myStress: Unobtrusive Smartphone-Based Stress Detection

Life is becoming increasingly stressful in many aspects, e.g., due to technology-induced stress and stress in organizational context. The assessment of stress experienced by individuals enables stress management and prevention with the long-term aim to avoid psychological and physiological harm from excessive stress. Commonly this assessment is performed through questionnaires on perceived stress or physiological measurements evaluating body reactions to stress. We explore a third assessment method: Our design science approach aims to unobtrusively assess perceived stress based on smartphone data while waiving additional devices and explicit user input. The presented design artefact, myStress, reads 36 hardware and software sensors to infer users’ perceived stress levels. A prototypical instantiation of myStress for the Android platform is distributed to test users. For evaluation purposes, the stress level additionally is determined by a questionnaire consisting of the Perceived Stress Scale. By analyzing data from test users, we gain first insights into the feasibility of unobtrusive, continuous stress assessment considering exclusively data from smartphone sensors. We find that several sensors seem to correlate with perceived stress, e.g., the frequency of switching the display on/off. For future research, behavioral and situational prevention measures can build on this method of unobtrusive stress assessment.