Marcus Andrey Study
Ler em português Library

UX Research · Step 2 · 9 min read

Research data and privacy

Almost everything a study produces is personal data: names, faces, voices, screens, notes. What Europe's GDPR and Brazil's LGPD say that matters for designers doing research, the difference between anonymous and pseudonymised data, and a simple life cycle from collecting to deleting.

A recording of an interview is a recording of a person. Treat it like their house keys, not like a screenshot. Research produces more personal data than most designers realise: the screener with someone's job and health, the video with their face and voice, the screen share with their inbox in the corner. The law in Europe and Brazil says the same things here, and they're mostly common sense: collect less, keep it safe, use it only for what people agreed to, and delete it on time.

What you'll be able to do after reading:

  • Recognise which research data is personal, and which is special or sensitive
  • Apply the principles of the GDPR and the LGPD to a study, in plain words
  • Tell anonymous data from pseudonymised data, and why it matters
  • Set up a study's data so you can find, protect and delete one person's data

1. Research data is personal data

The GDPR defines personal data as "any information relating to an identified or identifiable natural person", including identification by "a name, an identification number, location data, an online identifier" (GDPR, Art. 4(1)). Brazil's LGPD is shorter: "informação relacionada a pessoa natural identificada ou identificável" (LGPD, Lei 13.709/2018, Art. 5, I), information related to an identified or identifiable person.

The GOV.UK Service Manual gives the practical rule: "treat as personal data all the notes, recordings and other data you collect during your research" (GOV.UK, Managing user research data and participant privacy).

What you holdPersonal data?What to do
Names, e-mails and slots from recruitmentYesKeep apart from the notes. Delete when the study ends.
Screener answersYes, and sometimes special: health, ethnicity, religionAsk sensitive questions only if the research needs them, with explicit consent.
Video with face and voiceYesRestricted folder, set deletion date. Clips in reports only if they agreed.
A screen share that shows their inbox or their clientsYes, and other people's tooAsk them to close other tabs first. Blur or cut what you don't need.
Notes labelled "P3, plumber, Porto, 2 employees"Yes: pseudonymised is still personal if someone can work out who it isDrop details the finding doesn't need. In a small market, "plumber in Porto" can be one person.
"4 of 6 copy last month's invoice by hand"No: an anonymous findingShare it widely. This is the part of the study that should travel.
Almost everything a study produces is personal data, until it becomes a finding. Pseudonymised data, like a participant number with enough detail to guess who it is, still counts.Diagram: Marcus Andrey, CC BY 4.0. After GDPR, Art. 4, Art. 9 and Recital 26.

2. Special and sensitive data

Some data needs more protection. The GDPR's special categories include data revealing "racial or ethnic origin, political opinions, religious or philosophical beliefs, or trade union membership", genetic and biometric data, "data concerning health" and data about a person's sex life or sexual orientation (GDPR, Art. 9(1)). Processing them is prohibited unless an exception applies, such as explicit consent. The LGPD calls them dados pessoais sensíveis, with a similar list (Art. 5, II), and when consent is the basis it must be given "de forma específica e destacada, para finalidades específicas", specifically and prominently, for specific purposes (LGPD, Art. 11, I).

For a designer, this shows up in two places:

  • Screeners. Asking about health, ethnicity or religion to build a diverse sample makes the screener sensitive data. Ask only if the research needs it, and say why.
  • Sessions. People volunteer sensitive things in interviews: an illness, a debt, a divorce. You can't always prevent it, but you can keep it out of notes and reports unless it matters to a finding.

3. The life cycle, and the principles behind it

  1. 1Collect

    Only what answers the research questions. Every screener question has to earn its place.

    Data minimisation
  2. 2Store

    One approved, restricted place. Name files so you can find one person's data.

    Security, and the right of access
  3. 3Use

    Only for the purpose people agreed to. A new purpose needs new consent.

    Purpose limitation
  4. 4Share

    Findings and anonymised quotes. Raw recordings stay with the research team.

    Anonymous data is outside the law
  5. 5Delete

    On a date set before the study, or as soon as someone asks.

    Storage limitation
The life of research data in five stages, with the rule that governs each. The same ideas are in Europe's GDPR (Art. 5) and Brazil's LGPD (Art. 6).Diagram: Marcus Andrey, CC BY 4.0. After GDPR, Art. 5 and Recital 26, and GOV.UK Service Manual, Managing user research data and participant privacy.

The GDPR's principles in Article 5 map neatly onto a study:

  • Data minimisation: data must be "adequate, relevant and limited to what is necessary" for the purpose. The LGPD calls it necessidade: "limitação do tratamento ao mínimo necessário" (Art. 6, III). In practice: don't record video if audio will do; don't ask for a surname if a first name will do.
  • Purpose limitation: collected "for specified, explicit and legitimate purposes" and not used in incompatible ways. LGPD: finalidade (Art. 6, I). A recording made to improve invoicing isn't material for a marketing video.
  • Storage limitation: kept identifiable "for no longer than is necessary". The LGPD says personal data is deleted when processing ends (Art. 16). GOV.UK suggests "2 or 3 years as reasonable retention periods for user research data", and sooner if you don't need it, because "most raw user research data loses its value quite quickly" (GOV.UK).

