Language Learning Strategy: Comparison of Online and Face-to-Face ESL Classrooms

Awareness of students’ LLS is highly useful for teachers to make informed decisions in their teaching practice. Although numerous researchers on LLS have been done, there has been no research comparing Malaysian students’ LLS in online lessons and face -to-face lessons. It is of great importance to explore this because the practice of incorporating blended learning that combines both online and face-to-face lessons has been on the rise. As such, this study explored Malaysian students’ language learning strategies during both online and face -to-face lessons. Mixed-method research approach was employed in which the first is a 5-point Likert scale survey questionnaire that was adapted from Oxford’s (1990) SILL questionnaire. 60 upper form ESL students were selected through convenience sampling to respond to the questionnaire. Aside from that, semi-structured interviews were carried out on 6 respondents from the questionnaire survey in order to obtain comprehensive responses from the participants. The research instrument aims to answer two research questions; 1) What are the language learning strategies used by Malaysian ESL learners in online classrooms? 2) What are the language learning strategies used by Malaysian ESL learners in face-to-face classrooms? The data was analysed using IBM SPSS Statistics version 27 and the findings were analysed descriptively by interpretation of mean scores and standard deviation. The findings suggest that students mostly use metacognitive and cognitive skills in both online and face-to-face classrooms. However, students use more direct learning strategies including memory and compensation strategies in online classrooms. Whereas, in face-to-face classroom, students generally lean more towards indirect learning strategies such as affective and social strategies. These findings implied that both online and face-to-face classrooms allow students to employ different learning strategies which can be beneficial for their language development. Hence, learning and teaching instructions that include both traditional and technologically enhanced ESL lessons would positively impact students. Future research of wider scale that includes diverse populations particularly from the quantitative view point would greatly help in adding onto the findings of this research.


Introduction
As the world is experiencing rapid globalisation, the English language plays a significant role as a mode of communication. Many countries worldwide speak English to communicate since it yields the power to connect people of different cultures, backgrounds, and points of view. According to Caplan (2020), "English is regarded as the most spoken language in the world, fluently used by 1.13 billion speakers worldwide". This includes non-native speakers who are learning the language as either a foreign language or a second language (L2). Rao (2019) stated that "the English language, which is said to be the first global lingua franca, has become the fastest-increasing language in this modern world and it occupies the status of a commercial language by connecting the East and the West and the North and the South". There is a perception that individuals with good English language command are educated, knowledgeable and trustworthy. Therefore, English language acquisition is crucial among individuals who wish to thrive in their future endeavours.
In the Malaysian context, English is regarded as a second language (L2) and Malaysians are required to learn the language for at least eleven years of schooling (Ansari, 2015). The Malaysian Ministry of Education (MOE) has implemented many novel policies as part of the mission to prepare school leavers with excellent command of the English language. A few years back, the Ministry of Education (MOE) launched the English Language Reform: The Roadmap 2015-2025 as an indication of the nation's determination to substantially improve students' English proficiency (Zuraidah & Mardziah, 2019). In higher educational institutions, the English language is used as a medium of instruction, and students are expected to at least be able to communicate basic information using the language. English in education is deemed fairly significant since it is used to measure students' academic achievement. Hence, Malaysian students must learn English to be on par with the demands of globalisation and improve their quality of life by securing good job opportunities.
English teachers are responsible to understand that each learner learns to acquire the English language differently. Subconsciously, they have their preferences for language learning strategies that help them learn and perform better in acquiring the English language. Weinstein and Mayer (1986), defined language learning strategies (LLS) as "behaviors and thoughts that a learner engages in during learning, intended to influence the learner's encoding process" (p.315). Many pieces of research have been carried out to show the significance of understanding language learning strategies to improve second language learners' English acquisition. These researches were conducted to study the strategies that second language learners (L2) deploy either when learning a language or when using a language or both (Cohen, 1998). The rising awareness to conduct research on this topic has resulted in many attempts among teachers, educators, and stakeholders to apply the knowledge as a step forward to understand the student's needs and plan for better lessons.
However, the certainty that the majority of ESL students being able to learn successfully through one language learning instruction is still questionable. Since different students learn to acquire the language differently, it is unfair to generalise the learning strategy students commonly used as a guide for teachers to plan their lessons. One learning strategy might not be applicable to different students, especially students from different levels of proficiencies. Nevertheless, younger generations are born into the world of technological advancement as digital natives, making them well-versed in multimedia and technological tools. Due to the recent pandemic, Hoq (2020), suggested that "the adoption of online learning has become the new regulation, and technology has emerged as the primary need". Thus, this study is specifically focused on identifying the language learning strategies employed by ESL secondary school students to enhance their language learning in online classrooms and face-to-face classrooms.

