Flexible covariate adjustments in RDD
Fuzzy Design
The simulations are based on the make_simple_rdd_data-DGP with \(2000\) observations. The simulation considers data under a fuzzy regression discontinuity design.
Metadata
DoubleML Version 0.9.dev0
Script rdd_fuzzy_coverage.py
Date 2025-01-09 20:41:20
Total Runtime (seconds) 19885.911859
Python Version 3.12.8
Method | Learner g | Learner m | fs specification | Bias | CI Length | Coverage |
---|---|---|---|---|---|---|
rdflex | Global linear | Global linear | cutoff | 2.536 | 12.547 | 0.977 |
rdflex | Global linear | LGBM | cutoff | 2.608 | 12.744 | 0.982 |
rdflex | Global linear | Linear | cutoff | 2.585 | 12.716 | 0.977 |
rdflex | Global linear | Stacked | cutoff | 2.610 | 12.539 | 0.972 |
rdflex | LGBM | Global linear | cutoff | 0.465 | 2.413 | 0.968 |
rdflex | LGBM | LGBM | cutoff | 0.503 | 2.431 | 0.972 |
rdflex | LGBM | Linear | cutoff | 0.484 | 2.475 | 0.991 |
rdflex | LGBM | Stacked | cutoff | 0.472 | 2.399 | 0.968 |
rdflex | Linear | Global linear | cutoff | 2.529 | 12.549 | 0.977 |
rdflex | Linear | LGBM | cutoff | 2.618 | 12.794 | 0.977 |
rdflex | Linear | Linear | cutoff | 2.591 | 12.777 | 0.977 |
rdflex | Linear | Stacked | cutoff | 2.489 | 12.374 | 0.977 |
rdflex | Stacked | Global linear | cutoff | 0.499 | 2.495 | 0.968 |
rdflex | Stacked | LGBM | cutoff | 0.463 | 2.371 | 0.991 |
rdflex | Stacked | Linear | cutoff | 0.451 | 2.503 | 0.982 |
rdflex | Stacked | Stacked | cutoff | 0.445 | 2.401 | 0.972 |
rdflex | Global linear | Global linear | cutoff and score | 2.543 | 12.537 | 0.977 |
rdflex | Global linear | LGBM | cutoff and score | 2.619 | 12.850 | 0.982 |
rdflex | Global linear | Linear | cutoff and score | 2.580 | 12.744 | 0.977 |
rdflex | Global linear | Stacked | cutoff and score | 2.618 | 12.899 | 0.977 |
rdflex | LGBM | Global linear | cutoff and score | 0.495 | 2.557 | 0.982 |
rdflex | LGBM | LGBM | cutoff and score | 0.528 | 2.617 | 0.968 |
rdflex | LGBM | Linear | cutoff and score | 0.477 | 2.610 | 0.977 |
rdflex | LGBM | Stacked | cutoff and score | 0.506 | 2.535 | 0.982 |
rdflex | Linear | Global linear | cutoff and score | 2.575 | 12.653 | 0.977 |
rdflex | Linear | LGBM | cutoff and score | 2.678 | 12.900 | 0.986 |
rdflex | Linear | Linear | cutoff and score | 2.565 | 12.749 | 0.977 |
rdflex | Linear | Stacked | cutoff and score | 2.625 | 12.811 | 0.972 |
rdflex | Stacked | Global linear | cutoff and score | 0.495 | 2.588 | 0.972 |
rdflex | Stacked | LGBM | cutoff and score | 0.495 | 2.557 | 0.968 |
rdflex | Stacked | Linear | cutoff and score | 0.501 | 2.629 | 0.991 |
rdflex | Stacked | Stacked | cutoff and score | 0.488 | 2.674 | 0.982 |
rdflex | Global linear | Global linear | interacted cutoff and score | 2.515 | 12.522 | 0.977 |
rdflex | Global linear | LGBM | interacted cutoff and score | 2.643 | 12.944 | 0.982 |
rdflex | Global linear | Linear | interacted cutoff and score | 2.562 | 12.701 | 0.977 |
rdflex | Global linear | Stacked | interacted cutoff and score | 2.581 | 12.557 | 0.972 |
rdflex | LGBM | Global linear | interacted cutoff and score | 0.496 | 2.544 | 0.977 |
rdflex | LGBM | LGBM | interacted cutoff and score | 0.534 | 2.670 | 0.982 |
rdflex | LGBM | Linear | interacted cutoff and score | 0.520 | 2.622 | 0.986 |
rdflex | LGBM | Stacked | interacted cutoff and score | 0.522 | 2.529 | 0.986 |
rdflex | Linear | Global linear | interacted cutoff and score | 2.593 | 12.699 | 0.977 |
rdflex | Linear | LGBM | interacted cutoff and score | 2.723 | 13.189 | 0.982 |
rdflex | Linear | Linear | interacted cutoff and score | 2.600 | 12.847 | 0.977 |
rdflex | Linear | Stacked | interacted cutoff and score | 2.629 | 12.776 | 0.972 |
rdflex | Stacked | Global linear | interacted cutoff and score | 0.528 | 2.647 | 0.972 |
rdflex | Stacked | LGBM | interacted cutoff and score | 0.512 | 2.629 | 0.977 |
rdflex | Stacked | Linear | interacted cutoff and score | 0.513 | 2.662 | 0.963 |
