Date of Award
2025
Degree Type
Thesis
Degree Name
Master of Science in Computer Science
Department
Computer Science and Statistics
First Advisor
Shaun Wallace
Abstract
Limited research has explored how English as a Second Language (ESL) individuals interact with United States voter pamphlets. This study examines the effects of artificial intelligence (AI) based summarization on reading speed and comprehension among native and non-native English speakers. Results show that ESL participants read baseline passages with images faster than AI-summarized versions. Similarly, non-ESL participants also demonstrated better performance with baseline passages, especially those containing images, compared to AI-generated summaries. To account for individual variation, we incorporated the Object-Spatial-Verbal Cognitive Style Questionnaire (OSIVQ) to assess whether cognitive style moderates these effects. Our findings reveal that individuals with object and verbal cognitive styles benefited most from baseline materials, particularly when images were included. These results underscore the importance of tailoring voting materials to both linguistic background and cognitive preferences to improve how effectively information is conveyed.
Recommended Citation
Khadka, Kushas, "ESL AND NON-ESL READERS ACROSS COGNITIVE STYLES READ AI-SUMMARIZED VOTER PAMPHLETS SLOWER" (2025). Open Access Master's Theses. Paper 2649.
https://digitalcommons.uri.edu/theses/2649