ABDUCTIVE ANALYSIS STRATEGY

The data collection is focused around a pragmatic mixed-method approach as the primary epistemological foundation. The PhD project is, therefore, based on both a qualitative and quantitative approach. The following section describes the project analysis approach, which is based on an abductive coding strategy.

The project’s data collection focuses on the students’ reflective experiences, thoughts and feelings in connection with the use of Game-Based Learning as a facilitating framework in a semester project. Thus, the primary focus will be qualitative, in which the quantitative data are processed descriptively and used as a supplement. Thus, the focal point of data collection aims to collect and analyse data through iterative processes, where theoretical assumptions are continuously modified through the three iterative phases. Therefore, data collection is primarily characterised as exploratory observations as well as conversations in open interviews with an understanding and explanation perspective.

The purpose of the analysis is the construction of theoretical ideas based on empirical data through a continuous pragmatic process of puzzling pieces together (Timmermans

& Tavory, 2012). The theoretical constructs of the analysis thus occur through a dialectic between theory and empirical finding (Timmermans & Tavory, 2012). Timmermans and Tavory (2012) describe the analysis process as follows: “Asserting that unexpected theoretical formulation and categories emerge in relation to data locates a social practitioner within a meta-theoretical debate about the relation between data and theory” (Timmermans & Tavory, 2012, p. 168). In doing so, they point to a pragmatic analysis process where coding strategies are the focal point. Here, grounded theory in particular has been a widely used qualitative methodology within social sciences such as education and sociology (Ong, 2012; Timmermans & Tavory, 2012). Grounded theory has traditionally been built around a controlled methodological and step-by-step approach, where the categorisation of data is central while preserving an inductive interpretation perspective. Grounded theory is therefore based on an open and emergent coding of data where the category formation occurs as a result of the inductive approach (Charmaz, 2014; Ong, 2012; Thomas, 2010). Brinkmann (2014) states: “What we call data are always produced, constructed, mediated by human activities, or ‘taken’ as Dewey wanted us to understand through his pragmatism” (Brinkmann, 2014, p. 721).

This perspective means that grounded theory’s inductive foundation is challenged as data will always be constructed rather than given (Brinkmann, 2014; Haig, 2008;

Timmermans & Tavory, 2012).

Consequently, critics have attempted to use ground theory’s methodological approach to promote alternative approaches without necessarily buying entirely into the inductive premise (Timmermans & Tavory, 2012). In particular, the issue of coding through a theoretical lens has been a point of contention. Therefore, several researchers have argued that “grounded theory is epistemologically much closer to what pragmatist Peirce called abduction: a central concept in his theory of logic and inference

that denotes the creative production of hypotheses based on surprising evidence”

(Timmermans & Tavory, 2012, p. 168). If grounded theory’s systematic is considered in the light of a pragmatic and abductive approach, coding becomes much more of a tool that makes it possible to work with what Brinkmann (2014) calls “abductive breakdowns” of the data material. What is being coded, and how, becomes secondary, and it is instead about creating a process of inquiry and reasoning. He writes, among other things:

I have since become sceptical not just of coding but also of the very idea of data as such. The concepts of coding and data often go together as twins. Qualitative researchers who talk about data tend to want to code them and those who do coding usually want to solely code data.

(Brinkmann, 2014, p. 720)

Brinkmann thus argues that data lead to theory and that data “speak for themselves”

(Brinkmann, 2014). Other researchers have justified the lack of theoretical breakthroughs in grounded theory projects with incomplete or inaccurate application of grounded theory principles (Timmermans & Tavory, 2012). As Timmermans and Tavory (2012) write, induction contains a practical dilemma in that it is challenging to generate new theory without being sensitive to existing theory (Timmermans & Tavory, 2012). This means that analysing through an inductive approach has been criticised by several researchers for not being able to offer any “new findings” contained within the logic of the argument (Brinkmann, 2014; Kolko, 2009; Timmermans & Tavory, 2012).

The desire to form new theories is also something that concerns the research field of Educational Design Research, as one of the central points being discussed.

Alternatively, a deductive approach where a theoretical framework guides the analysis is also challenged by theories often being created with little connection to substantive social life (Brinkmann, 2014; Timmermans & Tavory, 2012). Brinkmann (2014) frames this dilemma by talking about the inductive collector and the deductive framer (Brinkmann, 2014). Another point of view could, therefore, be to understand the coding process as abduction, where the coding process alone is a tool for creatively understanding and brooding the patterns of the phenomena (Brinkmann, 2014;

Timmermans & Tavory, 2012). Based on a pragmatic understanding inspired by Peirce, Brinkmann (2014) describes abduction as a “form of reasoning that is concerned with the relationship between a situation and inquiry. It is neither data-driven nor theory-driven, but breakdown-driven” (Brinkmann, 2014, p. 722). Another argument for applying a pragmatic abductive analysis strategy is the association between abduction and Design Thinking. The process of design synthesis is based on the same fundamental premises that characterise the abductive analysis process.

As the methodology of the PhD project is based on Design Thinking (see Chapter 8 and Section 3.4), the subsequent analysis of the individual iterations will advantageously be based on cultural coding of patterns that later act as an argument for the best

explanation (Kolko, 2009). Kolko (2009) describes it as follows: “It is the idea of putting together what we had never before dreamed of putting together that flashes the new suggestion before our contemplation” (Kolko, 2009, p. 21). Thus, in the same way as the initial design process, in which the theoretical basis was reformulated into current principles, the coding of data will, in an analytical context, create breakdowns in existing understandings, thereby leading to new realisations. Examples of some of these breakdowns can be found in Section 10.1.1 of the analysis, where quests and levels are discussed as a catalyst for motivation, and in Section 10.1.2, where the development of the students’ autonomy challenges the common understanding of whether a learning game is something that needs to be played through in order to reach a learning outcome, as well as in Section 10.3.1, where the concept of “game over”

is reinterpreted into a higher education context, and where the concept of “dying” is understood as creating disruptions in the students’ project.

