Transcribing Interviews for a Dissertation or Thesis
When to transcribe interviews in full, how much time to budget, whether to do it yourself, use a service or use software, what supervisors expect, and where your recordings go along the way.
For many dissertations and theses, interviews are the core data. Before you can analyse, code and quote them, they have to be turned into text, and that step is almost always underestimated in the project plan. Dissertation interview transcription takes time, forces decisions about conventions and format, and raises data protection questions that are far easier to settle before the first interview than after it.
This article works through the questions that come up in academic transcription. It covers when you need to transcribe everything, how much time to budget, which routes are open to you, what supervisors and examiners typically expect, and what a workflow looks like when it starts from an automatic draft. For a broader overview of every option and what it costs, see the guide on how to transcribe interviews.
When to Transcribe Dissertation Interviews in Full, and When Not To
Not every research design requires every interview to be transcribed word for word. That depends on what you plan to do with the material.
- Transcribe in full as a rule when your analysis is qualitative, for example qualitative content analysis, grounded theory or conversation analysis. These methods work with the whole text, and examiners will want to follow how you got from the data to your codes and categories.
- Transcribe selectively when interviews play a supporting role, such as expert interviews used to put documents or figures in context. You transcribe the relevant passages and summarise the rest in interview notes.
- Notes instead of a transcript are sometimes enough for preliminary conversations or field notes. The difference between the two is explained in Transcript vs. Minutes: What's the Difference?.
Whether selective transcription is accepted in your discipline is not a decision to make alone. Raise it with your supervisor early, and record the decision and your reasons in the methods chapter.
How Long It Takes to Transcribe an Interview
Transcribing by hand takes many times the length of the recording. How many times depends on audio quality, the number of speakers, how fast and how clearly people talk, accents and technical vocabulary, and the conventions you transcribe to. A verbatim transcript that marks pauses, false starts and overlaps takes noticeably longer than a clean one.
For your timeline, add up the total length of your interviews and budget a multiple of that figure for transcription. With twenty or thirty interviews, transcription quickly becomes a work package of several weeks in its own right. Time yourself on the first interview before you commit to a schedule.
Dissertation Transcription: Do It Yourself, Use a Service, or Use Software
There are three basic routes, and they can be combined.
Transcribe it yourself. It costs nothing but time. You get to know your data very closely, and many researchers value that as a first round of analysis. The price is the hours, which with many interviews can crowd out the analysis itself.
Hire a transcription service. The work becomes someone else's. You usually pay per minute of audio, and your recordings go to a third party. That has to be covered by your participants' consent, and you normally need a data processing agreement with the service.
Use automatic transcription software. An automatic transcript is ready far sooner than a manual one. It is still a draft. You read it against the recording, correct the errors and bring it into the form your transcription conventions require. Where the software runs matters. With a cloud service, recordings are processed on someone else's servers. Self-hosted software processes them on a server belonging to your university or institute.
What each route costs and how to compare them is covered in Transcription Cost per Minute: What Actually Drives the Price.
What Supervisors and Examiners Expect from Academic Transcription
Requirements differ between disciplines, faculties and individual supervisors, so always check your faculty's requirements and ask when in doubt. Some expectations come up again and again:
- Transcripts in the appendix. Many faculties expect transcripts to accompany the thesis, printed, as a digital appendix, or made available to examiners only. Find out early which applies.
- A named set of conventions. The methods chapter states which transcription conventions you followed, with a reference. The common systems and how to choose one are covered in Transcription Conventions: Choosing a System for Your Interviews.
- Consistency. Every transcript follows the same conventions, even if several people or tools were involved.
- Anonymisation. Names, places, employers and other identifying details are replaced in the transcripts before they leave your hands.
- Traceable quotations. Quotations in the thesis point to an exact place in a transcript, for example by interview code and line number. How you cite them follows your discipline's style guide.
Transcription for a Master's or Undergraduate Thesis
The same principles apply to transcription for a master's or undergraduate dissertation, just on a smaller scale. There are usually fewer interviews, and the expectations around conventions are often simpler. Because the deadline is tight, though, it pays to put transcription into your plan from the start. Here too, your department's handbook and your supervisor's guidance take priority over any general advice.
Data Protection: Where Do Your Recordings Go?
Interviews contain personal data, often of a sensitive kind such as health, political opinions or details about someone's workplace. As soon as you hand a recording to a service or a cloud tool, it leaves your computer and your university. That can be lawful, provided it matches what participants consented to and is covered by a contract.
Many universities and ethics committees prefer raw data not to leave their own infrastructure. What a consent form should cover and which questions to settle are set out in Interview Recording Consent and GDPR: What to Cover. The specific rules for your study are a matter for your university's data protection officer.
A Workflow: Automatic Draft, Then Correction
A workflow that saves time without sacrificing care looks like this:
- Fix your conventions. Before the first interview, decide which transcription conventions you will use and agree them with your supervisor.
- Generate a draft. The software produces a first draft with speaker separation, so questions and answers are already apart.
- Correct against the recording. Listen to the whole interview and correct words, names, technical terms and speaker attribution.
- Bring it into the required form. Apply your conventions for pauses, emphasis, filler words, line numbers and whatever else your system asks for. No software does this step for you.
- Anonymise. Replace identifying details and keep the key in a separate, secure place.
Scriber is self-hosted transcription software that runs on a Linux server belonging to your university or institute. You upload a recording, record directly in the browser, or drop files into a watched folder, and you get a speaker-separated transcript without the recording leaving your own network. It is licensed per user, and transcription itself is unlimited. That makes most sense for a research group or institute where several people run interviews regularly. Details are on the page about transcription for research.
If you want to see what such a draft looks like first, use the free hosted demo at scriber.determin.de with a recording that is not confidential. Real interviews belong in your own installation. For questions or a quote, write to [email protected].
Frequently Asked Questions
Do I have to transcribe all my dissertation interviews in full?
Not always. With qualitative methods such as qualitative content analysis or grounded theory, full transcription is the norm. If interviews only play a supporting role, for example expert interviews, selective transcription may be enough. Agree this with your supervisor early and justify the decision in your methods chapter.
How long does it take to transcribe an interview?
By hand, transcription takes many times the length of the recording. Exactly how long depends on audio quality, the number of speakers, speaking pace, accents and your transcription conventions. True verbatim takes longer than intelligent verbatim. Time yourself on the first interview before you set your schedule.
Can I use software for academic transcription?
In many disciplines, yes, provided you check the output against the recording, bring it into the form your conventions require, and disclose the process in your methods chapter. Check your faculty's requirements, and find out where the software processes the recordings, because that has to match your consent form.
Do transcripts go in the dissertation appendix?
Often, yes, whether printed, as a digital appendix or for examiners only. Requirements differ between faculties and disciplines, though. Check your faculty's requirements and anonymise the transcripts before they leave your hands.
Ready to See Scriber on Your Own Servers?
Tell us about your infrastructure and workload, and we will answer with a concrete recommendation.