Repository navigation
Add the article "The cd model in practice" - #5
Merged
Merged
Conversation
docs/articles/cd-results.qmd documents how to use cd, its Proposition 99 results, and four disagreements: - cd vs sar: both models are fitted on all three samples (the 39-state 1988 panel and Cao & Dowd's 51/41-unit panels) to separate model from sample. The sign of Nevada's spillover is the model's on every sample; the size of California's effect depends on the model and, for cd, on Nevada's weight in California's synthetic control. - The paper's inference vs the authors' code: rejection years and interval shapes. - The kappa_A test rejecting the paper's structure: which states carry the unexplained residual, larger structures, and leave-one-out references. - The structure-selection statistic vs mlsynth's: same rankings here, different definitions and scales. The sar fits on the Cao-Dowd panels use rook-contiguity weights built by tools/build_cd_contiguity.py from the same Census TIGER/Line 2024 shapefile and rule as the bundled 39-state weights. The script checks that its 39-state subset reproduces them exactly. The weights live with the article (docs/articles/data/), not in the package. Co-Authored-By: Claude Opus 5.5 <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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
A new website article,
docs/articles/cd-results.qmd, in the Articles menu. It covers thecdmodel's usage and its Proposition 99 results, then four disagreements, with every number computed live.cdvssar: model or sample? Both models are fitted on all three samples, using new 51-unit weights forsar(described below).cdand negative undersaron every sample, so the sign comes from the model.sareverywhere, and −2.8 / −9.4 / −10.1 undercd. Thecdspread follows Nevada's weight in California's synthetic control: 0.25 / 0.07 / 0.16.New spatial weights.
tools/build_cd_contiguity.pybuilds rook contiguity for 50 states + DC from the Census TIGER/Line 2024 shapefile, the same source and rule as the bundled 39-state weights. It self-checks that its 39-state subset reproduces the bundled weights exactly. The output lives with the article (docs/articles/data/), and there is no package API change.The model page, the catalogue and the case study link to the article, and the CHANGELOG and
llms.txtare updated.Validation
make docsrenders without errors.make lint typecheck test: 306 passed, 3 skipped.🤖 Generated with Claude Code