Qualitative Research as Stepwise-Deductive Induction

Aksel Tjora

• Publisher: Routledge
• Year: 2018
• Topic: Social Sciences
• eBook ISBN:
  9781351396950
• Subtopic: Sociology
• Index: Social Sciences

This book provides thorough guidance on various forms of data generation and analysis, presenting a model for the research process in which detailed data analysis and generalization through the development of concepts are central. Based on an inductive principle, which begins with raw data and moves towards concepts or theories through incremental deductive feedback loops, the 'stepwise-deductive induction' approach advanced by the author focuses on the analysis phase in research. Concentrating on creativity, structuring of analytical work, and collaborative development of generic knowledge, it seeks to enable researchers to extend their insight of a subject area without having personally to study all the data generated throughout a project. A constructive alternative to Grounded Theory, the approach advanced here is centred on qualitative research that aims at developing concepts, models, or theories on basis of a gradual paradigm to reduce complexity. As such, it will appeal to scholars and students across the social sciences with interests in methods and the analysis of qualitative data of various kinds.

1 The basis for qualitative research

Contents of this chapter

• the importance of curiosity in research;

• an introduction to the stepwise-deductive inductive approach;

• the scientific basis of qualitative research;

• contrasts between quantitative and qualitative research approaches;

• the role of theory in qualitative research;

• considerations regarding sampling, case studies and empirical limitations;

• ethical aspects of qualitative approaches;

• the impact and potential of the Internet for qualitative studies.

Qualitative research is subject to significant sensitivity to the context in which it is carried out. The researcher is often in close proximity to the subject, whether the latter has signed up as a participant in an interview-based study or takes part in situations under observation. This proximity makes qualitative research exciting and intense, but also yields special challenges. For instance, researchers need to be prepared to adjust their own project, actions or perhaps even ideas upon contact with the research field. It is not uncommon to discover that things are not as they had been envisaged beforehand. For instance, hospital fieldwork may entail becoming aware of practices and situations that are specific to the hospital as an organization, even if the research issue is concerned with information technology, and as a result field-specific theories and knowledge need to be taken into consideration. Therefore, it will often be appropriate in qualitative studies to deal with data generation (often called ‘data collection’) at a relatively early stage in the project in order to have the opportunity to revise sampling, or to adjust the use of theory and perspectives to those that turn out to be interesting in the empirical analysis.

1.1 Research as systematic curiosity

When, halfway through my doctoral work during the period 1993–1997, I was obliged to change my field of research, from regional industrial development around a major Norwegian shipyard and oil platform construction complex to the coordination of medical emergency call services. I suddenly needed to become familiar with the literature around areas such as medical sociology, health services research and studies of the medical profession, laying aside much of the network sociology and industry studies that until then had been relevant. However, I was still able to keep hold of perspectives from interactional organization sociology and technology studies.

The situation reveals something of the complex connection between research methodology, topics, perspectives and theory, which is the subject of this chapter. Promising research ideas often arise from a concrete problem, question, or phenomenon that awakes one’s curiosity. A phenomenon often becomes an interesting research topic through a theoretical perspective. Both the industrial network around the shipyard and the coordination of medical emergency services invited study on the basis of my interest in organizational studies, networks, communication and technology. These perspectives made both the cases interesting for my sociological dissertation work. Perspectives, however, also make demands of the choice of methodology. To put it in general terms, interactionist perspectives form better tools for analysing observational data than interview data. Irrespective of perspective, a measure of methodological creativity is always beneficial.

