• No results found

3. Methodology

3.4 Beyond Validity & Reliability

Both validity and reliability constitute important criteria to assess the knowledge claims produced by research; i.e. the integrity of the findings (validity), as well as the extent to which the results can be repeated (reliability) (Miles and Huberman, 1994; Sandberg, 2005; Bryman and Bell, 2011; Guest et al., 2012). However, as I indicated in Section 3.1, these criteria assume that the world and knowledge about it is objective, meaning that truth claims can accurately be assessed. In other words, as measures ‘imported’ from positivism, they are problematic in qualitative and particularly interpretive research (Sandberg, 2005; Shah and Gorley, 2006). To conclude this methodology chapter, therefore, I briefly comment on validity and reliability notions, which as important issues I hinted at throughout the chapter, and how they can be perceived in an interpretive study.

Some qualitative scholars nonetheless argue for maintaining validity and reliably regardless of the methodology applied and in spite of its underlying positivistic assumptions (e.g. Morse et al., 2002; Silverman, 2011; Guest et al., 2012). However, if as proposed earlier truth is multiple instead of absolute, how can we say that a truth claim is ‘valid’, and that the way we arrived at it is completely ‘reliable’? It is for this reason that I echo the view of applying alternative measures to assess and justify knowledge produced via interpretive approaches (e.g Lincoln and Guba, 1985; Auerbach and Silverstein, 2003; Sandberg, 2005; Shah and Gorley, 2006; Bryman and Bell, 2011). Indeed, as Sandberg (2005: 62) suggests, “although objective knowledge is untenable, it is still possible to make truth claims consistent with the ontological and epistemological assumptions underlying the approach to interpretative research.” To assess said truth claims, Lincoln and Guba’s (1985) landmark publication is particularly useful. Lincoln and Guba propose four criteria for ensuring rigour in qualitative research: credibility, transferability, dependability, and confirmability (see also Shah and Gorley, 2006; Bryman and Bell, 2011).

To achieve credibility, prolonged engagement in the field is required, which, as suggested in Section 3.2, constitutes a key characteristic of ethnography and was achieved in this thesis. However, some argue that mere engagement is not sufficient (see Silverman, 2011). Based on Lincoln and Guba (1985), Shah and Gorley (2006) therefore suggest the use of multiple methods, which is commonly referred to as

triangulation (Agar, 1980; Sandberg, 2005; Jonson and Jehn, 2009; Bryman and Bell, 2011; Silverman, 2011; Guest et al., 2012). Since people may say and do different things, I indeed deployed both participant observations and supplemented these with several forms of interviews. A further major technique to ensure credibility is peer involvement, i.e. discussing findings and interpretations thereof with colleagues (Lincoln and Guba, 1985; Armstrong et al., 1997; Sandberg, 2005; Shah and Gorley, 2006; Bryman and Bell, 2011). To assess prevalence of my themes, for instance, it was not only my judgement at play (Sandberg, 2005; Braun and Clarke, 2006; Guest et al., 2012), but also that of fellow students who also engaged in ethnographic work and thematic analysis, as well as my supervisors. A similar and popular means is taking the findings back to the research participants, called ‘respondent validation’ (Lincoln and Guba, 1985; Sandberg, 2005; Shah and Gorley, 2006; Bryman and Bell, 2011; Silverman, 2011). To do so I regularly discussed my interpretations of things with my research participants during my time in the field (e.g. with Jack; see Section 3.2), and also went back to the research site to present and discuss my themes in order to ensure that they were credible (see Step 4, Section 3.3). Finally, an important issue is transparency of the research process (Miles and Huberman, 1994; Guest et al., 2012), which this methodology chapter itself is a testimony to. In fact, transparency in terms of an ‘audit approach’ – i.e. keeping records of steps taken and decisions made, data, etc. – also pertains to dependability, which represents a parallel reliability (see Bryman and Bell, 2011).

Transferability, meanwhile, accounts for the virtue of the contextual richness and uniqueness of qualitative findings (Bryman and Bell, 2011). Qualitative researchers are therefore encouraged to produce accounts that contain ‘thick descriptions’ (Geertz, 1973) of the data. I referred to this necessity in relation to the centrality of textwork in ethnographic endeavours in order to invoke a sense of participation and immersion (Section 3.2). For Lincoln and Guba (1985), thick descriptions provide “others with what they refer to as a database for making judgements about the possible transferability of findings to other milieux” (Bryman and Bell, 2011: 398). Such descriptions of social encounters at and of TexCo, therefore, are offered in Chapter 4 in the form of several vignettes that illustrate my immersion in the setting.

Finally, confirmability is concerned with ensuring that the researcher “has not overly allowed personal values or theoretical inclinations manifestly to sway the conduct of the research and findings deriving from it” (Bryman and Bell, 2011: 398). At the same time, however, I echoed several scholars, arguing that as researchers we inevitably operate within certain beliefs, assumptions, and theoretical preconceptions (e.g. Cunliffe, 2003; Sandberg, 2005; Yanow, 2012). To ensure confirmability, therefore, I proposed reflexivity to be all the more important in order to assess my values and impact on the setting and interpretations (Cunliffe, 2003; Sandberg, 2005; Alvesson et al., 2008; Alvesson and Sköldberg, 2011; Silverman, 2011), such as through the use of my sensitising framework, regularly keeping a diary, as well as embracing contradictions and tensions (see Section 3.2, 3.3; Sandberg, 2005; Nadin and Cassell, 2006).

3.5 Summary

In summary, the preceding three sections – research philosophy, data collection, data analysis – illustrate my overall research design applied to answer my research questions, which I illustrated also in terms of my ‘thematic map’ towards the end of Section 3.1.

