Part I: Background and Context
2.4 Methodological Strategies
2.4.3 Key Informant Interviews
Qualitative interviews with key informants are also used in this project in order to supplement and support the background and historical analysis of Chapter 3, as well as the CADS analysis of the news texts. The qualitative interview is a common feature of interpretive approaches to research (Kvale, 2007), often combined with fieldwork and document analysis (Kvale, 1996; Rubin & Rubin, 1995; Warren, 2002), and also often deployed in the service of discourse‐theoretical analysis (Hansen & Sørensen, 2005; Howarth, 2005). According to Kvale (2007), the qualitative interview is “a specific form of conversation where knowledge is produced through the interaction between an interviewer and an interviewee” (p. xvii). It is a “professional interaction, which goes beyond the spontaneous exchange of views [...] and becomes a careful questioning and listening approach [...]. The qualitative interview is a construction site for knowledge” (Kvale, 2007, p. 7).
An important decision in qualitative interviews is who, and how many, does one interview? For survey‐style research, interviews are conducted with a sample population selected in order to then generalise the findings (Warren, 2002). In qualitative interviews, however, the emphasis is on the importance of personalised subjective experiences in explaining social phenomena, and the contextualised “thick descriptions” of social reality produced through the interview process (Howarth, 2005, p. 338). Respondents are therefore specifically targeted for their connection to the social phenomena under investigation. They may be chosen “based on a priori research design, theoretical sampling, or ‘snowball’ or
74 convenience design, or particular respondents may be sought out to act as key informants” (Warren, 2002, p. 87).
This project adopts a ‘key informant‘ approach (Holstein & Gubrium, 1995; Spradley, 1979). Used often in anthropology and ethnographic research, key informants are defined as “native speakers” (Spradley, 1979, p. 8), in the sense that they are in possession of unique inside knowledge of some social world (Warren, 2002), and are ideally en‐ culturated into the phenomena under investigation, currently involved in it, and non‐analytical in their disposition (Spradley, 1979). Spradley’s (1979) criteria for a good key informant, however, have been criticised by Ervin (2005) as too rigid for many circumstances. This thesis, therefore, adopts Ervin’s (2005) definition of a possible key informant as a ‘frontline worker’, and chooses for its candidates journalists who were involved in the production of news media content found in the sample of news archives outlined above.
The ‘front line’ news media workers interviewed for this project are Tina Rosenberg and Donald G. McNeil Jr. of TheNewYorkTimes. They were identified based on findings from the CADS analysis, where a combination of frequency of articles and an interpretive assessment of the relative importance of the individual articles, were used to identify a short‐list of ten candidate key informants. These ten individuals were then approached, following the required ethics approval from the researcher’s university. Only two however, Rosenberg and McNeil Jr., accepted the invitation and were successfully interviewed, in both cases via telephone on 5 October 2010 and 4 June 2010 respectively. The recorded interviews were then transcribed, with transcriptions provided to the interviewees for approval. Once approval was attained, the transcribed texts were included in the collection of primary documents under consideration in this research. In the process of interviewing McNeil Jr., the
75 interviewee also provided comments in an email exchange, as well as an unpublished manuscript outlining his experiences covering the medicines access dispute. Both of these additional sources have been included, with the interviewee’s approval, for consideration in this thesis.
To conclude, it is worthwhile to briefly consider the strengths and weaknesses of including interview data in this project. On the positive side, qualitative interviews are widely recognised to provide novel first hand accounts of social phenomena. As Hansen and Sørensen (2005) noted:
… interviews present discursive images of the world that are much less sanctioned, formalized, and rationalized than documents [...] Furthermore, knowledge is obtained about the ways individuals in different positions within a given discursive structure construct meaning and identity ‐ and thus knowledge ‐ about themselves and the other actors. (Hansen & Sørensen, 2005, p. 99)
On the negative, however, the reliance on personal perspectives conjures questions of accuracy, or rather, as Hansen and Sørensen (2005) put it: “how do we know that the informants do not lie in order to please or tease the interviewer?” (Hansen & Sørensen, 2005, p. 99). However, as discourse theorists Hansen and Sørensen (2005) and Howarth (2005) have argued, this aspect, while potentially disruptive for positivist research, is of little concern for discourse theory where the distinction between material truth and symbolic discourse is collapsed. For Laclau and Mouffe (1985), what is ‘said’ and what is ‘done’ are both practices of articulation. Any differences between an interviewee’s statements and putatively objective accounts are therefore not differences between truth and falsity, but rather differences in discursive contexts where the rules for the
76 production of statements are most likely distinct (Hansen & Sørensen, 2005). As discussed above, the task of the discourse‐theoretical analyst is not to draw out the distinction between distorted false representation of ‘reality’ and reality itself, but to examine the always‐already distorted constructions of reality in specific contexts (Torfing, 2005).