Table 3.

The response to SARS by information source.

Dependent variable:	Individual visits
SARS case definition:	Reported		Probable
	(1)⁎	(2)⁎⁎	(3)	(4)
Local SARS incidence
–	− 0.14⁎⁎⁎	− 0.082⁎⁎⁎	− 0.29⁎⁎⁎	− 0.19⁎⁎⁎
	(0.029)	(0.029)	(0.054)	(0.055)
× N	− 0.050	− 0.0037	− 0.10	− 0.018
	(0.040)	(0.039)	(0.076)	(0.074)
National SARS incidence
–	− 0.87⁎⁎⁎	− 0.76⁎⁎⁎	− 3.80⁎⁎⁎	− 3.62⁎⁎⁎
	(0.16)	(0.16)	(0.64)	(0.62)
× N	− 1.29⁎⁎⁎	− 0.85⁎⁎⁎	− 3.52⁎⁎⁎	− 2.16⁎⁎⁎
	(0.21)	(0.20)	(0.82)	(0.79)
Change in peer visits
–		0.13⁎⁎⁎		0.13⁎⁎⁎
		(0.0040)		(0.0040)
× SARS		0.015		0.015
		(0.0093)		(0.0093)
× N		0.043⁎⁎⁎		0.043⁎⁎⁎
		(0.0058)		(0.0058)
× SARS × N		0.084⁎⁎⁎		0.085⁎⁎⁎
		(0.012)		(0.012)
Lagged individual visits	0.16⁎⁎⁎	0.16⁎⁎⁎	0.16⁎⁎⁎	0.16⁎⁎⁎
	(0.0040)	(0.0040)	(0.0040)	(0.0040)
Observations	79.7 mil	79.7 mil	79.7 mil	79.7 mil
R2	0.10	0.10	0.10	0.10

Note: Standard errors appear in parentheses. Standard errors are clustered by the patient's modal township. Individual visits are observed at time t, lagged individual visits are observed at time t − 26, and all other regressors are observed from time t to t − 2. SARS connotes Quarters 2–4 of 2003. N indicates that the person is a non-mover. All regressions include network × N, year, and two-week period fixed effects.

^⁎p < 0.1.

^⁎⁎p < 0.05.

^⁎⁎⁎p < 0.01.