Chapter 2: The Context of Equity in Skills Policy
4.11 Grounded Theory Guidelines
4.11.4 Theoretical Sampling, Abduction, Saturation and Sorting
4.11.4.1 Theoretical Sampling
Theoretical sampling is a necessity for undertaking grounded theory although it is often overlooked (Charmaz, 2014). Charmaz (2014) writes “Initial sampling in grounded theory gets you started; theoretical sampling guides where you go” (p. 197). The aim of theoretical sampling is to
“…saturate your categories with data and subsequently sort and/or diagram them to integrate your emerging theory” (Charmaz, 2014, p. 192). Theoretical sampling does not occur to address initial research questions, to reflect population distributions, to find negative cases or to determine that no new data are emerging.
I needed a conceptual framework for theoretical sampling and selected the guidance of Charmaz (2014). The conceptual model was: “theoretical plausibility, direction, centrality and adequacy” (p. 87) and I kept these words in front of me during the process of sampling theory. I began to organise theoretical sampling at the point where the data, coding and memo’ing had produced tentative focused codes that could equally be categories. I then used various strategies of theoretical sampling which Charmaz (2014) writes can entail “… studying documents, conducting observations, or participating in new social worlds as well as interviewing or re-interviewing with a focus on your theoretical categories” (p. 206). I was interested to note that it was the “plausible” question that figured the most in the theoretical sampling. I was constantly asking myself of a theory: is this plausible? There were several occasions when although my intuitive answer was “no”, the evidence said otherwise.
The strategies that I used for theoretical sampling were re-interviewing and observations. Theoretical testing to some degree was circumscribed by the contingencies of access to people, especially learners, because they were too busy, in another place or not interested and Charmaz (2014) alerts us to this potential problem. I therefore dwelt at length on what I needed from theoretical testing and determined that it was to re-interview people who spoke not only from their own position but also with regard to the broader aspects of equity in VET. I re-interviewed three policy-makers, two teachers and one learner. For interview, I took with me the lists of codes and categories I wished to discuss (see Appendix K). I placed the list in front of each person and encouraged them to reflect, or to “think aloud”, about their reactions.
The process of theoretical sampling caused me to position in the substantive grounded theory (5.2) the category as “loving to learn and”. Subsequently, the category cascaded into other categories of
“accommodatingterror” and “beingamazed”. Significantly, my approach to theoretical sampling, also drove emerging categories in the basic social organisational problem of “silentvoices”. Beyond these category changes, the interviewed informed the properties of the theory which are set out in Chapter 5.
4.11.4.2
Abduction
It is difficult to pin down definitions of abduction, how to abduct, and what is its purpose. A
definition might include that a researcher inducts evidence and then abducts by asking the “what-if”
question (Neuman, 2003, p. 112). In conGT, abduction is part of the pragmatist theoretical perspective and is inherent in the process of inferring to make meaning. The process of abduction involves: seeing inducted evidence in a surprising new way; scrutinising, hypothesising and testing to confirm or disconfirm possible explanations; creating a new theory, or organising extant theories in a novel way (Charmaz, 2014, p. 201, p. 341). Abduction is also imaginative reasoning which remains grounded in induction and empirical analysis. Reichertz (2007) proposes that abduction was present but not systemised in classic grounded theory although Strauss, with Corbin, became more and more abductive (p. 215). Reichaertz points out that abduction can be risky because it is an intellectual exercise that takes one away from the conventional view of things. However, abduction does require a researcher to induce situations (p. 221), perhaps by taking distance from analysis and even
daydreaming. In this way, it may not be logically orderedor an exact method. However, Reichaertz (2007) contends that abduction does question previously developed knowledge and it does produce new knowledge by taking data very seriously.
I was well into theoretical sampling before forming a systematic analysis from which I felt confident to abduct. Abduction, the “inferentialleap” (Charmaz, 2014, p. 200) then caused a major return to
the analysis that had gone before. Abduction took me back to review the nature of my own axiology and caused me to represent and rewrite social justice in philosophical terms as well as theoretical perspective. Abduction was especially useful in connecting disciplines in this interdisciplinary study. I abducted when I realised that some concepts in disciplines that appeared different were very similar or the same. The most notable abduction involved hypothesis about the meaning of “traditions and
dilemmas” in political science with reference to concepts of “habitus” in philosophy and “learner disposition” in education. This abduction led to a significant breakthrough in the categorising in the substantive grounded theory of “absence of learner disposition” and was one device that opened up new theorising for the potential formal theory.
4.11.4.3
Saturation
Saturation is one of the purposes of theoretical sampling to help delineate the properties of a category and to clarify relationships between emerging categories (Charmaz, 2014, p. 212).
