Being learners with mental resilience as outcomes of metacognitive strategies in an academic context

Abstract

The primary objective of this paper is twofold; firstly, to analyse the relationship between metacognitive strategies and learning performance. Secondly, a new mediator is proposed, namely digital literacy. Mental resilience is an omnitemporal skill that enables individuals to gain resilience thinking to successfully adapt to life tasks. Although Researchers accentuate that metacognitive competency is necessary and crucial in enhancing mental resilience, research on inculcating resilience thinking through metacognitive skills is still undeveloped and warrants the urgent attention of the research community. Data was gathered from a cross-sectional study of 563 university students. To accomplish the goals of the research, data were analysed empirically utilising the variance-based partial least squares structural equation modelling (PLS-SEM) technique. The results revealed that learning performance was not significantly directly impacted by metacognition. This may be due to the fact that metacognitive strategies are used to help students to increase academic performance as opposed to learning capacity. However, through the mediation analysis, it was discovered that digital literacy was proven to be an effective mediator to promote learning performance. These new insights can assist academics to reflect and evaluate current practices to cultivate metacognitive practices in the classroom without having to alter the subject curriculum.

Keywords:

• Metacognitive strategies
• learning performance
• digital literacy
• mental resilience

1. Introduction

Although according to the fourth UNESCO Sustainable Development Goal target (SDG 4), youths are to have a substantial amount of relevant skills to promote sustainable development (SDG, Target 4.4, 4.7), mental health implications amongst youths have become a global topic. The United Nations defines youth as persons between the ages of 15 and 24. In comparison to the general population, youths, particularly university or college students, are increasingly regarded as a vulnerable demographic, suffering from higher levels of anxiety, societal pressure, and depression, among other things (Browning et al., 2021). Furthermore, as the world is increasingly threatened by the consequences of the COVID-19 epidemic, the load on the mental health of this vulnerable demographic is compounded when the character of their life and educational experience changes dramatically (Browning et al., 2021). According to recent meta-analyses, the prevalence of stress, anxiety, and depression in the general population as a result of the pandemic is over 30% (Luo et al., 2020; Salari et al., 2020). According to one study, 71% of youths reported tension and anxiety as a result of the COVID-19 outbreak (Son et al., 2020). Furthermore, a leading newspaper in Malaysia revealed that approximately 30% of youths in Malaysia are facing increased pressure and varying degrees of anxiety due to the pandemic (Sheng, 2021).

Mental resilience is an important component of lifelong learning because it allows people to develop resilient thinking in order to successfully adjust to life tasks in the face of social disadvantage or extreme circumstances. A lifelong learner is someone who continues to gain new skills and talents after completing their official schooling. Individuals who are resilient are more likely to embrace a lifelong learning attitude, which allows them to intentionally choose to disrupt themselves in order to navigate toward becoming the next best version of themselves. Although mental resilience is a skill that can be taught, unfortunately, there is a worrying decline of resilience thinking among youths (Gómez- Molinero et al., 2018; Robbins et al., 2018). Possible reasons why youths are still struggling with resilience thinking skills is because they have not been taught how to train to develop daily habits that can build mental muscle to gain mental strength.

Metacognitive competency is necessary and crucial in enhancing the mental resilience among young adults to foster a deeper level of thinking that includes the ability of an individual to think about their thinking; their understanding, adaption, change, control, and thought processes to become future resilient thinkers in their lives and careers. More precisely, metacognition is awareness and understanding of one’s thought process. Metacognition competence is a skill that can be learned and mastered. Several past works of literature have revealed that metacognitive skills can help individuals with the resilience skills needed for to improve learning capacity for lifelong learning (Capobianco et al., 2020; Philipp et al., 2020). Metacognitive practices can transform individuals to be active thinkers, self-regulate their behaviors, and make informed decisions on the use of strategies. Nevertheless, youths are not taught how to self-regulate their thinking processes to acquire higher-order thinking skills. Often, youths are being told what to think and not taught how to think. Nevertheless, several findings from the literature revealed youths experiencing deficits in their usage of metacognitive practices and skills (Anthonysamy et al, 2020; Anthonysamy, 2021; Davies, 2019). That could explain why youths are still lacking in resilience thinking skills.

Only a few studies have been undertaken in the context of metacognition research in Asia (Li et al., 2018), compared to Europe and America (Richardson et al., 2012). This demonstrates a growing trend and demand for study on rudimentary metacognitive skills. With mental health at the top of governments’ agendas around the world, research on metacognitive competency is critical, given its relevance and importance in ensuring that teenagers have the tools they need to think resiliently. The objective of changing nations into resilient and educated societies may be jeopardised due to a lack of resilience thinking among youths. Such a trend necessitates investigation into the issue in order to gain a deeper grasp of it.

