3.3 Research Strategy
3.3.1 Case study research
Yin (1981) describes the case study approach as a research strategy. There is considerable discussion and debate regarding the methodological positioning of case study research. This research adopts the contemporary perspective of case study research encompassing the logic of design, and the techniques for data collection and analysis (Patton 2002; Yin 2013). Case study research invariably involves one or several case studies. Stake (1995) suggests a case study represents the researchers choice of what is to be studied. Merriam (2014) defines a case study as an in depth description and analysis of a bounded system’. Merriam (2014) also describes qualitative case studies as being focused on a particular situation, event or
114 phenomenon (particularistic), where the end product becomes a thick description of the phenomenon under study (descriptive), and where the readers understanding of the phenomena under study is expanded (heuristic).
Case study research may be exploratory, descriptive or explanatory and Yin (2013) suggests that ‘you use the case study method because you want to understand a real-life phenomenon in depth, but such an understanding encompassed important context conditions, because they were highly pertinent to your phenomena of study’.
Case study research can also play an important role in advancing knowledge in a particular research theme or domain (Merriam 2014). Case study research provides a close link with the empirical setting and as a consequence it can be particularly useful for building theory (Eisenhardt 1989b). Unlike experimental research that seeks to remove phenomena from its context, and survey-based methods that have limited ability to investigate context, case study research seeks an in depth investigation of phenomenon within its real-life context (Yin 2013).
For the qualitative researcher case study research has particular strengths that make it an extremely pervasive research strategy in many business and information systems research studies. At a philosophical level case study research is closely aligned to research utilising a subjective-interpretive philosophy, however case study research need not be either (Walsham 1995).
3.3.1.1 Multiple case study design
Yin (2013) uses the term “multiple case study” to describe case study research design involving more than one case. Elsewhere in the literature multiple case study designs can be also be referred to as multi-case, multi-site, collective or comparative case studies (Merriam 2014).
Yin (2013) delineates between different types of basic case study design on the basis of whether it involves one or multiple cases, and then whether each case has one (holistic) or multiple (embedded) units of analysis. Yin (2013) illustrates the design variants as a matrix (see figure 3-1) comprising four types: (1) the single case holistic design, (2) single case embedded design, (3) multiple case holistic design, and (4) multiple case embedded design.
115 There are of course all manner of variations upon this scheme. Yin (2013) is also sensitive to the notion case study design will invariably ‘include a desire to analyse the contextual conditions in relation to the case’.
Figure 3-1. Four basic types of case study design (Yin 2013).
This research adopts a multiple-case design exploring cases of IT innovation in the organisational context. The decision to include one or more cases is important for case study research design (Benbasat, I, Goldstein, K & Mead, M 1987).
Multiple case study designs are applicable to research that involves description, theory building or theory testing (Benbasat, I, Goldstein, K & Mead, M 1987). Multiple case study designs also allow researchers to conduct cross-case comparison of data from different sources. From an interpretivist perspective this opens up the options to develop and refine concepts for further study, and allows concept to be evaluated in different contexts (Darke, Shanks & Broadbent 1998). Thus a multiple case study design is highly suitable for extending existing theory (Benbasat, I, Goldstein, DK & Mead, M 1987).
The selection of a multiple case design has implications for the data collection, analysis and interpretation procedures used in this research. The use of multiple case studies as opposed
116 to a single case requires an appropriate selection procedure covering the number, type and method of selection for cases. Case selection procedures are described and justified in Section 3.4.2. The sequence and protocols for data collection must also reflect dealing with multiple cases. Data collection procedures are described and justified in Section 3.4.3. Multiple case study design also introduces the option of cross-case analysis, either in conjunction with or following on from a within-case analysis. Procedures describing and justifying the data analysis are provided in Section 3.4.4. Having multiple cases also impacts the discussion and interpretation of the research findings. How the researcher justifies the applicability and context of theory from the particular to the general is outlined in Section 3.4.5.
3.3.1.2 Limitations of case study research
Case study research is subject to various criticisms ranging from the philosophical to the practical. At a philosophical level case study research is closely aligned to research utilising a subjective-interpretive philosophy and issues associated with (1) a lack of empirical
generalisability, (2) research bias, and (3) rigour and reliability are often cited as limitations related to case study research.
Case study research is often criticised for its lack of generalisability. Valid research seeks to extend the scope of its findings beyond the empirical context. It is commonplace in positivist research to have scientific facts established via multiple sets of experiments that replicate phenomena under different conditions (Yin 2013). The main purpose of this replication is to establish external validity and statistically generalise from the empirical context to the
broader population. It would be rare if not problematic for case studies to represent randomly selected samples aiming to statistically generalise findings to a given population (Cavaye 1996; Yin 2013).