Literature Review
Improved language learning necessitates the use of language learning strategies (LLS). The learners' factors and strategies have been defined in a variety of ways by scholars. According to Oxford (1990), LLS is a set of specific behaviours and strategies that students employ to improve their performance in incorporating, maintaining, restoring, and utilising a second language. In its language learning techniques, Oxford (1990) included cognitive, emotional, and social components that may assist individuals in improving their language learning abilities and trustworthiness. Similarly, Griffiths (2008) emphasised that LLS is the selfregulatory effort made by language learners. According to Ali, Zaman, and Khan (2018), learners were given autonomy to become more accountable for their own learning. As a result, language learning is solely dependent on the outcomes of language learners' deliberate efforts with LLS. There are numerous ways to classify the various methods of learning languages. Regardless, Oxford's classification is generally regarded as the most significant piece of research in the field of language learning. Oxford (1990) divided LLS into two groups: strategies, both direct and indirect. Cognitive strategies, compensation strategies, and memory strategies make up direct techniques. While cognitive strategies make use of and transform the target vocabulary, memory strategies are required to store and retrieve information or knowledge.
Language learning strategies have been the centre of research by many researchers particularly during the online distance learning during the pandemic. A study carried out by Mokhtar & Anuar (2021) examined L2 students' preferences for LLS when learning English online from a distance. 78 people participated in this study, and questionnaires were used as the research instrument. When learning English online, students utilised a greater compensation strategy when selecting a LLS. The findings of this study have educational implications for English instruction and learning (Mokhtar & Anuar, 2021). According to Oxford (1990), memory strategies are required to store and retrieve information or knowledge, and compensation strategies help learners' understanding and output levels. On the other hand, indirect strategies can be broken down into metacognitive strategies, affective strategies, social strategies, and strategies that deal with language indirectly. Metacognitive strategies link one's learning processes with one's emotions, motivation, and beliefs. Affective strategies apply to learners' feelings. This tactic encourages self-practice of the language and reduces anxiety. Last but not least, social strategies include the communication and interaction of learners with peers in the target language. For successful language learning, language learners could adopt and modify LLS to suit their learning style. In the context of language learning strategies used in the ESL contexts, Maniam & Ramanlingam (2020) conducted a study to identify whether there is a significant difference in the language learning strategies utilised by bilingual and monolingual students. The findings indicated that it is possible to assert that there is a positive correlation between the use of strategies and the number of languages that learners are familiar with. There may be a positive correlation between the number of languages the students are fluent in and the application of the strategy. The findings also suggest that both monolingual Malaysian and bilingual Indonesian university students use a wide range of language learning strategies in face-to-face lessons, with monolingual students employing more strategies than bilingual students.
Although a lot has been learned about students' learning strategies in recent years, little is known about how these strategies are utilised in today's online distance-learning environment (Mokhtar & Anuar, 2021). Based on Nguyen & Terry (2017) it is believed that individual differences, as well as environmental and contextual factors, have an impact on language learning strategies (LLSs). As a result, a number of factors, including learner and learning variables, influence how well LLSs are used. The qualitative results of a larger-scale study on English LLSs among Vietnamese tertiary students are discussed in detail in this paper. Ten English as a Foreign Language (EFL) teachers and ten EFL students at a university in Vietnam participated in semi-structured interviews to collect the data. The students' attitudes and actual experiences with LLSs were the primary subjects of the interview questions. In a similar vein, the general attitudes of the EFL staff were the focus of the interview questions. The interview data were analysed thematically, checked once more, and translated into English. The results are textual and interpretative, with emerging themes and issues related to the target learners' attitudes toward and actual use of LLSs.
Ghumaludin (2022) conducted a study to determine that students' Language Learning Strategies (LLS) must be understood in order to address the COVID-19 pandemic's impact on learning. There are few studies on the English language learning styles (LLS) of STEM students. The study's goals are to investigate the LLS used by STEM students in Malaysia who are learning English as a second language (ESL) and the LLS's underlying factors in relation to open distance learning (ODL). The results showed that cognitive learning, compensation, and metacognitive strategies were used the most, while affective and memory strategies were used the least. The memory strategy corresponds to the metacognitive strategy, for example, was included in the PCA. Some conventional approaches to grammar instruction may need to be reconsidered in light of the cognitive strategy concept of "practice." Social and emotional factors must also be taken into account. Additionally, ODL has demanded more.
In the context of technology-enhanced language learning (TELL), the sixty-six reviewed studies investigate language learning strategies, including strategic self-regulation (Zhou & Wei, 2018). Any language learning activity that makes use of technological means and/or tools for efficiency, motivation, and adaptability in learning styles is broadly referred to as TELL. According to Oxford's classification, L2 learning strategies are broken down into a set of interconnected but adaptable systems: 1) strategies for the language subsystems (such as vocabulary and grammar) and the language skills area (such as reading, writing, listening, and speaking, and related aspects). The reviewed strategies are arranged similarly. Strategies for learning L2 skills are discussed first, followed by language subsystems strategies, and finally, self-regulated L2 learning is the primary focus (Zhou & Wei, 2018). Eggers et al (2021) shows how SRL strategies are categorised. They looked at classification of the strategies with various human and nonhuman interactional methods. Their findings suggested that various interactional strategies can effectively promote SRL strategies and finding the best balance between human and non-human interactional methods to promote self-regulation strategies in various learning environments, especially in a blended learning environment. As such, their findings indicated that cognitive and motivational strategies received little attention in a blend of both online and face-to-face environments. They also suggested that interactional strategies that promote students' SRL are required, with focus on student-student interactions for collaborative and cooperative learning. The student tools interactional method has been touted as promising for the advancement of metacognitive strategies. It is recommended to conduct additional research into the ways in which student tools can enhance metacognitive and other self-regulation strategies.

