Table A3.

COVID-19 Case Growth, Prior Proximity to Cases, and Other Predictive Measures.

	log(Change in Cases per 10k Residents + 1)
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
2 Week Lag:	0.414***	0.321***	0.362***	0.277***	0.351***	0.270***	0.141***	0.141***
log(Change in Social Proximity to Cases + 1)	(0.041)	(0.037)	(0.047)	(0.039)	(0.047)	(0.039)	(0.050)	(0.045)
4 Week Lag:	-0.002	0.010	-0.008	0.022	0.001	0.027	-0.039	-0.014
log(Change in Social Proximity to Cases + 1)	(0.036)	(0.032)	(0.042)	(0.035)	(0.042)	(0.035)	(0.051)	(0.050)
Google searches related to Fever (% Change)			0.286*** (0.024)	0.231*** (0.022)
Google searches related to Cough (% Change)			0.158*** (0.021)	0.117*** (0.018)
Google searches related to Fatigue (% Change)			0.022 (0.019)	0.021 (0.019)
2 Week Lag:					0.165***	0.123***
Google searches related to Fever (% Change)					(0.021)	(0.018)
2 Week Lag:					0.189***	0.139***
Google searches related to Cough (% Change)					(0.021)	(0.017)
2 Week Lag:					0.013	0.019
Google searches related to Fatigue (% Change)					(0.019)	(0.018)
2 Week Lag:							0.269***	0.179***
log(Change in LEX Proximity to Cases + 1)							(0.025)	(0.022)
4 Week Lag:							0.006	0.006
log(Change in LEX Proximity to Cases + 1)							(0.023)	(0.022)
Share of Friends within 50 Miles	0.050	0.076	0.006	0.084	0.010	0.086	0.036	0.022
	(0.100)	(0.082)	(0.091)	(0.075)	(0.092)	(0.076)	(0.093)	(0.081)
Share of Friends within 150 Miles	-0.256**	0.143	-0.213*	0.168*	-0.220*	0.171*	-0.126	0.287***
	(0.124)	(0.109)	(0.110)	(0.097)	(0.112)	(0.098)	(0.115)	(0.101)
2 Week Lag:	1.244***	1.388***	1.116***	1.272***	1.117***	1.261***	0.971***	1.104***
log(Change in Physical Proximity to Cases + 1)	(0.118)	(0.176)	(0.105)	(0.167)	(0.105)	(0.168)	(0.104)	(0.170)
4 Week Lag:	-1.037***	-1.225***	-0.915***	-1.077***	-0.912***	-1.066***	-0.852***	-0.996***
log(Change in Physical Proximity to Cases + 1)	(0.121)	(0.187)	(0.108)	(0.178)	(0.108)	(0.179)	(0.107)	(0.180)
2 Week Lag:	0.372***	0.351***	0.498***	0.467***	0.484***	0.456***	0.536***	0.510***
log(Change in Cases per 10k Residents + 1)	(0.022)	(0.019)	(0.024)	(0.019)	(0.024)	(0.019)	(0.023)	(0.020)
4 Week Lag:	0.071***	0.056***	0.027	0.013	0.034*	0.019	-0.005	0.001
log(Change in Cases per 10k Residents + 1)	(0.019)	(0.017)	(0.021)	(0.017)	(0.021)	(0.017)	(0.022)	(0.019)
Time x Pop. Density FEs	Y	Y	Y	Y	Y	Y	Y	Y
Time x Median Household Income FEs	Y	Y	Y	Y	Y	Y	Y	Y
Time x State FEs		Y		Y		Y		Y
Sample Mean	2.177	2.177	2.279	2.279	2.279	2.279	2.333	2.333
R-Squared	0.725	0.757	0.768	0.800	0.767	0.799	0.795	0.827
N	47,040	47,025	38,520	38,520	38,520	38,520	30,210	30,195

Note: Table shows results from regression 5. Each observation is a county $\times$ two-week period (between March 30, and November 2, 2020). The dependent variable is log of one plus the number of new COVID-19 cases per 10,000 residents. Columns 1 and 2 are the same as columns 7 and 8 in Table 1. Columns 3 and 4 add the percent growth in Google searches related to fever, cough, and fatigue from the week prior to the period to the second week of the period. Columns 5 and 6 includes analogous measures lagged by one period. Columns 7 and 8 add LEX-based proximity to cases. Standard errors are clustered at the time $\times$ state level. Significance levels: *(p$<$0.10), **(p$<$0.05), ***(p$<$0.01).