Research

Research Program

I study how digital platforms shape collective behavior from the bottom up, tracing how local interactions among individuals aggregate into higher-order patterns such as coordination, shared understanding, emotional contagion, and emergent conventions.
Across projects I combine theory development, online experiments, agent-based simulation, and computational measurement, and I work in settings where the stakes of coordination are concrete, such as disaster response, protest mobilization, platform culture, and collective consumption.
Individual Interaction Group Crowd Society

My work traces the passage from left to right, and asks how technology changes it.

Coordinating Physical Work in Emergent Crowds

When people converge on a disaster site or a protest, they often have no prior roles, hierarchy, or shared history, yet the work in front of them is physical and interdependent. This program asks what digital platforms can do for groups like these. It began as my doctoral dissertation at Simon Fraser University and continues in three linked studies: one theoretical, one experimental, one computational.

Aligning Crowds on the Ground: Information Systems for Physical Emergent Collaboration

Ehsan Nouri, Nilesh Saraf, Terri Griffith
Preparing for submission  ·  Theory development
A theoretical model of how information systems support crowds that assemble without established roles or structure to carry out work bound by material resources, spatial constraints, and real-world interdependencies. The paper bridges research on open online collaboration, crowdsourcing, and embedded computing to specify the sociotechnical mechanisms behind self-organized group work on the ground.
Presented at the Academy of Management Annual Meeting 2024 and the ICIS 2024 Doctoral Consortium. AOM 2024 proceedings

Making Physical Work Visible: Associative Cues and Shared Situational Awareness in Crowd Coordination

Ehsan Nouri, Terri Griffith, Nilesh Saraf
Under review, Journal of Management Information Systems  ·  Online experiment
How can platforms help a crowd coordinate physical work by making that work visible to others? We built a purpose-built multiplayer experimental environment and ran a synchronous online experiment to test how cues associating people, places, and objects shape shared situational awareness and the coordination that follows from it.
Accepted to the Academy of Management Annual Meeting 2026, CTO Division. Presented at the Virginia Research Seminar Series in Information Systems (VRSS 2026).

Decentralized Physical Coordination under Uncertainty: A Simulation of Communication and Shared Awareness

Ehsan Nouri
Work in progress  ·  Agent-based simulation
A computational model of how shared cognition forms through digital communication when actors face uncertainty and differ in how willingly they share what they know. The simulation follows the passage from individual perception and local messaging to collective organization, and tests where that passage breaks down.
Presented at the Biocomplexity Institute, University of Virginia (2025), and the Theoretical Organizational Models Society Brownbag (2025).
Featured interview on the dissertation research, Beedie School of Business

The Emergence of Behavioral Conventions on Digital Platforms

A convention is a behavioral regularity that a population sustains because each person expects the others to follow it. Conventions are arbitrary, in that other regularities would have served just as well, and they are never announced. They emerge from the bottom up as people observe one another and imitate, until scattered acts of imitation lock a population into a shared way of doing things, whether a greeting, a format, or a phrase. This project studies how that process unfolds in digital environments, and why one candidate prevails over its equally serviceable rivals.

The Emergence of Interpretive Conventions on Short-Video Platforms

Ehsan Nouri, Brent Kitchens, Steven L. Johnson
Under review, HICSS  ·  Computational measurement  ·  Vision-language models
Using vision-language models to annotate a large corpus of TikTok videos, we identify the recurring visual and narrative signature elements that define micro-genres, and trace how conventional forms consolidate out of dispersed creative choices. The work also develops the measurement approach itself, treating foundation models as instruments for coding unstructured multimodal content at scale.
Presented at the Responsible AI in Varying Environments (RAVE) Lab, School of Data Science, University of Virginia (2026), and the Digital Technology for Democracy Lab (2026).

Group Formation and Shared Consumption

Much of what people consume, they consume together. Films, restaurants, and concerts are chosen by groups rather than by individuals, which means the decision depends on a group forming in the first place. That adds a second way for demand to fail, one that individual-level models cannot see, since a person can want to go and still not go because no group ever assembled around them. I study how social information travels through both channels at once, shaping what people want and whether the groups that would act on that want ever come together.

Effects of Negative Word-of-Mouth and Joint Consumption on the Box Office Performance of Motion Pictures. An Agent-Based Model Approach

Ehsan Nouri, Jason Ho, Charles Weinberg
Preparing for submission  ·  Agent-based simulation
Moviegoing is usually a group decision, but models of word-of-mouth typically treat it as an individual one. We simulate movie choice as a group-level process in which negative word-of-mouth does double damage, lowering a person’s own interest and making it harder to assemble a viewing group at all. The model also shows how front-loading advertising into the opening week can offset part of that effect.
Best PhD Paper Award, Beedie School of Business, Simon Fraser University, 2024.

The Spread of Negative Emotions on Social Networks

Emotional contagion is the transfer of affective states from one person to another through exposure alone, without persuasion or intent. Someone encounters an expression of anger or grief, takes on something of it, and passes it along. What begins between two people becomes collective once it runs across a network, because individual transfers compound until a population can shift in mood with no coordinated cause behind it. I study how different emotions move through these cascades, how far and how fast each one carries, and why some fade within hours while others hold for days.

Online Propagation of Emotions. A Study of Resharing Dynamics on Social Media Following Celebrity Suicides

Ehsan Nouri, Nilesh Saraf, Jing Mei Goh, Srabana Dasgupta, Dianne Cyr
Published, PLOS ONE, 2025  ·  Econometric analysis  ·  Text classification
Emotional contagion on social media following celebrity suicides, and how those dynamics may contribute to copycat outcomes. Across four events, we extract discrete emotions from over one million posts using a BERT-based classifier and model four properties of the resulting cascades, namely size, lifetime, speed, and burstiness. Disgust proves the most contagious, spreading widely, quickly, and durably. Anger and surprise produce fast but short-lived bursts, fear travels weakly despite its high arousal, and joy, though rarer, outlasts both neutral and negative content. Valence and arousal alone, we conclude, do not account for how emotion moves at scale.
PLOS ONE 20(12), e0336134. Earlier versions presented at AMCIS 2023 and the Pre-ICIS SIGDSA Symposium 2022.
Working manuscripts are available on request at ehsan@virginia.edu. A full record of publications, talks, and teaching is in my CV.