Use Eye Movement Features to Explore the Impact of Language Features on Text Difficulty

Abstract

Lifelong learning emphasizes continuous learning throughout one's life, with reading enriching elders' lives and aiding their adaptation to aging. However, aging is often associated with a decline in cognitive functions and an increased risk of impairment, making providing appropriate reading materials for elders a crucial matter. Scholars have explored the correlation between eye movement indicators and text difficulty. However, previous research has rarely explicitly linked eye movement indicators to linguistic features, which may limit authors' ability to comprehensively assess their texts' readability. Therefore, this study utilizes an eye tracker to observe elders' reading behavior when engaging with texts of varying difficulties, extracts eye movement indicators to examine challenges faced by elders with different reading abilities, and employs CRIE (Chinese Readability Index Explorer) to analyze the computational linguistic features. The aim is to understand why certain linguistic features pose reading challenges for elders. The findings revealed that when elders read difficult texts, their dwell time and regression count increased, particularly with longer sentences and unfamiliar vocabulary. Linguistic indicators exhibited moderate correlations with several eye movement indicators, suggesting that language structure significantly influences reading behavior. This study, beginning with the reader's reading process, identifies specific correlations between eye movement indicators and text difficulty, offering a novel perspective for research in eye movement and computational linguistics. Furthermore, it has implications for the cognitive health management of elders in medical and psychological fields, ultimately enhancing their quality of life and promoting lifelong learning.

Save as a Favourite

Author Information
Yi Ju Chan, National Taiwan University of Science and Technology, Taiwan
Hou-Chiang Tseng, National Taiwan University of Science and Technology, Taiwan
Yao-Ting Sung, National Taiwan Normal University, Taiwan