The Economist and YouGov, a pollster, have surveyed 1,500 Americans each week. They have built a statistical model to estimate the odds of how each respondent will vote in next week’s mid-term elections.
Polling of voting sub-groups can be misleading. City-dwellers are usually liberal. Is that because of where they live, or because they tend to be better educated and are less likely to be white than countryfolk? The model measures each variable in isolation. Even among people of the same race and schooling, urbanites are more Democratic-leaning. It also considers how variables affect each other. For example, single women are more liberal than married ones, whereas this gap is negligible among men. Of the 12 factors in the model, the most important is religion. Atheists are even more likely to be Democrats than evangelical Protestants are to be Republicans.
The model adds up the impact of each variable, like a set of building blocks. As a result, a group of weak predictors that point in the same direction can cancel out a single strong one. In theory, the model could identify a black voter as a Republican leaner, or a white evangelical as a probable Democrat—though it would require quite an unusual profile.
Source: –, How to forecast an American’s vote, The Economist, November 3, 2018, https://www.economist.com/graphic-detail/2018/11/03/how-to-forecast-an-americans-vote.