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

Learners’ metacognitive reading strategies support their attempts to draw meaning from texts and to overcome comprehension difficulties. For second language readers, such strategies may

compensate for lack of language proficiency while reading. Taking a sample from a country that ranked last in the PISA 2018 reading assessment, this study aims to investigate potential

discrepancies in how students evaluate the usefulness of specific reading strategies and how these conceptions are associated with related to the students’ reading proficiency. We explored

the association between metacognitive reading strategies with reading proficiency by analysing data from a nationally representative sample of 15-year-old students who participated in the

PISA 2018 (N = 6591). Awareness of different reading strategies was compared using repeated measures ANOVA; relationships with reading proficiency were examined using regression analysis.

Self-reports on metacognitive reading strategies accounted for a significant portion of the variation in Filipino students’ English reading proficiency, after controlling for SES, sex, and

number of books at home. The reading strategies perceived as most useful were not the most strongly associated with reading proficiency, suggesting that students may not be aware of which

reading strategies are helpful in learning to read in English. The results indicate variations in the students’ awareness of which strategies aid in their reading comprehension and point to

the need to better understand how effective reading strategy instruction is taught to and is engaged by Filipino students in their reading classes.

In 2018, Filipino students participated in the OECD’s Programme for International Student Assessment (PISA) for the first time, and the results revealed that Filipino students ranked last

among 79 countries/economies in the domain of reading (Organisation for Economic Co-operation and Development 2019b). Around 80% of the Filipino students who participated in the assessment

failed to meet the minimum reading proficiency level. A few studies (e.g., Bernardo 2023; Bernardo et al. 2021; Haw et al. 2021) have attempted to explore the factors that may be related to

variations in Filipino students reading proficiency in PISA, and in this study, we focus on one factor, metacognitive awareness of reading strategies, which has been associated with reading

proficiency of readers in different parts of the world (Alkhateeb et al. 2021; Pinninti 2016; Sheikh et al. 2019). We inquire into metacognitive strategies considering that the Filipino

students were tested on their reading proficiency in English, the official medium of instruction for most high school subjects in Philippine schools, but a language that was not used at the

homes of 94% of the students. Researchers have noted how reading strategies play a very important role in learning to read in a second language or foreign language (Chen and Chen 2015;

Friesen and Haigh 2018); in particular, metacognitive strategies help students attain better reading proficiency even as they have low language proficiency (Kolić-Vehovec and Bajšanski

2007). In this study, we use data from the PISA 2018 database drawn from a nationally representative sample of 15-year-old Filipino students to explore how awareness of 11 different reading

strategies relate to Filipino students’ reading proficiency in English.

Among the variables that affect the process of reading in the second language, reading strategy use is one of the most studied (Chen and Chen 2015; Friesen and Haigh 2018). Second language

and foreign language readers need to be proficient in the use of reading strategies to be able to understand a text (Hong-Nam and Page 2014; Schiff and Calif 2004; Sheorey and Mokhtari 2001;

Zhang et al. 2014). Reading strategies are actions undertaken by readers to support their comprehension and attempts to draw meaning from texts (Garner 1987; Yoshikawa and Leung 2020).

Reading strategies also involve readers’ perceptions of the reading tasks and how they respond to difficulties encountered while reading (Singhal 2001), as the students’ reading strategies

indicate how they are attempting to overcome their comprehension difficulties (Tercanlioglu 2004). Without the use of the appropriate reading strategies, comprehension difficulties are

likely to arise among second or foreign language readers, and the difficulties might result in detachment from reading activities (Kasemsap and Lee 2015). On the other hand, the use of

appropriate reading strategies can compensate for second language readers’ lack of language proficiency while engaged in reading tasks (Carell 1989; Kolić-Vehovec and Bajšanski 2007).

