Skip to main content
. Author manuscript; available in PMC: 2016 Feb 5.
Published in final edited form as: Psychol Aging. 2015 Sep 21;30(4):911–929. doi: 10.1037/pag0000046

Table 8.

Relative Efficiency, N = 500, Simulation B

Listwise
Log
CART
RF
MI
30% 50% 30% 50% 30% 50% 30% 50% 30% 50%
X¯2
 Linear 1.020 .924   .985   .984 1.022 1.038 1.057 1.340 .954 .816
 One split   .855 .930 1.158   .961 1.010 1.021 2.135 1.395 .813 .811
 Two splits   .924 .823   .925   .99   .990   .997 1.177 1.650 .805 .764
 Three splits   .912 .858   .892   .824 1.017   .975 1.056 1.102 .825 .727
Y¯2
 Linear   .972 .907   .945   .893 1.024   .985 1.020 1.226 .945 .728
 One split   .833 .979 1.297   .869   .982   .942 1.813 1.187 .729 .691
 Two splits   .920 .887   .809   .815   .960   .982 1.205 1.597 .749 .669
 Three splits   .921 .955   .854   .793 1.019   .937 1.019 1.272 .801 .714
σX22
 Linear   .925 .819   .952 1.042 1.034 1.055 1.037 1.301 .889 .785
 One split   .752 .931 1.166 1.048 1.022 1.019 1.843 1.268 .735 .85
 Two splits   .858 .769   .916   .993 1.089 1.037 1.294 1.408 .867 .725
 Three splits   .792 .855   .900   .956 1.096 1.049 1.138 1.096 .792 .830
σY22
 Linear   .971 .971   .985 1.176 1.030 1.083   .976 1.206 .809 .814
 One split   .694 .897 1.834 1.179   .983   .953 1.167 1.070 .631 .678
 Two splits   .736 .681   .794   .921 1.023   .962 1.173 1.015 .729 .610
 Three splits   .757 .773   .909   .948 1.008   .998 1.011 1.098 .707 .649
σX2,Y2
 Linear   .936 .887   .965 1.128 1.022 1.116 1.023 1.331 .848 .774
 One split   .695 .900 1.575 1.098 1.016   .962 1.576 1.077 .689 .723
 Two splits   .789 .727   .835   .939 1.037   .975 1.151 1.218 .794 .635
 Three splits   .733 .838   .853   .967 1.026 1.036 1.001 1.108 .741 .685

Note. Relative efficiency ratios computed over SDPrune. Log = logistic regression; CART = classification and regression trees; RF = random forests; MI = multiple imputation.