Student Engagement Among Undergraduate Students in Southeast Asia: Systematic Literature Review

: Southeast Asia has up to 7,000 higher education institutions with various qualities. The inequality may affect the learning process, such as student engagement. Student engagement is a multidimensional construct that is often positively related to academic achievement. This systematic review aims to get an overview and factors that influence undergraduate students' engagement in Southeast Asian countries. The article search method was carried out through five online databases; ERIC, ProQuest, Scopus, SpringerLink, and Taylor & Francis Online, and obtained 11 articles. The articles were extracted using PRISMA flow. As a result, student engagement in Southeast Asia can be described through several dimensions, namely cognitive, emotional, behavioral, and agentic. The theory most widely used is Self-Determination and Social Cognitive Theory. Besides, there are also Broaden-and-Build Theory, Conservation of Resources, and Tinto Integration Model used as the theoretical framework in some studies. For the predictors, undergraduate student engagement in Southeast Asia is more influenced by external factors that will lead to the appearance of internal predictors.


INTRODUCTION
Southeast Asia has up to 7,000 higher education institutions with approximately 12 million students.Indonesia has 4,537 universities with more than 7 million students, this is ranked first in the list of Southeast Asian countries.The second position is occupied by the Philippines which has 3 million students and 1,943 universities (SHARE-ASEAN, 2019).Based on UNESCO (2022), the ratio of tertiary enrollment in South and West Asia has increased by 200% since 2000.This is contrary to the rate of tertiary education completion, which is much lower.Singapore has the highest rate of 40 % and Indonesia with a level below 10% (Yeung, 2022).This is due to the quality of education which is still not optimal.Aspects of the quality of education include learners, learning environments, content, process, and outcomes.
In Indonesia, the quality of higher education is unequal between regions.This is also due to different household incomes and the majority come from middle-high-income households, which in turn have limited access to education (Wicaksono & Friawan, 2011).In Cambodia, the university curriculum is expected to be more relevant to the needs of students and it is hoped that there will be an increase in research activities (Nhem, 2022).In addition, the learning model that is often applied is teacher-centered instead of student-centered learning, where students play an active role in building their own knowledge (Voitovska & Tolochko, 2019).In Thailand, the majority of learning methods are still using rote learning and this hinders the learning process and minimally increases critical thinking (Polrak, 2019).The same thing also happened in Malaysia, where lecturers have been given an introduction and training on student-centered learning (SCL), but they prefer to use conventional methods in delivering material daily (Bakar et al., 2012).The quality of the learning process, high drop-out rates, and demotivation are common issues in student engagement.
Student engagement has become the focus of many educational studies because it correlates with academic achievement.The more engaged students will be more likely to succeed academically (Bond et al., 2020).According to Newmann et al. (1992), student engagement is a psychological effort and investment made by students to learn, understand, and master skills and knowledge.The student engagement concept started from Tyler's work in 1930 which mentioned "time on task on learning" and Pace's work in 1989 which mentioned "quality of effort" (Groccia, 2018).These terms lead Kuh to make the definition of student engagement as the time and effort students devote to learning activities that are empirically linked to the desired outcomes of college (Kuh, 2003).However, there are many terms related to student engagement; Skinner et al. (2009) emphasize students' participation and identification with school and school-related activities.Newmann et al. (1992) mention student engagement is related to students' psychological investment in learning.
Although student engagement is a complex term and relatively diverse in definitions, researchers have a common opinion to say that it is a multidimensional construct.However, there are many views regarding the number of dimensions contained in student engagement.Fredricks et al. (2004) mention three interrelated dimensions of student engagement; behavioral, emotional, and cognitive.Behavioral engagement is measured by observable behavior, such as participation, interaction, collaboration, and completion.Emotional engagement defines emotional reaction towards learning activities and environment.Cognitive engagement defines student motivation, self-efficacy, self-regulation, and using the right learning strategy.Reeve and Tseng (2011) add agentic engagement as another dimension, which measures the student's contribution to the flow of instruction received.It is expressed by giving inputs, expressing preferences, and also making suggestions on how problems are solved.
Student engagement can be influenced by various internal and external factors (Fredricks et al., 2004).Internal factors arise from within the individual, such as motivation (Khaing & Myint, 2020;Zhang et al., 2015), attention, interest, goal orientation, and self-efficacy (Ginting, 2021;Zhang et al., 2015).Meanwhile, external factors that can affect engagement come from the role of family, peers, and educators (Hofer et al., 2022).In addition, educational institutions and teaching methods applied in class can also bring impact to student engagement (Khaing & Myint, 2020).Khaing and Myint (2020) show that motivation is the strongest internal factor and teaching style is the strongest external factor affecting student engagement.Demographic factors also affect student engagement, such as gender (Alghanmi & Nyazi, 2022) and university major (Dika et al., 2022;Magallanes, 2022).
