Transcription Conventions: Choosing a System for Your Interviews

Why transcription conventions exist, how true verbatim differs from intelligent verbatim, which systems are widely used, and how to choose one and document it in your methods chapter.

By ·

Give the same interview to two people to type up and you will rarely get the same text back. One writes down every "um" and the other leaves them out. One marks a pause where the other puts a full stop. Transcription conventions define how speech is turned into writing, so that transcripts are consistent, traceable and fit for the method you are using.

This article explains why transcription conventions exist, how true verbatim differs from intelligent verbatim, which systems are widely used, and how to choose one and document it in your methods chapter. It does not replace the original sources. If you adopt a system, cite the version your supervisor requires and work from its full rules. How to get to a transcript in the first place is covered in the guide on how to transcribe interviews.

Why Transcription Conventions Exist

Speech is full of things written text does not have, such as pauses, emphasis, unfinished sentences, people talking over each other, laughter and accents. Every transcript has to decide what to keep and how to show it. Without conventions, each person makes those decisions differently, often inconsistently within a single transcript.

Agreed conventions give you three things:

  • Consistency. Every transcript in a project follows the same rules, even when several people or tools are involved.
  • Traceability. Anyone who reads or examines your work knows what a symbol means and what was deliberately left out.
  • Fit with your method. An analysis interested in content needs different detail from one that studies how people talk to each other.

True Verbatim vs. Intelligent Verbatim Transcription

The most basic decision is how close the transcript stays to what was actually said.

True verbatim (sometimes called strict or full verbatim) records everything as spoken, including filler words, repetitions, false starts, slips of the tongue and often non-verbal sounds such as laughter or coughing. Depending on the conventions, pauses, emphasis or overlapping speech are marked too. It takes longer, but it preserves information that some analyses depend on.

Intelligent verbatim (also called clean verbatim) lightly edits speech into readable written language. Fillers, stutters and false starts are removed, but the speaker's wording and meaning are kept, and nothing is paraphrased. The result reads smoothly and is quicker to produce, which is why it is the default for many interview studies and for most transcription services.

An invented example shows the difference. True verbatim: "So, um, we never, we never really, uh, talked about it." Intelligent verbatim: "So, we never really talked about it."

Jefferson Notation for Conversation Analysis

Jefferson notation, developed by Gail Jefferson, is the standard transcription system in conversation analysis. It marks in detail how things are said. Overlapping talk, timed pauses, intonation, stretched sounds, changes in volume and other details of delivery all have dedicated symbols. Transcribing to this level is slow and demanding, and it is worth it when the interaction itself is what you are studying. Work from the published key to the notation itself.

Transcription Conventions in German-Language Research

If you work with German-speaking colleagues or supervisors, you will meet three names. The simple transcription system by Dresing and Pehl is widely used in qualitative social research. It focuses on content, lightly smooths the language, and uses only a few simple markers, much like intelligent verbatim. Kuckartz's transcription rules are closely tied to computer-assisted qualitative analysis and aim at transcripts that are easy to search and code in analysis software. GAT 2, a conversation-analytic transcription system from German interactional linguistics, plays a role similar to Jefferson notation and offers several levels of detail. As with any system, cite the original source in the edition your supervisor asks for.

How to Choose Transcription Conventions for an Interview

Your method determines the right conventions. A few questions help:

  • What are you analysing? If you are interested in content, opinions, experiences and themes, intelligent verbatim or a simple system is usually enough. If you are studying interaction, turn-taking or delivery, you need a detailed system such as Jefferson notation.
  • What is usual in your field? Look at how comparable theses and papers in your department were transcribed.
  • What does your supervisor expect? Some supervisors prescribe a system. Ask before you start.
  • How much time do you have? The more detailed the system, the longer each transcript takes. Plan for that.
  • Do you need adaptations? Many projects adopt an existing system and add a few rules of their own, for example on anonymisation. That is normal, but it has to be documented.

Documenting Your Transcription Conventions in the Methods Chapter

Your methods chapter explains how your transcripts were produced. That usually includes:

  • which conventions you used, with a full reference in the edition your supervisor requires;
  • any deviations or additions, and why you made them;
  • how you anonymised or pseudonymised the material;
  • who transcribed and with what tools, for example whether an automatic draft was revised by hand;
  • where the transcripts can be found, such as an appendix.

Many theses also add a short key to the symbols used, so readers can follow the transcripts without looking anything up.

Automatic Transcription and Transcription Conventions

Automatic transcription software produces text that does not conform to any of these systems. It follows its own conventions, and it does not capture things like timed pauses or intonation contours. An automatic transcript is therefore a draft that still has to be brought up to a recognised standard.

The draft still saves a great deal of time, because most of the text and the speaker attribution are already in place. You then read it against the recording, correct it and apply your conventions by hand. How that works in a thesis project is described in Transcribing Interviews for a Dissertation or Thesis. Whether you need a transcript at all, or whether minutes would do, is covered in Transcript vs. Minutes: What's the Difference?.

Scriber produces these drafts with speaker separation on a server belonging to your university or institute, so recordings never leave your own network. More on the page about transcription for research. You can try it in the free hosted demo at scriber.determin.de with recordings that are not confidential. For questions, write to [email protected].

Frequently Asked Questions

What are transcription conventions?

Transcription conventions define how speech is turned into text, including whether filler words, pauses, emphasis, accents or overlapping speech are recorded, and how. They keep every transcript in a project consistent and make it clear to readers what the transcript shows.

What is the difference between true verbatim and intelligent verbatim?

True verbatim records everything exactly as spoken, including filler words, repetitions and false starts. Intelligent verbatim lightly edits speech into readable text, removing fillers and stutters while keeping the speaker's wording and meaning. Which one fits depends on your method of analysis.

Which transcription conventions should I use for an interview?

That follows from your method. For content-focused analysis, intelligent verbatim or a simple system is usually enough. For conversation analysis, you need a detailed system such as Jefferson notation. Ask your supervisor whether a system is prescribed, and cite the original source.

Does automatic transcription follow transcription conventions?

No. An automatic transcript does not conform to any academic system, so treat it as a draft. It saves time because the text and speaker attribution are already in place. You then correct it against the recording and apply your conventions by hand.

Ready to See Scriber on Your Own Servers?

Tell us about your infrastructure and workload, and we will answer with a concrete recommendation.