Question: I’ve said how much I like reading these. Here’s my contribution on the first paragraph of yesterday’s post. Mask wearing is definitely politicized (what hasn’t been?). However, I’d remind anyone who would listen that pulling correlations/trends from grouped data is tricky business (e.g. Simpson’s Paradox). Related to your referenced Gallup survey for example, since so many more of the people who live in cities are democrats (order of 2 to 1), one could expect Democrats to wear masks significantly more than Republicans for no other reason — much harder to social distance, more contact with strangers, etc. The same explanation — demographics in cities — can be made for age btw and the Gallup data supports it. So, for more accurate (albeit less entertaining) left v right comparison: Republican males aged 20–40 living in the city vs Democratic males aged 20–40 living in the city.

Answer: Thanks for your contribution! I agree that univariate analysis is insufficient and I also agree that we must be cautious when drawing correlations based on overly simplistic analyses, let alone mistaking correlation for causation. I definitely would not want to be sharing data for entertainment, especially if it contributes to political polarization. So your contribution got me to thinking — have any researchers conducted more complex analyses to examine whether partisanship plays a role in behaviors like mask wearing or social distancing (including by controlling for numerous potentially confounding factors — like geography, age, sex, COVID burden, etc.)? As it turns out, they have!

Based on what I’ve read (see brief synthesis bulleted below) even after controlling for other factors, partisanship is significantly correlated with COVID-related behaviors. Nonetheless, as Dr. Sandro Galea, an epidemiologist and the dean of the Boston University School of Public Health who studies the politics of public health recently shared with FiveThirtyEight, “Nobody should ignore the fact that people on the political extremes are embracing polarizing positions on health behavior that should not be polarized, but I think the evidence we have indicates that most people have tried to be responsible and adopt the recommended behaviors, even at a time of immense polarization and confusion and discomfort.”

• This paper published earlier this month in Journal of Public Economics, Polarization and Public Health: Partisan Differences in Social Distancing during the Coronavirus Pandemic,”use[s] location data from a large sample of smartphones to show that areas with more Republicans engaged in less social distancing, controlling for other factors including public policies, population density, and local COVID cases and deaths. [Authors] then present new survey evidence of significant gaps at the individual level between Republicans and Democrats in self-reported social distancing, beliefs about personal COVID risk, and beliefs about the future severity of the pandemic.” The authors conclude, “Our empirical results show that partisan gaps in beliefs and behavior are real. GPS evidence reveals significant partisan gaps in actual social distancing behaviors. Survey evidence shows substantial gaps between Republicans and Democrats in beliefs about the severity of COVID-19 and the importance of social distancing. The raw partisan differences partly reflect the fact that Democrats are more likely to live in the dense, urban areas hardest hit by the crisis, and to be subject to policy restrictions — in other words, to face stronger individual incentives for social distancing. Even after controlling carefully.” [Note: this paper’s introduction section also gives a nice overview of other related research]
• This preprint publication (not peer reviewed) from three scientists at Syracuse University, UC Irvine, and Cornell, Partisanship, Health Behavior, and Policy Attitudesin the Early Stages of the COVID-19 Pandemic, presents analyses of survey data collected from 3,000 Americans during March 20–23. After controlling for age, marital status, race, income, education, geography (state), interest in the news, and frequency of news consumption, researchers found that “Republicans are less likely than Democrats to report responding with CDC-recommended behavior, and are less concerned about the pandemic, yet are more likely to support policies that restrict trade and movement across borders as a response to it. Democrats, by contrast, have responded by changing their personal health behaviors, and supporting policies that socialize the costs of testing and treatment. Partisanship is a more consistent predictor of behaviors, attitudes, and preferences than anything else that we measure.”