Braun and Clarke (2006) note that, in essence, ‘thematic analysis is a method for identifying, analysing, and reporting patterns (themes) within data’ (p6) and differentiate between inductive and theoretical thematic analysis, with these being the ‘inductive or ‘bottom up’ way…or…a theoretical…‘top down’ way’ (p12). Fereday and Muir-Cochrane (2006) demonstrate the greater rigour that can be achieved in thematic analysis when a ‘hybrid approach’ (p80) is taken. In practice, this constitutes a combination of inductive and deductive thematic analysis, with the aim of harnessing the advantages of each. In effect, this constitutes the use of both pre-ordinate and responsive codes and categories (Cohen, et al., 2007). The pre-ordinate categories involve application of the explicit theoretical framework developed through engagement with the literature and in this sense the analysis includes a deductive element. However, the approach also entails a thematic analysis, involving the generation of themes from the data, and so
59 includes an inductive element (Braun & Clarke, 2006). This combined approach helps to ensure that the voice of the participants is valued, whilst simultaneously allowing for a more theory-led analysis. In respect to the sequencing of the analytical process, this follows the approach set forward by Fereday and Muir-Cochrane (2006), which, for the purposes of written clarity is ‘presented as a linear, step-by- step procedure’ though in fact ‘an iterative and reflexive process’ (p83). However, to avoid suppression of variation and theoretical bias, the inductive analysis is to be entered into first, with the integration of the theoretical framework at a more secondary stage of analysis (Clarke & Braun, 2013, p. 257).
In specific respect to the theoretical thematic analysis, this can be seen to possess real value, because Hayes (1997) notes, this ‘offers a…theory-driven approach…with the flexibility and richness of
qualitative analysis’ (p113). However, it is important that such a process carries rigour to avoid theoretical bias. For the theory-led element of the thematic analysis, therefore, Boyatis (1998, p36) describes three stages: (1) to establish the themes ‘through reading and contemplation [of] the theory’, (2) to check the ‘compatibility with the raw information’ through pilot coding and (3) ‘to determine the reliability of the coder’ through inter-rater reliability testing. Each of these stages has been adhered to, with pilot coding indicating an applicable theory and inter-rater reliability proving very encouraging. Using the approach outlined by Boyatis (1998), when applied to an interview transcript, this resulted in a percentage agreement of 93%, with PhD main supervisor as second rater. However, many, beginning with Cohen (1960), have noted that percentage agreement does not allow for chance agreements to be accounted for statistically. Therefore, a widely accepted approach for more robust estimation of inter- rater reliability can be attained with Cohen’s Kappa (again, with PhD main supervisor as second rater), as noted by the likes of McHugh (2012). Very encouragingly, this emerged as:
Table 4: Measure of Kappa Agreement
Value
Asymp. Std.
Errora Approx. Tb Approx. Sig.
Measure of Kappa
Agreement .874 .060 10.548 .000
Such a figure of .874 would be regarded as representing strong and robust inter-rater reliability (e.g. McHugh, 2012; Landis and Koch, 1977).
60 Other virtues of such an approach are noted by Hayes (1997), including an answer to the ‘question of reflexivity’ (p112) which can arise in relation to qualitative work, namely that the research will be unduly influenced (inadvertently or otherwise) by the perspective and actions of the researcher. If the
researcher is using an existing theory (in a critically-minded way), then this can work to provide a check to the subjectivity of said researcher. This said, Boyatis (1998) nonetheless contends that there is still a risk of ‘projection’ (p35) when using a theory-driven code. Therefore, as Hayes (1997) also
acknowledges, there is still a need for a second perspective in the form of a critical friend who independently analyses the transcript data and in the present research, this process is undertaken as part of the PhD supervision process and using the three stage approach of Boyatzis (1998) above. Hayes (1997) also noted the value of ‘pre-determined themes to structure the analysis [of]…large amounts of data’ (p112) derived from qualitative interviews. Hayes (1997) also contends that: ‘the
method…requires a dynamic, broad-ranging theory with real-world applicability’ (p113). Given the wide practical application of SDT delineated in the literature review section, this would seem an appropriately flexible and inclusive framework for such an approach.
However, a departure from Hayes (1997) perspective occurs in relation to the view that ‘novel material, inappropriate to the theoretical themes, will not be included in the analysis – unlike grounded theory approaches’ (p113). This is as a consequence of the critical realist approach to the study. The principles of abduction and retroduction are brought to bear when considering the applicability of theory, as noted in relation to critical realism in Section 3.2. This is because, as Meyer and Lunnay (2013, online) argue, when ‘used in conjunction [abduction and retroduction]…lead to the formation of a new
conceptual framework or theory’. Therefore, such an approach should be theory-generative, rather than simply theory-confirmatory, negating the charge of unquestioning and biased acceptance of existing constructs that might be levelled against a theoretical-thematic approach by a grounded theorist. Indeed, it could be argued that a theory-led approach has a greater degree of openness and
transparency as to its influences. It would be a widely acknowledged argument to say that grounded theory must always struggle to ‘bracket-out’ prior awareness of theory. Therefore, to acknowledge such theoretical influences explicitly, whilst also being critically minded as to their utility, offers an arguably more valid and open approach.
In summary, in each instance of qualitative analysis (the open survey responses and the qualitative interviews respectively), the first section of the findings presents the bottom-up inductive thematic analysis that foregrounds teachers’ voices and uses selected quotes to illustrate the themes generated
61 in order to provide some transparency to the interpretation of the data. The second section presents the application of the self-determination theory framework as a deductive top-down tool for analysis that foregrounds theory and again presents illustrative quotes to make explicit the interpretation of the data.