技能 数据科学 政策分析研究可复现性指南

政策分析研究可复现性指南

v20260724
jpam-transparency-and-data
本指南为政策分析期刊投稿作者准备的必备材料。它详细阐述了如何确保研究结果的可复现性,核心要求包括在公共仓库存档数据和代码、构建主脚本生成所有图表,以及处理受限数据的完整流程和说明。
获取技能
271 次下载
概览

Transparency & Replication Data (jpam-transparency-and-data)

JPAM publishes program evaluations that often inform real spending decisions, so the evidence must be reproducible. The journal expects authors to archive the data and code behind the reported results in a suitable public repository, with a clear data-availability statement. Build the package as you go so it does not stall acceptance — and plan early for the restricted administrative data common in policy work. Verify the current wording in Wiley Authors / Research Exchange before upload (检索于 2026-06-20;以官网为准).

When to trigger

  • Building the replication package (data + code + documentation)
  • A manuscript is heading toward acceptance and materials are requested
  • Data are restricted (administrative, IRB-protected, provider-licensed) and you need the exemption path
  • Writing the data-availability statement

What to prepare

  1. Public deposit. Place the replication materials in a recognized repository (e.g., the project's ICPSR/openICPSR archive, Harvard Dataverse, OSF, or a journal-designated repository) with a persistent identifier — not a personal website or transient cloud link. Recent JPAM data-availability statements often use a JPAM/Harvard Dataverse repository, while Research Exchange controls the exact repository prompt.
  2. Reproduce every reported number. A master script regenerates every table and figure from the raw/constructed data. Exhibit numbers in the manuscript match the package output exactly.
  3. Documentation. A README covering data provenance, construction steps, software/package versions, seeds for stochastic steps, and the exact command to reproduce each exhibit.
  4. Data-availability statement. State what is shared, where, and under what license; if data are restricted, state precisely why and how a replicator can obtain access.

Restricted administrative data (the policy-research case)

Policy evaluation often runs on linked administrative or survey microdata that cannot be posted. JPAM's transparency expectation is met honestly by:

  • Explaining the restriction (legal, IRB, data-provider license) in the data-availability statement.
  • Providing access instructions — the application process, provider contact, and approximate timeline so an independent replicator could obtain the same data.
  • Posting all code plus any shareable derived/aggregated files and, where feasible, synthetic or simulated data that let the code run end-to-end.

Build-as-you-go checklist

  • One master script regenerates every table and figure
  • README: provenance, construction, versions, seeds, per-exhibit reproduction steps
  • Software/package versions pinned (renv.lock / requirements.txt / recorded installs)
  • Manuscript exhibit numbers match the package output exactly
  • Public repository with a persistent identifier chosen (journal-designated where specified)
  • Data-availability statement drafted
  • Restricted data: restriction explained + access path + synthetic data where feasible

Anti-patterns

  • Treating the package as a post-acceptance afterthought (it can gate publication)
  • Depositing code that does not actually reproduce the printed tables/figures
  • A personal URL or expiring cloud link instead of a persistent repository
  • Claiming data are restricted with no access path or synthetic substitute
  • Undocumented, un-seeded, unpinned code that only "works on my machine"

Calibration anchors (hedged)

  • Policy evaluations lean heavily on restricted administrative data; an honest restriction note plus a real access path and runnable code on synthetic/derived data is the expected, accepted route — not a loophole to skip transparency.
  • The exact deposit requirement, repository prompt, and reproduction-check workflow can change; confirm the current data-policy wording in Wiley Authors / Research Exchange.
  • Build the package alongside the analysis: retrofitting reproducibility after acceptance is where policy papers stall.

Worked micro-example (illustrative)

A welfare-reform evaluation uses linked state UI and TANF records that cannot be posted. The package still meets the bar: a data-availability statement explains the licensing restriction and names the state agency's data-request process and timeline; all code is deposited with a master script; a synthetic dataset matching the variable structure lets a replicator run the full pipeline; and shareable aggregated tables are included. An independent researcher could obtain the real data and reproduce every number. (Illustrative.)

Output format

【Repository】public archive + persistent ID — chosen? [Y/N]
【Reproduces results?】master script verified locally? [Y/N]
【Documentation】README + provenance + seeds + pinned versions? [Y/N]
【Restricted data?】restriction explained + access path + synthetic data?
【Data-availability statement】drafted? [Y/N]
【Next】jpam-review-process

Supplementary resources

信息
Category 数据科学
Name jpam-transparency-and-data
版本 v20260724
大小 5.68KB
更新时间 2026-07-28
语言