Metacognitive reading strategies and its relationship with Filipino high school students’ reading proficiency: insights from the PISA 2018 data

The study investigates how Filipino 15-year-old students’ metacognitive awareness of reading strategies relates to their reading proficiency in English, using the PISA 2018 dataset. The Philippines ranked last in PISA 2018 reading, with around 80% of students below the minimum proficiency level. Given that English is the medium of instruction in high school but not the home language for 94% of students, metacognitive strategies may compensate for lower language proficiency. The research focuses on: (1) which reading strategies students perceive as most useful, and (2) which strategies are associated with overall reading proficiency and with three cognitive subscales (locate information, understand, evaluate and reflect). The study aims to clarify possible mismatches between perceived usefulness of strategies and their actual association with proficiency, informing instruction and policy for second-language reading in the Philippines.

Research on second/foreign language reading emphasizes the role of strategy use in comprehension, distinguishing low-level (e.g., skimming, underlining, rereading) from high-level strategies (e.g., interpreting, summarizing, evaluating). While higher-proficiency readers tend to use more strategies, awareness and appropriate use (metacognitive strategies) better predict proficiency. Prior Philippine studies (largely small-sample) found frequent use of problem-solving strategies but mixed links to comprehension. Discrepancies were noted between self-reported strategy use on inventories and strategies reported after tasks. The PISA 2018 reading framework defines cognitive processes (locate information; understand; evaluate and reflect) and highlights task-management strategies and motivation; it includes measures of metacognitive awareness for understanding/memorizing and summarizing. Using PISA’s nationally representative data enables a more robust examination of Filipino students’ metacognitive strategy awareness and its relation to reading outcomes.

Design and data source: Secondary analysis of the Philippines sample from OECD PISA 2018. Participants: Nationally representative sample initially N=7233 15-year-old students from 187 schools across 17 regions; analyses restricted to cases with complete data on all variables, yielding N=6591 (53.86% girls). Only 5.64% reported English as the main home language. Measures: Reading proficiency assessed via PISA plausible values; the first plausible value (PV1) was used for overall reading proficiency and for three subscales: locate information, understand, and evaluate and reflect. Distributions and intercorrelations across the 10 plausible values were examined and found highly similar, supporting the use of PV1. Metacognitive reading strategies: 11 items in two scenarios—six for understanding/memorizing and five for summarizing—rated on a 1 (not useful at all) to 6 (very useful) scale. Control variables: sex (1=female, 2=male), socioeconomic status via ESCS index (derived from parental education/occupation and household possessions), and number of books at home (0–10; 11–25; 26–100; 101–200; 201–500; >500). Analytic strategy: Addressed missingness by complete-case analysis. For RQ1 (perceived usefulness), computed means of all 11 strategies and conducted a repeated measures ANOVA (strategy as within-subject factor) with Bonferroni-adjusted post hoc comparisons. For RQ2 (associations with proficiency), ran four hierarchical regressions: Model 1 included controls (sex, ESCS, books at home); Model 2 added the 11 strategy ratings. Outcomes were overall reading proficiency and each subscale. Reported model R2, ΔR2, F-tests, standardized betas, and 95% CIs.

