Table 1.

1. Table showing how the 4 methods presented in this paper change the resulting estimates of the local effect ${\widehat{\beta }}^c$. The reported estimates are the health effect times 10³, ${10}^3{\widehat{\beta }}^c$, corresponding to percent increase in mortality per increase in 10 units of PM₁₀.

	Local		Adjusted		Shrunken		Full Bayes
	${\widehat{\beta }}^c$	SE	${\widehat{\beta }}^c$	SE	${\widehat{\beta }}^c$	SE	${\widehat{\beta }}^c$	SE
Syracuse	3.18	1.56	0.72	1.60	0.82	0.31	1.41	0.98
Boston	2.50	1.27	0.72	1.31	0.83	0.31	1.35	0.88
Providence	2.03	1.23	0.72	1.27	0.80	0.31	1.21	0.83
Jersey City	1.25	0.75	0.72	0.81	0.80	0.29	1.03	0.60
Baltimore	0.40	0.62	0.72	0.69	0.66	0.28	0.55	0.52
Newark	0.23	1.05	0.72	1.09	0.68	0.30	0.56	0.72
Philadelphia	0.09	0.66	0.72	0.73	0.61	0.28	0.38	0.56
Washington	0.01	1.53	0.72	1.56	0.69	0.31	0.58	0.87
Kingston	−1.20	2.19	0.72	2.21	0.68	0.31	0.45	1.02
Arlington	−1.60	5.96	0.72	5.97	0.72	0.31	0.71	1.17
Richmond	−2.24	2.74	0.72	2.76	0.68	0.31	0.42	1.09
Pooled Estimate	0.72	0.32	0.72	0.31	0.72	0.32	0.79	0.50