Theoretical Framework
This study is guided by constructivist learning theories. Using constructivist methods, tools, and strategies is essential for a productive classroom where teachers and students communicate effectively. Based on Powell's (2009), in the classroom, there are two main types of constructivism: 1) Social constructivism based on Vygotsky's theory or cognitive constructivism, depending on Piaget's theory. The use of inquiry-based teaching methods and the creation of relevant and meaningful concepts by students based on previously acquired information are similarities. Language development theory differs from cognitive constructivism's thinking-before-language theory to social constructivism's thinking-beforelanguage theory.
Teachers can develop individual learning methods like discovery learning and social interactive activities to foster peer collaboration with the help of communicative tools and strategies. Hence the theories identified are Piaget's Cognitive Constructivism and Vygotsky's Social Constructivism. The ability to learn at various ages in childhood based on logical development is at the heart of Piaget's stages of development. According to Piaget's theory, equilibration, assimilation, and accommodation, all concern the children's capacity to resolve conflicts and construct cognitively or individually their new knowledge within their stages. The teacher can facilitate constructivist learning by understanding that this process takes place at a different rate for each student. The importance of comprehending what each person requires to acquire knowledge and learn at their own pace is incorporated into the cognitive constructivism theory proposed by Piaget. This process relies heavily on observing students and comprehending their difficulty level.
When teaching complex concepts, for instance, some students may grasp them quickly while others may struggle. As part of the inquiry method to prevent misinterpretation, students are asked questions to determine where they may have difficulty. Teaching students to be able to comprehend concepts logically and intellectually is a primary objective for all teachers who are aware of these stages. When clarity begins, effective learning begins. Since collaboration and social interaction are incorporated, social constructivism is a highly effective teaching method that all students can benefit from. After Piaget had already talked about his theories about an individual or cognitive constructivism, this kind of constructivism emerged.
The founder of social constructivism, Lev Vygotsky, believed that social interaction was an essential component of education. The personal critical thinking process of each student and the social interactions they have in the classroom are the foundations of social constructivism. Social constructivism and language development are intertwined across all of Vygotsky's research and theories, including cognitive dialogue, the zone of proximal development, social interaction, culture, and inner speech (Vygotsky, 1962). Effective classrooms can be created by comprehending his theories and creating an environment that encourages interaction.

