Re-visiting the relationship between neighbourhood environment and BMI: an instrumental variables approach to correcting for residential selection bias

Table 4 Parameter estimates of the alternative structural model specifications for the risk of being overweight/obese

Independent variables	Instrumental variables odds ratios (95% CI)	Logistic regression odds ratios (95% CI)
AGE^a	1.09	1.09
AGE^a	(1.07–1.10)	(1.07–1.10)
AGE_SQ^a	1.01	1.01
AGE_SQ^a	(1.00–1.00)	(1.00–1.00)
EDUC	0.89	0.89
EDUC	(0.87–0.91)	(0.87–0.91)
MARRIED	1.11	1.14
MARRIED	(0.99-1.25)	(1.01–1.28)
BMIYR	1.04	1.04
BMIYR	(1.03–1.06)	(1.03–1.06)
SIBBMI	2.37	2.38
SIBBMI	(2.23–2.53)	(2.23–2.53)
RESWALK	0.90	1.00
RESWALK	(0.83–0.97)	(0.96–1.04)
McFadden R²	0.08	0.08

^aAge is centered before creating age squared so as to avoid any possible collinearity problems associated with entering both age and age squared as regressors.

ISSN: 1479-5868