Repository navigation
Keep variance for identical-stat treatment observations (#5304) - #5304
Closed
saitcakmak wants to merge 1 commit into
Closed
saitcakmak wants to merge 1 commit into
saitcakmak wants to merge 1 commit into
Conversation
meta-cla
Bot
added
the
CLA Signed
|
@saitcakmak has exported this pull request. If you are a Meta employee, you can view the originating Diff in D126118253. |
Summary: ## What Changed Use status-quo arm identity for the forward shortcut in nonconstant-control `Relativize`, matching the existing vectorized path. Equal treatment/control means and SEMs do not establish that they are the same estimate. With the default zero covariance, a distinct treatment's delta-method relative variance is `2 * SEM**2 / mean**2`, not zero. This narrowly scoped follow-up to D126108875 addresses review comment 3. It changes one forward guard, adds actual observation/vectorized parity tests for positive/negative controls and zero/positive SEMs, and retains the treatment round trip. Actual status quo variance stays zero. Constant-control behavior, vectorized formulas, and inverse formulas/restoration remain unchanged. ___ landed-with-radar-review Differential Revision: D126118253
meta-codesync
Bot
changed the title
saitcakmak
force-pushed
the
export-D126118253
branch
from
October 19, 2026 20:34
92f584b to
3752e7f
Compare
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #5304 +/- ##
=======================================
Coverage 96.59% 96.60%
=======================================
Files 621 621
Lines 71054 71075 +21
=======================================
+ Hits 68638 68659 +21
Misses 2416 2416 ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
|
This pull request has been merged in be103fa. |
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
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.
Summary:
What Changed
Use status-quo arm identity for the forward shortcut in nonconstant-control
Relativize, matching the existing vectorized path. Equal treatment/control means and SEMs do not establish that they are the same estimate. With the default zero covariance, a distinct treatment's delta-method relative variance is2 * SEM**2 / mean**2, not zero.This narrowly scoped follow-up to D126108875 addresses review comment 3. It changes one forward guard, adds actual observation/vectorized parity tests for positive/negative controls and zero/positive SEMs, and retains the treatment round trip. Actual status quo variance stays zero. Constant-control behavior, vectorized formulas, and inverse formulas/restoration remain unchanged.
landed-with-radar-review
Differential Revision: D126118253