Having considered the four criteria for the design of good empirical social research, these are applied within the case study design. Yin (2003) points out that good case studies are hard to do for four main reasons. These are a perceived lack of rigour, little basis for scientific generalisation, length of time and volume of data resulting in long, dense reports, and the absence of a defined skill-set for researchers. The challenge is to design a good case study that addresses ‘the traditional criticisms of the method’ (Yin,
94 2003:1) by being rigorously designed; with a sound basis for generalisation; relevant and timely reporting back to industry; and carried out by a self-aware researcher, resulting in research that is a credit to the discipline.
Yin (2003) provides seminal guidance on case study design that addresses the methodological requirements for research to be objective, valid and reliable. A good research design should incorporate five key components; research questions; propositions or purpose; unit of analysis; logic linking data to propositions; and criteria for interpretation of findings (Yin 2003: 21). These are discussed below and are represented in the case study design at Appendix 3.4.
3.15.1 Research questions
The research questions emerged from the literature review and are expressed as ‘what’ and ‘how’ queries, as discussed in Section 3.7.2. and reprised here as:
1 What is the impact of the Code for Sustainable Homes on the early stages of social housing development projects11?
2 How do social housing development projects innovate12 to meet the Code for Sustainable Homes?
3.15.2 Research propositions
Research propositions are essential for directing the detail of the research for explanatory case studies. However, research propositions are not appropriate for an exploratory case study, which instead benefits from a clearly stated purpose about what is to be explored, the aim of the exploration and the criteria for judging the success of the exploration (Yin,
11 Where project is defined as ‘a temporary, formal group of organisations with a specific
objective’.
12 Where innovation is defined as ‘the process of a successful application of a new idea
95 2003: 30). For this research, the purpose is to generalise the findings to the theories of innovation generated by the literature review, using a combined model of the innovation structure for Complex Product Systems (Winch, 1998) within the given and interaction environment of Sexton and Barrett’s model of the organisational factors of innovation (2003) as a starting point.
3.15.3 Unit of analysis
The unit of analysis defines the subject of the case study and is a fundamental element of thorough research design. A carefully considered unit of analysis implies a ‘system of action’ (Tellis, 1997: introduction) linking well-defined research questions to the collection of data by identifying information which is relevant to the case, including clear boundaries of both time and people, to define the limits for data collection and analysis. A sound unit of analysis will guard against any confusion between data collection sources (e.g. people, documents) and the subject of the research. Conversely, a weak or irrelevant unit of analysis implies that ‘everything’ may be studied, which is not possible (Yin, 2003: 23), nor is it good research. The unit of analysis is thus expressed as the innovation needed to achieve Code 4. This clarity links the research questions and hypothesis through to the sources of data, and enables the specific issues of the unique site location and its physical challenges, which are important to participants but not of primary relevance to the research questions, to be separated during data analysis.
3.15.4 Logic linking data to purpose
This logic is shown in the case study design at Appendix 3.4. Although this research, being exploratory, does not make use of a theory-testing proposition (as noted in Section 3.15.2) the purpose (as an alternative) and data are logically linked by the semi-structured interview questions which explicitly capture evidence of capacity for innovation from
96 organisations within the Complex Product Systems (CoPS) model, and the additional data which contextualises this capacity.
3.15.5 Criteria for interpretation of findings
The research design reflects three key principles noted by Yin. First, the design includes multiple sources of evidence. These are noted in Section 3.17 and the extent of these supports the generation of robust research conclusions. Secondly, the design makes use of a case study database. This research, including the case study, has made extensive use of NVivo both as content analysis software and for its capacity to capture and organise many sources of data, ideas, thought processes and links between these. At the same time, the researcher is aware that ‘software cannot...compensate for limited interpretive capacity’ (Bazeley, 2007; 3). The use of software, and specifically NVivo, both as a project database and to support content analysis, is discussed and justified in Section 3.19. Thirdly, the design should demonstrate a chain of evidence, to establish a clear link from the final case study report back to the research questions via the database of multiple sources of evidence, supporting the research as a coherent whole. The research design at Appendix 3.5 reflects this coherence.