Transparency & Data (eursr-transparency-and-data)
ESR's transparency expectations are stricter than a sharing norm: a Data Availability Statement
(DAS) is required for every manuscript, and for submissions received on or after 1 January 2025,
authors using statistical or computational methods must deposit a replication package as a condition
of publication (assembled by acceptance, typically required at conditional acceptance). Qualitative-
data work is exempt from the package requirement. Confirm the current wording and effective dates on
the live OUP/ESR page — this skill states the policy as verified in 2026-06; treat specifics as
volatile.
When to trigger
- Writing the Data Availability Statement (needed at submission)
- Assembling the replication package ahead of conditional acceptance
- Deciding what can be shared given confidentiality, DUA, or proprietary constraints (register data)
- A reviewer or editor asked how the analysis can be reproduced
What the ESR policy requires (verify current wording)
-
Data Availability Statement in every article — describing how the data can be accessed and the
conditions for access (open, controlled/enclave, on-request, or restricted).
-
Replication package for statistical/computational papers, due at conditional acceptance: the code,
the constructed/analysis data (or a documented access path when redistribution is barred), and
documentation sufficient to regenerate every reported result.
-
Exemption: research using qualitative data (interviews, participant observation) is not required
to submit a replication package.
-
Restricted data path: where register/administrative data (e.g., national statistical offices,
SOEP, EU-SILC scientific-use files) cannot be redistributed, share all code plus a precise access
route (provider, file version, DUA), so a qualified researcher could reproduce the results.
Build a package that reproduces (do this from the start)
-
One master script regenerates every table/figure from the (constructed) data, in order.
-
Set and report seeds for imputation, bootstrap, and MCMC; pin versions (
renv.lock,
requirements.txt, recorded ssc/net installs).
-
Archive the harmonization/recoding code (ISCED/CASMIN, ISCO/ISEI/EGP) — it is part of the result.
-
README documenting data sources and versions, run order, expected outputs, runtime, and any
access restrictions.
- Deposit in a citable repository (e.g., OSF, Zenodo, GESIS) and reference it in the DAS.
Transparency posture by data type (ESR)
| Data type |
In the package |
Restricted |
DAS framing |
| Public comparative survey (ESS, EVS) |
data + code |
none |
"openly available from [archive], version X" |
| EU-SILC / SOEP scientific-use file |
code + constructed-vars script |
raw microdata |
"available from [provider] under its access terms" |
| National register / administrative |
all code + access route |
raw records |
"accessible via [NSO/enclave] under DUA; code provided" |
| Qualitative (exempt) |
analytic documentation (optional) |
identifiable transcripts |
state the exemption + confidentiality basis |
Worked micro-example (illustrative)
A comparative scarring paper uses public ESS plus a restricted national register linkage.
DAS: "ESS Round data are openly available from the ESS Data Archive (edition cited). The linked
register data are accessible to qualified researchers via [national statistical office] under a data-
use agreement; all analysis and harmonization code is provided in the replication package."
Package (for conditional acceptance): master.R + harmonization scripts + constructed analysis file for
the ESS portion; full code for the register portion with a documented access path; seed = 2026;
renv.lock pinned; README with run order and runtime; deposited on Zenodo with a DOI.
Statistical method → not exempt → package required.
The posture shares what supports the claims, documents provenance, and is explicit about what cannot be
redistributed and how to obtain it — exactly what ESR's mandate expects.
Referee / editor pushback → ESR-specific fix
-
"Your register data can't be shared, so this isn't reproducible." → Provide all code plus a
precise access route (provider, version, DUA); the policy accommodates restricted data via an access
path, not an exemption from code.
-
"The DAS is vague." → Name the archive/provider, the data version/edition, and the exact access
condition; a generic "data available on request" is weak.
-
"Numbers don't match your code." → Re-run the master script end-to-end before depositing; mismatches
read as a credibility failure under the mandate.
Calibration anchors
-
A replication package is a condition of publication, not a courtesy. For statistical/computational
work submitted from 1 Jan 2025, plan it from day one — it is not an afterthought at acceptance.
-
Restricted data still requires shareable code. ESR's mandate is satisfied by code + a documented
access path, not by declaring the data private.
-
The DAS is required for everyone. Even exempt qualitative work needs a Data Availability Statement;
confirm the current wording and effective dates on the live page.
Anti-patterns
- Treating the replication package as optional or leaving it to the last minute
- A vague "data available on request" DAS with no provider, version, or access condition
- Claiming a qualitative exemption for a statistical/computational paper
- Omitting the harmonization/recoding code from the package (the result is not reproducible without it)
- Depositing code whose output does not match the manuscript's tables
Output format
【DAS】drafted, names archive/provider + version + access condition? [Y/N]
【Package required?】statistical/computational (yes) vs. qualitative (exempt)
【Package contents】master script + data/access path + harmonization code + seeds + README? [Y/N]
【Restricted data handled】code shared + documented access route? [Y/N]
【Policy check】current ESR DAS + replication wording/dates confirmed? [Y/N/待核实]
【Next】eursr-review-process
Supplementary resources
