Chapter 3. Methodology
3.1 Conversation Analysis Methodology
3.1.4 CA data collection methods and data analysis procedures
CA methodology features totally data-driven and bottom-up methods with emic perspectives. The procedure of conducting CA research starts with data collection, which differs from other research beginning with a problem of social phenomena or an assumption of certain event/activity/theory. Researchers first choose the research site; that is, what kind of interaction is the choice: either ordinary conversation or
institutional interaction. “Ordinary conversation means informal, casual conversation without specific institutional goals or tasks.…Many practices of ordinary conversation are ubiquitous in talk, and research material can hence be collected from almost
anywhere” (Peräkylä, 2004, p. 169). The naturally-occurring discourse data are the core necessity and serve as primary sources for the research (Markee, 2000; Wooffitt, 2005) and are usually recorded. Pomerantz and Fehr (1997) offer advantages for employing recorded data for analysis:
First, certain features of the details of actions in interaction are not recoverable in any other way. Second, a recording makes it possible to play and replay the interaction, which is important both for transcribing and for developing an
62
the materials, in all their detail, that were used to produce the analysis. Finally, a recording makes it possible to return to an interaction with new analytic interests. (p.70)
After recording, those recorded data should be transcribed in detail. The evolution of transcription conventions emerges from Gail Jefferson who “devised a system of transcribing which uses symbols available on conventional typewriter and computer keyboards” (Wooffitt, 2005, p. 11). Symbols involve “a wide variety of vocal and interactional phenomena, including pitch variation, prolongation of sounds, amplitude, overlapping speech and silences” (Peräkylä, 2004, p. 169). Researchers can create their own symbols for their CA transcription to fulfil their different necessities. Both those recorded data and transcriptions can be stored in various ways (e.g., USB, computer and cloud storage), which permits analysts to retrieve them repeatedly. The availability and convenience of employing latest technologies allow researchers not only to obtain
naturally-occurring data easily but also record and store the details of talk-in-interaction.
As for online text chat, there is no need to record the talk-in-interaction while the participants’ online conversation is under way. Researchers can retrieve participants’ text-based data from the computer screens after their talks are closed. Though both paralinguistic and prosodic cues (visual and auditory) underpinning the turn-taking organization in spoken data do not exist, participants produce other strategies to cope with the turn-taking problems due to the lack of visible cues (Negretti, 1999). In spite of the lack of kinesic and prosodic features, online text chat can attribute to “the real-time (synchronous) nature of chat communication which obliges participants to ‘think on their feet’ and co-construct online talk, as occurs in face-to-face conversation” (Tudini, 2010, p. 1).
No transcription is needed for online text chat, which also signals the data collection method of “non-tradition CA” (i.e. working with text-based data). The text-based data retrieved directly from the Internet are the production of participants’ collaboration and control. The retrieved online scripts are completely authentic and reveal certain
paralanguages different from face-to-face conversation. For example, they can also use emoticons, punctuations and search engines to facilitate their mutual understanding in their online talk-in-interaction. The features of online paralanguages as social actions create new aspects for research and enrich CA studies (see also section 4.4.2 online scripts).
63
As for the procedure of data analysis, Peräkylä (2004) suggests the stages of starting CA analysis. The first one is unmotivated exploration of the data, which means “being open to discovering new phenomena rather than searching the data with preconceptions or hypotheses” (Seedhouse, 2004, p. 38). The initial observation of the data starts with listening and watching the recordings repeatedly and examining the transcripts simultaneously. Analysts sometimes focus on very small and trivial parts and at other times on larger entities “trying to explicate the organization of what is happening in the recorded interactions” (Peräkylä, 2004, p. 170). The focus is to identify the phenomena for examination as the second stage. In fact, the phenomena can be something that is exciting or challenging of a specific practice or sequence in the data. The third stage is collection of instances of the phenomenon. “Once a candidate phenomenon has been identified, the next phase is normally an inductive search through a database to establish a collection of instances of the phenomenon” (Seedhouse, 2004, p. 39). Peräkylä (2004) advices that the collection of instances should be inclusive instead of exclusive because instances not fitted to the collection can always be got rid of later. The next stage is to determine the variation of the phenomenon and form the regularities and patterns related to occurrences of the phenomenon, which is “to show that these regularities are
methodically produced and oriented to by the participants as normative organizations of action” (Seedhouse, 2004, p. 39). As for deviant cases, they can be seen and serve as demonstration of the normativity of practices (Heritage, 1995; Seedhouse, 2004). The final stage of CA research is a fine and logical description of the various types of realization of sequences or actions under examination.