Skip to main content
. 2015 Mar 18;18(1):40–66. doi: 10.1111/ectj.12039

Table 6.

Results for designs with unequal expected unit sizes

n=1500, p=0.3 n=1000, p=0.1
bias rmse bias rmse
Design 1
D 0.022 0.032 0.083 0.091
D bc −0.003 0.027 0.018 0.050
D dc −0.010 0.029 0.009 0.047
Design 2
D 0.027 0.035 0.094 0.100
D bc −0.002 0.027 0.016 0.049
D dc −0.009 0.028 0.005 0.045
Rejection frequencies for tests of H0:D pop =0.292
Nominal size 0.10 0.05 0.10 0.05
Design 1
W 0.192 0.116 0.596 0.457
W pb 0.106 0.056 0.200 0.119
T pb 0.162 0.100 0.175 0.100
W dc 0.135 0.078 0.127 0.071
Design 2
W 0.246 0.154 0.726 0.596
W pb 0.115 0.064 0.167 0.096
T pb 0.161 0.095 0.167 0.095
W dc 0.134 0.075 0.134 0.076

J=50, D pop =0.292; 10,000 Monte Carlo replications, 599 bootstrap repetitions.