Research Gap
The benefits of using a strategy for learning a language fall into two categories: practically and empirically. Practically, it makes use of language learning strategies that could assist students in achieving learning outcomes and encouraging them to gain experience in understanding English language knowledge through daily online language learning. Empirically, it demonstrates how students' language learning strategies exchanged learning process outcomes and experiences to learn the English language while they were enrolled in online courses. Additionally, online learning is essential for assisting teachers and students in learning the target language during and after a pandemic.
The researcher realised that there is still a need for this research. As a result, it does not rule out the possibility that future researchers will continue their research on additional schools at various levels and may even employ a different instrument or questionnaire. In addition, conducting research at the student level in another large area or using research subjects in an English class can produce different results and present additional challenges for future researchers. Through the review of past literature, it is evident that despite the growing body of research pertaining to the topic, no studies have been found which focus on students' language learning strategies utilised to promote blended learning in the context of online and face-to-face ESL classrooms. In addition, as to the knowledge of the researchers, there has been no studies comparing language learning strategies between online and face-to-face classrooms. Therefore, to fill the gap in this research, the current research aims to investigate ESL Malaysian secondary school students' language learning strategies in the virtual and conventional face-to-face environments to promote blended learning in the secondary ESL contexts.

Research Design
By features, this research followed the tracks of a mixed-method research design in which the research problem identified was answered through an establishment of overall response tendency of the 60 participants involved with further clarification through semi-structured interviews. Specifically, this study is a explanatory mixed-method survey research that measures the dependent variables which is the language learning strategy employed by secondary Malaysian ESL students in response to the independent variable which is the different learning modes during and after the pandemic which is online home-based learning and physical face to face lessons. As stated by Cresswell (2014), the blend between quantitative and qualitative data enables an integrated understanding of a research problem.

Respondents
A total of 60 students who were selected conveniently based on availability as well as willingness had participated in this study. The participants were Malaysian secondary school students of the upper form level; Form 4 and Form 5. Although it could not be claimed that these participants were representatives of the population, convenience sampling can still provide valuable information to the questions of a researcher (Creswell, 2012)

Instruments
The research instrument adopted was the questionnaire by Oxford (1989) Strategy Inventory for Language Learning (SILL) questionnaire which is a highly-established questionnaire that had been utilised by numerous researches particularly in relation to studies on language strategy. The researchers adapted the questionnaire into a questionnaire of 100 items that is divided into two parts; LLS in online classroom and LLS in face-to-face classroom. Furthermore, the researchers assessed the questionnaire results using a five point Likert scale ranging from 5 to 1. Usually true 3 Somewhat true 2 Usually not true 1 Never Table 2 Questionnaire Grid Questionnaire Grid The questionnaire was divided into two sections. The first section (item 1-50) was about the language learning strategy of English during the pandemic. The second (item 51-100) was about the language learning strategy of English after the pandemic.
The First Section (item 1-50) assesses English language learning strategy utilized by students throughout the online learning phase during COVID-19 pandemic.
The Second Section (item 51-100) assesses English language learning strategy utilized by students for physical lessons after COVID-19 pandemic.

