Table 2

	θ⁽⁰⁾ =0.35
	Complete-case MLE
Mean	0.475	0.154	0.004	0.425	0.251	0.017	−0.145	0.407
Median	0.473	0.154	0.004	0.426	0.252	0.014	−0.144	0.411
MAD	0.184	0.134	0.18	0.166	0.209	0.204	0.218	0.212
Emp. SE	0.185	0.131	0.178	0.164	0.214	0.217	0.221	0.207
Est. SE	0.19	0.134	0.182	0.166	0.209	0.217	0.221	0.204
Bias	0.125	0.004	0.004	−0.075	−0.099	−0.133	−0.145	−0.093
B. Score	30.22	1.249	1.066	−20.372	−20.787	−27.353	−29.367	−20.149
RMSE	0.223	0.131	0.178	0.18	0.236	0.254	0.265	0.227
CP	0.91	0.954	0.96	0.93	0.917	0.911	0.9	0.92
MLE
Mean	0.474	0.153	0.006	0.428	0.252	0.012	−0.148	0.402
Median	0.474	0.15	0.005	0.428	0.252	0.008	−0.148	0.402
MAD	0.139	0.102	0.133	0.12	0.171	0.172	0.185	0.17
Emp. SE	0.144	0.101	0.136	0.126	0.171	0.176	0.182	0.168
Est. SE	0.147	0.102	0.139	0.126	0.171	0.178	0.181	0.166
Bias	0.124	0.003	0.006	−0.072	−0.098	−0.138	−0.148	−0.098
B. Score	38.546	1.148	1.919	−25.553	−25.665	−35.047	−36.248	−26.070
RMSE	0.19	0.101	0.136	0.146	0.197	0.223	0.235	0.194
CP	0.881	0.959	0.954	0.91	0.909	0.886	0.864	0.902
Pseudo-conditional Likelihood Method
Mean	0.381	0.16	0.001	0.476	0.327	0.119	−0.039	0.478
Median	0.379	0.156	−0.000	0.473	0.33	0.115	−0.039	0.479
MAD	0.156	0.108	0.138	0.131	0.177	0.173	0.19	0.178
Emp. SE	0.157	0.108	0.139	0.132	0.179	0.18	0.191	0.177
Est. SE	0.164	0.111	0.141	0.165	0.192	0.188	0.188	0.208
Bias	0.031	0.01	0.001	−0.024	−0.023	−0.031	−0.039	−0.022
B. Score	8.836	4.061	0.426	−8.187	−5.687	−7.594	−9.079	−5.497
RMSE	0.16	0.108	0.139	0.135	0.181	0.183	0.195	0.178
CP	0.952	0.964	0.957	0.962	0.952	0.955	0.947	0.963

Table 2: Simulation results for the complete-case MLE, the MLE, and the pseudo-conditional likelihood method. Here MAD, Emp. SE, Est. SE, Bias, B. Score, RMSE, and CP denote median absolute deviation, empirical standard error, estimated standard error, root mean squared error, bias, bias score, root mean squared error, and 95% coverage probability based on the Wald-type confidence intervals, respectively. The results were based on 2,000 runs. There were 2×3×5 = 30 disease subtypes. The model for the intercepts was misspecified. The missingness probabilities depended on the covariate. This is Scenario 2b.