rdflex | Stacked | Stacked | interacted cutoff and score | 0.483 | 2.594 | 0.986 |
rdrobust | linear | linear | cutoff | 2.523 | 12.356 | 0.972 |
Method | Learner g | Learner m | fs specification | Bias | CI Length | Coverage |
---|---|---|---|---|---|---|
rdflex | Global linear | Global linear | cutoff | 2.536 | 10.530 | 0.945 |
rdflex | Global linear | LGBM | cutoff | 2.608 | 10.695 | 0.945 |
rdflex | Global linear | Linear | cutoff | 2.585 | 10.672 | 0.940 |
rdflex | Global linear | Stacked | cutoff | 2.610 | 10.523 | 0.949 |
rdflex | LGBM | Global linear | cutoff | 0.465 | 2.025 | 0.935 |
rdflex | LGBM | LGBM | cutoff | 0.503 | 2.040 | 0.931 |
rdflex | LGBM | Linear | cutoff | 0.484 | 2.077 | 0.954 |
rdflex | LGBM | Stacked | cutoff | 0.472 | 2.013 | 0.908 |
rdflex | Linear | Global linear | cutoff | 2.529 | 10.531 | 0.945 |
rdflex | Linear | LGBM | cutoff | 2.618 | 10.737 | 0.945 |
rdflex | Linear | Linear | cutoff | 2.591 | 10.723 | 0.935 |
rdflex | Linear | Stacked | cutoff | 2.489 | 10.385 | 0.945 |
rdflex | Stacked | Global linear | cutoff | 0.499 | 2.094 | 0.926 |
rdflex | Stacked | LGBM | cutoff | 0.463 | 1.990 | 0.949 |
rdflex | Stacked | Linear | cutoff | 0.451 | 2.100 | 0.940 |
rdflex | Stacked | Stacked | cutoff | 0.445 | 2.015 | 0.935 |
rdflex | Global linear | Global linear | cutoff and score | 2.543 | 10.521 | 0.945 |
rdflex | Global linear | LGBM | cutoff and score | 2.619 | 10.784 | 0.945 |
rdflex | Global linear | Linear | cutoff and score | 2.580 | 10.695 | 0.945 |
rdflex | Global linear | Stacked | cutoff and score | 2.618 | 10.825 | 0.949 |
rdflex | LGBM | Global linear | cutoff and score | 0.495 | 2.146 | 0.954 |
rdflex | LGBM | LGBM | cutoff and score | 0.528 | 2.196 | 0.940 |
rdflex | LGBM | Linear | cutoff and score | 0.477 | 2.190 | 0.945 |
rdflex | LGBM | Stacked | cutoff and score | 0.506 | 2.128 | 0.963 |
rdflex | Linear | Global linear | cutoff and score | 2.575 | 10.618 | 0.931 |
rdflex | Linear | LGBM | cutoff and score | 2.678 | 10.826 | 0.954 |
rdflex | Linear | Linear | cutoff and score | 2.565 | 10.699 | 0.926 |
rdflex | Linear | Stacked | cutoff and score | 2.625 | 10.752 | 0.935 |
rdflex | Stacked | Global linear | cutoff and score | 0.495 | 2.172 | 0.926 |
rdflex | Stacked | LGBM | cutoff and score | 0.495 | 2.146 | 0.949 |
rdflex | Stacked | Linear | cutoff and score | 0.501 | 2.206 | 0.945 |
rdflex | Stacked | Stacked | cutoff and score | 0.488 | 2.244 | 0.959 |
rdflex | Global linear | Global linear | interacted cutoff and score | 2.515 | 10.509 | 0.945 |
rdflex | Global linear | LGBM | interacted cutoff and score | 2.643 | 10.863 | 0.949 |
rdflex | Global linear | Linear | interacted cutoff and score | 2.562 | 10.659 | 0.949 |
rdflex | Global linear | Stacked | interacted cutoff and score | 2.581 | 10.538 | 0.945 |
rdflex | LGBM | Global linear | interacted cutoff and score | 0.496 | 2.135 | 0.935 |
rdflex | LGBM | LGBM | interacted cutoff and score | 0.534 | 2.240 | 0.945 |
rdflex | LGBM | Linear | interacted cutoff and score | 0.520 | 2.201 | 0.931 |
rdflex | LGBM | Stacked | interacted cutoff and score | 0.522 | 2.123 | 0.945 |
rdflex | Linear | Global linear | interacted cutoff and score | 2.593 | 10.657 | 0.935 |
rdflex | Linear | LGBM | interacted cutoff and score | 2.723 | 11.068 | 0.949 |
rdflex | Linear | Linear | interacted cutoff and score | 2.600 | 10.782 | 0.949 |
rdflex | Linear | Stacked | interacted cutoff and score | 2.629 | 10.722 | 0.931 |
rdflex | Stacked | Global linear | interacted cutoff and score | 0.528 | 2.222 | 0.926 |
rdflex | Stacked | LGBM | interacted cutoff and score | 0.512 | 2.206 | 0.935 |
rdflex | Stacked | Linear | interacted cutoff and score | 0.513 | 2.234 | 0.917 |
rdflex | Stacked | Stacked | interacted cutoff and score | 0.483 | 2.177 | 0.949 |
rdrobust | linear | linear | cutoff | 2.523 | 10.370 | 0.940 |