Skip to main content
. Author manuscript; available in PMC: 2014 Apr 24.
Published in final edited form as: Int J Stat Med Res. 2014 Jan 31;3(1):32–43. doi: 10.6000/1929-6029.2014.03.01.5

Table 3.

Estimates of log(OR) in Simulation III

P* Method β1 = 0.8 β1 =0
Bias ESE ASE§ RMSE CP(%)** Err(%)††
0.1 Full 0.022 0.149 0.144 0.151 0.95 5.4
Half DL 0.022 0.149 0.144 0.151 0.95 5.4
Deletion 0.025 0.155 0.149 0.157 0.94 5.6
MI 0.003 0.153 0.163 0.153 0.96 4.6
ROS −0.034 0.144 0.144 0.148 0.95 4.0
Bayesian −0.024 0.147 0.144 0.149 0.95 5.6
0.3 Full 0.020 0.147 0.143 0.148 0.95 4.5
Half DL 0.023 0.147 0.143 0.149 0.95 4.6
Deletion 0.026 0.164 0.163 0.166 0.96 4.5
MI −0.042 0.162 0.194 0.168 0.97 2.4
ROS −0.185 0.131 0.139 0.226 0.73 2.4
Bayesian −0.107 0.139 0.136 0.175 0.87 4.6
0.5 Full 0.021 0.143 0.143 0.144 0.95 5.2
Half DL 0.034 0.145 0.144 0.149 0.95 4.7
Deletion 0.034 0.197 0.188 0.200 0.95 5.3
MI −0.102 0.192 0.230 0.217 0.96 2.7
ROS −0.348 0.101 0.125 0.362 0.18 1.7
Bayesian −0.189 0.121 0.127 0.225 0.68 5.5
*

P is the proportion of DL

Bias is defined as β-β

ESE is the empirical standard error

§

ASE is the average standard error

RMSE is the root mean square error

**

CP is the coverage probability for 95% HPD

††

Err is the type I error rate under null hypothesis.