A few days ago I started to transcribe the first group of my research interviews: I made a preliminary check for practical tips, I asked for advice on a transcription software in the channel #phdchat in Twitter, but overall I retrieved and remixed my notes and readings on the trascription process that I had studied in the MRes Qualitative Data Analysis module. The following notes are intended as a first attempt to build my ‘guide-on-the-side’ to carry out a transcription not as a mechanical practice, but as an a analytical act. The aim is to gain awareness that transcription is a mediated practice – as other research practices – and that “it is not an antecedent to analysis, but it is a central aspect of the ways that researchers analytically orientate to data” (Gibson and Brown, 2009: 125)
Generally speaking, transcription is a time consuming and labour-intensive process by which researchers start to generate data from a source of information/other data (i.e. an audio-recorded interview/focus group) and contextually produce an early analysis about them. Intended as an analytical resource for working through data, transcription is suitable for all kind of data, such as audio, video, photographs, documents. For instance, in the experience reported by Heath and Hindmarsh (2004 :19), transcription is related to “aspects of interaction” observed in a video recording, and allows researchers “to clarify what’s said, by whom and in what way, and to begin to explore potential relations between aspects of the interaction”.
To a degree transcription already activates an interpretive process, because “it involves making analytic judgements about what to represent and how to represent it” (Gibson, 2009: 31). In fact it enables researchers to begin to interrogate available documents to gain informants’ views, having a specific research aim and a methodology to be applied. The adoption of specific conventions, the selection of symbols as many ‘markers’ of the discourse, the choice of one type or another one of transcription implies that researchers are able to frame a flow of information according to procedures which make the provided information be readable as many data to be furtherly analyzed. The choice of a specific analytic method involves a selective kind of attention to the elements to be transcribed: for instance, attention to the range of tones and pauses in a recorded discussion can be important in a conversational analysis approach, but it is not due using other forms of analysis such as thematic analysis, in which the main way of conduct is to code content by iteratively generating a set of categories.
Moreover, representation of the original document is always partial, firstly because it is impossibile to get a perfect reproduction of the original discourse, even with the most sophisticated recording devices: our hearing ability is itself selective and prompts inferences. Secondly, transcription is guided by researchers’ perspective on the study being conducted, sifting “details of speech (and gaze) and gesture and action) or writing that are arguably deemed relevant in the situation and that are relevant to the arguments the analyst is attempting to make”(Gee, 1999:88). Even if some authors (Maykut and Morehouse) recommend to transcribe everything, doing a selective transcription can be appropriate, if a justification of the selection is provided.
Indeed, a transcription process can produce multiple transcripts, more or less detailed: from a practical standpoint, it is important that “the level of detail of the transcription matches the use to be made of the transcripts” (Brown and Dowling, 1998:78). Moreover, it is worth noting that validity of a transcription doesn’t deal with how detailed the transcript is; rather, it is related to the researchers’ purpose and to the other elements of analysis as a whole.
In his study examining three alternative transcripts derived from a same piece of discourse, Mishler (1991:31) comments stating that these alternative versions don’t testify any inability of researchers and neither they are sign of technical problems. Instead, the case highlights how transcription is inherently problematic and how to “search for a standard system that could apply to any form and to any purpose is a misguided effort”. Transcription is a researcher’s construction and in this sense implies decisions that “reflect theoretical assumptions about relations between language and meaning, and between method and theory”, that in turn influence findings and their interpretation. Above all, Mishler underlines that “there is no way not to make such decisions”.
Therefore, transcription can’t be neutral just because it is an analytical act. Irrespective to the typology of transcription being adopted, such as focused (in which the focus is on particular nuances of the discourse or action in the data), unfocused (in which the attention is to what is said), indexical (in which an index puts in relation general content of the data and analytical interests), transcription can be never intended as merely procedural, even if it embeds a set of procedures to be followed. Transcription “is about analysis, and all analysis involves contextually working through specific empirical/theoretical/conceptual problems” (Gibson, p. 33). In turn, a transcript can be considered as a “theoretical entity”, which “does not stand outside an analysis, but, rather, is part of it” (Gee, 1999:88).
Furthemore, just because transcription is at the heart of the iterative process of generation of data in qualitative research, it shouldn’t hide the original data, as it is at danger in a concept of transcription as ‘translation’ (Kress et al., as mentioned in Gibson, p.32). In this concept transcript is seen as a mere guide to data and not as a means to make analysis emerge. Rather, it should preserve a proximity to original data, in an attempt to gain understanding that is always to be thought as recursive and tentative. As Heath and Hindmarsh underline: “Transcription doesn’t replace the video recording as data, but rather provides a resource from which a researcher can begin to become more familiar with details of the participants’ conduct” (2004: 19).
