2nd February 2010 at 1:30 am #5094Transcription is “in no sense a comprehensive replication of the scene, the talk, or the actions that together constituted the event and that, together, gave it ‘sense’ to the participants.” – Freebody
I recently collected 20+ hours of interview data, all video/audiorecorded. I am disappointed to discover that almost all of the existing literature suggests transcription must take place before in-depth analysis, rather than the other way around. It seems to me there is a blind spot, an assumption that the transcripted text is the basis from which analysis springs. To me that seems outdated (we live in a digital age, after all). It also represents a movement one degree away from the original texts: the spoken words, facial expressions, and gestures of the participants.
Can anyone enlighten me why this seems to be a given? Or am I mistaken: are there voices out there challenging this idea? I understand why some scholars might want written transcripts for the purpose of presentation, but I cannot fathom why the written word is privileged in matters of analysis. From the perspective of a phenomenologist, the video record is much more evocative (and original) than my written record.
Eager to read your thoughts!2nd February 2010 at 1:03 pm #5112Francisco VieiraParticipant
You really pointed out a very good question! I don’t know references that approach this subject. In any way, it looks like the literature isn’t up to date with your concerns. On the other hand, maybe the written be more accessible. Besides, there are key issues related to validity and reability, for instance, which seems to be not completely solved in what refers to images.
Francisco Vieira2nd February 2010 at 2:04 pm #5111Lexine HansenParticipant
I have to agree, the literature on using interviews leans heavily on written transcriptions as a source of data. I believe this is because, like Francisco mentioned, there is concern for reliability and validity. If there is a transcript, you can quote sources directly: you can clearly show that “15 informants mentioned this key concern” and back it up (if necessary) with 15 quotes discussing that concern. Moreover, with a transcript you can remove the identifying characteristics of your informant yet still have the raw data available for review. I would guess that with videos this is not possible, so that raises privacy concerns for data checking.
I have had similar struggles with the “rules” for transcription in my work since I have been conducting interviews in an oral form of Arabic and most literature suggests that data should be first transcribed in original then into English. But if there is no written original? I have found thinking about my interviews in an anthropological/ethnological sense is helpful. I take detailed notes during and after about the “mood” of the interviews–the body language, pauses, misconceptions, etc. that communicate how the informant was feeling about the questions. Then I translate and transcribe to English (in one step). Finally, I combine both so that my notes are interspersed in the written transcript. Thus both actions and words are coded and “analyzed.” In fact, however, the data analysis began with the interview itself, when I noted how the informant responded, probed specific answers, and guided the informant to relevant topics. While the books try to divide data collection from analysis (which seems appropriate in many quantitative studies), such strict division is impossible in the qualitative work I’ve seen.
How would you prefer to see the analysis proceed in work based on videos like yours? Are you suggesting no transcripts or a different type of transcription process?2nd February 2010 at 8:40 pm #5110
I’d suggest that transcription itself establishes a distance that may compromise validity, Francisco, when held up against original recordings (either audio or video).
Obviously privacy is important as it relates to presentation.
One argument that does make some sense is the writing-as-analysis, but even there, there’s no reason the writing in that case needs to be transcription.
I like the steps you took to work in the oral tradition, Lexine. But you mentioned that through transcription “you can quote sources directly.” I don’t think that’s the case: it is by definition indirect because it’s your writing that’s being cited, not their speech. And it’s still a question of presentation, rather than analysis. I don’t see how transcribing after the fact would change the situation in terms of quotability.
I didn’t take detailed analytical notes during my interviews, which were presented to participants as in-depth conversations. I made reflective notes afterwards, and then re-viewed interviews 2x without making any notes (phenomenology, after all). On 3rd viewing, I started seeing emerging themes and noting striking metaphorical descriptions. Full transcription of these conversations will only take place as a precursor to presentation, because I really feel doing so earlier in the process risks tainting the raw interview data with a layer of distance/selection that–if possible–should be avoided. I’m just stunned to discover that working from transcripts seems to be the sine qua non of qualitative inquiry, even though to me it feels decidedly out of sync with that very approach. I’m also surprised the place of transcription in the methodology seems unquestioned.3rd February 2010 at 12:00 am #5109
I think that your ideas are right. Surely, the video record is much more evocative (and original) than the written record. But, It is necessary to write the record in order to create a clear idea. You can create a new format of language that mix video and written record.
Giuseppina3rd February 2010 at 12:43 am #5108Lexine HansenParticipant
I see now what you meant about leaving transcription for later–particularly if you feel that transcription would somehow cloud how people were communicating with what they were saying. I like what you did and if I were to read a paper of yours I think I’d accept it as a rigorous study. My only desire would be for some other form of data check. This could be perhaps sharing your interpretations with subjects and including their feedback in your final analysis. There are other ways to check the “truthfulness” of data analysis, but as a participatory researcher, that’s my favorite. How might you assure your readers that your interpretations of the emerging themes are accurate representations of the interviews?