This purpose of this study is to first examine the connection between metacognitive strategies and learning performance and secondly, to evaluate the connection using a mediator. In this study, a brand-new mediator, namely digital literacy is put forth to improve student’s learning performance. The paper is structured as follows: Section 2 provides a background on previous research on metacognitive strategies, digital literacy, and student learning performance, which subsequently frames the concepts developed in this study. Section 3 outlines the method used in the empirical research. The results of data analysis using multi-level structural equation modelling and discussion are reported in Sections 4 and 5 respectively. The last section presents the research conclusion, limitations, and suggestions for future work.

2. Review of literature

2.1. Underlying theories of study

The research framework for this study is based on two theories: metacognitive theory (Flavell, 1979) and resilience theory (Flavell, 1979) (Rutter, 2006). The metacognitive theory is a knowledge theory concerned with how humans can actively monitor and control their on-thought processes (Flavell, 1979). Some people, according to Flavell, the theory’s creator, are more capable of controlling their brains than others. Individuals with this talent are encouraged to increase self-awareness in order to better understand how they function and learn. Resilience theory, on the other hand, views resilience as a higher-order construct that draws on other constructs that people utilise to adapt and bounce back when confronted with adversity such as dissatisfaction and misfortune. Individuals with resilience gain the skills to deal with unavoidable uncertainty, allowing them to learn and grow in any environment. It also helps in taking risks, trying new things, and dealing with any difficult situations that life may offer at you. Mentally robust people make fewer excuses, procrastinate less, and look after themselves.

2.2. The role of metacognitive skills in enhancing students’ learning performance

John Flavell (1979) used the word “metacognition” to characterise the awareness and comprehension of one’s mental processes. The goal of metacognitive methods is to figure out what, how, and when to apply a certain strategy for a specific task. People use metacognitive approaches to organise, monitor, and manage their cognitive processes in order to attain a goal. The majority of metacognition models include three types of strategies: planning, monitoring, and regulating, which are utilised at various phases of the learning process (Zimmerman & Martinez-Pons, 1986). There are two main actions in metacognition planning. To begin, students must understand the learning goal they must attain, as well as how they will approach the activities and tactics they will employ. Skimming through online information prior to completing a task analysis of the problem (Pintrich, 1999), planning the sequence, scheduling, and completion of activities oriented toward the learning goals, and planning the sequence, scheduling, and completion of activities oriented toward the learning goals are all examples of planning activities (Zimmerman & Martinez-Pons, 1986). Because students may measure their comprehension against particular self-set goals, monitoring the learning process can lead to a positive learning result. Checking against defined goals, tracking progress while conducting an online activity, maximising attention and focus, and self-testing through questions about the online material are all examples of monitoring tactics (Pintrich, 1999). Learners might ask themselves questions like, “How am I doing with this?” or “What else do I need to know to complete this task?” or “Are these learning tactics helping me learn effectively?” Students’ self-regulation is aided by accurate monitoring, which provides feedback on what they already know and where they should focus their efforts. The process of keeping watch of and directing one’s mental processes is known as metacognitive regulation. Techniques of regulation and monitoring are intertwined, but self-reflection is essential. Students can employ regulatory tactics to re-align their academic behaviour in order to achieve their objectives (Zimmerman & Martinez-Pons, 1986). Recognizing that a strategy for solving an issue isn’t working and trying a different approach, as well as fine-tuning and continually refining accomplished activities, are examples of metacognitive regulation. Self-check questions should allow students to reflect on their online actions, since this will help them stay on track with their studies.

Students’ success in a digital learning environment may be evaluated using both objective and subjective criteria (Vo et al., 2017; Yang et al., 2016). Although grades, marks, and attendance are widely used as objective measurements of achievement, they may not accurately reflect excellent learning (Biggs & Tang, 2007). Students’ behaviour is typically perceived as a short-term fluctuation in objective indicators, which can give the appearance of competence (Soderstrom & Bjork, 2015). Subjective metrics, on the other hand, are an ideal approach to evaluate students’ overall attitude toward learning since they lead to students accumulating information that can be utilised in real-life situations (Biggs & Tang, 2007). Because it focuses on personal growth, subjective measures are more significant in the context of lifelong learning. Metacognitive approaches are more favourable to boosting learning performance, according to prior studies (Cho & Heron, 2015; Goradia & Bugarcic, 2017) (Cho & Heron, 2015; Goradia & Bugarcic, 2017). Similarly, Goda et al. (2015) discovered that students with metacognitive abilities were better at managing their time and submitting assignments on time, resulting in increased learning performance. In the context of digital learning, however, some previous research argued that metacognitive strategies weaken control (Ackerman & Lauterman, 2012; Lauterman & Ackerman, 2014).