- Descriptive reading proficiency (using sampling weights): Overall mean = 339.47 (SD=79.54); subscales—Locate information = 342.94 (SD=92.55), Understand = 334.42 (SD=82.16), Evaluate and reflect = 333.57 (SD=88.19). - Perceived usefulness of strategies varied widely; repeated measures ANOVA indicated large differences across strategies: F(10, 6649)=56463.08, p<0.001, partial η²=0.90. - Highest perceived usefulness (means; 1–6 scale): (1) Read through, underline important sentences, write in own words (M=4.09); (2) Underline important parts (M=4.08); (3) Check that important facts are reflected in summary (M=4.06). Lowest: (10) Copy as many sentences as possible (M=3.32); (11) Read text aloud to another person (M=3.15). - Overall proficiency regression (standardized betas, final model): Model 1 (controls) R²=0.22; adding strategies increased R² to 0.40 (ΔR²=0.18, ΔF significant, p<0.001). Controls: sex (male) β=-0.09***; ESCS β=0.27***; books β=0.08***. Strategies significantly positively associated: (1) β=0.05**; (2) β=0.04**; (3) β=0.11***; (4) β=0.11***; (6) β=0.08***; (7) β=0.08***; (9) β=0.08***. Non-significant: (5) β=0.01; (8) β=0.03. Negative associations: (10) β=-0.17***; (11) β=-0.19***. - Subscale regressions (final models): R²: Locate=0.37; Understand=0.38; Evaluate=0.33. Added strategies explained ΔR²=0.16 (Locate), 0.18 (Understand), 0.14 (Evaluate), all p<0.001. Patterns largely mirrored overall results. Notable exceptions: Underlining (2) was positively associated with Locate and Understand but not Evaluate; rereading many times (5) was associated with Locate only. Negative associations for copying (10) and reading aloud (11) were consistent and strong across subscales. - Misalignment observed: strategies perceived as most useful (e.g., underlining) were not the strongest predictors; more effective strategies for comprehension and evaluation (e.g., summarizing in own words; checking key facts in summaries) were not rated uniformly as useful by many students.

Findings show metacognitive reading strategies account for substantial variance in Filipino students’ English reading proficiency beyond SES, sex, and home books, indicating their importance even in a low-performing national context. However, there is a notable misalignment between students’ perceptions and the strategies most predictive of proficiency. Underlining—rated highly useful—had weak associations (and none for evaluate/reflect), whereas higher-order strategies such as summarizing in one’s own words and checking inclusion of key facts showed the strongest positive associations across overall proficiency and subscales. Conversely, lower-level strategies emphasizing surface encoding (copying sentences; reading aloud) were rated useful by a sizable minority yet were strongly negatively associated with proficiency. These patterns suggest gaps in students’ metacognitive awareness and possibly reflect instructional emphases. Curriculum analyses indicate task-management skills (including strategies) in the Philippine curriculum may be taught discretely and not well-aligned with PISA’s integrated framework, potentially limiting students’ understanding of when and why to apply higher-order strategies. It is also plausible that many students have not reached a threshold of second-language reading proficiency at which metacognitive strategies yield optimal benefits, and cross-linguistic reading skills in the first language likely interact with English reading outcomes. The consistent effects of SES, sex (girls outperform boys), and books at home align with international literature, underscoring broader equity and literacy-environment factors. Overall, the results argue for integrated, explicit instruction and modeling of effective strategies, aligning classroom goals with higher-order comprehension processes measured by PISA and supporting students’ strategic choice and monitoring.

Using nationally representative PISA 2018 data, the study extends limited Philippine research by demonstrating that metacognitive reading strategies significantly relate to English reading proficiency among Filipino high school students. A key contribution is identifying a mismatch between perceived usefulness and actual effectiveness: students tend to overvalue low-level strategies (e.g., underlining) and undervalue higher-order strategies (e.g., summarizing in own words; verifying key facts in summaries) that more strongly predict proficiency. These insights point to needs for integrated strategy instruction across grade levels, alignment with cognitive processes emphasized in PISA, and explicit modeling of how strategies support understanding, evaluation, and reflection. Future research should broaden the set of cognitive, non-cognitive, home, classroom, and first-language reading factors; examine how teachers teach and students enact strategies; and consider proficiency thresholds to sequence strategy instruction appropriately in second-language contexts.

- Focused primarily on metacognitive strategies with limited control variables (sex, ESCS, books), precluding analysis of a wider set of student-, family-, classroom-, and school-level factors. - Strategy measurement restricted to PISA’s 11 items; potentially important strategies specific to Filipino readers or English L2 contexts were not assessed. - Cross-sectional, correlational analyses using self-reported perceptions of strategy usefulness cannot establish causality; students’ ratings may reflect instruction, peer influence, or differing conceptions of reading proficiency. - Use of one plausible value (PV1) may omit some uncertainty inherent in the plausible value methodology, although distributions and correlations across PVs were highly similar. - Complete-case analysis may introduce bias if data are not missing completely at random.