The primary objective of this study is to investigate the scanpaths of attentive and inattentive readers at global and local scales during different reading tasks. This research holds theoretical significance in three aspects: (a) the application of a new algorithm to analyze global patterns; (b) identification of local pattern, local strategy, and local strategy transition through analyzing local scanpaths; © exploring the local differences in these three aspects between attentive and inattentive readers within each task. The practical implications of the findings include valuable advice for improving reading system design and identifying reader types for effective training.
The global analysis revealed that neither attentive nor inattentive readers exhibit distinct global patterns. However, similarity exists within whole-scanpaths of each group. Additionally, the study conducted a detailed analysis of local sequences to identify five types of local strategies including UQN, TevQ, UTN, Sys, and Con. Most of these align with previous findings by Pearson et al. (1992), which identified seven strategies employed by readers: ask questions, activate prior knowledge, monitor and repair comprehension, determine important ideas, synthesize information, draw inferences and navigate.
The results indicate that UQN, TevQ, and Sys are the most frequently applied local strategies by attentive readers at the beginning phase of both fact-finding and content understanding tasks. At later phases of both tasks, TevQ remains among top three while other strategies change depending on task type. These findings suggest that attentive readers locate information repeatedly before evaluating its correlation with questions throughout both tasks - a process referred to as comprehensive exploration.
In contrast, for inattentive readers only TevQ exceeds 20% at all three phases of both tasks indicating their preference for organizing questions with paragraphs of text throughout each task to evaluate relevance continually - referred to as iterative content evaluation. The study found no significant differences between groups regarding context analysis implementation.
Overall, these findings provide insights into the differences in local pattern strategies and transitions between attentive and inattentive readers. The practical implications of this research include improving reading system design and identifying reader types for effective training.
