Case study data collection

Yin (2009) identifies a number of skills required for a good case study researcher. One must be:

• a good listener
• able to ask good questions -
• adaptive and flexible
• have a firm grasp of the issues being studies and 
• unbiased by preconceived notions

A good way to approach this last point, is to report preliminary findings (to a conference or colleagues) and look for alternative explanations and suggestions. If these can be refuted then the likelihood of bias will have been reduced. 

In the case whereby a team is to conduct the case study research, Yin recommends training for the group as this can uncover a number of problems. Flaws with the case study design or research questions; incompatibilities within the investigating team; any impractical time deadlines or expectations of resources. Positive features may also be uncovered at a training session such as team members' complimentary skills.

Case study protocol. As mentioned before this is a major way of increasing the reliability of case study research. Recommened sections of this protocol include: an overview of the project; field procedures; case study questions; a guide for the case study report.

The overview includes background information such as the context and perspective. It also covers the substantive issues being covered, which would include the rationale for selecting the cases, any propositions or hypotheses being studied and the broader theoretical or policy relevance to the inquiry. Finally, the overview would include any relevant readings relating to the issues.

Field procedures refers to being prepared for the unexpected. Since the case study depends heavily on the interviewee, the researchers must be prepared to be flexible and responsive to the interviewee, must bring along all equipment to the field location and be prepared. This includes receiving ethics clearance where appropriate and providing information and consent forms for signing, from participants.

Case study questions. The questions in a case study protocol are different from those in a survey because they aim to reflect the actual line of enquiry and not a respondents point of view. The protocols questions are specific reminders of what information needs to be collected, and why so their purpose is to keep the researcher on track. There are different levels of questions ranging from the interview questions, to the questions to be asked of the case, to the questions to be asked of nthe pattern of findings across multiple cases, plus questions to be asked of literature or other published data related to the study and finally, possibly, to normative questions about policy documentation and conclusions. Yin uses a nice description for distinguishing particularly between the first two levels of questions in that 'The verbal line of inquiry is different from the mental line of enquiry' (p87).
Great care must be taken not to confuse the data collection source with the unit of analysis.

In my study, I am interested in an individual as the unit of analysis. However information should also be sourced (and questions asked) from the organisation - from reports, other employees and managers etc. I'm now thinking it may be a good idea to interview the project leaders not just the project managers. ie Matt, Mitch and Justin. And so need to come up with a list of questions to gather data from these sources. Also, I'm not crystal clear about my unit of analysis. I'm thinking it is the individual project but maybe I should be looking at the person as they are the ones exhibiting the behaviour, attitudes and perceptions about evaluation....

The final element is the guide for the case study report. One must think ahead to the audience for the report (as in a good evaluation) before data collection begins but also the format and outline of what will be included. Unlike other research methods where the sequence of events in the research are linear, a case study approach requires this out-of-step- method so that all possibilities for data collection are considered and reconsidered throughout the study, allowing for redesign as we go along. The other consideration for the report is how to include the large amount of documentary evidence that may form part of the data collection phase. This can be included as an annotated bibliography thus directing readers to locations if they require further evidence. One final note rom Yin here is that other research methods are often dictated by journal requirements but as case-study research is less likely to make it to a journal publication, the researcher is able to be more flexible in their approach to the method.

Yin completes this chapter with a short discussion on how to screen the candidates for a case study. Reasons for selection should be included in the protocol. Pilot case studies are also recommended whereby the pilot test offers the opportunity for formative feedback on the research questions and even some conceptual clarification for the research design.