Reading strategies have been classified by experts as low-level and high-level strategies, and both types need to be activated and coordinated by a reader throughout the reading process

(Grabe Stoller 2013). Low-level strategies refer to the basic strategies for literal interpretation of texts that include skimming, underlining, and rereading. On the other hand, high-level

strategies are those essential to be able to regulate and monitor one’s understanding of a text like interpreting, summarizing, and evaluating the text. In the case of second and foreign

language readers who are not yet skilled in the target language, many are forced to use low-level reading strategies which reduces their employment of the higher-level strategies (Zhang

2001).

The importance of reading strategies does not simply depend on the quantity (range and frequency) of strategies. While studies show that learners with higher reading proficiency tend to use

more strategies (Sheorey and Mokhtari 2001), some non-proficient readers also use many strategies (Hong-Nam and Page 2014). Instead, awareness and appropriate use of reading strategies seems

more important in predicting reading proficiency (Hong-Nam and Page 2014), which has shifted the focus on learners’ metacognitive reading strategies (Fitrisia et al. 2015; Hong-Nam and Page

2014). Metacognitive reading strategies refers to learners’ knowledge of their reading processes, and in particular, the self-controlled techniques they use while monitoring their reading

comprehension (Ahmadi et al. 2013; Mokhtari and Reichard 2002). Research has consistently shown that metacognitive reading strategies differentiates highly proficient readers from less

skilled ones (Mohseni et al. 2020; Pinninti 2016; Sheikh et al. 2019) particularly among second language readers (Meniado 2016; Sheorey and Mokhtari 2001; Singhal 2001; Tavakoli 2014).

Second language learners who know and understand their strengths and weaknesses and who know which controlled learning strategies work for them are better able to overcome the difficulties

they encounter in second language reading tasks.

The PISA 2018 assessment framework also underscored the importance of metacognitive research strategies in the overall assessment of reading proficiency (Organisation for Economic

Co-operation and Development 2019a). The PISA 2018 framework for reading proficiency features a “typology of cognitive processes involved in purposeful reading activities as they unfold in

single or multiple text environments” (Organisation for Economic Co-operation and Development 2019a, p. 36). Three categories of cognitive processes are defined with specific subprocesses

specified in each category: (a) locate information (access and retrieve information within a text, search for and select relevant text), (b) understand (represent literal meaning, integrate

and generate inferences), and (c) evaluate and reflect (assess quality and credibility, reflect on content and form, detect and handle conflict). But in addition to the cognitive processes

associated with text processing, the PISA reading framework also emphasizes the goal-driven and intertextual nature of reading proficiency (McCrudden and Schraw 2007; White et al. 2010). As

such, the framework also highlights the role of the learners’ strategies and motivations that drive the management processes of the reading task (Vidal-Abarca et al. 2010). In this regard,

PISA 2018 also assessed a range of non-cognitive variables associate with the learners’ beliefs, motivations, engagement, practices, and experiences in the reading classroom; one of the

variables they assessed was metacognitive awareness of reading strategies related to two important cognitive processes: (a) understanding and memorizing the text, and (b) summarizing the

text.

Various measures have been developed to measure metacognitive reading strategies including the Metacognitive Awareness of Reading Strategies Inventory (MARSI, Mokhtari and Reichard 2002),

and the Survey of Reading Strategies (SORS, Sheorey and Mokhtari 2001) which was developed for second language learners. In the PISA reading assessment, the measure of metacognitive reading

strategies was measured using two reading scenarios (Organisation for Economic Co-operation and Development 2019a). The first reading scenario involved understanding and remembering a text,

and the second scenario involved summarizing information in a text. For each scenario, the students were asked to evaluate a list of reading strategies and to indicate how effective each was

to fulfill the goal of the scenario. Two indexes of metacognitive awareness of reading strategies were computed from responses to a list of strategies for each scenario.

