Skip to main content
. Author manuscript; available in PMC: 2011 Dec 1.
Published in final edited form as: Genet Epidemiol. 2010 Dec;34(8):803–815. doi: 10.1002/gepi.20527

TABLE III.

Simulation results for studying the effect of an untyped SNP on the risk of disease under the case-control design

β MLE
 IMP-DOS
 IMP-MLG
Bias SE SEE CP PW Bias SE SEE CP PW Bias SE SEE CP PW
S1 0.0 0.001 0.100 0.099 0.989 0.011 0.001 0.101 0.101 0.990 0.010 0.001 0.082 0.082 0.990 0.010
0.3 0.000 0.103 0.101 0.990 0.651 −0.010 0.101 0.100 0.989 0.633 −0.093 0.082 0.081 0.922 0.483
0.6 0.000 0.108 0.106 0.989 0.999 −0.031 0.102 0.099 0.984 0.999 −0.190 0.082 0.081 0.580 0.993
0.9 −0.008 0.115 0.113 0.986 1.00 −0.070 0.104 0.100 0.961 1.00 −0.297 0.083 0.082 0.156 1.00
S2 0.0 −0.002 0.091 0.090 0.991 0.009 −0.001 0.093 0.092 0.989 0.011 −0.001 0.073 0.072 0.988 0.012
0.3 0.000 0.085 0.084 0.989 0.844 0.012 0.094 0.091 0.987 0.805 −0.067 0.075 0.072 0.934 0.753
0.6 −0.002 0.084 0.082 0.990 1.00 0.046 0.103 0.092 0.971 1.00 −0.117 0.083 0.073 0.757 1.00
0.9 −0.009 0.084 0.083 0.989 1.00 0.093 0.115 0.094 0.919 1.00 −0.157 0.097 0.075 0.566 1.00
S3 0.0 −0.002 0.106 0.106 0.992 0.008 −0.001 0.109 0.108 0.989 0.011 0.000 0.082 0.082 0.989 0.011
0.3 −0.001 0.099 0.098 0.991 0.693 0.012 0.110 0.105 0.987 0.651 −0.067 0.081 0.079 0.953 0.641
0.6 0.000 0.096 0.096 0.990 1.00 0.047 0.120 0.103 0.969 1.00 −0.116 0.082 0.078 0.841 1.00
0.9 −0.003 0.096 0.097 0.989 1.00 0.096 0.136 0.103 0.915 1.00 −0.153 0.087 0.078 0.682 1.00
S4 0.0 0.000 0.073 0.073 0.990 0.010 0.000 0.075 0.074 0.989 0.011 0.000 0.071 0.070 0.989 0.011
0.3 0.001 0.075 0.074 0.990 0.929 0.002 0.078 0.077 0.990 0.913 −0.016 0.074 0.073 0.986 0.909
0.6 0.000 0.077 0.077 0.989 1.00 0.007 0.082 0.081 0.990 1.00 −0.028 0.077 0.077 0.984 1.00
0.9 −0.006 0.082 0.081 0.990 1.00 0.011 0.088 0.087 0.989 1.00 −0.039 0.083 0.082 0.979 1.00
S5 0.0 0.000 0.079 0.079 0.989 0.011 0.000 0.082 0.081 0.989 0.011 0.000 0.072 0.071 0.989 0.011
0.3 0.001 0.086 0.087 0.990 0.837 −0.005 0.085 0.084 0.988 0.830 −0.040 0.074 0.074 0.977 0.830
0.6 0.002 0.103 0.102 0.989 1.00 −0.023 0.093 0.088 0.980 1.00 −0.092 0.078 0.077 0.912 1.00
0.9 0.009 0.130 0.127 0.988 1.00 −0.051 0.102 0.093 0.961 1.00 −0.154 0.082 0.081 0.744 1.00
S6 0.0 0.000 0.096 0.097 0.991 0.009 0.000 0.100 0.100 0.991 0.009 0.000 0.095 0.095 0.991 0.009
0.3 0.001 0.102 0.103 0.990 0.634 0.000 0.105 0.106 0.989 0.603 −0.022 0.100 0.101 0.989 0.578
0.6 −0.002 0.112 0.112 0.989 0.999 −0.007 0.114 0.114 0.990 0.997 −0.049 0.108 0.108 0.981 0.996
0.9 −0.004 0.122 0.121 0.989 1.00 −0.014 0.125 0.123 0.987 1.00 −0.077 0.118 0.117 0.964 1.00
S7 0.0 0.000 0.109 0.108 0.991 0.009 0.000 0.112 0.111 0.992 0.008 0.000 0.112 0.111 0.992 0.008
0.3 −0.001 0.103 0.102 0.989 0.632 −0.001 0.106 0.106 0.991 0.601 −0.002 0.106 0.105 0.991 0.599
0.6 0.000 0.097 0.098 0.990 1.00 0.001 0.102 0.102 0.990 1.00 −0.002 0.101 0.101 0.990 1.00
0.9 −0.003 0.096 0.095 0.989 1.00 −0.001 0.101 0.100 0.989 1.00 −0.005 0.100 0.099 0.989 1.00
S8 0.0 0.000 0.070 0.069 0.989 0.011 0.000 0.072 0.071 0.989 0.011 0.000 0.071 0.070 0.989 0.011
0.3 0.000 0.066 0.067 0.990 0.972 0.002 0.069 0.070 0.990 0.962 −0.003 0.068 0.069 0.989 0.962
0.6 −0.001 0.066 0.066 0.990 1.00 0.004 0.071 0.070 0.990 1.00 −0.006 0.069 0.069 0.989 1.00
0.9 −0.006 0.066 0.066 0.990 1.00 0.005 0.072 0.071 0.990 1.00 −0.008 0.070 0.070 0.991 1.00
S9 0.0 0.000 0.069 0.069 0.988 0.012 0.000 0.071 0.071 0.989 0.011 0.000 0.071 0.070 0.989 0.011
0.3 0.000 0.067 0.067 0.992 0.971 0.000 0.069 0.069 0.991 0.960 −0.001 0.069 0.069 0.991 0.960
0.6 −0.002 0.066 0.066 0.991 1.00 −0.001 0.069 0.069 0.992 1.00 −0.004 0.069 0.069 0.992 1.00
0.9 −0.007 0.067 0.066 0.989 1.00 −0.002 0.072 0.071 0.989 1.00 −0.007 0.071 0.070 0.988 1.00

S1–S9 denote the nine scenarios listed in Table I. Bias and SE are the bias and standard error of the parameter estimator. SEE is the mean of the standard error estimator. CP is the coverage probability of the 99% confidence interval. PW is the type I error/power for testing zero parameter value at the 0.01 nominal significance level. Each entry is based on 10,000 replicates.