It should also be taken into account that trascription is always a ‘situated’ activity, whose analytic attention is oriented “towards the investigation of activities and events within the contexts in which they occur” (Heath and Hindmarsh, 2004:22).
More in-depth, Bucholz (2000:1463) maintains that “the transcription of a text always involves the inscription of a context“. It is up to researchers to make interpretive decisions (focusing on what is transcribed) and representational decisions (focusing on how is it transcribed). Such decisions always also imply an “act of power” and “ultimately respond to the contextual conditions of the transcriptions process”. These contextual conditions encompass intended use, intended audience, surrounding community, interests and position of researcher toward the text. Every decision during transcription has political effects and the analysis of the resulted transcript can’t be separated from ‘its own history’, unless researchers keep on by mistake believe that “an objective transcription is possible”. So, in a perspective of reflexive discourse analysis, Bucholz (2000:1440) highlights that researchers have to assume responsibility “of her or his role in the creation of the text and the ideological implications of the resultant product”. The author concludes that given the nature of transcription – which has always a point of view and is always partial in every sense – a responsible practice involves to “state our relationship to our transcripts”, declaring our own choices and related limitations and keeping on interrogating tapes, colleagues and the same participants, in order to “discover other ways of hearing and transcribing” (Bucholz, 2000:1462). However, it is worth noting that Bucholz contextually mentions other authors (usually outside discourse analysis) who differently solve the problem of non neutrality of transcription, by explicitly intending transcription as the production of a script, in which researcher “is as central as the other participants and the style is more literary than linguistic”.
Others (Vigoroux, 2007) discussed the character of transcription as a social activity, illustrating in a study how a research team comes to an agreed interpretation during the process of co-transcription of a video recording, within an ethnographical investigation. Unlike a long tradition in ethnography, in which textualization is conceived as a prerequisite for interpretation, this author sees interpretation as interwoven with the ongoing production of the transcribed text. More specifically, transcription is intended as being produced in a dialogical tension between experience, authorship and authority, in which there is an effort towards a sharing of linguistic and cultural identities within the different roles in the research team. “The negotiation and actualization of power and identity go hand in hand as speech is entextualized within the speaking present and into the scripted past” (Vigoroux, 2007: 90). Vigoroux utilizes the distinction between the process of transcribing (transcription) and the product of this process (scription) in order to rethink the relation between the fieldwork and transcription as mutually influencing. In this sense, transcription with its theoretical assumptions becomes part of the ethnographical investigation and helps clarify how the understanding of the context being researched is achieved by research team and participants. Confronting their competing views on transcription and shifting the attention from the scription to the making of the scription, the involved researchers become aware of the process of data construction and interpretation within the ethnographic approach.
To sum up, transcription is never a neutral analytic act because it is integral to analysis and therefore modeled on the character of the selected form of analysis, to the researcher’s aim, to the context in which events and interaction are occurring. In any case it can’t be intended as a mere procedure, but as researchers’ methodological construction of which they have to assume responsibility through reflexivity and iterative interrogation of original data and other stakeholders. Beyond the reliability of the reproduction devices, transcription is an inherently problematic act, that necessarily implies interpretational and representational decisions having political effects. Finally, to furtherly refute a naive notion of objective transcription, interactive and social dimension of transcription calls attention to the fact that transcription is based on a negotiation process aiming to gain findings in turn subjected to further exploration and interpretation.
Brown, A. and Dowling, P. (1998) Doing Research/Reading Research A Mode of Interrogation for Education, Routledge, London and New York
Bucholz, M. (2000), ‘The politics of transcription’, Journal of Pragmatics, 32:1439-65
Gee, J.P. (1999), An Introduction to Discourse Analysis: Theory and Method, London, Routledge
Heath, C. and Hindmarsh, J. (2002), ‘Analysing interaction: video, ethnography and situated conduct’, in May, T. (ed), Qualitative research in Action, London , Sage
Mishler, E. (1991), ‘Representing discourse: the rhetoric of transcription’, Journal of Narrative and Life History, 1(1):255-80
Gibson (2009), ‘Transcription’, Unit 4 Lecture Pack Qualitative Data Analysis, Institute of Education
Gibson; W. and Brown, A. (2009), Working with Qualitative Data, London, SAGE
Vigoroux, C.B. (2007)’Trans-scription as a social activity: an ethnographic approach’, Ethnography 8(1)_61-97