Back to search

Rochester Institute of Technology

Stance Detection on Social Media: The Impact of Doping on Athlete Reputation

Abstract

dc:description.abstract

Doping scandals can have a lasting effect on both athletes' and organizations' reputations, causing athletes to lose money and fans to lose trust in the organizations that govern their favorite sports. This paper investigates the impact a doping scandal can have on an athlete's reputation by conducting a case study on Jannik Sinner, one of the top tennis players in the world, who ultimately accepted a suspension after testing positive for performance-enhancing drugs, despite being unintentionally exposed. Using data sourced from YouTube comments on tennis-related videos, a fine-tuned Mistral-7B-Instruct-v0.3 model with a classification head was used to carry out stance detection to investigate the stance of fans towards Sinner over time, classifying comments as guilty (believing Sinner intentionally cheated) or non guilty (supporting innocence, neutral, or irrelevant). Our model achieved strong performance on non guilty comments (F1=0.97) but only moderate precision and recall on guilty ones (0.71, 0.72), struggling with sarcasm, indirect accusations, and case-specific references. Analyzing 843,005 comments and focusing on those spanning the 20 months post-scandal, the results show a downward trend in the percentage of guilty comments (Kendall's [tau] = -0.829), dropping from 2̃7.9% to 1̃.15%, supporting our hypothesis that doubters decrease over time. However, key events like Sinner's settlement and certain major tournament appearances can temporarily reignite discussion. The results also confirmed that fans revisit older videos to discuss new doping allegations. These findings demonstrate that the effects of a doping scandal slowly fade over time but persist for an extended period, and this paper offers a generalizable framework for examining similar cases across sports. This study highlights the potential value of social-media-based stance detection in informing anti-doping policies and communication strategies.

Degree

thesis:*
Name thesis:degree_name
Software Engineering (MS)
Level thesis:degree_level
Thesis
Discipline thesis:degree_discipline
Software Engineering, Department of

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • O'Neill, Connor
Contributors dc:contributor
  • Ashique KhudaBukhsh

Subjects

dc:subject × 6

Identifiers

dc:identifier.*
Repository record dc:identifier
https://repository.rit.edu/theses/12633

Chain of custody

source
Harvested from
Rochester Institute of Technology
Base URL
repository.rit.edu/do/oai/
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
citation

O'Neill, Connor. Stance Detection on Social Media: The Impact of Doping on Athlete Reputation. Thesis thesis, https://repository.rit.edu/theses/12633