Extended text processing in bilinguals and monolinguals

Extended text processing in bilinguals and monolinguals

First Author: Guoqin Ding -- Utah State University
Additional authors/chairs: 
Kathleen Mohr; Ronald Gillam; Stephanie Juth
Keywords: Text Comprehension, Eye-tracking, Processing costs, children, Bilingualism
Abstract / Summary: 

Purpose: Eye tracking has become the gold standard for investigating sentence processing, however, extended-text reading has not been well researched (Hyönä & Kaakinen, 2019). Moreover, temporal measures are widely used, but spatial measurements have been limited. Extended-text reading requires readers to integrate the meaning of successive sentences to construct a representation of the whole text, which may be reflected in both temporal and spatial measurements with differential results. To investigate global text processing, this study examined text processing patterns of bilinguals and monolinguals using three eye-tracking measures with both temporal and spatial aspects: average fixation duration, average regression amplitude, and regression probability.
Method: Sixteen English monolingual and 16 Chinese-English bilingual children matched by age (9-14 years) were assigned to an online reading task with four passages (featuring predominantly canonical or predominantly noncanonical sentences) monitored with SMI eye-tracker. The three eye-tracking measures were calculated from raw data.
Results: Multiple regression models show that bilinguals presented longer average fixation duration, greater average regression amplitude, and higher regression probability for canonical and noncanonical texts. Compared to canonical texts, noncanonical texts evidenced higher regression probability in both bilingual and monolingual groups.
Conclusions: Different eye-tracking patterns between groups typically suggest different cognitive processes in reading. Significant differences across three eye-tracking measures show that bilinguals utilized greater cognitive load in text processing which aligns with prior studies. Increased rereading rate in noncanonical texts indicated that compared to canonical sentences, noncanonical sentences were more difficult to process and required more effort in meaning integration.