2.3. Metacognitive strategies for digital literacy enhancement

According to modern writers, digital literacy is described as “the attitude, capacity, and awareness to successfully access, recognise, manage, analyse, build new information, and interact with one another in digital contexts using digital technology” (Ng, 2012; Perera et al., 2016). Students with excellent metacognitive skills may incorporate a variety of components of attention, skills, working memory, and inhibitory control into their learning (Lee et al., 2015). Multiple types of digital literacies envisioned as undergraduate abilities are successfully emphasised by the digital literacy framework (Ng, 2012), which encompasses technical, cognitive, and socio-emotional literacy. Because they reflect one’s thinking and academic behaviour, two sections of metacognition methods, planning and monitoring, have been identified as crucial features of digital literacy (Greene et al., 2014, 2018) (Pintrich, 1999). According to Van Laer and Elen (2017), the adoption of metacognitive skills is useful since it may help students overcome their difficulties with digital literacy. Furthermore, in their study (Zylka et al., 2015), Zylka and colleagues observed a positive relationship between metacognitive procedures and digital literacy, showing that metacognitive processes may help in the acquisition of higher levels of digital literacy (Van Loon, 2001). Similarly, research shows that students who employ metacognitive abilities can manage different sources online, which helps to improve digital literacy, which is often poor, even among high-performing students (Cho et al., 2017).

2.4. The effect of digital literacy in enhancing learning performance

Digital literacy, according to Tang and Chaw (2016), is a need for students to learn well in digital contexts. It is simpler for students to participate in the learning process when they have a high level of digital literacy, which offers them a more positive attitude on their educational experience. Because students with a high degree of digital literacy are familiar with the interfaces, access choices, language, how to behave online, and the norms of new technologies, digital learning will be less cognitively taxing. In a digital learning environment, those with strong digital literacy abilities may have fewer difficulties completing tasks online. Furthermore, when learning is done using digital tools and technology, such as in a digital learning environment, greater emphasis must be paid to the function of digital literacy in order to facilitate a smoother adaption and experience in digital learning (Shopova, 2014). Prior et al. (2016) discovered that students have different levels of digital literacy, and that one’s level of digital literacy has a significant influence on and predicts one’s learning achievement (Greene et al., 2014; Scholastica et al., 2016; Wichadee, 2018). However, opposing evidence has been revealed, indicating that digital literacy does not appear to improve students’ learning outcomes (Pagani et al., 2016). As a result, the findings varied, prompting more study into the relationship between digital literacy and learning performance.

In light of the aforementioned, the following hypotheses were constructed to examine the relationship between metacognitive strategies, digital literacy, and learning performance. Hypothesis 1, Hypothesis 2 and Hypothesis 3 will be examining the relationship between metacognitive strategies and digital literacy. Hypothesis 4 will be testing the relationship between digital literacy and learning performance

Hypothesis 1

(H1): Metacognitive strategies positively influences university students’ digital literacy improvement in digital learning.

Hypothesis 2

(H2): Digital literacy positively predicts the learning performance of university students in digital learning.

Hypothesis 3

(H3): Metacognitive strategies positively predicts the learning performance of university students in digital learning.

Hypothesis 4

(H4): Students’ level of digital literacy mediates the relationship between metacognitive strategies and their learning performance.

Following a thorough examination of the literature, Figure 1 depicts the research model for this study. This paradigm emphasises how metacognitive strategies relate to learning performance. This model would look at the significance of individual differences as well as the part that context plays in determining how metacognitive strategies and performance interact. The concept also includes digital literacy, which serves as a bridge between metacognitive strategies and effective learning. This model explores the significance of acquiring digital literacy abilities to successfully employ metacognitive techniques in a digital world and enhance learning capacities to build mental resilience.

Figure 1. Research Model.

3. Method

3.1. Participants and design of the study

For this study, which employed a descriptive design survey technique, university students were purposefully recruited from specified private higher education institutions (HEIs) in Malaysia. The students that took part in this study were in their first or second year of university of any gender and age group between 19 to 24 years of age. For a number of factors, the choice was made. According to Shopova (2014), first- and second-year students have substantial difficulty in the learning process owing to a lack of resources to assist them in smoothly navigating through the process. Furthermore, research has shown that a student’s first year of university experience is crucial to their effective adaption to the higher education environment, and that it has an impact on their engagement throughout their degree (Kift, 2015). Malaysia is divided into six areas, with the central region housing 53 private higher education institutions (68 percent) (Malaysia Qualification Registration, 2020. The search was narrowed even more by selecting a bachelor’s degree programme at an institution with a digital learning environment. Following the screening, a total of seven universities that provide students with digital learning opportunities were discovered. G-Power software was used to compute the sample size, which had a 0.05 level of significance at a 95% confidence level. GPower software has recommended a minimum of 129 samples based on the framework’s complexity. Since the research instrument was a questionnaire, a web-based mechanism was used to disseminate the final questionnaire to the target sample. The data collection for this study commenced after obtaining approval from the university’s ethics committee. After the final questionnaire was distributed, 626 responses were obtained. Out of that, 63 questionnaires were rejected because of outliers and respondents not meeting the criteria. Normality and outlier analysis were performed to refine the data further. The balance data obtained after cleaning was 563.

