Double Robust, Flexible Adjustment Methods for Causal Inference: An Overview and an Evaluation
Double robust methods for flexible covariate adjustment in causal inference have proliferated in recent years. Despite their apparent advantages, these methods are rarely used by social scientists. It is also unclear whether these methods actually outperform more traditional methods in finite sample...
Main Author: | |
---|---|
Format: | Other/Unknown Material |
Language: | unknown |
Published: |
Center for Open Science
2023
|
Subjects: | |
Online Access: | http://dx.doi.org/10.31235/osf.io/dzayg |
id |
crcenteros:10.31235/osf.io/dzayg |
---|---|
record_format |
openpolar |
spelling |
crcenteros:10.31235/osf.io/dzayg 2024-09-15T18:03:53+00:00 Double Robust, Flexible Adjustment Methods for Causal Inference: An Overview and an Evaluation Hoffmann, Nathan Isaac 2023 http://dx.doi.org/10.31235/osf.io/dzayg unknown Center for Open Science https://creativecommons.org/licenses/by-nd/4.0/legalcode posted-content 2023 crcenteros https://doi.org/10.31235/osf.io/dzayg 2024-08-29T04:08:53Z Double robust methods for flexible covariate adjustment in causal inference have proliferated in recent years. Despite their apparent advantages, these methods are rarely used by social scientists. It is also unclear whether these methods actually outperform more traditional methods in finite samples. This paper has two aims: It is a guide to some of the latest methods in double robust, flexible covariate adjustment using machine learning, and it compares these methods to more traditional statistical methods and flexible "single robust" methods. It does this by using both simulated data where the treatment effect estimate is known, and then by replicating some prominent articles in sociology that use simpler methods. Double robust methods covered include Augmented Inverse Probability Weighting (AIPW), Targeted Maximum Likelihood Estimation (TMLE), and Double/Debiased Machine Learning (DML). Results suggest that some of these methods do outperform traditional methods in a wide range of simulations, but only slightly. In particular, the top performers are TMLE and AIPW in conjunction with flexible machine learning estimators. But G-computation with the same flexible estimators obtains almost identical results, and simple regression methods have only slightly higher bias and are much more computationally efficient. In the article replications, the application of double robust methods substantively changes some, but not all, of the results, highlighting the importance of comparing performance from multiple estimators. Other/Unknown Material DML COS Center for Open Science |
institution |
Open Polar |
collection |
COS Center for Open Science |
op_collection_id |
crcenteros |
language |
unknown |
description |
Double robust methods for flexible covariate adjustment in causal inference have proliferated in recent years. Despite their apparent advantages, these methods are rarely used by social scientists. It is also unclear whether these methods actually outperform more traditional methods in finite samples. This paper has two aims: It is a guide to some of the latest methods in double robust, flexible covariate adjustment using machine learning, and it compares these methods to more traditional statistical methods and flexible "single robust" methods. It does this by using both simulated data where the treatment effect estimate is known, and then by replicating some prominent articles in sociology that use simpler methods. Double robust methods covered include Augmented Inverse Probability Weighting (AIPW), Targeted Maximum Likelihood Estimation (TMLE), and Double/Debiased Machine Learning (DML). Results suggest that some of these methods do outperform traditional methods in a wide range of simulations, but only slightly. In particular, the top performers are TMLE and AIPW in conjunction with flexible machine learning estimators. But G-computation with the same flexible estimators obtains almost identical results, and simple regression methods have only slightly higher bias and are much more computationally efficient. In the article replications, the application of double robust methods substantively changes some, but not all, of the results, highlighting the importance of comparing performance from multiple estimators. |
format |
Other/Unknown Material |
author |
Hoffmann, Nathan Isaac |
spellingShingle |
Hoffmann, Nathan Isaac Double Robust, Flexible Adjustment Methods for Causal Inference: An Overview and an Evaluation |
author_facet |
Hoffmann, Nathan Isaac |
author_sort |
Hoffmann, Nathan Isaac |
title |
Double Robust, Flexible Adjustment Methods for Causal Inference: An Overview and an Evaluation |
title_short |
Double Robust, Flexible Adjustment Methods for Causal Inference: An Overview and an Evaluation |
title_full |
Double Robust, Flexible Adjustment Methods for Causal Inference: An Overview and an Evaluation |
title_fullStr |
Double Robust, Flexible Adjustment Methods for Causal Inference: An Overview and an Evaluation |
title_full_unstemmed |
Double Robust, Flexible Adjustment Methods for Causal Inference: An Overview and an Evaluation |
title_sort |
double robust, flexible adjustment methods for causal inference: an overview and an evaluation |
publisher |
Center for Open Science |
publishDate |
2023 |
url |
http://dx.doi.org/10.31235/osf.io/dzayg |
genre |
DML |
genre_facet |
DML |
op_rights |
https://creativecommons.org/licenses/by-nd/4.0/legalcode |
op_doi |
https://doi.org/10.31235/osf.io/dzayg |
_version_ |
1810441330480906240 |