The main objective of the current study is to explore how Filipino students’ metacognitive reading strategies is related to their reading proficiency in English as a second language. There

have not been many previous studies that inquired into Filipino students’ reading strategies. Mante-Estacio (2016) surveyed students from one Philippine university using the SORS (Sheorey

and Mokhtari 2001) and found overall high use of reading strategies, with problem-solution strategies used more than support and global strategies. Cirocki et al. (2019) surveyed high school

students from a school in a rural region of the Philippines, and also found them preferring to use more problem-solving strategies than the global and support types. But Filipino students’

responses to these quantitative scales do not always converge with qualitative inquiries into the reading strategies. For example, Mante (2009) administered the MARSI (Mokhtari and Reichard

2002) among university students and found that the top reported reading strategies were reading the sentence again, relating an unknown word to something you already know, using context

clues, and reading the text until it is clarified. But when given a reading task and then asked to respond to an open-ended question on what strategies they used in the task, the students

reported a different set of strategies: making previews, identifying relevant and useful learning strategies, relating one’s prior knowledge, and double-checking on comprehension. The study

also showed that the students reported using some ineffective reading strategies and not using some effective strategies (e.g., doing close reading as an initial strategy and throughout the

reading process despite difficulties encountered, and paraphrasing and checking comprehension during reading) (Mante 2009).

The above studies merely describe what strategies tended to be used by Filipino students, but two studies inquired into how the students’ strategies related to their reading comprehension

performance. Ilustre (2011) used the SORS and a researcher-made reading comprehension test with university students. Only problem-solving strategies subscale was positively associated with

text comprehension; text comprehension was negatively correlated with support reading strategies. Mante (2013) administered the MARSI among Filipino high school students and two

comprehension tests after reading four reading materials. Similar to previous studies, the students reported frequent use of all three types of metacognitive reading strategies, problem

solving, global reading strategies, and support reading strategies, and that the last two were strongly correlated with each other. But only support reading strategies predicted the reading

scores of the students’ unaided written recall.

While these few studies seem to suggest the use of metacognitive strategies (particularly, problem solving strategies) based on the quantitative scales, at least one study (Mante 2009)

showed that the responses to the scale did not correspond to the students’ self-reports of actual strategies use after completing a reading task, and there is inconsistency in results

showing which of these strategies relate to better reading performance. We note that these studies all involved small sample sizes of students drawn from one school or university.

As the Filipino students’ performance in the reading domain of PISA 2018 was disappointing, the PISA assessment provides data on reading proficiency and on metacognitive awareness of reading

strategies from a nationally representative sample. The PISA 2018 database provides a good dataset to inquire into Filipino students’ reading strategies and proficiency in ways that

previous Philippine research studies were unable to. More importantly, the inquiry allows for an investigation of a potentially important factor that explains reading proficiency, when that

proficiency is very low. We note that there are previous studies that attempt to identify factors to explain the low reading proficiency (e.g., Bernardo 2023; Bernardo et al. 2021; Haw et

al. 2021), those studies did focus on reading strategies. In the current study, we explored two related questions: (1) What strategies are perceived to be more useful by Filipino high school

students? (2) What strategies are associated with Filipino high school students’ total reading proficiency and with each of the three cognitive subscales of reading proficiency?

We will seek answers to these questions using data from the PISA 2018 survey, and the PISA definition and framework for assessing reading proficiency is adopted. In particular, the PISA 2018

framework for reading proficiency features a “typology of cognitive processes involved in purposeful reading activities as they unfold in single or multiple text environments” (Organisation

for Economic Co-operation and Development 2019a, p. 36). Thus, aside from the overall reading proficiency, we also explore how the perceived usefulness of the strategies related to the

three broad categories of cognitive processes described earlier: locate information, understand, and evaluate and reflect. As regards metacognitive reading strategies, we explore each of the

11 strategies measured in PISA 2018 instead of using the two indexes of metacognitive strategies computed in the database. We believe that using the 11 strategies will provide more detailed

analysis and answers to the main research questions.

Previous preliminary analysis of the Philippines PISA 2018 data (Besa 2019) indicated significant sex differences (i.e., girls outperform the boys), and across different socioeconomic

statuses (socioeconomically advantaged students outperformed socioeconomically disadvantaged ones). In this regard, we decided to include sex and socioeconomic status of the student as

control variables. We included one other home background variable from the PISA survey as another control variable in the analysis; students were asked the number of books in their home, and

this factor has been consistently identified as an important home variable that predicts reading proficiency in many different countries (Chiu and McBride-Chang 2006; Park 2008).