3.2. Instrument development and constructs

Several prior studies were examined to verify that the self-report instrument provided a comprehensive range of measures. To match the digital learning environment, all of the measures for each component were modified from previously verified equipment. The Online Self-regulated Learning Questionnaire (OSLQ) (Barnard et al., 2009) and Motivated Strategies for Learning Questionnaire (MSLQ) were adapted because they have been widely used to assess students’ self-regulatory behaviour (Pintrich et al., 1993) in an online environment (Zhu, Au, & Yates, 2016). Digital literacy metrics were supplied by Van Laer and Elen (2017) and Ng (2012). Learning performance was measured using learning outcomes, student interactive involvement, and student satisfaction. Student interactive involvement was taken from Eom et al. (2006), Trowler (2010), Yang et al. (2016), and student satisfaction was derived from Sun et al. (2008), Wu et al. (2010), and Eom et al. (2006), and learning outcomes were derived from Rovai et al. (2009). Appendix A presents the measurement items used in this study in order to achieve both research objectives.

3.3. Statistical analysis

In this study, data was analysed using SPSS (Statistical Package for the Social Sciences) version 25 and SmartPLS 3.0. SPSS was used to clean the data and run descriptive statistics on it before it was analysed. Structural Equation Modelling (SEM) is a multivariate statistical method that can evaluate a series of dependent connections at the same time and was used to explain the link between many variables (Hair et al., 2017). The constructs in this investigation were formative higher order structures. As a result, a two-stage strategy was applied, with the repeated indicator technique and a causal model based on latent variable scores.

4. Results

There were 167 female respondents and 396 male respondents, accounting for 29.7% and 70.3%, respectively. Respondents’ age categories are 19–24 years. The majority of respondents were freshmen.

4.1. Measurement model validation

It is necessary to analyse a measurement model that outlines the linkages between the constructs and the indicators in order to ensure that it is valid. As part of the measurement model evaluation, reflective and formative measurement models are evaluated. The first-order measuring methodology employed in this work includes internal consistency reliability, indicator reliability, convergent validity, and discriminant validity (Hair et al., 2017). On the other hand, the formative measurement model was assessed for convergent validity using the repeated indicator approach, indicator collinearity, statistical significance, and relevance of the indicator weights (Gaskin et al., 2018; Hair et al., 2017).

4.1.1. First-order measurement model

Internal Consistency Reliability, Indicator Reliability, Convergent Validity

Internal consistency is a measure of reliability. Reliability measures the consistency of the device being played. Reliability was measured using compound reliability (CR) and extracted mean variance (AVE). This study shows that the AVE score of all constructs is above 0.5 and the CR is 0.7–0.9, indicating that the internal consistency is reliable, as shown in Table 1. Higher values of indicate that latent variables better explain the variance of each indicator and thus indicate the achievement of internal consistency reliability.

Table 1. Item loading, CR, AVE, and VIF values for measurement model

Construct	Item	Loading	AVE	CR	VIF
Metacognitive Strategies	I analyse an academic task once before attempting it (E.g. understand the expected outcome, task objectives, etc).	0.863	0.750	0.857	1.334
	I set some standards for myself when accomplishing an academic task.	0.870			1.334
	When doing an academic task, I question myself to help me understand the task better.	0.920	0.658	0.790	1.135
	When doing an academic task, I check my progress with my instructor to ensure I am on the right track.	0.685			1.135
	If I get confused during an academic task, I use other approaches to learn (E.g. Youtube, Google, peer help, etc).	0.782	0.676	0.807	1.145
	I check over my academic task to make sure I did everything that was required before making the final submission.	0.861			1.145
Digital Literacy	I know most of the features of my digital device (E.g camera, settings, etc.)	0.808	0.699	0.903	2.000
	I know how to use my digital devices for everyday tasks (E.g. Emails, installation of apps, social media, etc).	0.863			2.398
	I communicate information or ideas effectively to friends using a variety of online tools.	0.869			2.302
	I collaborate online with friends to exchange ideas or generate meaning to complete a task.	0.801			1.836
	I search for information from a variety of online sources.	0.784	0.523	0.864	2.092
	I know how to select useful information to complete an academic task.	0.809			2.199
	I can interpret online information that is text-based and non-text-based (E. images, videos, maps, etc).	0.793			1.834
	I go online to clarify facts before I complete an academic task.	0.770			1.771
	I can critically judge the suitability of online information before copying/retrieving any data.	0.698			1.577
	I use an online platform (E.g blog, social media) to suggest new ideas for discussion to complete an academic task.	0.402			1.205
	I use the Internet responsibly for communicating, socialising and learning.	0.716	0.522	0.845	1.509
	I respect other cultures when I am communicating online.	0.743			1.573
	I reveal little of my personal information online to safeguard myself.	0.647			1.336
	I understand the potential risks that exist when I am online.	0.790			1.880
	I understand copyright rules when I copy or reproduce content from online sources.	0.709			1.450
Learning Performance	I can apply the content learned in the academic task for other courses/related tasks.	0.730	0.556	0.882	1.564
	I can intelligently critic the academic task I completed.	0.706			1.518
	I become more independent after completing an academic task.	0.771			1.914
	I become a better thinker after completing an academic task.	0.770			1.901
	I can use my hands-on skills during the accomplishment of an academic task for other related tasks.	0.742			1.794
	I can demonstrate to others the hands-on skills learned from an academic task.	0.750			1.818
	I acquire useful knowledge through interacting with other learners when accomplishing an academic task.	0.708	0.579	0.873	1.622
	I acquire useful knowledge by interacting with my instructor when accomplishing an academic task.	0.767			1.771
	I engage with Web 2.0 tools to enhance my skills.	0.787			1.721
	I am fully engaged when I am given an academic task.	0.785			1.812
	I am interested to complete an academic task as I enjoy learning.	0.755			1.683
	I am satisfied with the academic tasks I am given by my instructor.	0.811	0.654	0.883	1.731
	Accomplishing academic tasks help me improve my subject matter better.	0.772			1.732
	I acquired more knowledge from accomplishing academic tasks.	0.847			2.161
	I would gladly take up another academic task.	0.803			1.844

