Journal of Environmental and Public Health

Research Article

Estimating the Effectiveness of Health-Risk Communications with Propensity-Score Matching: Application to Arsenic Groundwater Contamination in Four US Locations

Table 1

Summary statistics for the unmatched data by sample type: pooled, nonaverters (POUFILTER = 0), and averters (POUFILTER = 1) among survey respondents in regions with arsenic contamination of groundwater.


Variable			Nonaverters		Averters
	Pooled		POUFILTER = 0		POUFILTER = 1
	Mean	Std. dev.	Mean	Std. dev.	Mean	Std. dev.

POUFILTER	0.485	0.501	0.000	0.000	1.000	0.000

PRIORKNOW	0.624	0.485	0.562	0.498	0.690	0.464
FRIEND	0.094	0.292	0.051	0.221	0.140	0.348
UTILITY	0.248	0.433	0.255	0.438	0.240	0.429
MEDIA	0.158	0.365	0.175	0.382	0.140	0.348
OTHERSOURCE	0.211	0.408	0.161	0.368	0.264	0.442

AGE	50.169	15.357	51.314	16.339	48.953	14.202
SCHOOLING	14.165	2.488	14.146	2.585	14.186	2.391
INCOME	68,224	37,100	64,708	36,387	71,957	37,625
HASCHILDREN	0.421	0.495	0.372	0.485	0.473	0.501

NM	0.143	0.351	0.226	0.420	0.054	0.227
NV	0.320	0.467	0.255	0.438	0.388	0.489
OK	0.241	0.428	0.314	0.466	0.163	0.371
WI	0.297	0.458	0.204	0.405	0.395	0.491

Number of observations	266		137		129