. Author manuscript; available in PMC: 2016 Apr 28.

Published in final edited form as: Can J Econ. 2013 Aug 20;46(3):791–810. doi: 10.1111/caje.12039

TABLE 2.

Effects of comtamination during first birth on the probability of moving before second birth

Interaction	High school			Hispanic	Over 35	Teen mom	Smoking
Interaction	None	or less	Black	Hispanic	Over 35	Teen mom	Smoking
A: Any chem.	0.0098	0.0248⁺	0.0133	0.0176	0.0049	0.0120	0.0094
	(0.0076)	(0.0137)	(0.0103)	(0.0108)	(0.0076)	(0.0083)	(0.0077)
* Mom has CHAR		− 0.0266⁺	− 0.0176	− 0.0211	0.0474^**	− 0.0244	0.0048
		(0.0142)	(0.0225)	(0.0154)	(0.0152)	(0.0182)	(0.0188)
B: Any contam.	0.0123⁺	0.0265^**	0.0155⁺	0.0178^**	0.0088	0.0139^**	0.0120⁺
	(0.0066)	(0.0108)	(0.0086)	(0.0089)	(0.0066)	(0.0069)	(0.0066)
* Mom has CHAR		− 0.0276^**	− 0.0174	− 0.0167	0.0307^**	− 0.0198	0.0042
		(0.0123)	(0.0206)	(0.0147)	(0.0146)	(0.0177)	(0.0167)
No. observations	189705	189705	189705	189705	189705	189705	189705

NOTES: The sample consists of women who had two births in New Jersey between 1997–2007. The dependent variable is an indicator equal to 1 if the mother moved between the first and second births in the sample. The key independent variable is contamination during the first pregnancy. Exposure to contaminants during pregnancy is instrumented with potential exposure during the 39 weeks following conception. All regressions include year*month fixed effects and the following indicators measured at first birth: Mother’s age (19–24, 25–34, 35+, missing); mother’s race: (non-Hispanic white, non-Hispanic black, Hispanic, missing); mother’s education (<12, 12, some college, college+, missing); mother smoked; risk factors for the pregnancy parity; mother married, married missing, child male, weather controls. Robust standard errors are clustered on the water system.

⁺

p < 0.10,

^**

p < 0.05,

^***

p < 0.001