4.1.1.1. Discriminant validity

The reflective constructs must be empirically distinct from other constructs in order to have discriminant validity (Hair et al., 2017). Hair et al. (2017) proposed the Heterotrait-Monotrait Ratio (HTMT), which verifies that each construct is actually distinct from the others. The HTMT values collected and given in Table 2 for this investigation were all below 0.85 and 0.90, showing that discriminant validity was not an issue.

Table 2. HTMT ratios

	Digital Literacy	Metacognitive Strategies	Learning Performance
Digital Literacy
Metacognitive Strategies	0.588
Learning Performance	0.582	0.542

4.1.2. Second-order measurement model

4.1.2.1. Repeated indicator approach

A causal model employing an iterative index technique and latent variable values was used due to the high-order formation structure of digital literacy and learning outcomes. The precursor can forecast higher-order formative measures because to its two-tiered technique, which avoids the “flooding effect” of recurrent signs (Gaskin et al., 2018). The latent structure of the iterative index technique comprises all of the major components inside that structure. Individual significance and mediation between independent, parametric, and dependent factors were tested in this study using latent variable values obtained in SmartPLS as an indicator of higher-order structural model analysis.

4.1.2.2. Indicator collinearity

In formative measurement models, the indicators that are the independent driving forces of the model must not be highly correlated with each other. This is because high levels of co-linearity affect the weight estimation method and its statistical power. In this regard, the model should be checked for the possibility of linearity as measured by Variance Expansion Factor (VIF) (Hair et al., 2017). Hair et al. (2017) stated that there are two generally accepted rules to indicate a significant co-linearity problem between indicators of formal measurement of composition. That is, (1) a VIF score of 5 or higher and (2) a VIF score of 3,3. And higher. In this study, the multicollinearity test showed that all indicators met the VIF values and were consistently below the thresholds of 5 and 3.3 shown in the configurations in Table 2 for the PLS pathway model in this study. It became clear that there was no problem with the estimation.

4.1.2.3. Statistical significance and relevance of the formative indicators

Outer weight is an essential measure for evaluating the contribution of formative indicators in determining their importance and relevance. The bootstrapping approach may be used to acquire the values of the outer weights. Bootstrapping is a statistical significance determination approach (Hair et al., 2017). By looking at the p-values, the outcome of bootstrapping must demonstrate that the outer weight from each indication is significant. If the indicator isn’t substantial, the outer loading is examined to see if the indications may still be kept due to content validity (Hair et al., 2017). When the outside loading is less than 0.5 and not significant, researchers can remove the formative indications (Hair et al., 2017). The bootstrapping results in this study revealed that all formative indicators were significant when the P-values obtained were less than 0.05 and the outer loading was larger than 0.5. As a result, the indicators were determined to be qualities that comprised the investigated constructions. As a result, the conceptions of digital literacy and learning performance have been validated as formative variables.

4.2. Structural model validation

4.2.1. Hypotheses testing

The strength of the association between the latent variables is shown by the path coefficient. Bootstrapping is a PLS approach for determining the significance of a route between constructs (Hair et al., 2017). Both t-values and R2 values are generated through bootstrapping. The suggested framework is used to evaluate T-values, which measure the magnitude and importance of route coefficients. When using the SmartPLS bootstrapping approach, 5000 bootstrap samples are advised (Hair et al., 2017). As a result, 5000 bootstrap samples were employed in this analysis, as indicated. Table 3 shows the results of the path coefficient evaluation.

