Stress pervades our everyday life to the point of being considered the scourge of the modern industrial world. The effects of stress on knowledge workers causes, in short term, performance fluctuations, decline of concentration, bad sensorimotor coordination, and an increased error rate, while long term exposure to stress leads to issues such as dissatisfaction, resignation, depression and general psychosomatic ailment and disease. Software developers are known to be stressed workers. Stress has been suggested to have detrimental effects on team morale and motivation, communication and cooperation-dependent work, software quality, maintainability, and requirements management. There is a need to effectively assess, monitor, and reduce stress for software developers. While there is substantial psycho-social and medical research on stress and its measurement, we notice that the transfer of these methods and practices to software engineering has not been fully made. For this reason, we engage in an interdisciplinary endeavor between researchers in software engineering and medical and social sciences towards a better understanding of stress effects while developing software. This article offers two main contributions. First, we provide an overview of supported theories of stress and the many ways to assess stress in individuals. Second, we propose a robust methodology to detect and measure stress in controlled experiments that is tailored to software engineering research. We also evaluate the methodology by implementing it on an experiment, which we first pilot and then replicate in its enhanced form, and report on the results with lessons learned. With this work, we hope to stimulate research on stress in software engineering and inspire future research that is backed up by supported theories and employs psychometrically validated measures.
Keywords: Biological markers; Effects of stress; Methodology; Psychological assessment; Software development; Stress.
© 2020 Ostberg et al.