Table 1.

Spearman correlation and linear regression results between the number of tweets per population and alcohol-related behavior and health indicators.

Outcome		Spearman correlation			Adjusted regression			Sample size, n
		ρ	P value	Coefficient β		P value
Metropolitan-micropolitan statistical area
	Alcohol consumption	0.526	<.001	1038		.01	179
	Binge drinking	0.355	<.001	184.0		.37	179
	Heavy drinking	0.387	<.001	244.8		.005	179
County and equivalent
	Excessive drinking	0.377	<.001	32.8		.03	2641
	Percentage of alcohol motor vehicle fatality	0.063	.002	110.0		.21	2641
	Drinking places (alcoholic beverages) per capita	−0.177	<.001	−2.18e–03		.23	1479
	Breweries per capita	0.263	<.001	1.86e–03		<.001	334
	Wineries per capita	0.130	.05	2.73e–02		<.001	228
	Beer, wine, and liquor stores per capita	−0.043	.11	0.0039		<.001	1444
US states and Washington, DC, all hashtags
	Wine, gallons of ethanol per capita	0.756	<.001	74.11		<.001	51
	Beer, gallons of ethanol per capita	−0.050	.73	9.911		.63	51
	Liquor, gallons of ethanol per capita	0.320	.01	62.54		.03	51
	All sources, gallons of ethanol per capita	0.437	<.001	146.6		<.001	51
US states and Washington, DC, hashtags stratified by alcohol category
	Wine, gallons of ethanol per capita (5 hashtags)	0.754	<.001	214.6		<.001	51
	Beer, gallons of ethanol per capita (19 hashtags)	−0.001	.99	16.05		.63	51
	Liquor, gallons of ethanol per capita (3 hashtags)	0.140	.33	338.0		.01	51