Methods, Themes; Data Analysis

Discourse analysis has been described as a ‘messy method’ (Tonkiss, 1998: 250) which is difficult to formalise into a standard approach. Instead of focusing on quantity of data, as is the case with other methods such as content analysis, discourse analysis attempts to uncover data that is rich in textual detail, providing fruitful insights into a particular issue under research (ibid: 246). Fran Tonkiss remarks that in this regard, even ‘a single speech or newspaper report or conversation can generate very fruitful themes for analysis’ (ibid:

252-253). This thesis made use of qualitative interviews (elite interviews) and case studies to reply the research questions.

Discourse analysis is not institutionalised and standardised as is the case with other methods with a longer sociological tradition. Given that the use of discourse analysis is not as mainstreamed as other more conventional methods, evaluation of studies using this research method may be difficult and maybe even not appreciated enough, hence leading to criticism. On the other hand, one should appreciate its labour-intensive and time-consuming qualities (Phillips and Hardy, 2002: 11), whilst acknowledging that it is one method amongst a plurality of equally important methods (ibid: 16). As Phillips and Hardy (2002) put it

‘discourses are not neatly packaged in a particular text or even in a particular cluster of texts. Researchers can only trace clues to them regardless of how much data they collect’ (74).

Keeping such challenges in mind, as an initial guide for data collection and analysis, I applied Virginia Braun and Victoria Clarke’s (2006), six phases of what they define as

“thematic analysis”. The phases consist of familiarizing oneself with one’s data;

transcribing data (if necessary), and formulating initial ideas; generating initial codes by systematically coding data deemed interesting from the entire data set; collating codes to form potential themes, whilst collecting data for each theme; checking whether the themes work with both the coded extracts and the entire data set; giving a name and definition to the themes, whilst refining the specificity of each; and, producing a final report.

In relation to the above, Howarth (2000) proposes three basic operations for the analysis of empirical data through discourse analysis. First, ‘the “translation” of information into textual form’ (141). Here, linguistic and non-linguistic data is treated as text which can be analysed through techniques that are in line with the ontological assumptions of discourse theory. Second, ‘the application of constructed theoretical frameworks to the problematized object of investigation’ (ibid). Here, abstract logics and concepts are applied to a specific case. Third, ‘the deployment of the various techniques of discourse analysis to the problem investigated’ (ibid). In each case, theoretical flexibility takes place through the application process, as theoretical concepts are articulated during the research process itself (Howarth and Stavrakakis, 2000: 5).

This study applied Howarth’s proposals as follows. As regards the translation of information into textual form, different forms of data were collected (Geertz, 1973;

Phillips and Hardy, 2002; Wetherell, 2001). The qualitative elite interviews and case studies were treated as text for analysis. Collection of information from newspapers did not involve accumulation of as much data as possible with the intention of discovering discursive truths through numerical presentation of key text, but rather, involved the unearthing of ‘essences with sufficient context’ (Wolcott, 2001: 44). Notwithstanding this, the number of newspaper articles collected was extensive.

As regards the construction of theoretical frameworks, this was carried out through key concepts related to hegemonic struggles namely articulation, antagonism, equivalence, hegemonic formations, overdetermination and dislocation (Laclau and Mouffe, 1985;

Howarth, 2000; Howarth and Stavrakakis, 2000), and, where applicable, through substantive, procedural, structural and sensitizing impacts of ENGOs (Giugni; 1995; van der Heijden 1999; Carter 2001; Rootes 2007a). After being initially explored through the narratives of ENGO respondents, thus raising queries on their potential application, these were analytically applied to each case study.

Discursive techniques were used to analyse the empowerment of ENGOs. Data was sorted and coded into keywords and themes (Tonkiss, 1998: 253-254) in relation to ENGO politics, namely ideology; institutionalisation; State power; Europeanization; and empowerment. Variation in the text (Tonkiss, 1998: 256) was analysed, for example by looking at how social agents, such as ENGOs with different ideologies, discursively relate to specific concepts and issues. I also read for emphasis and detail and attended to silences (Tonkiss, 1998: 257-258). In the case of the former, I carefully analysed how discourse was used by different social agents in strategic ways. For example, I noted how at times the Government made use of ‘rational arbiter’ discourse to seemingly depict itself as being a fair judge on environmental issues. In the case of the latter, I noted how, for example, social agents such as ENGOs and political parties were not always equally active on all issues. For example, in the case of bluefin tuna, Maltese ENGO tended to be much less active than they were in other issues. Hence, less discursive evidence for analysis for produced, leading to a relatively shorter case study.

Ideologies, keywords and arguments utilised by social agents in their articulatory practices were identified through their respective discourses. Such a method is similar to

that proposed by Volkmar Lauber and Elisa Schenner (2011) and by the “new political ecology” which focuses on environmental struggles and power relations and on discourses which are articulated by the State and movements (Hannigan, 2006: 55).

The next section will focus on the first of the research methods being used within the framework of discourse analysis, namely elite interviewing.