The main categories
Official statistics
Published by the national statistical authority and by sectoral ministries. Aggregate figures are generally publicly available; microdata from national surveys usually requires a formal request and stated purpose.
Practical note: definitions change between survey rounds. Comparing across years without checking definitions is a common source of error.
Administrative data
Held by agencies as a by-product of operations — health facility records, school records, business registration. Rich but access is negotiated case by case, typically through a formal request and often through an institutional partner.
Survey data collected for research
Either collected by your project or obtained from an earlier study. Sharing practices are developing; asking the original authors directly is often the quickest route.
Enterprise and organisational data
Held by companies or organisations. Requires written permission specifying use and publication scope. This is where most disputes arise, so settle it in writing before collection.
Access: what to expect
- Requests generally go through an institutional partner. A foreign researcher approaching an agency directly usually gets further with a domestic institution's endorsement.
- Allow time. Data access approvals are a common cause of project delay. Start early and have a fallback.
- State the purpose precisely. Vague requests are harder to approve than specific ones.
- Expect conditions on use, on publication, and sometimes on review of outputs before publication. Negotiate publication rights explicitly — a condition that outputs must be approved before publication can conflict with journal requirements and with academic freedom obligations at your own institution.
Ethics approval
Research involving human participants requires ethics review, and requirements vary by institution and by the nature of the research. Two points that catch partners out:
- You may need approval on both sides — from your own institution and from the Vietnamese partner's committee. Establish early who obtains which and how long each takes.
- Approval must precede data collection. Retrospective approval is generally not available, and a journal may refuse a paper without it.
Personal data is protected by law, and both collection and sharing must comply with current regulations as well as with what participants were told in the consent process.
Limitations to state in your papers
Stating these honestly strengthens a paper rather than weakening it:
- Definitional changes across survey rounds if you use time series.
- Coverage gaps — which populations or areas are under-represented in the source.
- Administrative data limitations — collected for operational purposes, so variables may not measure the concept you need.
- Sampling frame issues, particularly for populations that are hard to enumerate.
- Access conditions that constrained what you could analyse or report.
The recurring analytical error
Applying a measurement instrument or a definition developed elsewhere without checking whether it means the same thing in this context. Household composition, employment status, informal sector activity, and educational attainment categories all differ in ways that make direct transfer unsafe.
This is precisely where a local partner's contribution is substantive rather than logistical — and a reason to involve them in design, not only in collection.
Data sharing after the project
Increasingly required by journals and funders. Two constraints to respect: personal data protection obligations, and conditions attached by whoever supplied administrative or enterprise data. Where full sharing is not possible, options include anonymised or aggregated versions, controlled access on request, or sharing code with synthetic data.
Agree the sharing plan at the start, in the project agreement — not when the paper is being submitted.
Note: regulations, funding schemes and institutional arrangements described here are set by the competent authorities and change over time. Verify against current official sources and ask the relevant institution before relying on any of this for a decision.
