If possible, when conducting interviews you should plan to record and transcribe the conversation.

Well no. Of course not. You don’t have to. But you might want to… and here’s why.

One of the great pleasures about researching is that you get to talk to lots of different people. They’ll usually have something interesting to say. It’s rarely the case for example that I leave an interview thinking that it was a complete waste of time. And if I do, this is almost always because of something I did. I didn’t set up the interview properly, the person didn’t know me well enough, I asked the wrong questions, I interrupted or I didn’t listen enough for clues about where the conversation was going…

Now the important thing in research is that you must re-live this interview experience afterwards. Rather than rely on an imperfect memory, researchers always, if at all possible, find ways to record their conversations. This might be in the form of notes taken at the time, or after the event. Or it might be an audio recording or even a film/video.

Each of these methods has their own strengths and weaknesses of course – it’s hard to look at people and follow the conversation if you are taking notes, if you leave the notes till after the conversation you’ve forgotten half of it, and it’s hard with any form of note-taking to get down extended verbatim speech. Recording devices, on the other hand, can be pretty off-putting to the participant(s), and it may take a while for them to forget that they are there. My hunch is that takes longer to ignore a camera than an audio recorder, but either bit of research-kit can be a serious conversation stopper. So no method of keeping records of conversations is perfect.

But I reckon that recordings are preferable to notes. For a start, we often don’t hear things when we are in the middle of a conversation. We miss bits, we take up one thread of a possible line of thought and leave another dangling. We don’t notice that the person we are talking to has used a particular kind of language, or they have always laughed when a particular topic came up, or they paused for a very long time when considering one of our questions. Recordings capture all of these. So I love good records of conversations for the simple reason that you can pick up a lot of the things that you missed the first time round. Did the other person really say that? Wow. I didn’t hear that at all and it does seem to be pretty important in the context of the whole conversation.

My preference is to use an audio recorder, and wherever possible I get the recordings transcribed in full. I used to do this myself, and indeed I did transcribe all forty or so of the interviews I did for my own PhD. I think DIY transcript is actually preferable to anything else, although a lot of people don’t. It certainly is time-consuming, but if you are transcribing a conversation in which you were a participant, then the physical act of transcription is the first opportunity to listen to it again – in great detail. You not only hear things you didn’t hear at the time, but you also have to think, when typing out what you are listening to, about what to record and what not to. Was that ‘you know’ important? What might you lose if you leave it out? Was that silence meaningful? Because transcription may be the first opportunity to re-think the conversation, it is rarely time ‘wasted’.

These days, because I have no time and I sometimes have the money, I pay someone else to do transcripts. T, who does them, takes care with pauses and laughs, with the ums and ahs … these are put in rather than edited out(1). When I get the transcripts back from him, it is always with a sense of anticipation. I generally read through them straight away before I’ve filed them. That’s because I prefer to engage pretty intensively with research conversation records – I see each reading as creating an opportunity for reflection, remembering, and re-thinking. I also listen to the audio recordings to check what T has done, and I can fill in the words that he can’t pick up but I can pick out, because I know how the conversation went. T also puts in my own questions and comments, so that I can see how I responded at the time, and whether there is something I could/should have done differently. This does create some cringe-worthy reading at times, but each of these moments is important as I can see in the transcript how much I steered the conversation in a particular direction. I get some sense of how trustworthy the conversation is, and how much I need to think about my own influence on it.

Right now one of the research projects I’m doing has no budget for transcription and no time for me or the co-researcher to transcribe in detail. I’m worried. The best we can do is to listen to the recordings and note the general points that are being made. We can then go back and transcribe for ourselves the bits that are most relevant. But this does make me pretty uncomfortable. I am very concerned about what we might miss. I’m also feeling sad that I will miss an enjoyable part of the research process.

I always get great pleasure from the process of analyzing transcripts. A lot of people use NVivo or some other kind of software to deal with transcripts and that’s a choice I could make. But I confess that I’m a bit of a technophobe when it comes to transcript analysis. I tend to analyse using the equivalent of long-hand; gradual understanding is gained by reading and re-reading and reading again. I not only look for themes in and across transcripts, but also key words, repetitive phrases, places where people pause or laugh. I often use some kind of discourse analytic approach or a Foucauldian approach to ‘get at’ the ways in which meanings are being framed and made. I do cut and paste things into files but sometimes I also use markers and stickers. It’s not uncommon for me to make three or four passes through transcripts to see what happens when different analytic approaches are used.

So should you transcribe? Well no. Of course not – but really, the question perhaps should be rephrased – why would you NOT want to?

Note:
(1) I don’t use numbered lines in transcripts, but some people like to have this for recall purposes. It’s also important to note that some conversation and discourse analysis methods require particular transcription conventions. I’m not talking about those here.