Furthermore, it is necessary to understand that there are changes in the characteristics of students along with the times.Currently, the majority of students are Generation Z (Gen Z), namely those born in the period 1996-2012 (Schwieger & Ladwig, 2018).Gen Z was born when access to technology has become more widespread and sophisticated, where everything can be accessed through one smart device.They are selfdirected learners and tech-savvy.However, when compared to Millennials, Gen Z's attention span has decreased to 8 seconds (12 seconds for Millennials) (Vizcaya-Moreno & Pérez-Cañaveras, 2020), and lack of critical thinking, especially in terms of validating information (Shatto & Erwin, 2016).The majority of lecturers and staff at tertiary institutions are of previous generations, namely Boomers, X, and Millennials.
Previous research has found that there are differences in learning preferences from each generation.According to Oblinger (2003), Millennials prefer collaborative learning processes and use technology.Meanwhile, Gen X and Boomers prefer coursework as a learning tool (Hampton & Pearce, 2016).Gen X and Boomers also have different preferences, namely structured and detailed (Gen X) and tactile (Boomer) learning environments.This shows that the characteristics of the generation should also be considered to increase student engagement in class.From this background, the researchers want to dig deeper into the dynamics of student engagement among undergraduate students in Southeast Asia.The researchers hypothesize that student engagement in this context can present unique characteristics, especially in technology integration.It is hoped that this study can help related parties in higher education to understand more about teaching and learning situations and improve the quality of learning.Therefore, the researchers aim to know the overview of student engagement studies and explore the factors that influence student engagement in undergraduate students in Southeast Asia.

METHODS
The article selection process used in this systematic literature review was carried out based on the Preferred Items for Systematic Review and Meta-analysis (PRISMA) (Moher et al., 2009).PRISMA can be used as a guideline to ensure the completeness of studies when conducting and reporting systematic reviews and meta-analyses (Tam et al., 2019).The first step was articles searching in several journal databases, such as SpringerLink, ERIC, ProQuest, Scopus, and Taylor & Francis Online.To get the right articles, a keyword search process was carried out with the following combinations: 'student engagement' AND/OR 'academic engagement' AND/OR 'learning engagement', 'undergraduates' AND/OR 'college students', 'Southeast Asia' AND/OR 'Asia'.The researchers also included the names of countries in Southeast Asia as filter and search keywords and used several inclusion and exclusion criteria.The inclusion criteria include; First, the article must discuss student engagement as the main variable.Second, the articles should have been published in international journals, through a peerreview process, written in English, available in full-text, and published from January 2018 to October 2022.Third, the studies reviewed are empirical studies and quantitative in nature.Fourth, the research samples were undergraduate students from various countries in the Southeast Asian region.Meanwhile, the exclusion criteria include; First, the article does not discuss student engagement as the main variable.Second, articles discuss psychometric studies, qualitative research, mixed methods, or literature reviews.Third, articles that are not available in full-text, are not written in English and were published before 2018.And fourth, the participants involved are not undergraduate students and are not from the Southeast Asian region.The flow of PRISMA can be seen in Figure 1.

Figure 1. Flow Diagram of Study Selection Process
Based on the search results from the five journal databases, only quantitative studies related to student engagement within the scope of undergraduate students and in Southeast Asia were selected.Studies in forms other than quantitative, measuring instrument validation studies, or involving a sample of international students were excluded.Initial search results across all five databases yielded 1,495 articles.After the reduction process based on duplication, title, and abstract, 19 full-text articles were obtained and only 11 articles were analyzed thoroughly.The search results on the five databases can be seen in Table 1.The data analysis process was carried out based on the general characteristics of the study and student engagement variables.Records excluded based on title and abstract: 1.451 Full-text articles excluded (n = 8), reasons for not meeting the inclusion criteria, such as non-student engagement in university students, qualitative/mixed design, published before 2018, not peer-reviewed articles, not published in English, full-text not available, and not SouthEast Asian sample.

