Four Stages of Analysis

I return to my research question: “What does the language of representation of the BP crisis ‘look like’?” Such a broad aim called for an open- ended and interactive approach, which would allow the data themselves to suggest areas of particular linguistic interest. This emergent approach to analysis can be summarised as having four stages:

  • 1. A contextualisation stage. A broad description of the data sets. What sort of texts were they, and where did they come from?
  • 2. An immersion stage—Preliminary analysis by reading and rereading the texts in order to identify language features of interest for analysis, at Barthes’ four semiotic levels.
  • 3. A depth analysis stage—Investigating these language features and analysing them for frequency, function and change.
  • 4. A holistic analysis stage. What does an in-depth analysis of a single text tell us about how these features interact? Can this exercise describe the “language map” of the representation?

Table 5.1 summarises the stages of the analysis process for BP:

Table 5.1 Four stages of data analysis

BP data



Analysis aim

Stage 1



Full data sets:

  • 2010 (169 texts)
  • 2011 (94 texts)
  • 2012 (31 texts)


Overview of country of origin, genres, salience of BP story

Stage 2

Preliminary analysis stage (immersion)

Data subsets:

  • 2010 (20 texts)
  • 2011 (20 texts)
  • 2012 (20 texts)


"Immersion" in data to identify significant features of language use

Stage 3

Depth analysis stage

Data subsets:

  • 2010 (20 texts)
  • 2011 (20 texts)
  • 2012 (20 texts)



Analysis of frequency and type of significant features.

Analysis of language usages in context

Stage 4

Holistic text analysis

Data subsets: Single texts as required


Single text analysis in-depth—language usages in combination

Chapters 6, 7, 8, 9, 10, 11, and 12 describe explicitly how each of these four stages of analysis can be applied to data, using the example of the BP texts. At each stage, findings from the data analysis are outlined in brief. This not only gives an idea of the kinds of information generated by the analysis approach, but also sets out a source of data which can be further interpreted using semiotic concepts.

< Prev   CONTENTS   Source   Next >