Validity and Reliability
Validity of the questionnaire was established through expert validation of three experts within the ESL field which consisted of 1 senior lecturer and 2 senior teachers. Aside from that, experts' checking for validity of the translated items were also acquired through 1 lecturer and 2 senior teachers of Bahasa Melayu major. Additionally, prior to the data collection through the distribution of the questionnaire the adapted questionnaire was first piloted with 10 non-participating respondents to test its reliability in obtaining the students' inventory of language learning skills as established in the research aims. Other than that, Cronbach's Alpha was used by the researcher in this study to assess the internal consistency reliability of all the items in the adapted questionnaire. Based on the results in Table 3, r value is 0.956, as such it can be deduced that the values were significantly correlated because the r values above 0.70 is considered significantly strong and it is sufficiently reliable to be included in the data for descriptive research (Surucu & Maslakci, 2020).  (2010) was conducted on 6 respondents who had participated in the survey questionnaire. According to Creswell & Creswell (2018) semi-structured interviews are time-efficient, flexible and straightforward to be administered. As such, Semi-structured interview was chosen to obtain further understanding on the students' responses in relation to the SILL questionnaire survey.
The interview was recorded through audio recording and each session lasted for around five minutes.

Data Collection Method
Data for this study was collected in accordance with the steps in a quantitative research method as presented by (Creswell & Creswell, 2018). Firstly, the researchers selected the samples from a secondary school in Melaka based on convenience. Availability and willingness to participate were the criteria that fit the study participant selection. The next step required the researcher to prepare the questionnaire that was adopted and adapted from Oxford (1989). After that, both reliability and validity tests were conducted on the finalised survey questionnaire items. Then, the questionnaire was distributed by the researcher online through Google Form by providing students with the links in the school Telegram channel. Online questionnaire was opted because it is both time efficient and convenient. After that, Statistical Package for the Social Science (SPSS) version 23 was used to analyse the data to fulfil the research objectives. Aside from that, the data obtained was tabulated for clear presentation of the statistical patterns from the responses collected.

Data Analysis Method
Statistical data of the responses collected from the participants were analysed through Statistical Package for the Social Science (SPSS) version 27. The data was later tabulated and descriptive analysis of the mean and standard deviation were done to explain the findings in relation to the research questions. Additionally, data was interpreted according to the key understanding of LLS frequency as stated in Oxford (1989) SILL questionnaire. The scale is explained in table 4. Aside from that, the quantitative data is supported with excerpts from the interview transcriptions. Pseudonyms are used instead of the actual names of the participants in order to maintain participants' confidentiality. Removing details that could lead to identification of the participants is an crucial step to adhere to the research ethics (Adnan et al., 2019) Hence, the students were labelled as S1 to S6.

Demography of Respondents
Gender Table 5 shows the distribution of respondents according to gender. Female students constituted the highest frequency with 66.7% (n=40). Meanwhile, the remaining 33.3% of the respondents who participated in this study were male (n=20).  Table 6 shows the respondents' distribution according to their first languages. With reference to the table, it can be seen that the majority of the respondents who participated in this study used Bahasa Melayu as their first language. It accounts for 96.7% of the total respondents (n=58). The remaining 3.3% of the respondents used other language as their first language which is Iban (n=2).  Table 7 indicates how respondents perceived their English language proficiency level. More than half of the respondents perceived their English language proficiency level as good with 53.3% (n=32) which is the highest among the others. The second highest percentage comes from those who perceived their proficiency level as fair with 28.3% (n=17). Meanwhile, 15% of the respondents perceived themselves as having excellent English proficiency level (n=9) and the remaining 3.3% of the respondents perceived their proficiency level as poor (n=2).