In connection with teaching and supervising research methods I have found that qualitative methodology literature makes too little attempt to convey creative diversity. This can lead researchers and students towards standardized approaches, which up to a point can be justified in terms of the methods literature but which is insensitive to particular contextual issues. For instance, in-depth interviews have ended up as something of a standard choice for data generation within much of qualitative social research, irrespective of research issue. I regard this as very unfortunate in that it stretches the limits of what can be illustrated by interview data (or its ontological status). In this book I am therefore anxious to support a greater methodological openness in the sense that choice of method needs to reflect what one is actually trying to find out. This may, for instance, mean careful consideration of whether observation or Internet-based data generation might produce relevant empirical material more effectively than, for instance, in-depth interviews, even though at the time it may seem simpler just to get on with it, follow the path of least resistance and get hold of a few interview subjects. As well as greater creativity in terms of data generation there is a strong argument for a good deal of stringency in conducting analysis. The stepwise-deductive inductive (SDI) approach on which this book rests represents the very opposite of following the path of least resistance and in addition embodies this kind of stringency. When planning and carrying out studies using this method, all choices are to be taken with the support of academic reasoning, without taking shortcuts or drawing premature conclusions. A high academic standard, or academic probity, is to be observed at all stages of the research.

In addition to this probity and ethical responsibility, research methods must be deemed efficient. The data generation must yield the most relevant and reliable information possible without wasting time or resources on the part of the researcher or of the subjects. I will therefore deal primarily with methods that can yield an adequate volume of good empirical data within a limited time scale. For instance, I introduce focused interviews as a variant of the in-depth interview, as an effective method of exploring people’s experiences with particular social phenomena. The analytical work should also be efficient, not least through a systematic approach. The SDI approach is a viable model for the whole of the qualitative research process. A significant premise for this model is a constant inductively driven curiosity, in which the researcher is using the empirical material as the basis on which to define what is to be regarded as interesting topics, questions and concepts (Glaser 2002). Another goal for the model is to work towards generalizing and the development of theories – in other words, with explicit theoretical ambitions – and using theoretical input, especially in the later stages of the work.

1.2 Stepwise-deductive induction as Grounded Theory

In the stepwise-deductive inductive model (Figure 1.1) we are working in stages from raw data to concepts or theories. The ‘upwards’ process is to be understood as inductive, in which we work from data towards theory. The ‘downwards’ feedback connections are to be understood as deductive, in the sense of checking from the ‘more theoretical’ towards the ‘more empirical’.

The model may give the impression that the research process is completely linear, but as a rule this will not be the case in reality. Nevertheless, the model forms a good starting point for a systematic approach and progress in a qualitative research project, not least through offering the possibility of devising a stepwise plan with various mileposts along the way. The model’s deductive back associations rest on six different tests, visualized in Figure 1.1 as curving downward arrows, forming a basis for iterations between two adjacent stages in the model. For instance, when conducting code-grouping tests, work will continue on grouping until all codes have been grouped.

Figure 1.1 The stepwise-deductive inductive (SDI) research model

While deductive tests 3–6 are described in the various subsections of Chapter 2 regarding analysis, sampling and data tests (tests 1–2) are more implicitly described in Chapters 2 to 6 regarding different types of data generation. To put it simply, the sampling test is based on a constantly recurring question of to what extent the most suitable selection criteria have been used for participants in a study, whether the case has been well chosen, whether the snowball method is suitable, whether the selection is too broad or too narrow, and so on. The sampling test consists of a number of different questions, which are discussed in sections 1.8, 3.2 and 5.7. Similarly, the data test consists of a number of questions regarding the extent to which empirical data generated is fit for purpose (in other words, relevant to the question being asked), whether it is sufficiently detailed but not unnecessarily so and whether data is suitably recorded. The data test forms the basis for the whole of Chapter 4 (on observational data) as well as sections 5.3 and 5.6 (regarding interview data). When carrying out document studies (Chapter 7), equivalent questions regarding sampling and data quality (tests 1–2) will consistently be predominant issues.

In some cases, this deductive back-coupling can take place over several stages, for instance where concepts are developed before finding out that more empirical data is required on specific phenomena, requiring a return to the first stage. This kind of iterative process is termed ‘theoretical sampling’ within the Grounded Theory (GT) research tradition (Glaser and Strauss 1967; Strauss and Corbin 1990). The type of ‘large’ iteration seen in GT, however, requires a more spacious time scale than many research projects can accommodate. For this reason, SDI provides a more linear process in this respect than GT, with iterations generally limited to backward coupling between two adjacent stages in the model.