We use data from the Philippine sample in the OECD PISA 2018 database. The complete nationally representative sample comprised 7233 15-year-old Filipino students, who were randomly selected

using a two-stage stratified random selection system. First, stratified sampling was used to select 187 schools from the country’s 17 regions, and then students were randomly sampled from

each school to participate in the PISA assessment (Besa 2019). Because English is the official medium of instruction in most subjects in high school, reading proficiency was assessed in

English, although only 408 (or 5.64%) reported that the main language they used at home was English.

To assess reading proficiency, we referred to the plausible values provided in the PISA 2018 dataset. To clarify, the PISA 2018 assessment does not provide actual achievement scores for each

student; instead, it assesses cognitive learning in the reading domain using ten plausible values that represent ten random values drawn from the posterior distribution of the student’s

scores for reading (Organisation for Economic Co-operation and Development 2019b). In addition to the plausible value for the overall reading proficiency, PISA 2018 also provided plausible

values for three cognitive process subscales of reading: (a) locate information, (b) understand, and (c) evaluate and reflect. For the current study, we used the first plausible for the

overall reading proficiency and for the three cognitive subscales. Previous studies on PISA data have used only one plausible value (e.g., Bernardo et al. 2023; Gomez and Suarez 2020;

Spiezia 2010; Trinidad 2020) based on the assumption that one plausible value is said to provide unbiased estimates of population parameters. Prior to deciding to use only one plausible

value, we examined the distribution and correlations among the ten plausible values for overall reading proficiency and for the three cognitive subscales and we found the means and standard

deviations of the ten distributions are almost identical and are highly correlated with each other. Thus, it is unlikely that an analysis with only one of ten plausible values would lead to

loss of information.

The student questionnaire of PISA included 11 items that referred to different strategies that students use in their reading and writing tasks (see Table 1 for the items); 6 items referred

to strategies to help them understand and memorize the text that they read, and five items referred to strategies to help them write summaries of the text that they read. The students were

asked whether they perceived each strategy as being useful for the different reading tasks indicated, and they answered using a scale from 1 (not useful at all) to 6 (very useful).

Several indexes of the students’ SES were computed in the PISA 2018, and for this study the index of economic, social, and cultural status (henceforth, ESCS) was used. The ESCS was derived

from the students’ report on the availability of 16 household items (e.g., a room of one’s own, air-conditioning unit, and three country-specific items), other possessions in the students’

homes (e.g., cell phones with internet access, computers), education and work status of the students’ parents.

The students were also asked to estimate how many books there were in their home. They were instructed to exclude magazines, newspapers, and their schoolbooks, and they responded by ticking

one of six options: 0–10 books, 11–25, 26–100, 101–200, 201–500, and more than 500 books.

There were numerous missing data across the different variables, and we conducted analysis only on participants with complete data on all the main variables and control variables. The final

sample for the analysis comprised 6591 students (53.86% were girls). To answer the first research question, mean scores of the 11 metacognitive reading strategies were computed and then

analyzed using a completely repeated analysis of variance. The analysis of the main effect of type of learning strategy was followed by a post hoc multiple comparison of means using

Bonferroni test. To answer the second set of questions, four hierarchical regression analyses were conducted. In each analysis, the reading proficiency score or subscale score was first

regressed to three control variables (sex, ESCS, and number of books at home). In the second step of each hierarchical regression, the 11 learning strategies were added to the model. At each

step, the overall model and change in R2 of the model was assessed.

The first research question of the study investigates which learning strategies for reading are perceived to be most useful by Filipino 15-year-old learners. The results are summarized in

Table 1, which lists the different learning strategies according to their perceived usefulness. As the standard deviations indicate, the students varied in their perceptions of the different

strategies, and the students’ responses spanned the full range of options for each strategy. The repeated measures ANOVA (with learning strategy as within group factor) indicated a

statistically significant difference across the means, F(10, 6649) = 56463.08, p