Skip to main content
. Author manuscript; available in PMC: 2016 May 17.
Published in final edited form as: Demogr Res. 2015 May 17;32:1081–1098. doi: 10.4054/DemRes.2015.32.39

Table 4. Effect of weighting and propensity scoring on mean body mass index and prevalence of obesity (panel A) and mean waist circumference and prevalence of central obesity (panel B), ages 24-32 Years.

Panel A: BMI and Obesity

Add Health [Home], Wave IV (2008) NHANES [Clinic] (2007-2008) Odds of obesity in Add Health vs. NHANES

n=13762
BMI ≥ 30 BMI ≥ 30 Crude Adjusted

n BMI % (95% CI) n BMI % (95% CI) OR (95% CI) AORb (95% CI)
Unweighted 13049 29 37 (36, 38) 713 28 31 (28, 35) 1.3 (1.1, 1.5) 1.4 (1.1, 1.6)
Weighted 13049 29 36 (35, 38) 713 28 29 (24, 35) 1.4 (1.1, 1.8) 1.4 (1.1, 1.8)
Propensity-scoreda 13049 29 36 (34, 38) 713 28 29 (23, 37) 1.4 (1.0, 1.9) 1.4 (1.1, 1.8)

Panel B: Waist Circumference and Central Obesity

Add Health [Home], Wave IV (2008) NHANES [Clinic] (2007-2008) Odds of Central Obesity in Add Health vs. NHANES

n=13828
Central Obesity Central Obesity Crude Adjusted

n Waist % (95% CI) n Waist % (95% CI) OR (95% CI) AORb (95% CI)

Unweighted 13138 98 48 (47, 49) 690 94 41 (37, 45) 1.3 (1.1, 1.6) 1.4 (1.2, 1.7)
Weighted 13138 98 47 (45, 49) 690 94 38 (32, 44) 1.5 (1.1, 1.9) 1.5 (1.2, 2.0)
Propensity-scoreda 13138 98 47 (45, 49) 690 94 37 (30, 45) 1.5 (1.1, 2.0) 1.5 (1.2, 1.9)
a

Adjusted for the predicted probability of being in the Add Health (versus NHANES) population conditional on age, sex, race/ethnicity, nativity, education, and income

b

Logistic regression model included all above-listed covariates