Current checklists for empirical software engineering cover either experimental research or case study research but ignore the many commonalities that exist across all kinds of empirical research. Identifying these commonalities, and explaining why they exist, would enhance our understanding of empirical research in general and of the differences between experimental and case study research in particular. In this report we design a unified checklist for empirical research, and identify commonalities and differences between experimental and case study research. We design the unified checklist as a specialization of the general engineering cycle, which itself is a special case of the rational choice cycle. We then compare the resulting empirical research cycle with two checklists for experimental research, and with one checklist for case study research. The resulting checklist identifies important questions to be answered in experimental and case study research design and reports. The checklist provides insights in two different types of empirical research design and their relationships. Its limitations are that it ignores other research methods such as meta-research or surveys. It has been tested so far only in our own research designs and in teaching empirical methods. Future work includes expanding the comparison with other methods and application in more cases, by others than ourselves.
|CTIT Technical Report Series
|University of Twente, Centre for Telematics and Information Technology
- IS-Design science methodology