Beware of scams impersonating Jump Trading Group. We only communicate through our official accounts.
- Firedancer
- Thinking
- Connect
Machine Learning for Compliance-Weighted Wald Estimators
Nihar ShahDominic Coey
In experiments with noncompliance, some subjects assigned to the treatment group may be unlikely to actually receive treatment, presenting difficulties for obtaining precise estimates of treatment effects. In this study, we seek to address the issue by introducing a class of compliance-weighted Wald estimators in which those unlikely to be compliers are downweighted and observe the effects.
