Anonymize research data and interviews
Interview transcripts, survey responses, and field notes often contain names, job titles, organizations, and other details that could identify a participant. For a thesis, a publication, or just to meet an ethics board's requirements, that usually needs to come out before you store, share, or run the material through an AI tool.
Your text stays in your browser on your own computer. Nothing is sent, stored or shared.
How to anonymize a transcript
- Upload your transcript (PDF, docx, txt, or md) in step 1 of the main tool, or paste the text.
- Dikke Stift automatically flags names, addresses, organizations, and other sensitive data.
- Review the highlights carefully: research material often needs slightly different judgment calls than a standard text.
- Click 'Next' and download the result as clean text.
Why this is more than just removing names
Real anonymization means a participant also can't be identified indirectly, for example through a combination of job title, organization, and age. Dikke Stift helps you quickly find the obvious things (names, addresses, contact details), but you're still responsible for the final check on indirect identifiability. Always read through the result yourself.
Tip: Run into the same organization names or project codes across multiple interviews? Add them to 'My words' once, so you don't have to flag them again every time.
More ways to use Dikke Stift
- Remove personal data from a PDF
- Anonymize a document before sharing it with AI
- Anonymize a resume or CV
Frequently asked questions about anonymizing research data
Do I need to anonymize interview transcripts for my thesis?
Usually, yes. Most ethics boards and supervisors expect you to remove names, job titles, organizations, and other identifying details from a transcript before storing, sharing, or analyzing it, especially if participants were promised anonymity.
Can I upload Word files and plain text files too?
Yes. Besides PDF, Dikke Stift also supports .docx, .txt and .md, the formats transcripts are usually delivered in.
Does it also catch names that are not obviously names, like nicknames?
Automatic detection works on patterns (capitalization, job titles, known name lists) and can miss things. Add specific names, project names, or abbreviations that keep coming up in your research under 'My words' so they get flagged every time from then on.