Introduction
The following is a summary of the FGB Implementation
of the National Guidelines for Archiving of Academic Research for
Faculties of Behavioural and Social Sciences. This summary
highlights the major concerns when archiving data; in-depth explanation
is found in the full guidelines.
NB: The guidelines summarized here refer specifically to archiving data for preservation purposes; they do
not address the complexities involved with the reuse of data by third parties.
Why Archiving?
Data and supplementary materials from your research should be
archived because:
- Archiving promotes research integrity and transparency by:
- Ensuring that research results can be verified, reproduced and,
where possible, replicated
- Providing assurances that data have not been inappropriately
tampered with after collection
- Archiving preserves valuable data and the materials necessarily for
the proper interpretation of this data well into the future
- Archiving data is important for the safety of your participants when
conducting medical research
- The archived data serves as a record of the medical interventions
that each participant experienced
- This information is important for safety monitoring and long-term
follow-up
Summarized from guideline sections: 1 and
1.1
What to Archive?
You should archive whatever would be necessary to properly
interpret your data. This includes data, but also documentation about
the data and the research process, as well as code scripts used in the
process of your research
- There should be sufficient documentation in the archived materials
so that another researcher could reanalyse your data and reproduce your
research results without ever contacting you.
- Section 2.1.1 of the full guidelines lists the kinds
of materials you are expected to archive. Not everything listed will
necessarily apply to your research.
When to Archive?
A data package must be archived whenever you publish a
research article. Researchers are expected to manage their data
efficiently during the course of their research so that they can archive
the data and supporting materials as soon as their research
article is accepted for publication.
Additionally, archiving is recommended, but not required, when:
- Data collection for a research project has finished, so that the raw data
can be stored in a way that prevents (unintentional) modification
- A research project is complete, even if the data are
not used in any research publications
Summarized from guideline sections: 1.1
and 2.2
Where to Archive?
The vast majority of data used within the faculty is considered personal data and therefore should be stored in a
secure archive. The default archive at FGB for most situations is YODA.
- If your data can be de-identified enough that they fall under the “Blue”
category from the Privacy Risk categorization, then you can use other
archives, such as DataverseNL or an external data repository.
It is very important, however, that when de-identifying
your data that you do not irrevocably modify the raw data. Irrevocable
modification defeats the purpose of archiving unadulterated raw data.
See How Do I Meet Privacy and Security
Requirements? for further explanation.
- You should check with the FGB
Privacy Champion that your data are indeed “Blue” data before
archiving them in a less secure archive than YODA.
- If your data cannot be de-identified to the “Blue” category level,
it may still be possible to archive the supporting materials
(e.g. research code, codebooks, interview scipts etc.) in archives or
data repositories other than YODA. Just make sure these supporting
materials don’t contain any personal or confidential information.
Who Needs to Archive?
Who do the guidelines apply to?
The FGB archiving guidelines apply to all researchers conducting
research within the faculty.
It also applies to Bachelor’s or one-year Master’s students
if their research results in a research
publication.
- For Bachelor’s and one-year Master’s students who don’t publish any
research articles, it is still recommended that they provide an informal
data package of their work to their supervisor as a way to practice
archiving. The supervisor can decide whether these materials needs to be
preserved.
Who is responsible for archiving?
If you are the first author on a research paper, you are
responsible for archiving. If the first author works at another research
facility, you should make sure that the first author will ensure that
the data are archived.
If data are archived for preservation purposes after data
collection or upon completion of a research project, the lead researcher
is responsible for archiving.
For research from PhD candidates, Master’s and Bachelor’s
students, the supervisor is responsible for archiving; they may delegate
the task to be completed by their student, but they remain ultimately
responsible for this task.
The final responsibility for all archiving in the faculty lies
with the dean.
Who should have access to the archived data?
- Whoever is responsible for archiving should have access to the
archived data and supporting materials.
- There should also be at least one other person who also has access
to the archived materials.
Summarized from guideline sections: 1.2,
2.3 and 2.4
How Long Should Archiving Last?
