Chapter 3 Methodology
D. Documentation intern Internal documents (e.g draft reports, internal memos, briefing papers)
3.8 Data Collection
Establishing methodological rigour is a lot like building a legal case: the more supporting evidence you have, the more solid the case. In case study research, many different sources of evidence are used to build the case. The purpose of this section is twofold: firstly, to discuss the various collection techniques used in this study; and secondly, to highlight the principles of data collection. It is a combination of these techniques and principles that allow validity of research to be attained. To minimise any misinterpretation, researchers employ
commonly two main procedures: redundancy of data gathering and procedural challenges to explanations (Goetz and LeCompte 1984; Denzin 1989). This triangulation method is used when there are multiple perceptions and the researcher is trying to clarify meaning and verify and the repeatability (as discussed earlier) of observations and/or interpretations (Stake 2000). Using a triangulation approach, the techniques being used in this piece of research are documentation, non-participant observation and semi-structured interviews.
Documentation can take many forms and tends to be relevant to every case study topic (Yin 1994). In this instance, 291 documents have been collected—Table 3.10 shows the types of documents collected, the number of each types, and examples of the nature of the corresponding types of documents:
Table 3.10 – Types of Documentation
Document Type Examples Number Date Range
Communiqués Emails, proposals, invoices. 17 2002–2007
Meeting Information Minutes, agendas, presentations, supplementary reports. 61 2002–2008
Internal Documents Internal reports, scenario drafts, analytical summaries, scenario planning interviews. 183 2002–2008
Formal Studies &
Reports Community Plans, Service Reviews, Structure Plans, Improvement Plans, Government Acts. 30 2001–2008 Documentary analysis allows the information to be reviewed repeatedly and contains exact names and references that can corroborate and augment other arguments as well as being a starting point for further research. It is also not created as result of the case study, which helps the researcher maintain a certain distance from the case. Finally the use of documentation affords insight into the process and development of the element under study that could not be witnessed firsthand. However, this also poses a problem for the researcher because some information may be deliberately blocked or ignored to manipulate a specific conclusion. With this taken into account extra precaution was taken to avoid committing any reporting biases.
To maintain the desired separation from the case, this research used non-participant observation to examine and record the steering group’s workshops. Non-participant
observation (NPO) affords the researcher the same contextual reality and insights into the interpersonal motives and behaviours as participant observation except the researcher has little or no direct interaction with them (Maylor and Blackmon 2005), thus leaving the discussions (in this case) free of manipulation and any biases on the part of the research. Although it has criticisms (for example, its likeness with ‘snooping’), NPO does give an accurate picture of what happens (Easterby-Smith et al. 2002), serving as a preliminary collection of data and as a casual enquiry as to the nature of the workshop (Thietart 2001). In this case study it was used as a preliminary and complimentary data source.
The third and arguably most important method of data collection is interviews, which “are an essential source of case study evidence” (Yin 1994: 85). By using semi-structured or open- ended interviews, it is possible to sharpen the focus towards the exact research questions while gathering a large quantity of insights and perceived inferences (Yin 1994) at the relevant stage of the process. However, interviews are subject to many criticisms on many fronts. Often biases in the questions and in the responses can give a false representation of the case; also, respondents can sometimes give interviewers what they think they want to hear (Maylor and Blackmon 2005). To counter this, the interview-questions were constructed carefully, and with guidance from polling and interviewing literatures (for example, Payne 1951), and were tested for biases several times; also, given the exploratory nature of the case study, there is not the same onus on the researcher to produce a set of desired results. The interviews were constructed to extract the most value from the subjects as pertaining to their specific stage and role in the case. A list of all persons interviewed and the outline of the questions used in the interviews can be found in appendix A and B, respectively. While using group interviews can help reduce individual biases, the research is dependent on in- depth explorations, which is seldom achieved in group situations (Maylor and Blackmon 2005).
To maximise the benefit of these sources of evidence, Yin’s (1994) three principles of data collection were followed closely. First of all, this research uses multiple sources of evidence, capitalising on the benefits of data triangulation to address a wider range of issues. Findings are more likely to be accurate if they are the conclusion of multiple sources of corroboratory information. It is also through triangulation that the issue of construct validity may be dealt
with. The second principle is to create a case database, an example of which was shown in Table 3.8. Essentially, it is a method for organising and documenting data, and is intended to help increase the reliability of the entire case study. The third principle is to maintain a chain of evidence to help achieve construct validity and thus increase the reliability of the information contained within the case study. As was discussed above in section 3.6.3, the “principle is to allow an external observer…to follow the derivation of any evidence from initial research questions to ultimate case study conclusions” (Yin 1994: 98). The chain of evidence should allow the reader to examine the case itself, being able to traverse easily between findings and evidence that all possess consistent methodological procedures.