Skip to main content
. Author manuscript; available in PMC: 2010 Aug 27.
Published in final edited form as: J Am Stat Assoc. 2010 Jun 1;105(490):683–691. doi: 10.1198/jasa.2010.tm09302.

Table 2.

Simulation results for covariates-dependent censoring.

β01 = −1.0 β02 = 1.0

Model Design Prop. Bias MSE SE/SD CP Bias MSE SE/SD CP
ζ = 0 I 20% 0.004 0.059 0.955 0.944 0.025 0.159 0.942 0.940
40% −0.026 0.087 0.946 0.938 0.027 0.227 0.943 0.946
II 20% −0.059 0.073 0.943 0.930 0.080 0.174 0.963 0.928
40% −0.065 0.102 0.939 0.916 0.050 0.244 0.947 0.956
Avg. MCSE 0.012 0.005 0.032 0.020 0.014 0.032

ζ = 1 I 20% −0.022 0.168 0.927 0.934 0.047 0.444 0.938 0.938
40% −0.045 0.189 0.961 0.930 0.018 0.536 0.936 0.936
II 20% −0.055 0.184 0.927 0.918 0.043 0.478 0.929 0.938
40% −0.065 0.195 0.957 0.936 0.048 0.531 0.945 0.936
Avg. MCSE 0.019 0.012 0.030 0.031 0.030 0.028

ζ = 0.5 I 20% −0.027 0.113 0.925 0.932 0.018 0.288 0.937 0.934
40% −0.036 0.129 0.972 0.946 0.026 0.369 0.936 0.934
II 20% −0.060 0.123 0.936 0.940 0.043 0.306 0.940 0.934
40% −0.068 0.139 0.976 0.944 0.048 0.364 0.967 0.940
Avg. MCSE 0.016 0.008 0.031 0.026 0.020 0.030

SD, sample standard deviation; SE, mean of estimated standard error via resampling method; CP, empirical coverage probability of 95% confidence intervals; MSE, mean squared error of estimates; Prop.: censoring proportion; Avg. MCSE, averaged Monte Carlo standard errors of Bias, MSE and SE/SD for each model.