Research Ethics, Transparency & Data (newms-transparency-and-data)
NM&S studies people through digital media, so research ethics for online and platform data is
central, and transparency is calibrated to the method, not to a one-size mandate. NM&S follows
SAGE/COPE publishing ethics and asks authors to handle ethics, consent, conflicts, funding, and data
availability deliberately. The posture: clear ethics, anonymization that actually protects people, and
documentation that lets a stranger judge your claims without overstating a replication requirement.
When to trigger
- Planning or reporting consent, anonymization, IRB/ethics approval
- Scraping or using platform data and unsure about ToS, public/private, and consent
- Documenting qualitative or computational work so the claims are credible
- A reviewer asked how others could verify or build on your analysis, or flagged an ethics concern
Research ethics for online/platform data (the NM&S core)
-
Consent and the public/private blur. "Publicly available" is not the same as "consented." Posts in
a semi-private group, identifiable users, or vulnerable communities need a defensible ethics rationale,
not just "it was online." State your IRB/ethics-board position.
-
Anonymization that protects. Remove handles, paraphrase searchable quotes (verbatim posts can be
reverse-searched to a user), blur faces and identifying detail in screenshots, and consider aggregation.
-
Scraping and platform ToS. Be transparent about how data were collected (API vs. scraping), and
acknowledge ToS and legal/ethical constraints; do not over-claim that data are freely redistributable.
-
Vulnerable users and harm. Extra care for minors, marginalized groups, or sensitive topics;
minimize harm in what you collect, report, and republish.
Transparency by method (NM&S / SAGE-COPE norm)
| Data type |
Typically shareable |
Restricted |
Credibility documentation |
| Interviews |
coding scheme, anonymized excerpts |
identifiable transcripts |
memo trail, excerpt-to-claim table |
| Digital ethnography |
analytic memos |
informant identities, raw fieldnotes |
positionality + IRB conditions note |
| Content / discourse corpus |
codebook, sample |
ToS-restricted raw posts |
coding reliability + sampling frame |
| Computational / scraped |
code + seeds, derived measures |
raw API/ToS-restricted data |
collection log, validation, model versions |
| Any quantitative survey |
data + code where ethical |
identifiable records |
codebook + master script + pinned versions |
Good practice
-
Document provenance and construction: a README/codebook describing sources, collection window,
sampling, variable/category definitions, and analysis steps — so the work is checkable.
-
Quantitative/computational: keep a master script + pinned versions + seeds; share code even when
raw platform data cannot be redistributed (give a documented access path).
-
Qualitative: protect informants first; share coding schemes and anonymized excerpts that support
the claims without exposing participants.
-
Confidentiality outranks sharing: state clearly when and why data cannot be shared, and what can.
Worked micro-example (illustrative)
Study: digital ethnography of a courier forum + interviews + scraped public posts.
Ethics: IRB-approved; forum posts paraphrased (not verbatim) to defeat reverse-search; interview
informants pseudonymized; minors excluded; employer unnamed.
ToS: posts collected via the platform's API within ToS; raw corpus not redistributed; codebook + derived
counts shared instead, with an access path on request.
Transparency statement: "Coding scheme, anonymized excerpts, and analysis code are available;
identifiable data are withheld to protect participants and per platform terms."
Referee pushback → NM&S-specific fix
-
"You quote identifiable users verbatim." → Paraphrase searchable quotes, strip handles, justify consent.
-
"'It was public' isn't an ethics argument." → State IRB position, public/private rationale, and harm minimization.
-
"How could anyone verify the qualitative claims?" → Provide a coding scheme and excerpt-to-claim table.
Calibration anchors
-
Public ≠ consented. The ethics question is harm and reasonable expectations, not mere availability.
-
Anonymize against reverse-search. Verbatim posts and visible handles can re-identify users — paraphrase.
-
Transparency, not a replication gate. SAGE asks authors to include a data-availability statement
and encourages sharing data where ethical/legal constraints allow; do not promise raw platform data
that cannot be shared.
Anti-patterns
- Treating "publicly available" as automatic consent
- Verbatim, searchable quotes that re-identify users; unredacted screenshots
- Silence on scraping method, API/ToS constraints, or IRB/ethics review
- Promising open data that consent/ToS/confidentiality forbid
- Over-stating NM&S's policy as a mandatory verified replication deposit
Output format
【Ethics】consent, IRB/ethics approval, public/private rationale handled? [Y/N]
【Anonymization】reverse-search-resistant; screenshots redacted? [Y/N]
【Data collection】API/scraping + ToS disclosed? [Y/N]
【Sharing posture】what can be shared (code/codebook/excerpts) and what cannot, and why
【Policy check】ethics and data-availability statement aligned with current SAGE guidance? [Y/N]
【Next】newms-review-process
Supplementary resources