Language Learning Strategy Use in the Online Classroom
To answer the first research question regarding the language learning strategies used by Malaysian ESL learners in the online classroom, analysis of mean scores and standard deviations of each strategy was conducted. The most and least frequent strategies used for each category were also identified. Table 8 shows the strategies employed by the respondents to learn the English language in the online classroom. Generally, the results indicated medium usage of all six strategies in the online classroom with a mean score of between 2.50 to 3.49. Based on the table, it can be seen that the respondents mostly used metacognitive strategies in the online English language classroom with the highest mean score of 3.45 (SD=1.352). Among the metacognitive strategies, the results indicate that the respondents usually put in effort to identify the way to be a better English language learner (item 33) with the highest mean score of 3.78 (SD=1.367). Meanwhile, with the lowest mean score of 2.65 (SD=1.388), it is indicated that the respondents only planned their schedule to study English sometimes (item 34).
With the second highest mean score of 3.25 and a standard deviation of 1.371, cognitive strategies were the second most used strategies among the respondents. The results also suggest that the respondents frequently watched English language programmes on the television or English movies (item 15) (M=4.07, SD=1.326) as part of their cognitive strategies. It is also indicated that making summaries of information that they heard or read in English was the least used strategy among the respondents (M=2.83, SD=1.486).
Memory strategies were the third most frequently used strategies in the online ESL classroom with a mean score of 3.00 (SD=1.381). The results also suggest that the respondents tried to remember a newly learnt English word by mentally associating it with a situation in which it might be used as part of their memory strategies (item 4). It was indicated by the highest mean score of M=3.28 (SD=1.427). As for the least used memory strategy, the respondents generally did not use flashcards to remember new English words (item 6) with the lowest mean score of 2.03 (SD=1.221).
Meanwhile, compensation strategies were used moderately by the respondents in the online classroom (M=2.99, SD=1.362). Among the compensation strategies used by the respondents, the results indicate that the respondents most frequently made guesses to understand unfamiliar English words (item 24) (M=3.50, SD=1.444) and least frequently made up new words to replace the ones they did not know (item 26) (M=2.55, SD=1.443). In addition, the results also indicate a medium use of social strategies in the online ESL classroom (M=2.95, SD=1.454). The respondents most frequently asked the other person to speak slowly or repeat when they did not understand (item 45) (M=3.18, SD=1.490). Three social strategies that were least frequently used by the respondents were items 47, 48 and 50. The lowest mean scores indicate that they only practiced English with other students (M=2.87, SD=1.359), asked for help from English speakers (M=2.87, SD=1.501) and tried to learn about the culture of English speakers (M=2.87, SD=1.535) occasionally.
As the least frequently used strategy in the online ESL classroom, the respondents reported a medium use of all the affective strategies (M=2.68, SD=1.435) in the SILL. Among them, the most frequently used strategy was item 40 by which the respondents occasionally encouraged themselves to speak English although they were afraid of making mistakes (M=3.83, SD=1.392). Meanwhile, the least frequently used affective strategy with the lowest mean score was item 44. The results indicate that the respondents generally did not talk to another person about their feelings when learning English (M=2.32, SD=1.295). The interview data seems to be consistent to the findings of the questionnaire data. Two questions were asked to further explore LLS use in Online classrooms. 1) Generally, what techniques do you use in understanding English during online learning? And 2) Is there anything in particular you do in order to practice English during online learning? Three of the respondents said: S1: "…when learning English online, I try to recall things I have learnt from teacher before…" S5: "…sometimes when I have to answer questions, I will arrange ideas in my mind before giving my answers…" S6: "...when I need to say a new word when I have to read a text during googlemeet I will try to say the word based on similar words I have learnt before…"

