Causal Inference for Aggregated Treatment
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909983559385088 |
|---|---|
| author | Caetano, Carolina Caetano, Gregorio Callaway, Brantly Dyal, Derek |
| author_facet | Caetano, Carolina Caetano, Gregorio Callaway, Brantly Dyal, Derek |
| contents | In this paper, we study causal inference when the treatment variable is an aggregation of multiple sub-treatment variables. Researchers often report marginal causal effects for the aggregated treatment, implicitly assuming that the target parameter corresponds to a well-defined average of sub-treatment effects. We show that, even in an ideal scenario for causal inference such as random assignment, the weights underlying this average have some key undesirable properties: they are not unique, they can be negative, and, holding all else constant, these issues become exponentially more likely to occur as the number of sub-treatments increases and the support of each sub-treatment grows. We propose approaches to avoid these problems, depending on whether or not the sub-treatment variables are observed. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_22885 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Causal Inference for Aggregated Treatment Caetano, Carolina Caetano, Gregorio Callaway, Brantly Dyal, Derek Econometrics In this paper, we study causal inference when the treatment variable is an aggregation of multiple sub-treatment variables. Researchers often report marginal causal effects for the aggregated treatment, implicitly assuming that the target parameter corresponds to a well-defined average of sub-treatment effects. We show that, even in an ideal scenario for causal inference such as random assignment, the weights underlying this average have some key undesirable properties: they are not unique, they can be negative, and, holding all else constant, these issues become exponentially more likely to occur as the number of sub-treatments increases and the support of each sub-treatment grows. We propose approaches to avoid these problems, depending on whether or not the sub-treatment variables are observed. |
| title | Causal Inference for Aggregated Treatment |
| topic | Econometrics |
| url | https://arxiv.org/abs/2506.22885 |