All that said, there is no way that my PhD committee, as forward-thinking as they are, would accept that approach as rigorous enough for my dissertation. Basically, unless my chair sat with me (or someone else did) and watched the process of theme development. That person (or people) would review my emerging themes and provide a check that I am not just making up themes that aren’t there or that I want to see. In my case, I will use 3 peers to review my coding lists and code a minimum of 4 transcripts to be sure that the list is being applied consistently. I think for scholars more “expert” and along in their careers, it would be more logical to trust their conclusions without more supervised data review–which is what transcripts and peer-coding schemes allow.
I love Giuseppina’s idea to create a new format of presentation which would include both video and written record, or video alone provided the subjects’ confidentiality was protected as they understood it would be when they consented to participate.3rd February 2010 at 2:56 am #5107
A dual-media presentation would be feasible, were it not for the privacy concerns.
My PhD committee wouldn’t go for it as described, either, but I was only listing those factors relevant to the issue of transcription for the purpose of analysis. Obviously, some means of triangulation must take place (in my case: through peer evaluation, participant reflection and the application of other data sources).
As a phenomenologist, technically no coding is permitted in my study, as it goes against the spirit of the discipline.3rd February 2010 at 5:06 pm #5106Jenny HallMember
firstly there are some authors on qualitative research who state that it is not necessary to transcribe but to record the ‘impressions’ of the interview- listening over and over again until the themes emerge. Sorry I haven’t time to look but it may be worth you doing a search for that. Secondly there are theses that have been done using digital media including video that you could explore as an option. But you would have to consider carefully the ethical implications for this- have your participants agreed to you presenting their video recording for example? Could you instead create digital sound recordings and use those in your thesis presented on a CD-rom? However you may have to disguise the voices to ensure the participants cannot be recognised. What has to be clear is that these will haev to be available in some way in the thesis
jenny3rd February 2010 at 7:07 pm #5105
Hi Jenny & thanks,
It’s important to note that I have no beef with transcription: it’s a viable means of presenting research. What I’m hoping to hear is that someone, somewhere has said that transcribing is not a necessary precursor to analysis, and may in fact be a detriment to it, in that it distances the researcher from the original texts (i.e. interviews).3rd February 2010 at 11:05 pm #5104
there are a lot of means of presenting research, but only you know the right methods for your search.
Giuseppina4th February 2010 at 7:50 pm #5103Jenny HallMember
Forum:Qualitative social research Vol 9, No 3 (2008): Visual Methods Interpretative Visual Analysis. Developments, State of the Art and Pending Problems
Bernt Schnettler, Jürgen Raab
Is verbatim transcription of data always necessary? Applied Nursing Research, Volume 19, Issue 1, February 2006, Pages 38-42
Elizabeth J. Halcomb and Patricia M. Davidson
Carter B (2004) How do you analyse qualitiative data? Chapter 6 in Demystifying qualitiative research in pregnancy and childbirth Lavender, Edwards and Alfirevic Quay books
I think you will find in a number of sources that people write that qualittaive interpretaion is an iterative process,going on right from the moment you start creating your research, during your data collection with the notes you are making, well before you start transcibing or whatever you are going to do with your data! It is interesting what you say about ‘distancing from the original’. It sounds like you are seeking to have a reflexive approach to your data and explore what you ‘feel’ and ‘know’ from it before or instead of actually deeply analysing the actual ‘words’. The above papers may give you some more thoughts on transcribing. The other thought I have is Bordieu’s concept of refelxive sociology- I am not a Bordieu scholar but there may be some help for you in his apporach?
jenny4th February 2010 at 9:03 pm #5102
I think that your ideas and concepts expressed in your message are right. Also I think that it is necessary that the qualitative search is interactive process, because it is a process that starts creating new discoverys. In fact the Bordieu’s concept and his approach could help you.
With best wishes
Giuseppina7th February 2010 at 12:57 am #5101
Thank you once again, Jenny. Your comments are helpful: I’ll look the references up asap.7th February 2010 at 12:59 am #5100
Upon re-reading your post I have a question. You say you will “code a minimum of 4 transcripts.” May I ask why that number?
My inquiry involves 10 interview subjects, and I will have to examine every single interview. I do not have the option of leaving some out of the inquiry: do you?10th February 2010 at 11:20 pm #5099
I agree with you that video record is much more evocative than written record.
- The forum ‘Default Forum’ is closed to new topics and replies.