In terms of my research philosophy, I adopted a perspective that argues practices are constitutive of the social world and hence social reality. At the same time, since this world comes into being via everyday enactments of practice, it is at the same time socially constructed. Since this means that actors cannot be separated from the world, the ontology adopted is subjective. Indeed, as actors, including researchers, we are always entwined rather than separate from the world. Concurrently this means that there are always multiple interpretations about the world, and that it can only be known through participation in it – exhibiting implications for my epistemological position. That is, in line with other practice scholars, I adopted an interpretivist approach to knowledge, suggesting that as part of my research I am mainly interested in the possible and indeed multiple meanings of social phenomena. The implications of such a constructionist-interpretivist philosophy, for instance, pertained to working inductively and reflexively in both data collection and analysis.

A further vital implication of this philosophy is a qualitative methodology. Whilst possibilities are multiple, I adopted an ethnographic approach because of its power to engage the mundane nature of everyday life in organisations and its interest in local knowledge, practices and culture. Moreover, ethnography ‘violates’ the fundamental idea of positivism, ontological objectivity, through externality and detachment. I therefore immersed myself for five and a half months in a British textiles’ organisation, TexCo.

The core methods of ethnography, meanwhile, constitute daily and ongoing participant observations. Fieldnotes (n=250 pages in a Word document), therefore, were my primary source of data. As common practice though, ethnographers are fond of supplementing observations with several interview techniques. I thus conducted 39 semi-structured, unstructured, as well as ‘interviews to the double’ (a technique aimed at foregrounding moral and normative dimensions of practices). Because of my close

participation in the local setting, moreover, I also kept a reflexive diary. This became a habit so much so that the diary and memos were a tremendous aid during data analysis.

To analyse my data subsequent to gathering it, I opted for a thematic analysis, since it works particularly well with the ethnographic virtue of induction and flexibility. Also, I found thematic analysis to be overly systematic, helping to ensure comfirmability of the study – which is particularly delicate in interpretative philosophy and qualitative methodologies (see Guest et al., 2012). I hence followed a six-step process model, which was nonetheless fluid and iterative rather than rigid and linear. In my particular case I in fact added a step between write up and defining the different themes: namely to conceptualise and analyse the relations between the themes. The process I undertook was as follows. As a first step I familiarised myself with the data set, which occurred mainly during data collection (e.g. via transcribing interviews) already. Next, I inductively coded the entire data set, after which I analysed it for patterns and similarities in order to draft a preliminary set of themes. Once these themes were in place, I continuously reviewed the themes for internal homogeneity and external heterogeneity before eventually establishing the (semi) final themes. Normally, one would now write up the report of the themes (i.e. the findings section of a thesis). However, as mentioned, I inserted the step of analysing the relations of the themes. This yielded three interrelated thematic groups instead of as previously five distinct themes. From these groups several tensions appeared, which I explored further and narrowed down to three, and interrogated vis-à-vis existing theory and my research questions; which eventually constitute the way I present my findings in the next chapter; namely, ‘Old & New’, ‘The ‘Boundedness’ of Empowerment’, as well as ‘Inclusive Practice & Elitism’. At the end of each section, moreover, I comment on how each theme indicates the relationship of leadership and power dynamics.

4. FINDINGS

In this chapter, I present my empirical account based on the methodology outlined in the foregoing chapter. This findings chapter in turn consists of three main sections. In the first section, Contextual Background (4.1), I set the scene for the ethnography, and primarily discuss the ways in which TexCo’s practices have evolved through said change initiative – in order to shed light on the practices of the present. In the second section, The ‘Boundedness’ of Empowerment (4.2), I explore the ways in which empowerment was enacted at TexCo, how it is produced to begin with and how it might be limited. This involves examining a puzzling contradiction: is empowerment in fact generated through notions of authority, accountability, and legitimacy? Lastly, in the third section, Inclusive Practice & Elitism (4.3), I delve into the ways in which empowerment translates into opportunities to participate in certain practices, and how simultaneously notions of authority and hierarchy shape these opportunities. In other words, I examine a tension between working inclusively and collaboratively on the one hand, and how merely a handful of ‘powerful’ actors shaped the ways in which work gets performed on the other.

These three sections combined comprise 11 vignettes, which as suggested in Chapter 3 are supplemented with further observational and interview material, as well as recollections of events. Also, I continue my inductive and data-driven approach for the time being, by presenting the data as largely un-theorised. Whilst not engaging explicitly with the literature yet, I do return to a more theory-driven approach at the end of each section, in order to illuminate how power and leadership dynamics were evident in the data presented beforehand.

As also mentioned previously, my analysis reveals three major ‘expressions’ of power: shaping norms, enablement and constraint, as well as inclusion and exclusion. Although these expressions are evident in each of the sections, they map nicely on 4.1, 4.2, and 4.3, respectively. Therefore, at the end of each section, I concentrate on one expression in particular, discussing how it is evident in the data presented. For future reference and clarity, Table 6 lists every main character that appears in this chapter, including the position held by the individual.

Character Position held

Jack CEO

Maria POD Manager

Patrick POD Manager

Jacob Kelly

POD Manager POD Manager

Igor Product Development Manager

Phil Product technician, reporting to Igor

Madeleine POD assistant to Patrick

Keith In charge of Composites’ POD shop floor

Garry In charge of machine maintenance

Linda Director of operations

Lucy Production/Purchasing scheduling

Donald Product development

Adrian Assisting Lucy with scheduling

Arianne Receptionist

Abigail Formerly in charge of production

scheduling, now quality testing Kacy

James

Customer service representative Quality engineer