Table 3 Average Percentage Differences Between the Estimated Treatment Effects and True Treatment Effects from the Causal Forest Algorithm within the Generalized Random Forests Application (CFA-GRF) Under Fully Observed Heterogeneity Across Simulated Populations Which Differ by the Extent That Treatment Effect Influences Treatment Choice
A | B | C | D | E | F | G | H | I | J | |
|---|---|---|---|---|---|---|---|---|---|---|
Simulation | Proportion of true (TEi) influencing (ETEi) at Treatment Choice – (Kj)a | % of Treatment Choice Variation Explained by (TEi)b | % of Patients Overlappedc | Average Percentage Difference Between True and Estimated Treatment Effects | ||||||
Full Population | Treated | Untreated | Overlapped with Fully Observed Heterogeneity | Non-Overlapped with Fully Observed Heterogeneity | ||||||
Treated | Untreated | Treated | Untreated | |||||||
1 | 0 | .0006 | 100 | -0.56% | -0.64% | -0.44% | ||||
2 | .10 | .18 | 100 | -0.36% | -0.39% | -0.28% | ||||
3 | .20 | 1.4 | 100 | 0.56% | 0.38% | 0.76% | ||||
4 | .30 | 5.3 | 100 | -0.84% | -1.12% | -0.54% | ||||
5 | .40 | 11.9 | 100 | 0.48% | -1.21% | 2.83% | ||||
6 | .50 | 20.1 | 100 | -0.24% | -3.55% | 4.75% | ||||
7 | .60 | 27.8 | 97.0 | 2.76% | -1.84% | 10.31% | -1.15% | 8.62% | -14.98% | 11110.00% |
8 | .70 | 34.5 | 90.7 | -0.96% | -9.78% | 14.51% | -7.07% | 8.35% | -26.27% | 508.00% |
9 | .80 | 39.8 | 84.5 | 2.84% | -7.74% | 22.07% | -3.13% | 11.48% | -23.85% | 293.41% |
10 | .90 | 44.3 | 78.3 | 0.08% | -11.66% | 22.11% | -5.53% | 8.79% | -26.11% | 192.88% |
11 | 1.00 | 48.0 | 68.8 | 0.84% | -14.74% | 30.99% | -5.33% | 10.47% | -28.58% | 159.22% |