What Ethical Responsibilities Come With Using Archived Research Data?

Does Archiving Research Data Make Them Free to Use?

A dataset may sit in a university repository, disciplinary archive, research center, or long-term project database years after the original study has ended. You find it, obtain access, and realize that it could answer your research question without recruiting a single new participant.

It is tempting to think that the difficult ethical work happened during the original study. But archiving changes where the data are stored, not necessarily what researchers owe the people represented in them. Ethical responsibilities can follow data into their second, third, or twentieth research life.

Archived Data Still Carry Ethical Conditions and Responsibilities

Archiving Preserves Data, but It Also Preserves Their Ethical History

Start with provenance: where did these data actually come from?

Data provenance is the documented history of a dataset: how the information was collected, from whom, for what purpose, using which instruments and procedures, and what happened to it before it reached you.

For secondary researchers, provenance is both a methodological and an ethical issue. You may need to know whether participants consented to future reuse, whether records came from research or another context, whether identifiers were removed, whether variables were recoded, and whether particular participants or observations were excluded before archiving.

A clean downloadable file can conceal a rather untidy history. The absence of visible identifiers or consent information in the spreadsheet does not establish that those issues never existed.

Repository access does not necessarily equal ethical authorization

Being technically able to obtain a dataset does not answer whether your particular use is permitted. Archives operate under different access models. Some datasets are publicly downloadable. Others require registration, institutional affiliation, an application, a data-use agreement, ethics documentation, or approval by a data-access committee.

WHO's policy and implementation guidance on sharing and reuse of health-related data explicitly treats data reuse as requiring attention to ethical, legal, and technical issues, including privacy and appropriate data-management and sharing arrangements.

Read the repository's terms rather than inferring permission from the presence of a download button.

Original participant authorization still matters

Archived research data may have been collected under study-specific consent, broad consent, another authorized arrangement, or an older consent process that said little about future reuse. CIOMS states that institutions storing health-related research data should have governance systems for authorizing future research use and that researchers must not adversely affect the rights and welfare of the individuals from whom the data were collected.

The secondary researcher therefore needs to establish whether the proposed use is covered by the original authorization or whether another legitimate pathway applies. Whether archived data require new participant consent depends on the circumstances rather than on their archival status.

Restrictions can travel with the dataset

The original study or repository may impose restrictions concerning research topics, commercial use, attempts at re-identification, onward sharing, linkage with other datasets, geographical transfer, publication, security, or destruction after the approved project.

These are not decorative terms attached to a dataset. If you obtain access under a data-use agreement, repository license, or other formal arrangement, those conditions define what you are permitted to do. A scientifically interesting analysis does not automatically override them.

De-identification reduces some risks but does not erase every responsibility

Removing direct identifiers can substantially reduce privacy risk and may change the regulatory status of a project. It should not, however, be treated as a universal ethical exemption.

Datasets can contain unusual combinations of demographic, geographic, temporal, clinical, educational, or behavioral variables that make individuals easier to distinguish than a superficial inspection suggests. New external datasets can also create re-identification opportunities that did not exist when the archive was created.

More fundamentally, de-identification does not automatically make every secondary use ethically acceptable. Commitments to participants, repository restrictions, group harms, and inappropriate uses can remain relevant even when researchers cannot readily attach a person's name to a row.

Do not attempt re-identification simply because it is technically possible

Archived data sometimes contain enough detail to tempt a researcher to determine who particular participants are, especially when combined with publicly available information. Unless re-identification is specifically authorized and ethically justified within the research design, such attempts can violate participant expectations, repository rules, contractual obligations, or privacy requirements.

The same caution applies to linkage. Combining an archived dataset with another source can transform the informational content of both. The ethics of linking datasets without recontacting participants therefore requires its own assessment rather than being treated as ordinary secondary analysis.

Context can disappear when data are separated from their original study

Archiving often preserves variables more successfully than it preserves context. A secondary researcher may receive a codebook and dataset without having observed how questions were explained, which practical problems occurred during recruitment, how participants interpreted sensitive items, or why investigators made particular coding decisions.

This creates a risk of technically correct but substantively misleading analysis. Before interpreting archived variables, consult the original protocol, instruments, codebook, metadata, publications, data-cleaning documentation, and repository notes where available.

If the information needed to interpret a variable responsibly has been lost, uncertainty should be acknowledged rather than reconstructed from guesswork.

Secondary researchers inherit the limitations of the original design

You cannot redesign the sampling frame, add a missing confounder, change how a construct was operationalized, or repair a measurement that was unsuitable for your new purpose simply because you now possess the raw observations.

The methodological question remains whether the archived dataset can legitimately answer the different research question you want to ask. Ethical data stewardship includes resisting claims that the archived evidence cannot support.

Privacy protection continues after access is granted

Once archived data enter your possession, your responsibilities may include secure storage, access controls, approved computing environments, encryption, limits on local copies, restrictions on cloud services, destruction schedules, and breach-response procedures.

Under the Philippine Data Privacy Act and its implementing rules, personal-data processing must remain fair and lawful, compatible with declared and legitimate purposes, proportionate to those purposes, and subject to appropriate safeguards. The law also contains specific provisions concerning scientific and statistical research and confidentiality.

Repository approval therefore does not relieve the secondary research team of its own data-security responsibilities.

Data minimization still applies to secondary research

Researchers sometimes request every available variable because the dataset already exists. That logic is weak. If your study needs 20 variables, acquiring 200 may create unnecessary privacy and security exposure without improving the science.

Where applicable, request or retain only the information necessary for the approved research purpose. The Philippine implementing rules, for example, articulate proportionality by requiring processing to be adequate, relevant, suitable, necessary, and not excessive in relation to the declared purpose.

Archived data can create harms beyond individual identification

Some analyses can stigmatize communities, expose small populations, reinforce discriminatory classifications, or generate conclusions with consequences for groups even when no participant is named. These concerns can be particularly important for rare conditions, small geographic populations, Indigenous communities, marginalized groups, or datasets containing highly sensitive attributes.

Individual de-identification therefore should not exhaust the ethical analysis. Consider what can be inferred about populations as well as individuals and how findings will be communicated.

When a Publicly Discoverable Dataset Still Requires Ethical Work

Common Ethical Mistakes When Working With Archived Data

Treat Archived Data as Entrusted Research Material, Not Found Property

Before opening your statistical software, reconstruct the dataset's ethical and methodological history. The key question is not merely whether you can obtain the file, but whether you understand enough about its origin and conditions to use it responsibly.

Before Using Archived Research Data

Frequently Asked Questions About Archived Research Data

Archived Data Remain Connected to the People and Conditions That Produced Them

How to Cite This Guide

This guide is intended to be read, shared, and used in research, teaching, and academic work. If you draw on its ideas, explanations, or other content, please acknowledge the source by citing the guide. Doing so gives appropriate credit and helps your readers locate the original resource.