Yilin Yu’s research while affiliated with Education University of Hong Kong and other places

Guided by the family investment model, family process model, and componential model of reading, this study examined underlying mechanisms linking family socioeconomic status (SES) to reading comprehension. Participants were 682 Chinese third graders (Wave1, M age = 9.31 years, 338 girls) randomly recruited, and they were followed up after 1 year. Structural equation modeling revealed that higher parents’ education was related to more books in the home and higher levels of positive family relationship, which was associated with better language skills directly or indirectly through stronger reading interest—better language skills ultimately linked to better reading comprehension. No significant indirect pathways were found from family income to reading comprehension. The mediating pathways were similar for children with different parental migration statuses.

How brains of children of different ages extract and encode relational patterns contained in orthographic input remains elusive. We investigated this question by measuring event-related potentials (ERPs) of 99 Grades 1-3 children while performing an artificial orthography statistical learning task that comprised logographic components embedded in characters with high (100%), moderate (80%), and low (60%) positional consistency. The behavioral results indicated that across grades, children more accurately recognized the characters with high than with low consistency. The neurophysiological results revealed that in each grade, the amplitude of some ERP components differed across the positional consistency, with a larger P1 effect in the high consistency condition and a larger N170 and left-lateralized P300 effect in the low consistency condition The grade effect was significant, i.e., the smaller N170 amplitude in Grade 3 than Grade 1 and the larger P300 amplitude in Grade 1 than either Grade 2 or 3. These findings suggest the dynamic nature of statistical learning by showing that at least two mechanisms, i.e., neural adaptation associated with N170, and attention and working memory related to P1 and P300 regulate different types of structural input, and that children’s abilities to prioritize these mechanisms vary with context and age.