Who Tweets for the autistic community? A natural language processing-driven investigation

Area of Science:

• Autism advocacy
• Social science research
• Computational linguistics

Background:

• Autism advocacy organizations led by family members have faced criticism from the autistic community.
• Autistic individuals formed self-advocacy groups, emphasizing leadership by autistic people.
• Recent debates question the representativeness of self-advocacy groups within the broader autistic community.

Purpose of the Study:

• To computationally analyze the language used by autism advocacy organizations and individuals on Twitter.
• To investigate claims of partial representation within different autism advocacy groups.
• To provide empirical insights into the long-standing conflict surrounding representation in autism advocacy.

Main Methods:

• Application of machine learning and natural language processing techniques.
• Analysis of a large dataset of Tweets from autism advocacy organizations and individuals.
• Specification curve analysis to assess language similarity and representation.

Main Results:

• Evidence supports claims of partial representation for both self-advocacy groups and family-led organizations.
• Both types of organizations show a stronger alignment with the interests of specific subgroups within the autism advocacy community.
• Computational analysis revealed differential representation patterns in online communication.

Conclusions:

• The study introduces a novel computational approach to understanding representation in autism advocacy.
• Findings suggest that no single type of organization fully represents the diverse interests within the autism community.
• The research aims to equip advocacy groups with tools for evidence-based discussions on representation.