Psychometric instruments in software engineering research on personality: Status quo after fifty years

Abstract

Context:Although software development is a human activity, Software Engineering (SE) research has focused mostly on processes and tools, making human factors underrepresented. This kind of research may be improved using knowledge from human-focused disciplines. An example of missed opportunities is how SE employs psychometric instruments. Objective:Provide an overview of psychometric instruments in SE research regarding personality and provide recommendations for adopting them. Method:We conducted a systematic mapping to build an overview of instruments used within SE for assessing personality and reviewed their use from a multidisciplinary perspective of SE and social science. Results:We contribute with a secondary study covering fifty years of research (1970 to 2020). One of the most adopted instruments (MBTI) faces criticism within social sciences, and we identified discrepancies between its application and existing recommendations. We emphasize that several instruments refer to the Five-Factor Model, which despite its relevance in social sciences, has no specific advice for its application within SE. We discuss general advice for its proper application. Conclusion:The findings show that the adoption of psychometric instruments regarding personality in SE needs to be improved, ideally with the support of social science researchers. We believe that the review presented in this study can help to understand limitations and to evolve in this direction.