Evaluating the predictive power of metacognitive reading strategies across diverse educational contexts

Abstract

Abstract Background Metacognitive reading strategies, such as monitoring and regulating one’s reading processes, are critical for enhancing reading comprehension and achievement. However, their effectiveness can vary across different educational contexts and student populations, necessitating flexible instructional approaches. This study examines the predictive utility of metacognitive reading strategies on metacognitive reading skills and reading achievement scores using data from the PISA 2018 assessment. Specifically, it compares the effectiveness of these strategies across four distinct educational contexts: the United States, OECD countries, non-OECD countries that participated in the assessment (Partner Countries), and the census data (Full Sample). Methods A multi-stage analytical approach was used, combining Random Forest models for feature importance analysis with survey-weighted linear regression to assess the relationships of reading strategies on metacognitive skills and reading achievement. The analysis utilized data from the PISA 2018 Student Questionnaire, which included responses from 612,004 students across 79 countries. The study focused on three key metacognitive reading skills: Understanding and Remembering, Summarizing, and Assessing Credibility, as well as reading achievement scores based on ten plausible values. Results The Random Forest analysis identified several metacognitive reading strategies as consistent predictors of metacognitive reading skills, with some variations in relative importance by educational setting. Notable metacognitive reading strategies included underlining important parts of the text, carefully checking the representation of important facts in summaries, and verifying information on websites. These strategies showed a positive correlation with higher metacognitive skills and overall reaching achievement. The direction of the relationships between metacognitive reading strategies, metacognitive skills, and reading achievement scores was consistent across contexts, indicating a modest yet significant association between strategy use and both outcomes. Conclusions This study enhances our understanding of how metacognitive reading strategies influence reading skills and achievement in diverse educational settings. The findings highlight the potential benefits of reinforcing specific strategies to support students' reading comprehension. Given the consistent positive relationships between strategy use, metacognitive skills, and reading achievement, targeted interventions to promote these strategies could effectively improve reading comprehension across diverse educational contexts.

Keywords

• Metacognitive reading strategies
• Metacognitive skills
• Reading achievement
• PISA 2018
• Random Forest
• Survey-weight linear regression