Shlomo Hareli
Professor of Social Psychology · School of Business Administration · University of Haifa
My main research interest is the social perception of emotions — how people read others' emotional expressions and draw inferences from them — which I study both theoretically and empirically. I head the Laboratory for the Study of Social Perception of Emotions at the University of Haifa.
I am interested in the social information that emotions convey: what observers learn from another person's reaction about that person's traits, intentions, and social standing, and how the surrounding context shapes those inferences. This work spans the dynamics of how expressions unfold over time, the way context and emotion reciprocally inform one another, and the attributions and personality impressions people form from what others feel and show.
Part of this work focuses on developing tools and approaches for using artificial intelligence in emotion research — including methods for creating and validating photorealistic AI-generated facial-expression stimuli that give researchers fine-grained experimental control that photographic methods cannot.
Research Areas
Social Information in Emotions
What types of social information do people extract from others' emotional reactions, and how do these inferences shape behavior in achievement, motivational, and relational contexts?
Key papers: Hareli & Hess (2010), Cognition and Emotion; Hareli & Hess (2012), Cognition and Emotion
Emotion Expression Dynamics
How emotional expression unfolds over time shapes what observers perceive. I developed the "Frozen Dynamism" methodology to examine temporal aspects of expression in controlled experiments.
Key papers: Hareli, David & Hess (2016), Cognition and Emotion; Hareli, Halhal & Hess (2018), Frontiers in Psychology
Context Effects on Emotion Perception
Context and emotion expressions reciprocally inform one another. My work includes identifying the concept of situative informativeness — the finding that different emotions vary in how strongly they override contextual information.
Key papers: Hareli, Elkabetz & Hess (2019), Emotion; Hess, Dietrich, Kafetsios, Elkabetz & Hareli (2019), Cognition and Emotion
Personality Inference from Emotions
Emotional expressions are rich cues for trait attributions. Research in this area covers inferences of modesty, arrogance, dominance, and social power from what people feel and show.
Key papers: Hareli, Shomrat & Hess (2009), Emotion; Hess, David & Hareli (2016), Emotion
Attribution Theory
Building on Weiner's framework, I examine how causal attributions shape social emotions — guilt, shame, envy, pride, sympathy — and how these emotions, in turn, communicate attributional content to observers.
Key papers: Hareli & Weiner (2002), Educational Psychologist; Hareli (2014), Emotion Review
AI-Generated Research Stimuli
I develop and validate AI-generated facial expression stimuli (Flux diffusion models, custom LoRA training, Py-Feat validation), providing fine-grained experimental control previously impossible with photographic methods.
Key papers: Hareli & David (2026), Behavior Research Methods; Hareli, Hanoch, Elkabetz & Hess (2025), Journal of Nonverbal Behavior
Shapes and Social Judgment
Sharp versus rounded shapes in the environment influence social perception: exposure to sharp shapes leads people to see others as more aggressive, and sharp-leaved vegetation around a house raises its perceived value.
Key papers: Hess, Gryc & Hareli (2013), Social Cognition; Hareli, David, Lev-Yadun & Katzir (2016), Journal of Environmental Psychology
A note on methodology: Generating and validating AI-produced stimuli has become a distinct research contribution. Our dual-validation framework — combining computational facial action coding with behavioral human validation — has achieved recognition accuracy above 97% across multiple studies. Validated image datasets and workflow specifications are openly available on OSF.
I also use these tools to develop interactive classroom demonstrations that bring emotion perception research to life — students manipulate expressions and contextual variables in real time, bridging theory and observation directly in the classroom.