Language Learning Strategy Use in the Face-to-face Classroom
In answering the second research question 'What are the language learning strategies used by Malaysian ESL learners in the face-to-face classroom?', a descriptive analysis of the mean scores and standard deviations for SILL in the face-to-face classroom was conducted and the results are presented in tabular form. The strategies with the highest and lowest frequencies of use were also identified for each category. Table 9 presents the language learning strategies used by the respondents in the face-to-face classroom. The results indicate high usage of metacognitive strategies and medium usage memory, cognitive, compensation, social and affective strategies. With reference to the table, it can be said that the respondents mostly used metacognitive strategies in the face-to-face classroom with the highest mean score of 3.50 (SD=1.350). From the findings, it is indicated that the students usually thought about their progress in learning English (item 88) which had the highest mean score among the metacognitive strategies (M=3.85, SD=1.400). Meanwhile, the respondents occasionally planned their schedule to get sufficient time to learn English (item 84) which was the least used metacognitive strategy with a mean score of 2.83 (SD=1.392).
The second highest mean score was recorded for cognitive strategies at 3.25 and a standard deviation of 1.335. The most used cognitive strategy in the face-to-face classroom was item 65 which indicates that the respondents usually watched English language TV shows and movies (M=3.85, SD=1.338). In contrast, the least frequently used cognitive strategy was trying to identify patterns in English (item 70) (M=2.85, SD=1.424). It is indicated that the respondents only used the strategy sometimes. This is followed by affective strategies (M=3.12, SD=1.430). The results suggest that the respondents usually tried to relax when they felt afraid to use English (item 89) which had the highest mean score of 3.75 (SD=1.373). Meanwhile, the respondents talked to someone about their feelings when learning English (item 94) only occasionally in the face-to-face classroom. It recorded the least frequency of use among the affective strategies with a mean score of 2.58 (SD=1.510).
Following that, the respondents indicated medium use of social strategies (M=3.02, SD=1.384) and compensation strategies (M=3.00, SD=1.421). Looking into the use of social strategies in the face-to-face ESL classroom, it is indicated that when they do not understand something in English, the respondents most frequently asked the other person to either slow down or repeat it (item 95) (M=3.35, SD=1.482). The results also show that the respondents only asked questions in English sometimes (item 99) which was the least used social strategy (M=2.78, SD=1.263). In addition, with the highest mean score of 3.40 (SD=1.429), it is indicated that the respondents most frequently used another word or phrase with similar meaning to replace the English word that they did not know (item 79) as part of their compensation strategies. Meanwhile, the least frequently used compensation strategy among the respondents in the face-to-face classroom was making up new words for unknown English words (item 76) (M=2.63, SD=1.484).
Memory strategies were the least used strategies in the face-to-face classroom with a mean score of 2.96 (SD=1.327). Among the memory strategies the highest frequency of use was recorded for item 52. It is indicated that the respondents most frequently used new English words in a sentence so they can remember them (M=3.53, SD=1.396). Meanwhile, it is also indicated that the respondents generally did not use flashcards to remember new English words (item 56) in the face-to-face classroom with the lowest mean score of M=2.30 (SD=1.344). Similarly, the interview data for face-to-face classrooms also seems to be consistent to the findings of the questionnaire data. Two questions were asked to further explore LLS use in Face-to-face classrooms. 1) Generally, what techniques do you use in understanding English during face-to-face classroom after the pandemic? And 2) Is there anything in particular you do in order to practice English during face-to-face classroom after the pandemic? Three of the respondents said: S2: "…when learning English face-to-face, I try to repeat the things I learnt…for example if I learnt a new word, I will try to use the words with friends many times…" S3: "…for example sometimes when I am doing writing questions with my friends, I will use Wh-questions to get more ideas on what to write…" S4: "...face-to-face learning makes me more nervous, so when learning English face-to-face I will try to relax by breathing in and out slowly especially when doing speaking activities…"

