Utilizing Large Language Models to Monitor Social Media for Disability: An Analysis of Sentiment and Disability...

Area of Science:

• Artificial Intelligence
• Computational Social Science
• Disability Studies

Background:

• Online discourse significantly shapes public perception of disability rights.
• Large language models (LLMs) offer potential for analyzing large-scale digital conversations.
• Understanding the framing of disability (social vs. medical model) is crucial for advocacy.

Purpose of the Study:

• To evaluate the efficacy of LLMs in analyzing tweets concerning disability rights.
• To determine LLM capabilities in sentiment analysis of disability-related content.
• To assess LLM performance in differentiating between the social and medical models of disability.

Main Methods:

• Collection of 5,000 tweets related to disability rights.
• Training a large language model for content analysis.
• Quantitative evaluation of model accuracy for sentiment and model classification.

Main Results:

• The trained LLM demonstrated high accuracy in identifying tweet sentiment (positive/negative).
• The model achieved promising accuracy in classifying tweets according to the social versus medical model of disability.
• Results indicate LLMs are effective tools for analyzing nuanced online discussions on disability.

Conclusions:

• LLMs are viable tools for automated analysis of online discourse on disability rights.
• This technology can aid researchers and advocates in understanding public opinion and framing of disability.
• Further research can refine LLM applications for more complex social science inquiries.