wber-replication-package

Replication Package & Data Policy (wber-replication-package)

When to trigger

• The paper is empirical and headed toward acceptance (or you are building the package early)
• You need to write the data availability statement and the DOI repository link for the title page
• Some data are restricted, proprietary, or governed by a national statistics office and you must plan the deposit around that
• An R&R asks you to make data and code available
• You want to know what "condition of publication" actually requires here

Why this is a WBER signature

WBER's most distinctive, stable policy is transparency as a condition of publication: authors must, where ethically and legally possible, publicly release all data and the software code underlying a published paper. By repute WBER is the only development field journal that routinely publishes data and code, and this fits the broader World Bank reproducible-research push. The manuscript must carry a data availability statement — on the title page / author-affiliation footnote — with a persistent (preferably DOI) link to the data in a repository (检索于 2026-06；以官网为准; verify exact wording on the OUP author guidelines). Build the package as you analyze; do not treat it as an acceptance-day formality.

What the package must contain

Handling restricted developing-country data

This is the WBER-specific hard part — much development data is governed by national statistics offices, ministries, or data providers (LSMS, DHS, census, tax records):

• You cannot post restricted microdata, but you must still deposit all code and document the exact access procedure: provider, application steps, any cost, approximate wait, and the data version/round used.
• Provide a small synthetic or public extract so the code runs end-to-end and a verifier can check the logic where the real data cannot be shared.
• For DHS/LSMS-type data, cite the specific survey round and version and the registration step a replicator must complete; do not assume "DHS data" is self-locating.
• Declare any restricted-data or exemption situation at the earliest opportunity, not at the final check — arrangements are at editor discretion.

Reproducibility hygiene (build as you go)

• Pin versions: Stata version + recorded ssc/net package versions; requirements.txt/conda env (Python); renv.lock (R).
• Set and report seeds for every simulation, bootstrap, and randomization-inference step.
• No absolute paths — one root macro/variable; relative paths thereafter.
• Exhibit-to-code map in the README: Table 3 → code/05_main.do, Figure 2 → code/06_event_study.R.
• Run it clean on a fresh checkout before depositing.
• Currency/PPP transparency: document deflators, PPP conversion factors, and exchange-rate sources — replicators in development work need these to reproduce real-value results.

Adapt the vendored skeleton in ../../resources/code/ (master → clean → descriptive → DiD/IV/RD → mechanism → robustness → tables) as the package backbone.

Checklist

• One run_all master script regenerates every table and figure from raw data
• All data deposited (or restricted-data access path documented + synthetic/public extract provided)
• README has a complete exhibit-to-code map and computational requirements
• Data Availability Statement on the title page with a persistent DOI repository link
• Specific survey rounds/versions (DHS/LSMS/census/admin) cited with the registration step
• Software/package versions pinned; seeds set and reported; no absolute paths
• Deflators / PPP / exchange-rate sources documented
• Restricted-data / exemption situation declared early; field instruments included

Anti-patterns

• Treating data/code release as optional or as an acceptance-day task — it is a condition of publication
• A title page with no data availability statement or no persistent DOI link
• "DHS/LSMS data" cited with no round, version, or registration path (not reproducible)
• Restricted data with no documented access procedure and no synthetic extract
• Absolute paths, unrecorded package versions, or unset seeds that break reproduction
• Undocumented deflators/PPP factors so real-value results cannot be regenerated

Worked vignette (illustrative)

A poverty paper uses LSMS consumption data (restricted, registration-gated) plus a public price index. The team cannot post the LSMS microdata, so the package contains: every cleaning and analysis script; a README naming the specific LSMS country, round, and version and the exact registration steps a replicator must complete; a small synthetic extract with the same variable structure so run_all executes end-to-end; documented deflators and PPP factors; and pinned Stata package versions with seeds set for the bootstrap. The title page carries a data availability statement with a DOI link to the openICPSR-style deposit holding the code and synthetic data. Because the public price index can be shared, it is deposited in full. The package would let a verifier reproduce the logic, and reproduce the numbers entirely once they obtain LSMS access.

Output format

【Master script】run_all regenerates all exhibits from raw? [Y/N]
【Data】all deposited, or restricted path + synthetic extract? [state]
【DAS + DOI】title-page statement with persistent repository link? [Y/N]
【Survey provenance】DHS/LSMS/admin round + version + access step cited? [Y/N]
【Reproducibility】versions pinned + seeds + no absolute paths + clean run? [Y/N]
【Currency】deflators/PPP/FX documented? [Y/N]
【Restricted/exemption】declared early? [Y/N/NA]
【Next step】wber-referee-strategy

Supplementary resources

• ../../resources/code/ — reproducible Stata + Python skeleton to adapt as the package backbone
• ../../resources/official-source-map.md — official OUP / WBER URLs behind the data policy