Many qualitative researchers claim to ‘use’ or ‘be inspired by’ GT without actually employing this iteration between data generation and conceptual development. This is partly caused by a certain degree of inflation in references to GT when it would be more accurate to speak in terms of a more general inductive method, and signifying a broad acceptance of the Glaser/Strauss/Corbin references above. The SDI model’s description of qualitative research as stepwise underlines the systematic design of the research. In practice, it is possible to be at different stages of the SDI model at the same time in respect of different aspects of a project (topics, empirical data or theoretical areas). An ideal behind the development of the model is to make use of the potential in the empirical data that has been generated. A conscientious empirically based initial phase in the analysis is therefore essential.

The SDI model has a stronger theoretical motivation than much other qualitative research. It has a great deal in common with Glaser’s emphasis on inductive conceptualization within GT (1978, 2002), but attempts to represent a more distinct and unambiguous inductive approach than is currently represented by a kind of GT ‘ecosystem’. The many practitioners with their own individual variants of GT, have proven the usefulness of the approach, but with a blurring of the concept as a result. In some sense, the SDI approach takes a step back to the inductive premise of GT, with a fundamental agreement in scientific-theoretical terms – the inductive principle – but with a stricter a terminology.

The fact that the SDI approach operates on inductive-empirical principles by no means negates the significance of theory, but contrasts with research that is centred on evidence and documentation (which is attempting to prove or document a proposition) or on ideology or politics (which is intended as a critique of defined issues/institutions or to liberate specific bodies/positions). This means that there are situations, in which the SDI framework does not work well,

•when using qualitative research to test theories;

•when applying research mainly as political emancipation;

•when doing ‘research’ (i.e. contract research, evaluation) with no intention of generalization;

•when doing research which is purely narrative, i.e. person-centred rather than theme-centred.

In short, the SDI model has curiosity as its genesis and generalizable understanding as its goal. It features a critical rather than dogmatic standpoint (Popper 1959:50). The inductive method of conducting research is by no means a new one; it has been fundamental in the development of all the sciences, with well-known representatives including physicists such as Newton, Einstein and Bohr, the biologist Darwin and the psychologist Freud.

Within the social sciences a large number of concepts and ways to understand various aspects of society have been developed on basis of qualitative studies. Thomas and Znaniecki (1918–1920) established a seminal ethnographic attitude in the study of Polish immigrants to the US, on the basis of letters sent between relatives, and contributed significantly to what has later been known as the Chicago School of Sociology. As part of the same research community Whyte (1943) contributed to the understanding of immigration and urban gang culture. Lipset, Trow and Coleman (1956) published a seminal single-case study about the International Typographical Union, exploring in length how prospects of democratic conduct were developed within the private governments of industrial society. A few famous studies were developed in relation to healthcare. Becker, Geer, Hughes and Strauss (1961) explored student culture at medical school. Strauss, Fagerhaugh, Suczek and Wiener (1985) studied hospital organization and the practice of the medical profession. Of the most famous sociologists we find Goffman, who applied various forms of qualitative studies to explore social interaction in public situations, self-presentation, stigma, mental illness and much more (Goffman 1959, 1961, 1963, 1967, 1971), and contributed to a large number of useful concepts. More recently, Wacquant (2008) has explored urban marginality in the black ghetto of Chicago and the deindustrializing banlieue of Paris for rethinking processes related to increasing social problems in urban areas. What ...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Dedication
• Table of Contents
• Preface
• 1 The basis for qualitative research
• 2 Analysis as stepwise-deductive induction
• 3 Observational studies
• 4 Field notes and recordings
• 5 Forms of qualitative interviews
• 6 The practice of interviewing
• 7 Document studies
• 8 Quality and presentation of research
• Index