Experiences from conducting semi-structured interviews in empirical software engineering research

Many phenomena related to software development are qualitative in nature. Relevant measures of such phenomena are often collected using semi-structured interviews. Such interviews involve high costs, and the quality of the collected data is related to how the interviews are conducted. Careful planning and conducting of the interviews are therefore necessary, and experiences from interview studies in software engineering should consequently be collected and analyzed to provide advice to other researchers. We have brought together experiences from 12 software engineering studies, in which a total of 280 interviews were conducted. Four areas were particularly challenging when planning and conducting these interviews; estimating the necessary effort, ensuring that the interviewer had the needed skills, ensuring good interaction between interviewer and interviewees, and using the appropriate tools and project artifacts. The paper gives advice on how to handle these areas and suggests what information about the interviews should be included when reporting studies where interviews have been used in data collection. Knowledge from other disciplines is included. By sharing experience, knowledge about the accomplishments of software engineering interviews is increased and hence, measures of high quality can be achieved