Table 3. Hypotheses testing

Hypothesis		Std Beta	Std Error	t-value	p-value	Decision
H1	Metacognitive Strategies -> Digital Literacy	0.554	0.029	19.792**	0.000	Supported
H2	Digital Literacy -> Learning Performance	0.520	0.034	13.376**	0.000	Supported
H3	Metacognitive Strategies -> Learning Performance	0.029	0.044	0.664	0.251	Not Supported

Note: ** represent significance at 95%, p<0.05, one-tail

All examined hypotheses were found to have a t-value ≥1.645, with a 0.05 level of significance except for Hypotheses 3 (H3). Based on the path coefficient as shown in Table 4, it is evident that metacognitive strategies (β = 0.554) have the strongest effect on digital literacy, followed by digital literacy on learning performance (β = 0.520), and subsequently, metacognitive strategies on learning performance (β = 0.029). Metacognitive strategies (MS) show no relationship with learning performance (LP).

Table 4. R – Squared (R²) criterion

Construct	R Squared
Digital Literacy (DL)	0.318
Learning Performance (LP)	0.270

4.2.2. R² assessment

The predictive potential of independent variables towards dependent variables in the structural model was assessed using R-squared values to determine model fitness. R² values of 0.26, 0.13, and 0.02, according to Cohen (1988) recommendations for predictive accuracy, are considered considerable, moderate, and weak, respectively. Table 4 shows that the structural model in this study had an R² of 0.318 for Digital Literacy and 0.270 for learning performance. The R² value is more than 0.26, as stated by Cohen (1988), indicating a significant model.

4.3. Mediation analysis

Mediation, also known as the “indirect effect,” is the result of an independent construct acting on a dependent construct via one or more intervening or mediating constructs, all of which are backed by strong theoretical or conceptual evidence (Hair et al., 2017; Preacher & Hayes, 2008). The indirect impact strategy is used in this study, which is bootstrapped (Hayes, 2009). Metacognitive methods were investigated in this study to determine if they were mediated by digital literacy in terms of learning performance. With a t-value of 4.178, the bootstrapping analysis revealed that metacognitive methods ( = 0.116) are significant. The 95 percent bootstrap confidence interval bias-corrected indirect effects [UL = 0.071, UL = 0.172] do not straddle a 0 in between, demonstrating mediation (Preacher & Hayes, 2008). As a result, the mediation effect may be judged to be statistically significant. The study’s findings of the mediation analysis are illustrated in Table 5.

Table 5. Hypotheses testing on mediation

				Confidence Interval (BC)
				LL	UL
	Std Beta,	Std Error	t- value	2.50%	97.50%	Decision
H4: MS->DL -> LP	0.116	0.028	4.178**	0.071	0.172	Supported

**p<0.05, t-value>1.96, BC=Bias Corrected, UL=Upper Level, LL=Lower LevelMS=Metacognitive Strategies, DL=Digital Literacy, LP= Learning Performance

4.3.1. Discussion

During the COVID-19 confinement in Malaysia, this study looked at the impact of metacognitive methods on digital literacy and learning performance in digital learning among university students. The hypotheses that were established in this study helped to address the research questions. The study’s findings indicated two positive relationships: one between metacognitive methods and digital literacy, and the other between digital literacy and learning performance. Surprisingly, while there was a negative association between metacognitive methods and learning performance, the mediator found that metacognitive tactics are favourably associated to learning performance through digital literacy.

Hypothesis 1(H1):

Metacognitive strategies positively influences university students’ digital literacy improvement in digital learning (Supported).

With a t-value of 19.792 and a significance level of 95 percent using a one-tail test, the metacognitive techniques construct was shown to have the strongest link with digital literacy. This study’s findings are consistent with previous research, which shows that students with metacognitive skills have the ability and awareness to improve their digital literacy skills (Greene et al., 2018; Lee et al., 2015; Shopova, 2014). To become digitally literate, students must acquire function skills, values, attitudes, and behaviour in the context of digital learning. To utilise digital technology efficiently and successfully, this would also necessitate the development of multiple literacies, since digital literacy necessitates a grasp of the numerous forms of information as well as an integrated understanding of these sorts. The outcomes of this study show that metacognitive skills can help university students improve their digital literacy abilities. As a result, metacognitive tactics may be important for boosting university students’ digital literacy skills. Metacognitive abilities play an important role in improving digital literacy because they can lead to more efficient and critical use of digital resources, which can lead to a more enjoyable and productive educational experience. Students that are digitally literate will be able to complete digital activities more effectively, have a better learning experience, and consequently perform better in class. As a result of the findings, there is a compelling need for initiatives and the promotion of metacognitive abilities in order to ensure that students become more technologically literate. Providing metacognitive skills to pupils. Students must be equipped with metacognitive abilities in order to become resilient lifelong learners. Metacognitive abilities are required to foster digital literacy and prepare students for the digital future, given the availability of technology and tools in digital learning.

