Tom Pepinsky

Author: tompepinsky

  • Americans Have Become Much Less Supportive of Delaying Elections since March 2020

    American politics is in a fragile state. Projects such as Bright Line Watch have been charting how American politics is changing over the past several years, but it doesn’t take much expertise to see how COVID-19 comes at what might be the worst possible time for a sharply divided country. As part of an ongoing project on the politics of COVID-19 in the United States, Shana Gadarian, Sara Goodman, and I have been collecting data on a range of opinions and attitudes about COVID-19. These data give us insights about how Americans are experiencing the pandemic, and how our partisan politics governs our responses.

    Amid all of this, we have found a bit of cause for optimism: Americans on the whole have become much less supportive of delaying elections since March 2020.

    We’ve reached this conclusion by comparing the first two waves of our surveys. Waves 1 and 2 of our survey came in late March and late April, and American politics seems to have undergone at least a couple of revolutions during that period.

    In March, Americans were still getting used to lockdown conditions, whereas by late April protesters were starting to press for reopening. In between, mid-April saw the hotly contested election in Wisconsin, which made very plain that for many Americans, voting in 2020 will have public health consequences.

    So what to Americans think about delaying elections? The figure below compares responses to the question “To what extend to you agree with the following statement? Elections should be delayed if it means protecting people,” which we asked in both waves of our survey.

    In March 2020, the modal response was to strongly agree with this statement. Today, the opposite is true: the modal response is to strongly disagree with this statement.

    Now, this is good news, but where are these changes coming from? You might have expected that the question of whether or not to delay elections would end up being a partisan issue, just like everything else is in American politics these days. Or perhaps it is driven by other demographic features. Below, we calculate the average response (1 = strongly disagree, 5 = strongly agree) across these and a wide range of other variables, and show differences in average responses between Wave 1 and Wave 2.* The vertical dotted and dashed lines are the sample averages in Waves 1 and Wave 2, so you can see how each subgroup compares to the average in that wave.

    These results are striking. There are modest differences in average support for delaying elections across different populations of Americans, but we see substantial declines across all of these groups.

    We can also look at differences across groups based on their attitudes or geographic characteristics. It could be that we see more support for delaying elections in contested states, or among those who express greater anti-immigrant sentiments, or among those who are opposed to voting by mail, or among those who consume right-wing news. We are able to measure all of these things and more, and here is what we find.

    These patterns are more interesting, and there’s a lot to dig into here. But note that we see substantial decreases in support for delaying elections across all of these groups, even if we do still see, for example, more support for delaying elections among respondents in states that Trump lost (regardless of these respondents’ individual partisanship).

    All of these details about differences across groups are important, and suggest that there’s more to explore. But the topline finding is unambiguously good news from my own perspective: Americans of all types are becoming ever more eager to make sure that elections continue as planned.

    NOTES

    * These figures are all weighted to be representative of the U.S. population via YouGov’s weighting scheme.

  • What Predicts COVID-19 Rates in the Advanced Capitalist Democracies?

    A couple days back I posted a scatter plot that suggested that social democracies and developmental states had lower levels of COVID-19 prevalence. Here it is again (updated by several days).

    I got a lot of feedback on this, asking if there were other factors that I’m leaving out. I’m certain that we don’t have good models of the epidemiology of COVID-19, so I’m sure that the answer is “yes.” But was able to add in a couple of control variables to see if they affect those results.

    • Perhaps I should measure “island” comprehensively, coding Japan and the UK and Ireland as islands too instead of just Australia and New Zealand.
    • What about the differences in seasons between northern and southern hemispheres?
    • What about population density explaining low rates in Scandinavia and Australia?
    • No one said this explicitly, but lots of people think that South Korea and Japan are just monoethnic societies that are better able to enforce social compliance.
    • Could this be about level of economic development, or spending on health?

    I dumped all of those variables in a series of OLS regressions with COVID-19 rates as the dependent variable, predicted by the Index of Social Democracy from Lane Kenworthy, a developmental state dummy, and various combinations of the rest. Here are coefficients plus 90% confidence intervals for all of these models.

    We shouldn’t interpret too much from these results. The data are incomplete on case loads, and the sample size is just 21. But the findings are still quite provocative.