Identification
Records identified through database searching (n = 1.495)Records screened: 1.470 Full-text articles assessed for eligibility: 19 Articles included for systematic review (N = 11)

RESULT AND DISCUSSION
After the selection process of the studies, only eleven were relevant to be included in this review.These came from five over eleven Southeast Asian countries, namely the Philippines (1), Indonesia (1), Malaysia (7), Thailand (1), and Vietnam (1).The total number of participants from eleven countries is 5.069, which is 460 for the average sample.The majority of the studies included participants in the range of 17 to 26 years old, but there were four studies that included participants above 26 years old (Baloran et al., 2021;Benlahcene et al., 2022;Cahyadi et al., 2021;Rahim, 2022).The participants came from various types of universities, such as public or state universities (Benlahcene, 2022;Benlahcene et al., 2022;Benlahcene et al., 2020;Benlahcene et al., 2020;Rahim, 2022), private universities (Baloran et al., 2021;Cahyadi et al., 2021), global university, which has international branch campuses (Tang et al., 2018), and the combination of private and public university (Handagoon & Varma, 2019;Lim et al., 2022).Meanwhile, Hoi (2021) in Vietnam did not mention the type of the university.In addition, there are several samples conducted in an online learning environment (Baloran et al., 2021;Hoi, 2021;Rahim, 2022) and also with working students (Cahyadi et al., 2021;Rahim, 2022).
All studies in this systematic review are quantitative research with survey methods.Various scales were used to measure student engagement in undergraduate students, such as Engagement Versus Disaffection with Learning Engagement Scale (Skinner et al., 2009), Metacognitive Strategies Questionnaires (Wolters, 2004), Agentic Engagement Scale (Reeve, 2013), School Engagement Measure (Fredricks et al., 2004) (Appleton et al., 2006).
The analysis of eleven studies related to student engagement can be seen in Table 2.All studies were cross-sectional.Two studies examined student engagement in online learning (Baloran et al., 2021;Rahim, 2022).There were a variety of theories used in the research, namely Self-Determination Theory (Benlahcene et al., 2022;Benlahcene et al., 2020;Benlahcene et al., 2020), Broaden-and-build Theory (Benlahcene, 2022) Three studies used Self-Determination Theory (SDT), that students will be more engaged when social context satisfies individual needs for autonomy, relatedness, and competence (Deci & Ryan, 1986;2000).In Benlahcene et al. (2020), it is found that teacher autonomy support (TAS) fully predicts three dimensions of engagement (behavioral, emotional, and cognitive) through the mediation of Personal Best Goals (PB Goals).PB Goals are personal academic targets that students create to surpass themselves as progress (Martin, 2006).This is also in line with the research by Benlahcene et al. (2022) where TAS predicts agentic engagement through the mediation of PB Goals.Research on the basic psychological needs satisfaction variable was carried out in-depth by Benlahcene et al. (2020).The basic psychological needs dimensions studied are autonomy, competence, relatedness, and novelty.The result is that satisfaction of competence and relatedness needs predicts four dimensions of engagement, namely behavioral, emotional and cognitive, and agentic.Novelty needs satisfaction predicts only three dimensions, except agentic.Whereas the satisfaction of autonomy needs only predicts the agentic dimension.
Two studies used Social Cognitive Theory (SCT) as their framework.SCT defines an individual sense of belonging to an institution and participation in academic and non-academic activities (Lim et al., 2022).Both of these studies (Rahim, 2022;Hoi, 2021) were carried out in an online learning environment.Hoi (2021) discussed that the use of Facebook as a learning medium can increase student cognitive engagement.This happens because Facebook supports innovative and interactive learning which makes students more active in sharing information.Rahim's research (2022) was conducted on third and fourth-academic-year students where teacher competence in teaching online did not predict engagement.The same moderator variable was used in both studies, that is self-efficacy.Hoi (2021) shows that the benefits of Facebook for learning on cognitive engagement will be even greater if students have self-efficacy in sharing information.Meanwhile, Rahim (2022) explains that the relationship between lecturer competence in online teaching will predict engagement when students have selfefficacy in learning.The third theory used in the included literature is the Tinto Integration Model.This theory conceptualizes the intention to persist in college as an outcome of deeper integration between student characteristics and their academic institution (Cabrera et al., 1992).Handagoon and Varma's research (2019) discusses the integration of internal and external factors that influence academic engagement.The result is that social support and selfefficacy predict psychological distress negatively.However, these two variables did not predict academic engagement.Likewise, the mediation model of a sense of belonging and psychological distress does not work in predicting academic engagement.The researcher assumes that there are Thai cultural factors that play a role in this research sample.
