Separating Work-Related and Personal Emotional Signals in WELD

Ascertain how to separate work-related emotional responses from personal emotional influences (e.g., family stress, health concerns) within the WELD dataset’s facial-expression-derived emotion measurements, enabling attribution of observed emotional dynamics specifically to workplace factors.

Background

WELD provides longitudinal, naturalistic facial expression data from workplace cameras, capturing authentic emotional dynamics. However, the dataset lacks contextual annotations that distinguish whether recorded emotions are elicited by workplace stressors or by external personal events.

Disentangling these sources is essential for valid interpretation of emotional metrics, for designing targeted organizational interventions, and for ensuring that analytics derived from WELD do not conflate personal wellbeing issues with workplace conditions.