Hypothesis 2 (H2):

Digital literacy positively predicts the learning performance of university students in digital learning (Supported).

The outcomes of this study confirm that digital literacy improves students’ digital learning ability significantly (Greene et al., 2014; Scholastica et al., 2016; Wichadee, 2018). In line with Wichadee (2018), the researcher hypothesised that students with greater levels of digital literacy will get higher scores than students with lower levels of digital literacy. Students with a poor degree of digital literacy are also less inclined to continue studying in this manner. In summary, students who have a high degree of digital literacy are more likely to persist with digital learning and achieve higher learning outcomes despite the hurdles they confront because they are more confident and competent in their use of digital learning resources. Furthermore, comparable findings have been reported in the literature, indicating a significant link between digital literacy and learning performance (Scholastica et al., 2016). As a result, digitally literate students are more adaptable while dealing with digital devices and technologies in a digital learning environment, allowing them to enhance their learning outcomes. Those who are digitally illiterate may be at a significant disadvantage. Since a result, there is a need to strengthen students’ digital literacy in order to speed up the progress of the education delivery system, as students’ digital literacy levels influence the balance that digital learning developers set on content production against dissemination.

Hypothesis 3 (H3):

Metacognitive strategies positively predicts the learning performance of university students in digital learning (Not Supported).

Despite prior studies finding a positive relationship between students’ metacognitive skills and their learning performance (Cho & Heron, 2015; Goradia & Bugarcic, 2017), the findings of this study suggest the contrary. According to the study’s findings, metacognitive approaches had no significant direct impact on learning capacity. One possible explanation for the lack of a significant relationship is that metacognitive strategies involve the use of strategies that help students become increasingly effective at learning to improve academic performance (e.g. grades, exams, tests) rather than learning performance, which measures subjective outcomes (e.g. student engagement, student satisfaction). As a result, students who understand metacognition will be able to make better use of their knowledge and talents as they go through their studies. These findings also show that university students, particularly when presented with novel types of online learning, are incapable of looking inside themselves to examine how they learn and evaluate if strategies are effective. This might be related to a lack of self-awareness and reflection skills. As a consequence, metacognitive strategies don’t appear to affect students’ learning results. Furthermore, the outcomes of this study might indicate that students who are not taught metacognitive strategies are unable to understand the bigger picture of the job at hand by planning, monitoring, and directing academic assignments online as successfully as those who are. Metacognitive approaches have been found to have a direct and positive impact on students’ learning performance while using digital learning in many nations (Tian et al., 2018). The findings of this study, on the other hand, are similar with those of other Malaysian researchers who found that the use of metacognitive strategies by Malaysian university students is still fairly low (Anthonysamy et al., 2020; Anthonysamy, 2021). Despite the fact that previous research has shown that students in higher education may monitor and reflect on their strategy usage (Roth et al., 2016), the results of this study show that metacognitive strategies are still weak among Malaysian students. Despite the fact that metacognition is a tool that not only involves students in the learning process but also allocates responsibility for learning to them, academics play an important role in assisting students in developing metacognitive skills. Students use metacognitive approaches to organise, monitor, and govern their cognition process in order to attain a certain goal. These various cognitive processors are intended to help students regulate their learning in terms of personal functioning, academic behaviour, and learning environment. To increase personal control, students might use techniques like self-evaluation, record keeping, monitoring, and self-consequences.

Hypothesis 4

(H4): Students’ level of digital literacy mediates the relationship between metacognitive strategies and their learning performance (Supported).

The addition of digital literacy as a potential mediator was shown to be relevant, since metacognitive abilities were found to have a substantial link with learning performance in digital learning through the mediator. With a t-value of 4.178, students’ digital literacy seemed to be an important antecedent to learning success in this study. Furthermore, this research shows that metacognitive methods can help students enhance their digital literacy, and as a result, learning performance as metacognitive practises improves students’ ability to be critically aware of their thinking and learning. For example, in a classroom setting where students are expected to use digital tools to access and analyze information, students with higher levels of digital literacy will be better equipped to apply metacognitive strategies in their learning, leading to improved performance. On the other hand, students with lower levels of digital literacy may struggle to effectively use these strategies, hindering their learning outcomes. Metacognitive practise is a beneficial process in which students learn to manage and regulate their cognition, motivation, and behaviour in order to achieve their learning objectives. Furthermore, metacognitive abilities may help students better comprehend their own thought processes, which is commonly accomplished through self-reflection and the use of planning, monitoring, and regulating procedures. As a result of the study’s findings, it is suggested that students be provided with situations in which they may practise these metacognitive skills. Different technologies or interactive media can be used to guarantee that students have the opportunity to practise employing metacognitive tactics under the observation of their instructors.