Saturation is not repetition of empirical evidence. The adept grounded theorist may generate theory on small numbers which is superior to theory generated by other researchers bounded by using large numbers (Glaser, 2001). Among grounded theorists, the debate is not about saturation through numbers but more about “theoreticalsufficiency” (Charmaz, 2014, p. 215). This gives rise to how does one prove saturation in grounded theory when it is a subjective determination.
Charmaz (2014) suggests that the solution lies in using grounded theory guidelines to “… give you a handle on the material, not a machine that does the work for you” (p. 216). I determined that I had reached saturation when no new theory began to emerge and when I was satisfied that my
categories were genuinely expressing the information raised before theoretical testing.
4.11.4.4
Sorting
Sorting in grounded theory is a strategy to serve the purpose of theorising and will include systematic diagramming and integrating – of which I did a great deal from the beginning of fieldwork. Charmaz (2014) suggests that sorting in grounded theory is different from the processes of other qualitative research because it is the framework for the theoretical guidelines of memos and clustering.
I sorted by extracting printed copies of memo’ed and coded observations and categories – on a desk. I stood back and looked at it, and wrote on butcher’s paper and a whiteboard to cluster what I was seeing. I would return to the computer and diagram again. When the first major categories emerged, I began to sort by sketching the thesis contents page and to outline chapters. But I wrote as an analyst for many drafts; I did not undertake final thesis writing until much later.
4.11.4.5
Clustering
Constant clustering is a method of conGT although clustering is generally used in qualitative research. Clustering that is constant and involves “playing” with ideas” (Charmaz, 2014, p. 184). Clustering provides an “image of how your topic fits together” (Charmaz, 2014, p. 184). Clustering is a form of mind-mapping involving circling and linking multiple although potentially connected thoughts. I clustered using “boxes” drawing on: concepts, especially . productivity, learner
disposition; memos, especially those I categorised as “big ideas arising”; gerunds, especially when I wanted to compare these across macro, meso and micro levels of inquiry e.g. “regretting”,“loving”;
emerging categories, especially when I was beginning to move tentative categories, and later sub- categories, towards core categories. I constantly clustered by hand or on the computer and most often on butcher’s paper.
Clustering allowed for free-wheeling expression and organising of ideas. One of the concepts most subjected to clustering was “emotion”. I explored multiple aspects of emotion from each participant group. In the first instance, the clustering was around “emotion”; then it was among apparently unconnected ideas such as policy-makers remembering the past, or teachers and trainers loving their work. The various clusters were eventually subjected to comparison and from there began to emerge properties and ultimately sub-categories of the substantive grounded theory (Chapter 5.) Constant clustering drove the theorising a great deal.
4.11.5
Constant Comparison
Constant comparison is the cornerstone to traditional grounded theory and to conGT. If the
researcher’s philosophical position recognises multiple realities, there must follow a commitment to constant comparison to make numerous realities visible (Charmaz, 2014). The constant comparison method compares everything with everything: data with data, data with code, code with category and so on (Charmaz, 2014).
I suggest that a significant contribution of this research is to show that comparison across macro, meso and micro levels of inquiry is a way of situating interpretation of evidence e.g. the micro within the macro, or the meso; or the meso within the macro. The contribution shows that macro, meso and micro interact with one another; each informs the other by comparing similarly or differently and makes analysis rigorous. Constant comparison is subjective and is a key to developing theoretical sensitivity. Constant comparison helps to identify where one has pre-judged or assumed; it provides the process for seeing the world through the eyes of participants and “… understanding the logic of their experience …” (Charmaz, 2014, p. 133). Constant comparison micro-analyses empirical evidence
(Glaser & Strauss, 1967) and shifts obdurate codes into categories and theory. Constant comparison encourages interrogation of silences and invisibles (Star, 1995) all with a focus on analysis and theory generation.
I began comparison on the first day of interviews and came to realise that interspersing learners, teachers and trainers and policy-maker interviews, is its own form of theoretical testing. For example, if I interviewed a learner one day, I could theoretically test with interviews with teachers the following day. I compared the internal statements of, for example, a learner interview each with the other. The following day, I had a with a heightened sense of how to situate my interpretation when I interviewed a teacher, or a policy-maker. I compared gerund codes freely and in an open- ended way.
Consistent with conGT, the analysis took “interpretiveleaps” (Charmaz, 2014, p. 200) because of constant comparison in two ways. First, new theory emerged when comparison began to be practicable with reference to the relations between codes and emerging theory. I hand-wrote a comparison chart with separate boxes with, for example, initial codes of learners and policy-makers in the two outer boxes and with a blank box in between to highlight the relations between each group. The middle box was often the focus of intense concentration. A small example of the process is set out below (Figure 4.3). Boxes 1 and 3 represent the first comparisons of initial codes while Box 2 records the interpretive leap towards properties that would ultimately lead to focused coding and categories. The language of learners was expressive of deep feeling; the policy-makers’
interpretation was distant and objective. The interpretive leap was a view of policy as euphemistic as it used language which glossed over the lived experience of learners experiencing disadvantage.
