Chapter 4: Research design and methodology
4.6 Data analysis
4.6.3 Modes of inference
Critical realism uses four modes of inference: deduction, induction, abduction and retroduction. These modes of inference are complementary (Danermark et al., 2002). In this study I used inductive, abductive and retroductive modes of inference. According to critical realist research studies, any combination of these three modes of analysis is valid. Abduction and retroduction provide methodological approaches that allow explaining events and processes in critical realism research (ibid.). Induction takes place at the empirical level of critical realism (Danermark et al., 2002). This mode of inference attempts to “draw conclusions about an entire population from studies of a sample of investigated units (such as people, organizations and tests); or from studies at a particular point of time, draw conclusions about other points of time” (Danermark et al., 2002:86). The limitations of the inductive mode of inference are that it can never be either analytically or empirically certain and it is restricted to conclusions at the empirical level (ibid.). Induction in this study happened through generation of themes and sub-themes from the data (primarily used in Chapters Five, Six and Seven). However, as much as the inductive method is a valid mode of inference, it does not give guidance on how knowledge of the underlying structures is to be inferred from the observable characteristics (Danermark et al., 2002). Induction is therefore not able to go beyond the empirical level of critical realism. Therefore, it was necessary in this study to use abductive and retroductive modes of inference to find answers to some of the research questions. According to Danermark et al. (2002), abduction and retroduction are the most important forms of inference when the aim of the research is to explain events and processes. This is because:
... to explain something implies (from the perspective of critical realism) first describing and conceptualizing the properties and causal mechanisms generating and enabling events, making things happen ... and then describing how different mechanisms manifest themselves under specific conditions. (p.74)
The fundamental structure of abduction is “to interpret and recontextualise individual phenomena within a conceptual framework or a set of ideas”; this is to understand them “in a way by observing and interpreting this something in a new conceptual framework” (Danermark et al., 2002:80). Abduction therefore allows one to gain new knowledge on the existing phenomena. In the case of this study, I used an analytical conceptual framework from critical realism and capabilities (see Chapter Three) to look at the data in the new way. The concepts of structure, agency, functionings, capabilities, and conversion factors were used to look at the data in a new way. Tao‟s (2013a) analytic framework (Figure 3.3) was used to analyse the causal mechanisms (for example, Life Sciences teachers‟ motivation to participate) in PLCs, their tendencies, and how these tendencies
operate within the social system (the PLCs) among counter-tendencies.As discussed in Section 3.7, Tao‟s (2013a) framework that combined the concepts of capability, issues of structure and agency and underlying mechanisms and tendencies were used to shape the analytic tools for the study (Figure 3.3). As noted in Chapter Three, Tao‟s framework was suited for the study, as it has potential to identify the relationships between the conversion factors of Science PLCs and the conversion factors in the Fundisa for Change programme, and how those conversion factors act as enablers or constrainers for teachers‟ capabilities. These were able to inform the Fundisa for Change programme on what support to offer to Life Sciences teachers in the PLCs related to biodiversity. I analysed how these counter tendencies act as enablers or constraints, and how they relate to causal mechanisms affecting teachers‟ knowledge, for example, on biodiversity. I further analysed how the social structures from both the Science teachers PLCs and Fundisa for Change contribute to that, by being enablers or constraints. I then analysed the decisions teachers make as part of their reflective deliberations and how those decisions result in chosen outcomes (structure and agency dialectic process develop and define the causal mechanisms possessed by an individual). This study was also recontextualised with the existing understanding of the concepts of professional development, particularly the PLCs in the context of teachers‟ biodiversity knowledge provided in Chapters One, Two and Three of the study.
“The fundamental tenet of critical realism is that we can use causal language to describe the world” (Easton, 2010:119). The most fundamental aim of critical realism is to explain answers to the question: “what caused those events to happen?” (Easton, 2010:121). The explanatory power of the substantive theory and theoretical perspective given in Chapters One, Two and Three were used to provide a language of re-describing what the teachers and the Fundisa for Change partners said in the interviews, as well as what I read in the teachers evaluations, reflections, questionnaires, portfolios of evidence, other documents and what I observed in the PLCs. To construct explanatory accounts, critical realist research uses retroduction as a method of inference through which the generative mechanisms are most likely to be identified. This contrasts with induction which aims to reach a reliable generalisation (Reed, 2009). Deploying retroductive logic as its main method of inference, this study aimed at providing a theoretical explanation of how Science PLCs function in the context of teachers‟ continuing professional development in South Africa. The strength of using retroduction in research such as this is that “it provides knowledge of transfactual conditions, mechanisms that cannot be directly observed in the domain of the empirical” (Danermark et al., 2002: 80). Retroduction facilitated the identification of causal mechanisms which expanded and/ or constrained teachers‟ functionings within the PLC. Data on institutionalisation of PLCs of Life Sciences teachers in South Africa was analysed by means of a retroductive mode of inference,
which is a thought operation that enables the understanding of social reality beyond what is empirically observable or experienced (Danermark et al., 2002).
I was further interested in the relationship that exists between the conversion factors, agency, structures and functionings of continuing professional development programmes with associated conversion factors, structures, agency and functionings in Life Sciences teachers‟ PLCs. The study thus required analytical and explanatory processes that took account of causality (not to produce correlational forms of causality, but rather explanatory analyses of causality within a critical realist framework). This is the focus of Chapter Eight: to make explicit the relationship between continuing professional development with teachers in PLCs functionings as they specifically relate to the teaching and learning of biodiversity. Finally I drew from the critical realism concept of contingencies (Sayer, 2000) to analyse the relationships between the conversion factors of continuing professional development programmes with the associated conversion factors in the Science teachers‟ PLCs.