• No results found

Chapter 3 Research Design

3.3 Case Study Protocol

3.3.4 Data Analysis

Data analysing is the most important stage when building theory from a case study, however, since the qualitative data derived from PO and interviews usually come with plenty of unstructured textual materials, the analysis of data becomes the most difficult and the least codified part in the research process (Eisenhardt 1989; Yin 2009; Bryman 2012). Unlike statistical analysis with fixed formulas to guide the analysis, in case study research, data analysis mostly depends on the researcher’s own style of empirical thinking and evidence, without fixed analysis models (Bryman 2012). However, some guidelines can genuinely help researchers to undertake their qualitative data analysis.

The first guideline is from the perspective of the data analysis strategy. According to Yin (2009), there exist four strategies for analysing data in case study research, including relying on theoretical propositions, developing a case description, using both qualitative and quantitative data, and examining rival explanations. In this research, due to the constructs derived from the literature review, and the qualitative data collected in the data collection phase, this research adopts the strategy of relying on theoretical propositions to guide the data analysis activities. Differently from the other three strategies, analysing the data relying on theoretical propositions can be helpful to ensure concentration on “useful” data whilst ignoring the “useless” data, and it also helps to organise the entire case study and to update the proposed theoretical framework to be employed (Yin 2009, p.130).

The second guideline is from the perspective of data analysis technique. Based on the data analysis strategy, relying on theoretical propositions, pattern matching is the technique adopted in this study. According to Saunders et al. (2007, p.489), pattern matching “involves predicting a pattern of outcomes based on theoretical propositions to explain what you expect to find”. When adopting this data analysis technique, based on the constructs developed in the literature review, the data are classified into patterns, and all these empirically-based patterns are compared with other researchers’ existing findings (Yin 2009). If the patterns of the data appear to match other research, the result will be helpful in strengthening the internal validity of the case study (Saunders, Lewis et al., 2007).

The third guideline is from the perspective of data analysis methods. This study adopts within-case analysis to deal with the data as the first step. Since the research questions for a case study are usually open-ended, the research usually

comes with a massive volume of data, which makes within-case analysis one of the key steps in the research process, to cope with the volume of data. With the help of within-case analysis, this study created familiarity with each case, and accelerated further cross-case comparison (Eisenhardt 1989). The subsequent step in the data analysis work is cross-case analysis. For the data comparison, as previously mentioned in the case selection section, four cases were selected from the three case sites concentrated on in this research. According to Eisenhardt (1989), there are two strategies to undertaking cross-case analysis: firstly, dimensions are selected to look for within-group similarities as well as inter-group differences; secondly, different pairs of cases are chosen, and the similarities and differences between each pair are listed. Following the above data analysis methods, in the current research, the selected R&D department’s capabilities as identified in the literature chapter were catalogued into each activity of the conceptual framework for the in-case analysis, and referring to Rohrbeck’s research work (2011), all the capabilities were marked from level zero to level three in each case, where the general measurement criteria can be discussed as follows:

 Level 0: The R&D department did not concentrate on the capability for developing its radical innovation.

 Level 1: The R&D department had some concentration on the capability but not much.

 Level 2: The R&D department concentrated on the capability but neglected a few significant perspectives.

 Level 3: The R&D department concentrated on the capability from all the significant perspectives.

Moreover, for the specific measurement criteria of each individual capability, they are further developed in details (see Appendix 2) in the current research for undertaking the cross-case analysis more efficiently.

By undertaking the cross-case analysis comparing the capabilities involved in each activity with the measurement criteria among the four cases, the reasons for the differences between the marking levels of the R&D departments’ capabilities can be explained with reference to the six contextual factors in relation to categorising radical innovations, which are helpful to discover which contextual factors played the most significant roles in affecting the R&D departments’ capabilities at each stage of the radical innovation development cycle.

The fourth guideline is from the perspective of the data analysis process. Referring to Eisenhardt (1989), this research divides the case study data analysis into two steps. In the first step, the constructs are redefined with the building evidence in each case. In the second step, this study verifies that the emergent relationships between constructs fits with the evidence in each case, and the cases that confirm the emergent relationships can enhance confidence in the validity of the relationship, whilst the cases that weaken the relationships can inspire an opportunity to refine and extend the theory.

The final guideline for data analysis in the case study research is from the perspective of comparing the proposition with the literature. In this study, when the research findings were proposed, the capabilities and contextual factors in each radical innovation developing activity were compared with the literature to find theoretical support from the sources used. According to Eisenhardt (1989), this perspective can be helpful in enhancing the emergent theory from the

perspectives of internal validity, generalisability, and the theoretical level of theory building.

3.4 Summary

This chapter focuses on the research design and methodology that are used in this research work in order to answer the research questions of this study. After discussing the research-related philosophical issues underpinning the study and the related qualitative or quantitative considerations, to examine the topic case studies were selected as the main research approach in this research. The case study research design is divided into four stages, consisting of getting started, identifying cases, data collection, and data analysis. Subsequently, the four cases in the three target telecommunications firms were selected, and the data collection and data analysis methods were discussed, which guided the data collection and further data analysis activities in this study. The summaries of the data collection work in this research were also presented in this chapter.