Coping Through Precise Labeling of Emotions: A Deep Learning Approach to Studying Emotional Granularity in Language
(A2024-119336)

AUTHORS

Ali Faraji-Rad, University of Maryland; Ali Tamaddoni, Deakin Business School; Atefeh Jebeli, University of Maryland, Baltimore County

ABSTRACT

When describing their emotions, people may demonstrate high granularity by differentiating between emotions when using emotional labels, or low granularity by using emotion labels interchangeably to indicate general valance. We develop a deep-learning-based computational method to investigate whether the granularity with which a person describes their negative emotions from an unpleasant experience, in an online review, predicts how well they cope with that experience. Using review ratings as a proxy for coping, we observe that describing negative emotions more granularly in the review text is associated with better coping. Thus, beyond trait-level variances in negative emotional granularity, situation-level variances in how negative emotions are described also impact emotion regulation and coping. The computational method developed to assess emotional granularity in language use may be applied to unobtrusively measure the construct in settings beyond online reviews.