LLS in Online ESL Classroom
At the beginning of the study, it has been set out that one of the researchers' primary aims was to determine the language learning strategies used by Malaysian ESL learners in online classrooms. From the analysis of the data obtained, it is apparent that the learning strategies used by Malaysian ESL learners in online ESL classrooms follow the order of metacognitive, cognitive, memory strategies, compensation, social and affective strategy. Referring to Oxford's taxonomy of LLS (1990) it can be said that in online English lessons, Malaysian students learn by centering their learning through overviewing and connecting new language information to their existing knowledge by utilising metacognitive strategy. Aside from that, they also utilise cognitive strategy such as repeating and formally practising the new language forms they learn. This finding is different from that of Mokhtar & Anuar (2021) in their study that examined L2 students' preferences for LLS when learning English online wherein it was reported that the strategy mostly used by the 78 participants involved were compensation strategy. However, when analysed according to gender, it was discovered that their female participants preferred the Metacognitive approach. Although this current study did not put focus on gender, the differences between these two studies could be due to the number of female participants involved. Nonetheless, this finding is consistent with that of Ghumaludin (2022) who reported that cognitive learning, compensation, and metacognitive strategies were used the most amongst the respondents of a study on LLS for online learning. Strikingly, one finding that emerged from the data is the fact that Malaysian ESL students utilize more direct learning strategies compared to indirect learning strategies in online ESL classrooms. As established by Oxford (1990) direct strategy include cognitive strategies, compensation strategies, and memory strategies. For instance, in order to self-regulate in an online ESL classroom, students may utilise cognitive strategies to make use of and transform the target vocabulary. Other than that, they may also store and retrieve information or knowledge on the target language through memory strategies.

LLS in Face-To-Face ESL Classroom
The second aim of this study was to determine the language learning strategies used by Malaysian ESL learners in the face-to-face classrooms. The data analysis indicates that, it is clear that the learning strategies used by Malaysian ESL learners face-to-face classrooms follow the order of metacognitive, cognitive, affective, social, compensation and memory strategy. Similar to online classrooms, in face-to-face English lessons, Malaysian students also learn by centering their learning through overviewing and connecting new language information to their existing knowledge by utilising metacognitive strategy (Oxford, 1990). However, different from online ESL classrooms, Malaysian students utilise more indirect learning than direct strategies during physical face-to-face lessons. Indirect learning strategies include metacognitive strategies, affective strategies, social strategies (Oxford, 1990). One situational example is in face-to-face ESL classrooms, Malaysian students may self-regulate language learning processes through metacognitive strategies by linking the language input received with emotions, motivation, and beliefs. Other than that, their self-regulation may also include the utilisation of affective strategies by responding to the language input with their feelings. Other than that, they may also indirectly approach language input through social strategies for example participating in communication and interaction with peers in the target language. This study confirms the study conducted by Maniam & Ramanlingam (2020) that explored the connection between language learning strategies utilised and the number of languages the students are fluent in. Similar to the current study, it was identified that both monolingual Malaysian and bilingual Indonesian university students use a wide range of language learning strategies in face-to-face lessons.

Implications and Recommendations
It is evidently clear that although the highest LLS utilised by Malaysian ESL students in both online and face-to-face English lessons are similar, which are cognitive and metacognitive strategies, students use direct learning strategies to a greater extent during online learning and indirect learning strategies during physical face to face lessons. This may suggest that due to the amplified utilisation of online lessons during the pandemic, students had developed self-regulatory learning in which they continue to use when going back to face-to-face lessons. As such, these findings imply that both online and face-to-face lessons possess benefits to the student's language development, one of which is increased awareness of their individual LLS utilisation.
Researches are all restricted to several limitations and as for that, every finding should be considered cautiously. Due to the limited demography of this study which was conducted in a semi-rural location, future research that is inclusive of diverse population groups should be considered. Aside from that, this study is subjected to the limitations of quantitative survey research. As such, it is suggested that for future studies researches that employ experimental research would greatly build upon the existing knowledge around the issue of Malaysian ESL students language learning strategies.

Conclusion
In conclusion, the study focused on language learning strategies used by Malaysian ESL learners in online classrooms during the pandemic and to determine the language learning strategies used by Malaysian ESL learners in face-to-face classrooms after the pandemic. The comparison showed mere differences and this may suggest that due to exposure of online lessons during the pandemic, Malaysian students may have developed self-regulatory learning strategies that they continue to employ in current post-pandemic face-to-face lessons. Nonetheless, current literature on this generally focuses on higher education and traditional face-to-face classrooms in which the emphasis on self-regulatory learning was rather limited. Other than that, the findings drawn in this study suggests that online learning may have fostered Malaysian students to be more aware of their own use of LLS.