As Timmermans and Tavory (2012) explain it, abduction seeks a theory where induction seeks facts (Timmermans & Tavory, 2012). Since the PhD project rests on a pragmatic epistemology that rejects a research process that is said to lead to facts, an abductive analysis approach that seeks to develop new theory or models is a logical choice.

The abductive process

The abductive analysis process relies on elements from both induction and deduction (Timmermans & Tavory, 2012). Abduction is thus a form of reasoning based on a creative inferential process that produces new theories based on the surprising element found in the data (Brinkmann, 2014; Timmermans & Tavory, 2012). The abductive method looks at theories and data as developing entities and thus, according to Haig (2008), becomes a method for theories in the making (Haig, 2008).

Through its coding process, the abductive analysis seeks to engage in imaginative thinking about intriguing findings. Through iterative loops in the data collection, the goal is to create inferencing creatively, and a double-check of these assumptions (Timmermans & Tavory, 2012). Thus, the abductive analysis process works from the premise of moving back and forth between data and theory iteratively, as is known from the traditional grounded theory (Timmermans & Tavory, 2012). The essential difference between abductive coding and grounded theory is the importance of existing theory that helps to make possible empirical anomalies visible:

Abductive analysis constitutes a qualitative data analysis approach aimed at theory construction. This approach rests on the cultivation of anomalous and surprising empirical findings against a background of multiple existing sociological theories and through systematic methodological analysis. As such, it requires a fundamental rethinking of the core ideas associated with the grounded theory, specifically the role of existing theories in qualitative data analysis and the relationship between methodology and the theory generation. (Timmermans &

Tavory, 2012, p. 169)

As described in Section 3.1, knowledge creation from a pragmatic perspective is focused on the discovery of anomalies and breakdowns that can contribute to the development of new understandings, theories and models (Caldwell, 1983; Godfrey- Smith et al., 2015; Kjær, 2010; Lehmann-Rommel, 2000; Timmermans & Tavory, 2012). Abductive reasoning is about more than just describing patterns of data, and is, therefore, looking for plausible explanations of phenomena (Haig, 2008).

The inquiry concept is used to describe how the analysis phase attempts to frame the creation of breakdowns in one’s understanding through an explorative approach (Brinkmann, 2014). Haig (2008) describes it as follows:

Sets of data are analysed to detect robust empirical regularities or phenomena. Once detected, these phenomena are explained by abductively inferring the existence of underlying causal mechanisms.

Here, abductive inference involves reasoning from claims about phenomena, understood as presumed effects, to their theoretical explanation in terms of underlying causal mechanisms. (Haig, 2008, p. 1019)

Haig’s (2008) description points in the direction of putting pieces of related information together in order to make a story (Thomas, 2010). Coding data in a pragmatic and abductive understanding is, therefore, more than just recording and categorising. It is a kind of doings that create the mentioned breakdowns, through the series of events being merged into a story ( Brinkmann, 2014; Thomas, 2010). According to Thomas (2010), the analysis process must question (1) which elements are woven together, (2) how the elements fit together, (3) how/whether they contradict, and (4) whether paradoxes arise in the process (Thomas, 2010). The coding process, as it is known from grounded theory, can thus advantageously support finding the “sequence of steps” (Thomas, 2010). The abductive coding differs in that it consciously seeks to challenge and question data by combining the individual categories ‒ a process similar to Feyerabend’s (1993) suggestion that science creation occurs through the use of counter rules and hypotheses that contradict well-established thinking. Thomas (2010) also points to Kuhn’s (1970) concept of the “awareness of anomaly” (Thomas, 2010).

Findings anomalies and the unexpected

The risk of traditional coding is that only ready-made categorisations are created (Timmermans & Tavory, 2012). The ability and opportunity to discover new theories depends on the ability to frame, modify and extend the empirical data within an existing theoretical framework. In this understanding, the theoretical framework becomes crucial as in-depth knowledge of multiple theorisations is the key to being able to see the missing link or anomalies in an area of study (Timmermans & Tavory, 2012).

Haig writes, for instance: “Some phenomena are detected that are surprising because they do not follow from any accepted hypothesis or theory” (Haig, 2008, p. 1020).

Theoretical insight can also stimulate innovative and original theoretical knowledge

creation, especially in the light of a research design based on the application of Design Thinking (Timmermans & Tavory, 2012). In this regard, Timmermans and Tavory (2012) argue that the prerequisite for an abductive analysis is that the research process is based on methodological approaches that allow the coupling of theory and empirical.

While such anomalies are opportunities to develop new theoretical insights and modify existing theories, researchers need to foster an environment that allows doubt to develop. This conducive environment is predicated on a series of pre-established steps through which the researcher revisits the phenomenon – in other words, a method.

(Timmermans & Tavory, 2012, p. 175)

Thus, through intensive coding, the aim is to link empirical data with existing theories so that it is possible to identify changed circumstances and thereby find new dimensions of the problem. The abductive analysis process requires the researcher to access it with the deepest and broadest theoretical base possible and from there develop the theoretical repertoire through empirical findings (Timmermans & Tavory, 2012).

9.3.1. ABDUCTIVE CODING OF THE QUALITATIVE DATA

The methodological assumptions that apply to grounded theory can thus stimulate abductive reasoning (Timmermans & Tavory, 2012). The foundation of data processing is the following four main features: the selection of relevant parts of the data material, the degree of transcription, the coding procedure and the writing-up process. The individual steps are described in the sections below.

1) Selection of data: All interviews and reflective conversation are transcribed to