GOV.UK adds the practical steps: store data only in places your organisation has approved for personal data, restrict access, move files off phones and recorders quickly, and name files (date, round, participant number) so you can find one person's data if they ask to see or delete it.


There are two different "consents" in research, and it helps to keep them apart:

  • Research consent is the ethical agreement from Research ethics and consent: people understand the study and agree to take part. You always need it.
  • The legal basis is the reason the law allows you to process the data. The GDPR lists six in Article 6(1); consent is one of them (a), and "legitimate interests" is another (f). The LGPD lists ten in Article 7, including consent (I) and legitimate interests (IX). Which one your organisation uses for research is a decision for its data protection officer or legal team, not for the designer.

When consent is the legal basis, the bar is high. GDPR: "freely given, specific, informed and unambiguous", given "by a statement or by a clear affirmative action" (Art. 4(11)), and "it shall be as easy to withdraw as to give consent" (Art. 7(3)). LGPD: "manifestação livre, informada e inequívoca" (Art. 5, XII), which "pode ser revogado a qualquer momento", can be revoked at any time, through a free and easy procedure (Art. 8, §5).

Either way, people have rights over their data. Under the LGPD, for example, they can ask for confirmation that you hold it, access, correction and deletion (Art. 18). Your file naming is what makes those rights possible to honour.


5. Anonymous or pseudonymised

This is the distinction that most often goes wrong.

  • Anonymous data can't be linked to a person by any means "reasonably likely to be used", and the GDPR says its principles "should therefore not apply to anonymous information" (GDPR, Recital 26). The LGPD agrees: anonymised data isn't personal data, unless the anonymisation can be reversed with reasonable effort (Art. 12).
  • Pseudonymised data has the names swapped for codes, with the key kept separately (GDPR Art. 4(5); LGPD Art. 13, §4). It's still personal data: Recital 26 says it "should be considered to be information on an identifiable natural person".

Replacing "Ana Souza" with "P3" is pseudonymisation, not anonymisation, especially if the notes still say "plumber in Porto with two employees". In a small market or inside a company, a job title and a city can be one person. Before a quote leaves the research team, read it as someone who knows the participant would.

That's also why GOV.UK says to use "fully anonymised extracts whenever possible" in outputs: then there's nothing to do if someone later asks for their data or withdraws (GOV.UK).


6. Tools, vendors and AI

Every tool that touches the data is part of the study: the video platform, the transcription service, the whiteboard, the repository. GOV.UK says to include third-party suppliers, such as transcription, in your consent materials. Before a study:

  • Check which tools your organisation has approved for personal data, and where they store it.
  • List them in the information sheet.
  • Treat AI transcription and summarising tools the same way. Uploading a recording to an unapproved AI tool is sharing personal data with a new party. See AI governance and data protection and AI in UX research.

GOV.UK also recommends having your research data practices reviewed by your organisation's data protection expert, which may include a data protection impact assessment (DPIA) (GOV.UK).


7. Common mistakes

  • Recordings on personal phones or laptops, synced to a personal cloud.
  • Raw videos shared in a public team channel "so everyone can learn".
  • Calling data with "P3" instead of a name "anonymous".
  • Sensitive questions in a screener "just in case".
  • No deletion date, so nothing is ever deleted.
  • Uploading recordings to AI tools nobody approved.

8. Vocabulary

Term Meaning
Personal data Any information about a person who can be identified, directly or indirectly
Special categories / sensitive data Health, ethnicity, religion, politics, sexual life, genetic and biometric data, and similar
Data minimisation Collecting only what the purpose needs (LGPD: necessidade)
Legal basis The reason the law allows the processing: consent, legitimate interests and others
Pseudonymised Names replaced by codes, with a key kept elsewhere; still personal data
Anonymous Data that can no longer be linked to anyone; outside data protection law
Retention period How long data is kept before deletion, agreed in advance

9. Try it (no tools needed)

  1. List every place data from your last study lives: tools, folders, chats, laptops.
  2. For each, write: who can see it, and when it will be deleted.
  3. Pick one quote from a past report and read it as the participant's manager would. Could they tell who said it?

Reflection questions: If a participant e-mailed you today asking to delete their data, how long would it take you to find all of it? What would you change so it took five minutes?


10. If you already practise

  • Write a one-page data policy for research with your data protection officer: approved tools, folder structure, retention, who can access what. NN/g lists a research data policy as a sign of an ethically mature practice (Rosala, NN/g, 2019).
  • Put a deletion date in every study's folder name, and a calendar reminder to act on it.
  • Separate the participant tracker (names, contacts) from the research data (notes, recordings), and give them different access.

Checklist

Learn more

Books, courses and videos to go deeper

Book links go to Amazon. As an Amazon Associate I earn from qualifying purchases, at no extra cost to you.

Research ethics and consent · Notes, recordings and transcripts · Research repositories · AI governance and data protection · AI in UX research · ResearchOps

Saved only on this device.

↑ ↓ to move · Enter to open · Esc to close