Kyle F Butts

Results 35 comments of Kyle F Butts

BTW, probably need to make a few more changes to make it faster, so I'll push and circle back when I think it's in a good state!

@lrberge I think we are good to merge! Some dev notes for HC2/HC3 implementation: 1. Small-sample adjustments are a bit awkward with HC2/HC3 standard errors (since they are a form...

@grantmcdermott I think memory usage will be about the same in most cases. `model.matrix.fixest` will return just a single vector for fixed effects (so not the matrix of 0/1s). That's...

Figured out a faster way to implement this using block formula for projection matrix: https://en.wikipedia.org/wiki/Projection_matrix#Blockwise_formula - `cov.iid` already has solved for (Xtilde' Xtilde)^{-1} where Xtilde is demeaned by fixed effects...

@lrberge I put a couple things in this one PR. Want me rebase onto `main`? I'll also adjust the work with data.table commits based on your changes - [b992cd7](https://github.com/lrberge/fixest/pull/418/commits/b992cd7b09907ee09cac007e3ee5c372956e4396) is...

The marginal effects package is very versatile for this FYI: https://vincentarelbundock.github.io/marginaleffects/articles/hypothesis.html#hypothesis-tests-with-the-delta-method

@mikenguyen13 I think that you're right that `sunab()` can't take multiple cohorts at once. If you're willing to make assumptions (parallel trends for all combinations of treatment and additive separability...

I think what's causing the error is how `reframe` is being called under the hood and how that interacts with how `feols` objects are being transformed in `broom::tidy()`. There's no...

@coatless you can use a Quarto [`pre-render`/`post-render`](https://quarto.org/docs/projects/scripts.html#pre-and-post-render) command for this, storing your script in `altdoc/`. `altdoc` will set up `_quarto/` with all the files and then call `quarto::quarto_render()` on the...

@hsantanna88 Have you seen this paper? It also proposes a bias correction method, https://eml.berkeley.edu/~pkline/papers/KSS2020.pdf. With PR #418, the `hatvalue` function produce, $P_{ii}$, which would let do a bias correction a...