. Author manuscript; available in PMC: 2012 Jul 1.

Published in final edited form as: Health Serv Outcomes Res Methodol. 2011 Jul 1;11(1-2):54–78. doi: 10.1007/s10742-011-0071-9

Table 4.

Effects of Cigarette Smoking before pregnancy on OFC under Alternative Model and Instrument Specifications

Model	Probit Regression Coefficient	Odds Ratio [95% CI]	1st IV condition X²/F [df]	2nd IV condition X² [df]	Exogenous Cigarettes X² [1]
Excluding OTHER_BEHAVIOR
Probit	0.013^* (0.007)	1.21^* [0.99,1.51]
CML-IV probit – Instrument Specification 1	0.099^*** (0.030)	4.2^** [1.01,9.13]	24.52^*** [8]	3.08 [7]; p=0.88	5.33^**
Two stage IV probit – Instrument Specification 1	0.11^** (0.04)	3.82^** [1.09,7.13]	3.3^*** [8, 829]		4.98^**
CML-IV probit under alternative instrument specifications
Instrument Specification 2	0.11^*** (0.03)	4.73^** [1.13,8.17]	21.8^*** [5]	2.45 [4]; p = 0.65	6.24^**
Instrument Specification 3	0.11^*** (0.03)	4.59^* [0.94,8.19]	19.19^*** [6]	2.88 [5]; p= 0.72	5.00^**
Instrument Specification 4	0.11^*** (0.03)	4.82^* [0.96,8.32]	17.42^*** [4]	2.45 [3]; p= 0.48	5.14^**

Note: The models are based on a sample size of 856 observations. Standard errors of regression coefficients are in parentheses. The 95% confidence bounds of the odds ratio estimates are in brackets. The first IV condition column reports a chisquare test of the joint significance of the instruments in the cigarettes function (degrees of freedom [df] in brackets). The second IV condition reports the over-identification chi-square test (df in brackets).

p < 0.1,

^**

p < 0.05,

^***

p < 0.01.