The Washington PostDemocracy Dies in Darkness

Those least likely to worry about getting covid (Republicans) have been those most likely to get covid

President Donald Trump after speaking at a campaign rally in Pennsylvania on Oct. 26, 2020. (Alex Brandon/AP)
Placeholder while article actions load

Since early in the pandemic, there have been a variety of ways in which views of the situation fall along partisan lines. For example, Republicans have consistently expressed less concern about contracting the coronavirus than have Democrats — a view that correlates to being less likely to get vaccinated, wear a mask or reduce public interactions.

We might not be surprised, then, by new data from YouGov polling conducted for the Economist: Republicans are significantly more likely to report having contracted the virus than are Democrats.

Sign up for How To Read This Chart, a weekly data newsletter from Philip Bump

That’s been the case for a while. The three-week average of responses to the question showed more reported positive tests among Republicans back in the fall of 2020. In recent weeks, though, the gap has widened. In YouGov’s most recent poll, fully a quarter of Republicans say they’ve tested positive at some point.

This is polling, obviously, and not confirmed data on infection rates. It’s possible that Democrats are warier of reporting infections or Republicans more likely to say that they were infected without negative effect. It’s hard to control for such things, should they factor into responses.

What we can say is that the data comports with other data looking at the spread of the virus.

The Washington Post has compiled county-level data on new cases and deaths since the outset of the pandemic. What that data shows is that counties that voted more heavily for Donald Trump in the 2020 election have higher cumulative case totals relative to their populations than do counties that preferred President Biden. That’s largely because during prior surges — though not the current one — those counties had higher adjusted infection rates than blue counties, month after month.

You’ll notice that we included vaccination percentages on that chart, too. A similar pattern emerges on month-by-month data: Less-vaccinated places were more likely to see more cases per resident. But we’re basing those numbers on current vaccination rates, and the disparity in vaccination status precedes the actual broad rollout of vaccines! Why? Because vaccination rates correlate to the 2020 vote.

You can see that below. Each circle is a county. Farther to the left, more support for Biden. Higher, more vaccinations per person. The cloud goes from top left (lots of Biden/vaccines) to lower right (lots of Trump/unvaccinated). Correlation.

That chart also overlaps the most important figure: deaths. The smaller the circle, the lower the per capita death toll in a county. Those upper-left counties have a lot of very small circles.

But we can look at this in the aggregate, too. The places with the most deaths, even now, are places that voted more heavily for Trump and have lower vaccination rates.

Of course, death itself correlates inversely to vaccination: If you’re vaccinated, you’re better protected against serious illness.

This is the other grim caveat we might apply to that poll. What would the numbers look like if we could include those who contracted the virus and didn’t survive to record their infections?