Skip to main content
. 2010 Jul;11(3):572–582. doi: 10.1093/biostatistics/kxq007

Table 3.

Estimates of confounding bias (Δ) and nonlinearity effect (Δnl) in the low birth–weight study. Smoking confounds the log odds ratio relating uterine irritability and low birth–weight. Simple (5.1) and standardization (3.4) estimates are shown along with 95% CI based on bootstrap resampling. IPW estimates are identical to the standardization estimates

Inline graphic (CI) Inline graphicnl (CI)
Simple − 0.03 (− 0.21, 0.14)
Standardization − 0.07 (− 0.27, 0.09) 0.03 (− 0.02, 0.12)