Skip to main content
. Author manuscript; available in PMC: 2014 May 15.
Published in final edited form as: Stat Med. 2010 Nov 10;29(25):2592–2604. doi: 10.1002/sim.4016

Figure 1.

Figure 1

Percentage bias (PB) and average length (AL) vs. selection probability for the FAWE and the MI methods of the Cox Model. Baseline hazard function is 1, β = (−ln(2), ln(2)), where Zm, Zo ∼ N(0, 1), with correlation coefficient of 0.6; uniform censoring time with censoring rate 65%; and selection probability π(X, δ, Zo) = (1 + exp(a + bδ + cX + dZo))−1, where a,b,c,d were selected to generate 30% to 80% of cohort members to have missing Zm; the cohort size is 250. For FAWE, both π and E were obtained based on (X, δ, Zo) using the Nadaraya-Watson estimator with normal kernel and bandwidth h = 4σWn−1/3. For MI, imputation models contained (X, δ, Zo).