Skip navigation
Please use this identifier to cite or link to this item: http://arks.princeton.edu/ark:/88435/dsp01gt54kr21j
Title: Twitter Transmission: A Quantitative Analysis of Social Media Content Diffusion Through Controlling for Cascade Features
Authors: Nelsen, Jerod
Advisors: van Handel, Ramon
Department: Operations Research and Financial Engineering
Certificate Program: Program in Cognitive Science
Class Year: 2022
Abstract: Social media sites are being used more widely and frequently than ever before. Sites like Facebook, Twitter, and LinkedIn—originally conceived for social networking purposes—now play a critical role in the dissemination of news and information. The ubiquity and rapid content-sharing capabilities of social media make it a particularly appealing mechanism for political, economic, and broader social interest groups to disseminate their messages (sometimes for nefarious purposes) and shape opinions. Yet, despite the escalating popularity and social importance of these sites, relatively little is known about the underlying diffusion dynamics governing content spreading on social media. Why does an individual receive a particular form of content, and what does it tell us about macro-scale human consumption patterns? Moreover, why does misinformation appear so rampant on these sites? Due to the recency of social media’s inception, these questions are only beginning to be explored. This thesis seeks to quantitatively analyze the diffusion of content on social media, answering important questions about the differential spreads of true and false rumors on Twitter. Following three recent, seminal papers analyzing social media content diffusion and misinformation, this thesis makes three important contributions: first, Vosoughi et al.’s central finding that “falsehood diffuse[s] significantly farther, faster, deeper, and more broadly than the truth” is revisited and verified; next, by implementing a novel stochastic matching algorithm, Juul and Ugander’s discovery that “for false- and true-news cascades, the reported structural differences can almost entirely be explained by false-news cascades being larger” is observed to be regime-sensitive; finally, by analyzing alternative cascade statistics, controlling for size (as opposed to depth, maximum breadth, or structural virality) is found to best collapse cascade featural differences. Various mathematical techniques such as two-sample Kolmogorov–Smirnov (“K-S”) tests, Monte Carlo simulations, and multiple logistic regression models are utilized throughout. Implications of these results are explored with an eye towards the future usage and policy of social media. This thesis presents a novel analysis of social media that engages with the existing literature to help propel knowledge on content diffusion and misinformation forward, in hopes of creating a more truth-favorable social media ecosystem.
URI: http://arks.princeton.edu/ark:/88435/dsp01gt54kr21j
Type of Material: Princeton University Senior Theses
Language: en
Appears in Collections:Operations Research and Financial Engineering, 2000-2023

Files in This Item:
File Description SizeFormat 
NELSEN-JEROD-THESIS.pdf2.09 MBAdobe PDF    Request a copy


Items in Dataspace are protected by copyright, with all rights reserved, unless otherwise indicated.