How to extract short stacking weights from the dr did estimator when using the ddml wrapper

[cgriffin480]

I am using your Diff-in-Diff with Double/Debiased Machine Learning tutorial as a guide to imbed ML estimators in the did function. I've been able to produce estimates with the learners I want. However, I'm having difficulty finding a way to extract the short stacking weights for the learners. Is it possible to extract the short stacking weights when the ddml functions are imbedded in the agg_gt() function? Thanks!

[thomaswiemann]

The current implementations do not allow for extraction of the weights out of the box.

We're currently working on better support for the did package in https://github.com/thomaswiemann/did/tree/dev-ddml

In the meantime, you can try the following (non-ideal) solution: Initialize a global object (e.g., a list) before running the att_gt() command and then change the wrapper for the ddml_att() command to write the stacking weights to this global object.

[cgriffin480]

Hey Thomas, thanks for the advice! I'll try that out.