Fix dof + improve collinearity detection#279
Merged
eloualiche merged 2 commits intoJun 4, 2026
Merged
Conversation
matthieugomez
force-pushed
the
fix-dof-and-collinearity
branch
from
f79c15d to
c4f209b
Compare
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #279 +/- ##
=======================================
Coverage ? 97.97%
=======================================
Files ? 8
Lines ? 693
Branches ? 0
=======================================
Hits ? 679
Misses ? 14
Partials ? 0 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
matthieugomez
force-pushed
the
fix-dof-and-collinearity
branch
from
c4f209b to
4fb651f
Compare
matthieugomez
changed the title
Main fix: residual dof no longer subtracts the intercept twice — it is now N-K (and G-1 for clustered standard errors), matching Stata. The standard errors themselves were already correct. Also: - Collinearity detection is scale-invariant, so small-scale independent regressors are no longer wrongly dropped. - dof_fes ignores continuous-FE groups (fe(id)&x) whose interaction is all zero. - coeftable labels the confidence-interval columns with the requested level. - partial_out: deprecate method=:gpu and default tol to 1e-6 (so its residuals match reg). - Clearer errors (malformed formulas), new docstrings (fe, has_iv, has_fe, dof_fes, FixedEffectModel, collinear-omitted NaN convention), and small README fixes. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
matthieugomez
force-pushed
the
fix-dof-and-collinearity
branch
from
4fb651f to
7bde4e4
Compare
HC2/HC3 compute leverage from the partialled-out (demeaned) regressors only, omitting the absorbed fixed-effect contribution to leverage. This silently understates leverage and yields anti-conservative standard errors whenever fe() is used. Throw an informative ArgumentError for the combination, mirroring the existing IV guard (HC2/HC3 are likewise rejected for IV), and point users to Vcov.robust() (HC1) or Vcov.cluster(). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Fix degrees-of-freedom reporting and improve API/docs
Main fix: residual dof no longer subtracts the intercept twice — it is now
N-K (and G-1 for clustered standard errors), matching Stata. The standard
errors themselves were already correct.
Also:
regressors are no longer wrongly dropped.
match reg).
dof_fes, FixedEffectModel, collinear-omitted NaN convention), and small
README fixes.
🤖 Generated with Claude Code