The most important issues for transcribing qualitative interview: should be verbatim (you cannot change any word. If it does not make sense, put your explanations in parenthesis), use punctuation carefully (otherwise it will distort meaning. don't go with words only, try to capture and comprehend intonations) and record laughter, sobs, side talks (if relevant). Do not forget to add background information (date and time of interview, name of interviewer, name of transcriber, etc.) and interviewer's impression.

I am glad you asked this question as I struggle with this issue as well. I tend to err on the side of extreme caution, transcribing verbatim and then coding all transcripts, but since I do not have the luxury of a student assistant, this level of transcribing is not realistic. I need to find a way to transcribe effectively and quickly (i.e. I am running out of time and need to find a better approach).

The most important issues for transcribing qualitative interview: should be verbatim (you cannot change any word. If it does not make sense, put your explanations in parenthesis), use punctuation carefully (otherwise it will distort meaning. don't go with words only, try to capture and comprehend intonations) and record laughter, sobs, side talks (if relevant). Do not forget to add background information (date and time of interview, name of interviewer, name of transcriber, etc.) and interviewer's impression.

Although verbatim transcription is the most preferred, sometimes we use some techniques to quickly transcribe our data when we are in short of time. What we do: avoid repeated words and sentences, avoid sounds that do not add to the content and avoid recording echo probes or 'uh hm' probes. Sometimes in emergencies, we also make summary transcriptions, especially during outbreak investigations of infectious diseases. However, to avoid distortion in such situation, it is best to have interviewer's transcribe the interview. Quotations cannot be used from this data. For quotations, we transcribe selected parts verbatim.

Thanks! Regarding the transcription software - don't you have to "train" it to the various voices? Most software is geared toward using only two or three different voices. Given that I have at least 50 different voices on the recordings, using transcription software has been more work than just transcribing directly.

Thank you all for your suggestions. I will be starting translating some interviews in a few weeks... and the same question has already occurred to me. My University has a licence for a software that I find really helpful, since the interviews I made were audio and video recorded. Do you know Nvivo? Here is a review: http://www.qsrinternational.com/FileResourceHandler.ashx/RelatedDocuments/DocumentFile/818/NVivo-9-review-Tim-Macer-Quirks-July-2011.pdf.

I go with Justyna and Ashwin, take your time for the transcription. And do it all by yourself. Don't source it out. Transcription is already half the way to the analysis if you do it properly. I heard about a lot of qualitative research settings where there were professional transcribers involved which passed the transcriptions on to the researchers. I can't see in which way this would be helpful. It only leads to missinterpreted data.....

Ok, let's assume time is scarce and you, Justyna, needed help to complete it on time. But where lies the advantage for an in-depth qualitative analysis? Wouldn't you prefer to do it all (interviews-transcription-analysis) by yourself (you and your research group) if you had the time to?

Justyna, I know exactly what you mean. If you are really deep into something you need a break from time to time, some distance to all those weird and complex structures you have in your mind. Distance makes it simple again.

A neutral interviewer is also very much desireable, of course.

I just wanted to outline that in most of the cases a stringent setting of interviews-transcription-analysis is preferably all done by yourself or for bigger projects by a research group. You will more likely have the better outcomes for a qualitative analysis. This is my experience and this is also what I can hear from other researchers. Still, scarce time and ressources force you to split it up. That's all...

Continuing with the software affaire... I tried Dragon Natural Speaking, and it (more or less) works. Obviously, nothing like the research's hand and ears for transcribing, but it is a hugh effort sometimes. Other possibilities http://www.sonal-info.com/ or http://www.nch.com.au/software/dictation.html (did not try them)

I am from southwest louisiana the Dragon Natural Speaking does not work for me anyone have a suggestions for us Cajuns/Creoles