Despite the statistical limitations associated with case study research, it is possible to generalise the findings of cases studies to theory. This mode of generalisation is distinctly different in purpose, Yin (2013) describes the process as analytical generalisation whilst Stake (1995) describes the process as naturalistic generalisation. Lincoln and Guba (1985) propose a concept of transferability over generalisability and suggest the researcher must make the research process explicit when presenting their research. The technique specified by
117 Lincoln and Guba (1985) to allow the reader to judge the transferability of research into other contexts is to provide a thick description. This research has employed detailed presentation of the context and findings for each case and for the cross case analysis.
Researcher bias is also a common criticism of case study research. Where qualitative methods are employed the case study method is likely to involve subjective judgment on the part of the researcher (Flyvbjerg 2006). Thus the researcher’s background and experiences can affect the process of case selection, data collection and interpretation of the data (Yin 2013). Whilst researchers working with a positivist objective epistemology seek to remove such influences or bias, arguments can be made that this influence is an essential component of the subjective interpretive empirical domain. Eisenhardt (1989a) for example suggests that the process of reconciling contradictory or paradoxical evidence from cases actually unfreezes the thinking of researchers and that this in turn leads to theory that is generated with less bias than research built purely from deductive methods. Both the researcher and participant face this limitation in the empirical context, an issue that is acknowledged and embraced by this study.
The researcher can also influence responses from participants during data collection. The participant’s reaction to the interviewer and interview process can introduce bias. For example the length of the interview, the emotional demeanor of the interviewer or interviewee, the types of questions, and the interview setting itself can all contribute to possible research bias. To address this type of bias the researcher endeavored to collect case data from multiple sources of evidence (Miles & Huberman 1994).
The third major criticism of case study research is that it lacks rigour and reliability. Yin (2013) argues that a lack rigour in case study research may be symptomatic of there being no standardised set of methods or procedures for conducting case studies, something that Yin (2013) and other researchers have arguably set out to remedy. To address the issue of rigour and reliability Yin (2013) advises researchers that they should make as many steps in the research processes as operational as possible, and have multiple researchers conduct each case. The institutional arrangements for this research prevent the use of multiple researchers conducting each case study, however open and transparent techniques that detail the research procedures used in this research are reported in Section 3.4 of this chapter and in Chapters 4 and 5. This research has also followed the guidance for quality and credibility set out for
118 interpretative field studies by Klein and Myers (1999). A synopsis of how this guidance has been applied to this research is provided in Section 3.5.
At a practical level researchers have identified several limitations that sometimes make case study research difficult or unsuitable for specific investigations. For example case study research can be time consuming and complex (Darke, Shanks & Broadbent 1998). Yin (2013) suggests that complexity in the description and reporting of case studies can be addressed through innovative adjustments to methodology. However for the qualitative researcher, case study research has particular strengths that make it an extremely pervasive research strategy in many business and information systems research studies.
3.3.1.3 Justification case study research approach
The case study approach is particularly relevant to the interpretative-epistemology, where the investigator in the empirical context the must analyse data based on the interpretation of participants who have experienced or observed events (Easton 2010; Perry 1998). In this context, case study research provides a close link with the empirical setting and as consequence it is particularly useful for building theory. Unlike experimental research that seeks to remove phenomena from its context and survey-based methods that have limited ability to investigate context, case study research seeks an in depth investigation of phenomenon within its real-life context (Yin 2013). Case study research is also useful for extending research where the existing body of theory may be inadequate (Eisenhardt 1989b). The case study research strategy was primarily selected for this research because the purpose of the research was to explore and describe complex contemporary phenomena, where the research question was seeking a detailed understanding of IT innovation. Rosenberg (1994) for example suggests that to understand innovation beyond more general concepts inevitably involves drilling down into the domain to examine the common patterns and cases. The case study research strategy is also widely accepted within the information systems and management science disciplines as being relevant and valid for research (Cavaye 1996; Jans & Dittrich 2008; Ravenswood 2011).
Many of the criticisms of case study research that are consequently labelled limitations lie more with the interpretative epistemological stance and the use of qualitative methods. For
119 example, the case study literature discusses issues of generalisability, research bias, rigour and reliability that are invariably linked to the subjectivity of the investigator and other participants. These arguments are addressed, but not dismissed by Yin (2013), however Stake (1995) and Merriam (2014) see concepts such validity and reliability as being positivistic constructions and impossible to apply the in the context of an interpretative-qualitative research (Yazan 2015). For the interpretive-qualitative researcher these are not limitations at all, but features of the research strategy to be embraced and utilised to answer research questions (Merriam 2014).
Thus an alternative approach for interpretive researchers is to embrace principles of research quality derived from interpretive perspectives (Stake 1995). Klein and Myers (1999) provide one such scheme that is particularly relevant to qualitative fieldwork in information systems and addressed in the context of this research strategy in the following Section (3.3.2).