Transcribing is converting speech to text word for word. Transcribing is a common practice when conducting interviews because it enables you to perform analysis.

How to transcribe an interview in five steps:

  1. Choose your preferred transcription method
  2. Transcribe the audio (using transcription software)
  3. Add speaker designation and time stamps
  4. Clarify the transcript where needed
  5. Proofread the transcript

Transcription software comparison

Transcription methods

Before you start transcribing, you first need to determine what transcription method you want to use. The best method depends on the goal of your transcription.

Verbatim transcription

Write down every single word, including pauses, the expression of emotions such as laughter, stuttering, and hesitations such as “uh”.

This type of transcription is mostly used in the legal profession or in research where you’re not only interested in what is said but also how it is said.

Intelligent verbatim transcription (most common)

Write down every word, but without irrelevant fillers like “uhm”, “yeah”, “you know” etc. To improve readability, you can also fix grammar mistakes, broken sentences and long paragraphs.

This method is more readable than verbatim transcription, but some data — such as emotions, pauses and hesitation — is lost in the process.

Edited transcription

A summarized and edited version of an intelligent verbatim transcript. In addition to omitting fillers like “you know”, irrelevant sentences can be omitted if it doesn’t change the meaning of the story.

Altering the transcript

If the audio quality is bad or the conversation itself needs clarification, you are allowed to make changes in the transcript. For instance:

  • Adding a clarifying comment: “I showed him that this option [raising prices] would be beneficial for profitability.”
  • Marking unclear / missing audio with ellipses: “I showed him would be beneficial for profitability”
  • Emphasizing words:Increasing prices is needed for profitability”

Example transcript

There are no rules for formatting and structuring a transcript. However, most transcripts contain the following information:

  • Names of the interviewer and interviewee (can be anonymized)
  • Date and time when the interview took place
  • Location of the interview
  • Speaker designation (who says what?)
  • Line numbers and time stamps (optional)
Intelligent verbatim transcription of an interview

Interviewer: Raimo Streefkerk (RS)
Interviewee: Sales manager John Smith (JS)
Date and time: April 5th 2019 16:00
Location: Headquarter company X in Los AngelesRS: Thank you for taking the time for this interview.JS: You’re welcome! I’m happy to answer your questions, because the subject interests me too.

RS: I’d like to start with a question about your relationship with your customers. What does this relationship look like?

JS: I always strive for a relationship where I truly know what challenges my customer is facing and how I can help with them. We’re not just a supplier of products, but we actually try to help. The only way that’s possible is by understanding what they want to accomplish.

Analyzing interview transcripts

After transcribing the interview(s) it is time to start analyzing. There are several techniques for doing this – coding and categorizing is one of them.

This means that you link keywords (e.g. “understanding customer”) to the answers you’ve received to your questions. Based on these keywords you are able to find connections between the answers of different respondents.

You can also use methods such as content analysis, thematic analysis or discourse analysis.

If you quote from an interview in your paper, make sure you correctly cite the source. Learn how to cite an interview in MLA and APA.

Transcription software

Transcribing interviews takes a lot of time, but luckily transcription software is developing quickly! Using transcription software can help you speed up the process.

Accuracy

Most software is able to accurately convert English speech to text. However, the audio quality must be good in order for the software to work. That means a noise-free background, no over-talk, clear accents and good microphones.

If the audio quality is too poor for automatic transcription, you unfortunately have to dictate it or transcribe it manually.

Comparison

We tested and reviewed the transcription software below using the audio of a YouTube video in which Bill Gates is interviewed. The audio meets all the criteria listed above.

Transcription software comparison 2019
Hourly rate (pay as you go) Hourly rate (monthly plan) Free trial?
Happy Scribe $13.40 $11.18 30 minutes
Trint $13.33 30 minutes
Transcribe $6 1 minute

If possible, when conducting interviews you should plan to record and transcribe the conversation.

Pros

  • Speaker recognition
  • Clean and intuitive editor
  • Omits ‘uhs’ and stuttering
  • Correct capitalization and use of periods
  • 25% student discount

Cons

  • Doesn’t insert punctuation (except for periods)

If possible, when conducting interviews you should plan to record and transcribe the conversation.

Pros

  • Good speaker recognition
  • Simple but powerful interface
  • Comment and highlight feature
  • Ignores intro music from video
  • Easy to keep track of reviewing progress

Cons

  • Some missing spaces
  • Doesn’t insert punctuation (except for periods)

If possible, when conducting interviews you should plan to record and transcribe the conversation.

Pros

  • Solid speaker recognition
  • Very good capitalization and punctuation (including commas)
  • Much cheaper than other transcription software

Cons

  • Just a 1-minute trial
  • Dated editor with limited functionality
  • Doesn’t connect audio and transcript
  • $20 annual license fee