Table 2.

Results of the weighted least squares regression model, county level (N = 1960).

Variables	Coefficient	VIF	C.I. 2.5%	C.I. 97.5%
(Intercept)	1384.31		1187.59	1581.02
Political affiliation (Republican)	0.16***	2.33	0.14	0.18
COVID-19 vaccination rate	− 0.09***	1.73	− 0.12	− 0.07
Educational level (% of population with a college degree)	− 0.06***	3.63	− 0.10	− 0.03
Weather—National Risk Index	0.87	1.80	− 0.17	1.90
Weather—Mean temperature (2015–2019)	14.70***	1.16	10.30	19.10
Median household income	0.01	2.57	0.00	0.03
Adjusted R-squared: 0.47	***p < 0.001

Note: Total number of tweets per county was used as the universal weights in the model. Counties with less than 50 tweets were excluded. Variance Inflation Factor (VIF) < 5 indicates low collinearity of the variables used in the model.

C.I. confidence intervals of regression coefficients.

Note: Significant values are in bold.