Date of Award

5-11-2025

Date Published

June 2025

Degree Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Mass Communications

Advisor(s)

Dennis Kinsey

Keywords

big data;chaos theory;computational methods;crisis;PREDICT;social media

Subject Categories

Communication | Mass Communication | Social and Behavioral Sciences

Abstract

The vast scale of digital information generated daily provides unprecedented opportunities to analyze human behavior in ways that can improve crisis prediction and response. This dissertation introduces PREDICT, a theoretical framework that reconceptualizes crises as predictable phenomena rather than unknowable disruptions. Grounded in chaos theory and complex systems theory, PREDICT operationalizes crisis prediction through three constructs: Perceived Rupture of Equilibrium (PRE), Dynamic Interactions (DI), and Crisis Trajectory (CT). These constructs enable a shift from reactive crisis management to proactive crisis forecasting by identifying early signals of instability, modeling stakeholder interactions, and mapping potential crisis trajectories. To test PREDICT, this study employed an inductive multi-method computational approach, Layered Analytic Crisis Insights (LACI), integrating regression analysis, agent-based modeling, network analysis, time-series forecasting, and decision-tree modeling to provide quantitative results as a baseline for theory refinement. Applying LACI to 108,000 digital traces from TikTok related to the UnitedHealthcare/Luigi Mangioni crisis, this research demonstrated how computational techniques can be layered to detect patterns, simulate stakeholder responses, and identify critical windows for intervention. Findings support PREDICT’s underlying assumptions and offer insights into ways to operationalize social media big data to model system instability. This study challenges the dominant paradigm that treats crises as unpredictable events, arguing that advances in predictive analytics and data science enable researchers and practitioners to anticipate crisis conditions before they destabilize organizations. The findings provide a tailorable framework for both crisis scholars and practitioners, offering new avenues to extend existing theories such as contingency theory, social amplification of risk framework (SARF), situational crisis communication theory (SCCT), and discourse of renewal. Keywords: crisis, chaos theory, PREDICT, computational methods, social media, big data

Access

Open Access

Download

Share

COinS