TABLE A5.

Results with pre‐existing measures of polarization: re‐contacted sample

	Dependent variable: number of behaviors
Opposing	−0.321	−0.192
	(0.281)	(0.310)
Opposing x	0.441	0.272
polarized individual	(0.450)	(0.310)
N	146	146
R 2	0.232	0.234
Polarization measured in:	Nov 2019	Mar 2020

Note: This table shows the results from a regression of the number of preventative behavior an MTurk respondent engages in on an indicator for living in an opposing party state, and this indicator interacted with an indicator for being above the national median in a measure of affective polarization. The sample is a set of individuals who were initially surveyed for a separate project in November 2016, and who we were able to recontact for this project. These individuals do not necessarily live in close election states, but details of sample restriction and the regressions are otherwise identical to those in the main analysis tables.

***p < 0.01, **p < 0.05, *p < 0.1.