Byungmin Lee

While it has been observed that second language learners selectively read information to answer comprehension questions, the amount of textual information required to correctly answer a question remains unclear. This study sought to identify the effects of this strategic reading on reading comprehension performance and explore how proficiency modulates its effects. A total of 279 Korean high school students across four proficiency levels responded to 18 multiple-choice fill-in-the-blank items from the English section of College Scholastic Ability Test (CSAT). Reading passages were modified into four versions with varying amounts of textual information to enforce selective, strategic reading. A multilevel logistic regression analysis revealed that learners with the lowest proficiency performed poorest in the complete passage condition, while those with the highest proficiency performed best, suggesting that the sweet spot of reading for fill-in-the-blank items varies depending on proficiency. The findings are discussed in the context of second language instruction and assessment.

This study classifies profiles of readers by combining good readers, poor comprehenders, dyslexia, and generally poor readers based on the Simple View of Reading (SVR) model that individual readers have different profiles of reading skills. A total of 501 ninth-grade Chinese middle school students participated in the study. They engaged in five tests pertaining to English reading-related skills (decoding, vocabulary knowledge, syntactic knowledge, morphological knowledge, and reading comprehension) and completed a survey questionnaire to provide information on their SES and English learning background. Results showed six profiles (top readers with highest decoding and language comprehension, good readers with good decoding and language comprehension, average readers with average decoding and language comprehension, poor comprehenders with average decoding but low language comprehension, poor decoders with specific decoding difficulties, poor readers with mixed reading difficulties) were drawn from latent profile analysis. Multinomial logistic regression analysis revealed that each profile of reading skill was influenced by the education level of parents, number of English books, English private tutoring lessons, and English advanced prior learning.