| dc.description.abstract | New techniques for refining requirements in software development are frequently presented. However, these techniques are often not scientifically validated on their performance. In this study, we aim to validate two refinement techniques, Example Mapping and Feature Mapping. These two techniques are both Three Amigo session techniques, in which people from different work together to refine the software increment. Shared understanding is an essential aspect of these sessions. We present a new measurement tool with which the performance of refinement techniques can be investigated, which also incorporates shared understanding. We investi- gate the performance of the refinement techniques with a controlled experiment and four case studies. Based on the results, Example Mapping performs well under certain conditions. We observed a learning effect for the technique, which resulted in better performance of sessions after having used the technique several times. For Feature Mapping, results were inconclusive and additional research is required to establish its performance. When conditions are not right for Example Mapping or Feature Mapping, Three Amigo session principles can still be applied to refinements to result in a well-performing session. Further research could be undertaken to generalise this study’s findings further, studying the long-term effects of the techniques on the implementation of a user story and studying the performance of the techniques when team members are co-located. Additionally, future research can be conducted to compare the Three Amigo session techniques to other refinement techniques, to further investigate the validity of the performance measurement tool and to study the performance of shorter Three Amigo sessions. | |