Figure 4.3: Comparisons
Second, new theory emerged quickly when categories were developed and compared with the literature to which theorising had either led me or which had appeared in the perusal of new publications. Constant comparison of the categories of the substantive grounded theory frequently
Box 1: Learner codes:
Fearing everything Regretting/grieving the
past Fearing for future of
children Box 2: Compare: Language; Insight Euphemistic policy? Box 3: Policy-maker codes: Identifying lack of confidence
led to comparison of the literature in different disciplines. For example, one single category led into sociology readings about representing the “other”; educational readings about “self-capitalisation”
and transition of learners; political science readings and public administration readings about new public management and new public governance. Consistent with conGT, out of the process of constant comparison, I began on a path towards a potential formal theory (Charmaz, 2014, p. 343). In the final section, I set out the approach to rigour in this study.
4.12 Rigour
In the name of rigour, in this study I laboured over three concepts in theory and in practice. First, in Chapter 3, I conjoined discussion of subjectivity with reflexivity to show that rigour in conGT is dependent on reflexivity (Chiovitti & Piran, 2003; Darawsheh, 2014). Charmaz (2006) writes that it is reflexivity that grounds and justifies the research and its theory. Second, I used the Charmaz (2014) concept which is that rigour comes with consistency in philosophy, methodology and methods. I constantly referred to recognition of multiple realities to create a flexible methodology. In methods, for consistency, I relied especially on constant comparison. Third, because I made a priority of using methodology that was user friendly to policy-makers, I used cross-disciplinary methods to represent symbols and language that I present as more likely to already be in the public sector domain or to be more acceptable. I base these choices on my thirty years of experience as a bureaucrat (most of which was in the New Public Management paradigm), where it was my job to focus on symbols and language of public administration. I crossed disciplines (see Chapters 5 and 6), heeding Madison’s (2011) exhortation for researchers to pursue intersections between theory, fieldwork methods and critical practice. I sought out discussions of constructivism by political scientists, public policy theorists and organisational theorists (Hay, 2011; McAnulla, 2006; Rhodes, 2007).
The rigour of the analysis in this study is founded on the systematic emphasis with all the
characteristics of grounded theory defined by Charmaz (2014) on constant iteration and comparison within and across all sources of empirical material (Figures 4.1 and 4.2) at micro, meso and macro levels of theory. Figure 4.1 shows methods but I attempted also to show the intense process of constant comparison and reiteration – it was never-ending. Especially but not only in interviews, comparison and iterative analysis set up as much opportunity as possible to go to issues of trust, to expose and deal with inaccuracies, embellishments, minimalist, or deceptive accounts (Charmaz, 2014).
I worked also with a checklist for ensuring “rigour” which draws from Strauss and Corbin (2014, p. 351). I am conscious of the sometime post-positivist position of Strauss and Corbin but I do not see that their checklist for rigour is incompatible with the constructivism of Charmaz. I take the risk in light of my preference to discuss rigour in terms that policy-makers may find will fit and work with their tasks. Table 4.4 below is the checklist with my responses.
Table 4.4: Checklist for Rigour Additional to Reflexivity and Philosophical Consistency
Checklist Response
Did the target sample involve multiple comparative groups?
1. The study involves learners in equity programs. The target sample of learners and other
participants were drawn from participants in five learner centres.
2. All participants were from diverse training organisations: public; private; community. Multiple sources of data Interviews: micro; meso; macro
Observation
Documentation: policy; legislation; interdisciplinary Literature: international; national
Concepts driven by data collection (not literature);
Change of research design
The research design evolved during and after data collection. Literature review and literature comparison followed generation of theory. The final design came late in the analysis influenced by literature comparison. Theoretical sampling, with description Re- interviews; observation; (Chapter 4).
Researcher sensitivity Dealt with in theory (Chapter 3); evidenced in empirical data collection (Chapter 5).
Are there negative cases? There was constant exploration for negative cases, e.g. see Chapter 5 and Preliminary Sampling which focuses on enquiring into negative cases among learners experiencing disadvantage.
Descriptions of coding and methods? Core category, with explanation, and diagram?
Are methodological decisions clear?
Chapter 4 Chapter 5 Chapter 3
4.13 Chapter Summary
This chapter describes the approach to methodology and methods as a package to address the intellectual puzzle. In Figures 4.1 and 4.2, I set out key elements of research design and methodology together with the conGT methods, or guidelines. I affirm ethics approval and address: evolving research design; sources of evidence; selection of participants; process of recruitment; processes for interviews; interviewing; and grounded theory guidelines. I exemplify approaches to achieve rigour. I attempted to give insight into the concern of the study to deeply analyse methods as the tools to generate plausible substantive grounded theory and potential formal theory which comprise the following Chapters 5 and 6 respectively.