One piece of advice I have received a number of times is to hire someone to do my interview transcriptions for me. However, this has never felt quite right for a couple of reasons:
- These are personal conversations and I don’t feel comfortable sharing them with just anyone – in terms of what is shared by others AND what I share with them.
- Listening to these conversations is important because not only do I get to hear the content again, as a listener, and not a participant, but I also hear the tone, and teh giggles, and the pauses in a way that you cannot when reading a transcript.
- Listening to the conversation allows me to experience it again and relive moments that stand out – or moments that I failed to miss the first time.
Of course, the main reason for getting interviews transcribed is the time factor as it can take quite a while to do the transcription – especially when you are atrocious a typer as I am! So I have also developed a great transcription method using Microsoft Word or Google Docs, NVivo, and Speech recognition so I can to speech to text transcriptions. Here is how I do it:
Let’s start with NVivo. I first make a document for each interview where I can make notes about the interview itself (setting, weather, nearby activity, etc…). I then import the audio interview file. There are two important things I do in NVivo do make the transcription easier:
- Set the program into “Transcript mode” so it will index the text to the relevant time in the recording
- Slow the interview audio to 80% to make keeping up with the transcription much easier. This also allows me to imagine that me and the interviewee were both drunk during the interview!
The other tool is a combination of either MS Word or Google Docs and a speech to text generator. I use a high-quality microphone and a computer recording interface to ensure a good quality microphone connection.
In short, as I play the audio through NVivo, I speak through the Microphone to generate the text in Word or Docs. Then, as natural breaks occur in the interview. I cut and paste the text into NVivo where it is indexed for that time period.
So now you may ask, why don’t I just do the speech-to-text right in NVivo? In short, it doesn’t really work.
Overall, I have found Google Docs much better at the transcription. However, because this is sensitive material I cannot keep the material on Google Drive, so after each transcription session I need to delete that file (or remnants can be found through the document history showing previous versions).
This method has been working really well and has made the transcription much less tedious than typing. I do have to go through and make corrections afterward, but I use that as another chance to listen to the interview.
I think listening to interviews multiple times is quite important. Not only does it give you a chance to pick up things you didn’t notice before, but with each listen I feel my brain is working in the background and so I can gain insights from both the interview material and the inspiration it can bring as I process the information. And I also have to wonder how actually speaking the words of each interviewee helps me integrate those words and their experience into my own experience. But at this point, I’m still just wondering!