Phase 3: Survey design support for OLS-based standalone estimators#226

Merged

igerber merged 24 commits into

mainfrom

survey-enhancements

Mar 22, 2026

Commits on Mar 22, 2026

Add Phase 3 survey design support for 6 OLS-based estimators

igerber
and
claude
committed
Fix 5 P1 issues from PR #226 review (round 1)

igerber
and
claude
committed
Fix 4 P1 issues from PR #226 review (round 2)

igerber
and
claude
committed
Fix weighted-mass IF linearization in ContinuousDiD and TripleDifference from PR #226 review (round 3)

igerber
and
claude
committed
Fix ContinuousDiD bread normalization and panel-level survey metadata from PR #226 review (round 4)

igerber
and
claude
committed
Fix EfficientDiD fweight consistency and BaconDecomposition index safety from PR #226 review (round 5)

igerber
and
claude
committed
Fix ContinuousDiD fweight TSL scaling and BaconDecomposition empty-cell guard from PR #226 review (round 6)

igerber
and
claude
committed
Pass survey df to SunAbraham aggregate inference from PR #226 review (round 7)

igerber
and
claude
committed
Add P3 regression tests: SunAbraham survey-df and DoseResponseCurve p-values from PR #226 review (round 8)

igerber
and
claude
committed
Fix BaconDecomposition exact-weight survey unit_share and add regression test from PR #226 review (round 9)

igerber
and
claude
committed
Fix DoseResponseCurve survey df and strengthen P3 tests from PR #226 review (round 10)

igerber
and
claude
committed
Reject EfficientDiD covariates + survey_design combination from PR #226 review (round 11)

igerber
and
claude
committed
Add survey test coverage for TripleDiff covariates and EfficientDiD aggregations from PR #226 review (round 12)

igerber
and
claude
committed
Document EfficientDiD covariates+survey deferral in REGISTRY.md from PR #226 review (round 13)

igerber
and
claude
committed
Add within-unit survey validation to BaconDecomposition and regression test from PR #226 review (round 14)

igerber
and
claude
committed
Defer ContinuousDiD anticipation event-study filtering to TODO.md from PR #226 local review

igerber
and
claude
committed
Add survey resolution pattern inconsistency to TODO.md from PR #226 local review

igerber
and
claude
committed
Add duplicated survey summary formatting to TODO.md from PR #226 local review

igerber
and
claude
committed
Thread survey_design through bacon_decompose() and triple_difference() convenience functions from PR #226 review (round 15)

igerber
and
claude
committed
Fix SunAbraham survey weight floor in overall ATT aggregation from PR #226 review (round 16)

igerber
and
claude
committed
Reject fweight for StackedDiD survey and add regression test from PR #226 review (round 17)

igerber
and
claude
committed
Fix TripleDifference RA nuisance linearization to include survey weights in score terms from PR #226 review (round 18)

igerber
and
claude
committed
Relax Bacon survey validation to exact-only and document fweight exclusions from PR #226 review (round 19)

igerber
and
claude
committed
Also reject aweight for StackedDiD survey — only pweight supported from PR #226 review (round 20)

igerber
and
claude
committed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Phase 3: Survey design support for OLS-based standalone estimators#226

Phase 3: Survey design support for OLS-based standalone estimators#226
igerber merged 24 commits into
mainfrom
survey-enhancements

Commits on Mar 22, 2026

Add Phase 3 survey design support for 6 OLS-based estimators

Fix 5 P1 issues from PR #226 review (round 1)

Fix 4 P1 issues from PR #226 review (round 2)

Fix weighted-mass IF linearization in ContinuousDiD and TripleDifference from PR #226 review (round 3)

Fix ContinuousDiD bread normalization and panel-level survey metadata from PR #226 review (round 4)

Fix EfficientDiD fweight consistency and BaconDecomposition index safety from PR #226 review (round 5)

Fix ContinuousDiD fweight TSL scaling and BaconDecomposition empty-cell guard from PR #226 review (round 6)

Pass survey df to SunAbraham aggregate inference from PR #226 review (round 7)

Add P3 regression tests: SunAbraham survey-df and DoseResponseCurve p-values from PR #226 review (round 8)

Fix BaconDecomposition exact-weight survey unit_share and add regression test from PR #226 review (round 9)

Fix DoseResponseCurve survey df and strengthen P3 tests from PR #226 review (round 10)

Reject EfficientDiD covariates + survey_design combination from PR #226 review (round 11)

Add survey test coverage for TripleDiff covariates and EfficientDiD aggregations from PR #226 review (round 12)

Document EfficientDiD covariates+survey deferral in REGISTRY.md from PR #226 review (round 13)

Add within-unit survey validation to BaconDecomposition and regression test from PR #226 review (round 14)

Defer ContinuousDiD anticipation event-study filtering to TODO.md from PR #226 local review

Add survey resolution pattern inconsistency to TODO.md from PR #226 local review

Add duplicated survey summary formatting to TODO.md from PR #226 local review

Thread survey_design through bacon_decompose() and triple_difference() convenience functions from PR #226 review (round 15)

Fix SunAbraham survey weight floor in overall ATT aggregation from PR #226 review (round 16)

Reject fweight for StackedDiD survey and add regression test from PR #226 review (round 17)

Fix TripleDifference RA nuisance linearization to include survey weights in score terms from PR #226 review (round 18)

Relax Bacon survey validation to exact-only and document fweight exclusions from PR #226 review (round 19)

Also reject aweight for StackedDiD survey — only pweight supported from PR #226 review (round 20)