Transcription Hell

Transcribing is easily the worst part of the Phd process so far. It makes me tired, grumpy and antisocial. Here are my tips to avoid the pain:

  1. Invest in the right kit

Dragon Dictate v12 £30 – I speak, and Dragon turns it into a Word document. Saves your fingers.

Logitech headphones and microphone £25 – comfier than the free ones which come with Dragon Dictate, which is important when they are on your head for 12 hours a day.

Infinity IN-DB9 Footpedal £40 – essential to start/stop/rewind the audio files.

Express Scribe software £20 – the paid-for version allows easy start/stop/rewind/slow speaking/fast speaking. You can also organise your audio files into folders.

Whatever you do, try to avoid typing or handwriting, unless it is to start identifying themes.

  1. Allow yourself enough time

I spent so much time finding interview subjects, travelling to their site, and doing the interviews, that a backlog of 25 interviews to build up. Given that each hour of audio takes me 4 hours to transcribe, that meant a solid 3 weeks of sitting at my desk, transcribing, with breaks for food, sleep and the bathroom. My husband, family and friends forgot what I looked like. This was boring at best and nearly wrecked my relationships at worst. Learn from my mistakes and don’t let your transcriptions build up.

  1. Work out your transcription methodology

A transcript is not a full copy of the original interview and should not be taken as a record of objective reality. We make decisions about what to record and how to record it. It is therefore imperative to decide on a methodology, consistent with your wider methodology and in line with your research objectives. For example, ethnographers, working in grounded theory, may choose record and make analytical inferences from grammatical errors in speech, breathing patterns and other details. However, I have already brought Bourdieu’s theories of doxa, fields and habitus, and a hypothesis to my research. While I am recording laughter, pauses, gestures and tone of voice, which illuminate my pre-determined conceptual framework, I do not read further detail into the interviews.

“Transcriptions that have no methodology are neither useful for learning what people said, nor rigorous enough to serve in research” (Powers, 2005, p. 5)

  1. Start to identify themes as you go

I started to highlight key sections of the text and identify themes as I went along: symbolic violence, altruism, extraversion, learning…For my 6 pilot interviews with family firms, this was a good way to check that my hypothesis, that “familiness” is correlated with innovativeness, was valid. My questions could also be refined where they failed to identify any useful themes. I’m alert to the danger of relying on this initial thematic analysis, which is based on pattern matching, and which can lead to a superficial analysis (Johnston, 2006). I’m hoping that NVivo’s search tool and other functionality, can help me view the data in a more complex way.

So how are you coping with the transcription process? Any tips would be gratefully received!


Johnston, L. (2006). Software and Method: Reflections on Teaching and Using QSR NVivo in Doctoral Research. International Journal of Social Research Methodology, 9(5), 379–391. doi:10.1080/13645570600659433

Powers, W. R. (2005). Transcription Techniques for the Spoken Word (p. 128). Altamira Press.