The benefits of rich data provided through non-standardised interviews highlighted above are affected by concerns of quality. Saunders et al. (2009) identifies three data quality issues that could affect semi-structured interviews. They are:
- Reliability
- Forms of bias
- Validity and generalisability
Reliability is defined by Easterby-Smith et al. (1991) are concerned with the ability to duplicate results if the interview was conducted by a different interviewer. Even with ample time and money ideal conditions may not be achieved therefore careful notes were taken to documents the approach and design of the interview.
Bias can be found in both interviewer and interviewee. These can affect the results of the interview and therefore need to be overcome and Saunders et al. (2009) suggests a few steps to limit bias.
- Preparation and readiness for the interview- is needed “to establish your credibility interview of the research participant” (Saunders et al., 2009). This preparation comes from the researchers literature review, researching the Bitcoin community and general communication with these executives and business owners.
- Level of information supplied to the interviewee- “credibility may also be promoted to the supply of relevant information to participants before the interview” Saunders et al. (2009). A list of common concerns and answers, an outline of the research topic, and the interview schedule were provided to interviewees at least two days prior the interview to satisfy these conditions.
- Demonstration of attentive listening skills- exploratory questions or statements were written into the interview schedule to probe into the meanings of interviewee answers. Careful listening allows significant answers relevant to the research topic to be identified.
Even after successfully applying these steps it is still difficult to ensure that the interviewee or interviewer is totally unbiased. “There may be misinterpretation of responses because of cultural differences between the interviewee and interviewer” Marshall & Rossman (1989). The author recognises this issue and will take it into consideration in the analysis.
Validity and generalizability- this concerns the ability of findings from a small number of cases to represent a larger population. Marshall & Rossman (1989) argue that “where you are able to relate your research project to existing theory, you will be in a position to demonstrate that your findings will have a broader significance than the case or cases which form the basis of your work”.
The New Exchange 