Table 2.
Parameter | Method | Mean | SD | Mean SE | Power | |
---|---|---|---|---|---|---|
|
Oracle | 0.357 | 0.129 | 0.120 | 82.5 | |
Cond select | 0.334 | 0.137 | 0.126 | 73.3 | ||
MV‐IVW‐PCA | 0.299 | 0.126 | 0.118 | 70.8 | ||
MV‐LIML‐PCA | 0.349 | 0.150 | 0.130 | 76.3 | ||
|
Oracle | −0.012 | 0.133 | 0.121 | 7.2 | |
Cond select | −0.015 | 0.140 | 0.126 | 7.4 | ||
MV‐IVW‐PCA | −0.018 | 0.126 | 0.119 | 5.8 | ||
MV‐LIML‐PCA | −0.009 | 0.148 | 0.131 | 7.1 | ||
|
Oracle | −0.543 | 0.130 | 0.121 | 98.7 | |
Cond select | −0.515 | 0.142 | 0.127 | 96.5 | ||
MV‐IVW‐PCA | −0.475 | 0.130 | 0.119 | 95.9 | ||
MV‐LIML‐PCA | −0.539 | 0.151 | 0.131 | 96.6 |
Note: Results from oracle, conditional selection, MV‐IVW‐PCA, and MV‐LIML‐PCA methods across 1000 simulated datasets: mean estimates, standard deviation (SD) of estimates, mean standard error (mean SE) of estimates, and empirical power of the 95% confidence interval to estimate , , and .