کد مقاله | کد نشریه | سال انتشار | مقاله انگلیسی | نسخه تمام متن |
---|---|---|---|---|
283821 | 509117 | 2015 | 11 صفحه PDF | دانلود رایگان |
• Qualitative building-energy research often uses semi-structured interviews.
• The number of interviews required for reliable results is controversial.
• Theoretical literature on small sample qualitative interviews is here reviewed.
• A binomial framework is developed for robustly estimating minimum sample size.
• This offers guidance for qualitative researchers to improve result reliability.
Research in building energy consumption often uses semi-structured interviews to produce qualitative data on consumer beliefs, attitudes, practices and skills. A survey of 54 recent papers in six prominent building and energy journals shows that the samples are typically small, but inferences are often made for interventions in the light of the findings, on the assumption that these are somehow transferable to wider populations. It is often asked ‘how many interviews are enough’ to produce reliable results. Theoretical literature on this theme has avoided a straightforward statistical critique, and justified the practice with appeals to precedent, the special nature of qualitative personal data, and a limited pool of empirical work. This paper reviews this literature and presents a statistical approach, based on binomial logic, to critiquing and supporting the practice of semi-structured interview research in the building and energy field. The approach developed offers a set of straightforward criteria which researchers can use to estimate the reliability of their findings and inferences from the qualitative data produced in semi-structured interviews.
Figure optionsDownload as PowerPoint slide
Journal: Journal of Building Engineering - Volume 1, March 2015, Pages 2–12