5. Conclusion

The major goal of this study is to look at how university students used metacognitive skills to improve their digital literacy and how it affected their learning performance during the COVID-19 lockdown. The overall findings of the study showed a positive connection with H1 (β = 0.554), H2 (β = 0.520) and H4 (β = 0.116) except for H4 which yielded a negative relationship (β = 0.029). The results of this study showed that metacognitive abilities had a favourable impact on the development of digital literacy and, as a result, on students’ perceptions of their learning success. This result implies that kids who can properly analyse their ideas have more control over their emotions, thoughts, and behaviour, which is indicative of resilient lifelong learners.

This study has several limitations. Firstly, only data from local private institutions in Malaysia’s central area were obtained for this study. Future researchers may examine students from various geographical regions and different types of higher education institutions, which may provide a fresh perspective on the same study’s environment. Secondly, rather than depending on students’ perceptions through data mining and data logs, it would be fascinating to record the actual use of metacognitive abilities in digital learning. As a result, trends in metacognitive skill utilisation may be documented, and relevant recommendations can be established to aid students who are underutilizing online learning tactics. Furthermore, mixed-method research can be very beneficial in acquiring a better and more sophisticated grasp of this issue.

From a global perspective, this research aligned with the United Nations Sustainable Development Goal 4 (SDG 4), which emphasises the necessity of generating and nurturing generations of resilient youngsters who can meet the difficulties of the world’s uncertainties in the future. Although new technologies such as the Internet of Things, robots, artificial intelligence, and big data are progressing, youngsters must be psychologically mature and robust to deal with the future’s uncertainties. Youth are the future leaders, and their ability to cope well with stresses, pressures, and obstacles while performing to their full potential regardless of the situations in which they find themselves is critical. Mental resilience should be an institutional aim for every organisation that is linked to many support networks. Mental resilience must be considered as a core part of sustainability thinking at higher education institutions where there is an increased emphasis on sustainability. Although learning how to think and fostering a resilient mentality can be self-instilled, higher education institutions can play a significant role in inspiring and encouraging the application of learning methodologies.

Disclosure statement

No potential conflict of interest was reported by the author.

Data availability statement

The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request http://dx.doi.org/10.1080/2331186X.2023.2219497.

Appendix A

Table

Construct	Items (Questionnaire)	Sources
Metacognitive Strategies	I analyse an academic task once before attempting it (E.g. understand the expected outcome, task objectives, etc).	Pintrich et al. (1993); Barnard et al. (2009)
	I set some standards for myself when accomplishing an academic task.
	When doing an academic task, I question myself to help me understand the task better.
	When doing an academic task, I check my progress with my instructor to ensure I am on the right track.
	If I get confused during an academic task, I use other approaches to learn (E.g. Youtube, Google, peer help, etc).
	I check over my academic task to make sure I did everything that was required before making the final submission.
Digital Literacy	I know most of the features of my digital device (E.g camera, settings, etc.)	Van Laer and Elen (2017); Ng (2012)
	I know how to use my digital devices for everyday tasks (E.g. Emails, installation of apps, social media, etc).
	I communicate information or ideas effectively to friends using a variety of online tools.
	I collaborate online with friends to exchange ideas or generate meaning to complete a task.
	I search for information from a variety of online sources.
	I know how to select useful information to complete an academic task.
	I can interpret online information that is text-based and non-text-based (E. images, videos, maps, etc).
	I go online to clarify facts before I complete an academic task.
	I can critically judge the suitability of online information before copying/retrieving any data.
	I use an online platform (E.g blog, social media) to suggest new ideas for discussion to complete an academic task.
	I use the Internet responsibly for communicating, socialising and learning.
	I respect other cultures when I am communicating online.
	I reveal little of my personal information online to safeguard myself.
	I understand the potential risks that exist when I am online.
	I understand copyright rules when I copy or reproduce content from online sources.
Construct	Items (Questionnaire)	Sources
Learning Performance	I can apply the content learned in the academic task for other courses/related tasks.	Rovai et al. (2009) Eom et al. (2006); Trowler (2010); Yang et al. (2016) Sun et al. (2008); Wu et al. (2010);
	I can intelligently critic the academic task I completed.
	I become more independent after completing an academic task.
	I become a better thinker after completing an academic task.
	I can use my hands-on skills during the accomplishment of an academic task for other related tasks.
	I can demonstrate to others the hands-on skills learned from an academic task.
	I acquire useful knowledge through interacting with other learners when accomplishing an academic task.
	I acquire useful knowledge by interacting with my instructor when accomplishing an academic task.
	I engage with Web 2.0 tools to enhance my skills.
	I am fully engaged when I am given an academic task.
	I am interested to complete an academic task as I enjoy learning.
	I am satisfied with the academic tasks I am given by my instructor.
	Accomplishing academic tasks help me improve my subject matter better.
	I acquired more knowledge from accomplishing academic tasks.
	I would gladly take up another academic task.