The duration of archiving depends on the reason you are
archiving as well as on other policies and laws.
For research that isn’t subject to the WMO law, the Good Clinical Practice (GCP) Guidelines or the other
regulations mentioned in section
1.1 of the full archiving guidelines:
- The data and supporting materials used in the publication of a
research article must be archived for 10 years from the date of
publication
- If the data are reused for new research articles, this archiving
term should be extended for another 10 years from the new publication
date
- If you chose to archive data after data collection was completed or
upon the completion of a research project, even if the data have not
(yet) been used for a research article, you can determine how long the
data should be archived
If you are conducting medical research, the duration of archiving
will depend on which laws and regulations apply. To determine which
archiving duration applies, see this page from the CCMO or section
3.1 from the full archiving guidelines
Any data that falls under the “Blue” category from the Privacy Risk
categorization and/or supporting materials that don’t contain any personal and/or confidential information can be archived
indefinitely, unless any other contracts or agreements apply that limit
the archiving duration.
Summarized from guideline section: 3.1
How Do I Meet Privacy and Security Requirements?
It is important that the archived data are protected,
particularly when these data are considered personal, but you must also ensure that the
integrity of the raw data is maintained.
- Raw data can usually be de-identified, but it’s important that they are not
irrevocably altered in the process. Generally, you can de-identify the
data up to step 4 of this de-identification guide without irrevocably altering the raw data.
Only de-identify raw data to a point where it could be returned
to its original state.
- You will need to determine whether any directly
identifying personal data collected in the course of your research
need to be archived** and, if so, for how
long. Also determine whether it’s necessary to re-identify any
de-identified data.
- This will depend on the nature of your research and what regulations
apply. Section 3.2 of the full guidelines explains in
detail what is required.
- If raw data are separated into directly
identifying personal data and de-identified research data, you may:
- Archive the de-identified research data in a less secure archive if
the data are sufficiently low risk (see Where to
Archive). The personal data must be archived in a separate, highly
secure archive. Make sure to cross-reference these separate archiving
submissions to each other in your documentation.
- Archive both sets of data in a single, highly secure archive, but
submit the personal data separately from the de-identified research data
(in other words, create two submissions). Make sure to cross-reference
the two submissions to each other in your documentation
- If the raw data cannot be separated into personal data and
de-identified research data without irrevocably altering the raw data,
archive all of the data together in a secure archive (see Where to Archive)
- You may consider encrypting the personal data (or all of the data if
the personal data cannot be separated from the research data), but you
must ensure that you have a plan for the long-term management of the
de-encryption key. When using YODA as an archive, your best option is to
contact the YODA administrator to provide them with a copy of the
de-encryption key, and then print a copy of the key and store it in your
department’s paper archive.
**One reason you may need to preserve directly
identifying personal data is if you obtained consent from
participants for the use of their data in your research. You must save
these consent forms for as long as the data will be archived.
- The forms need to be saved to serve as evidence that consent was
legally obtained. Depending on the nature of the research, the link
between the consent form and the data itself may or may not need to be
maintained. Section 3.2 of the full guidelines explains this
further. Once the archiving term is complete, assess whether the
research data can be destroyed. Once the research data are destroyed the
consent forms should also be destroyed.
- Even if the consent forms do not include any personal data, it is
still necessary to save the consent forms for the same duration as the
archived research data.
- If paper consent forms were scanned into a digital form, it is not
currently allowed (according to both current VU policy and legal
requirements for medical research) to destroy the original paper copies.
See this statement in the full guidelines for more
information.
Anything else?
Archived data must be persistently findable when the data are used in
a research publication. Archived data must also be registered in PURE
for VU administrative purposes. FGB researchers can meet these
requirements by archiving their data in YODA and publishing those data,
even if the data are kept closed or restricted access to prevent/limit
data
reuse of the data. The metadata published in YODA will be
automatically uploaded to PURE, meaning the researcher is not required
to make a separate PURE registration.
Summarized from guideline section: 4