Abstract
This study examines the changing perceptions of autism as represented in scientific and journalistic articles from 1940 to 2023. The study employs a comprehensive dataset of more than 160,000 articles obtained from various digital libraries and news APIs to discern changes in perception and sentiment over time.The study utilizes sophisticated computational methods, including Latent Dirichlet Allocation (LDA) for topic modeling and Valence Aware Dictionary and sEntiment Reasoner (VADER) for sentiment analysis. The combination of these methodologies facilitates a comprehensive analysis of the language and ideas inherent in autism-related texts. Significant findings reveal that scientific papers constantly maintain a positive tone regarding autism, but news stories demonstrate a shift from neutrality to positivity.
This trend indicates an increasing public acceptance and comprehension of neurodiversity. Furthermore, topic modeling uncovers distinct themes in scientific and news stories, characterized by a research-oriented tone in scientific literature and an emphasis on human experiences, familial relationships, and communal events in news media.
Evaluation procedures include an expert interview, focus group discussion, and
feedback from a poster presentation at the NWO ICT OPEN 2024 conference. These qualitative insights complement the computational findings, strengthening the validity of the conclusions.
This thesis offers a thorough examination of evolving perceptions of autism, combining computational approaches with qualitative methods to elucidate intricate social processes. The results have considerable significance for researchers, practitioners, and policymakers, indicating a transition towards a more inclusive and comprehensive view of autism.
| Date of Award | 4 Nov 2024 |
|---|---|
| Original language | English |
| Supervisor | Clara Maathuis (Examiner), Sylvia Stuurman (Supervisor) & Ebrahim Rahimi (Co-assessor) |
Master's Degree
- Master Software Engineering