Other theories used are the Broaden-and-build theory and the Conservation of Resources (COR) Theory.Broaden-and-build explains that a positive psychological state and optimal well-being may provide a full variety of positive consequences (physically, mentally, and psychologically) in the education domain (Fredrickson, 2001).Benlahcene (2022) used one of the positive psychology variables that is rarely explored, which is flourishing.Flourishing tends to foster the goal-setting of students.The result showed that flourishing predicted four aspects of student engagement (behavioral, cognitive, emotional, and agentic) in the Malaysian context.Meanwhile, COR Theory conceptualizes that when an individual has to face actual threats of losing resources, they tend to experience psychological distress (Hobfoll, 2001).Cahyadi et al. (2021) explained that high levels of incivility at the workplace tend to influence one's behavior outside work, such as decreasing learning engagement and detachment from learning activities.
The second aim of this review is to explore predictors of student engagement in the context of Southeast Asian countries.Predictors found were from internal or external factors.Internal predictors that positively correlated to student engagement, such as flourishing (Benlahcene, 2022), locus of control (Cahyadi et al., 2021), personal best goals (Benlahcene, 2022;Benlahcene et al., 2022;Benlachene et al., 2020), basic psychological needs satisfaction (Benlahcene et al., 2020), course satisfaction (Baloran et al., 2021), knowledge sharing behavior (Hoi, 2021), and acculturation level (Tang et al., 2018).Environmental predictors that positively correlate to student engagement are teacher autonomy support (Benlahcene et al., 2020), teacher behavior (Lim et al., 2022), and the use of technology as a learning tool (Hoi, 2021).The negatively correlated predictor is workplace incivility (Cahyadi et al., 2021).In addition, there are several variables related to demographic factors, such as gender (Benlahcene et al., 2022), academic year (Benlahcene et al., 2022;Baloran et al., 2021), type of university (Lim et al., 2022), as well as the type of pre-university education (Tang et al., 2018) that influences student engagement.
In addition, several predictors give inconsistent results, namely social support (Benlahcene et al., 2022;Benlahcene et al., 2020;Handagoon & Varma, 2019) and self-efficacy (Rahim, 2022;Hoi, 2021;Handagoon & Varma, 2019).The study by Benlahcene et al. (2022) and Benlahcene et al. (2020) explained that teacher support predicts engagement.However, this is different from what was obtained by Handagoon and Varma (2019), where social support does not predict engagement.In addition, the self-efficacy variable does not predict engagement when positioned as an independent variable (Rahim, 2022;Handagoon & Varma, 2019).However, a significant correlation appears when self-efficacy acts as a moderator in the model (Rahim, 2022;Hoi, 2021).This systematic review examined 11 studies in the range of 2018 to 2022 to explain the overview of student engagement in Southeast Asia.As mentioned earlier, student engagement is a multidimensional construct that can be explored by many aspects of engagement.Several researchers use similar terms to describe student engagement, such as academic engagement (Handagoon & Varma, 2019) and learning engagement (Cahyadi et al., 2021).In addition, there are various theories used to understand the dynamics of student engagement in depth.There are two theoretical frameworks used in more than one study, namely Self-determination Theory (SDT) (Benlahcene et al., 2022;2020) and Social Cognitive Theory (SCT) (Rahim, 2022;Hoi, 2021).SDT explains that students will be involved if their basic psychological needs are met.Benlahcene et al. (2020) explained that engagement arises because of a learning environment that supports students' interests and volition.Furthermore, Benlahcene et al. (2020) found that students who had the opportunity to overcome learning difficulties and feel connected to their peers and teachers were more engaged overall.Meanwhile, students' opportunities to learn new things in class will only affect their involvement in behavioral, emotional, and cognitive terms.This study also found that students who feel given the freedom to be themselves will more easily express opinions and suggestions (agentic engagement).This is in line with the research by Benlahcene et al. (2022) that autonomy support for students will have an impact on agentic engagement.
The second theory used to understand the mechanisms of engagement is Social Cognitive Theory (SCT) (Rahim, 2022;Hoi, 2021).SCT emphasizes the critical role played by the social environment on motivation, learning, and self-regulation (Usher et al., 2019).Rahim (2022) and Hoi (2021) both examine engagement in the context of online learning.Hoi (2021) used contemporary learning media, namely Facebook.In this study, students also go through an interactive and collaborative learning process with lecturers and friends.The combination of technology integration and a supportive environment increases students' cognitive engagement.In contrast to Rahim (2022) which the samples are adult learners, lecturer competence in online teaching does not affect engagement.Hence, their level of self-efficacy increases their engagement in the classroom.
Apart from the two previous theories, there are several other theories used to describe engagement, such as the Tinto Integration Model (Handagoon & Varma, 2019), the Broaden-and-build Theory (Benlahcene, 2022), and the Conservation of Resources Theory (Cahyadi et al., 2021).These theories are used to determine the predictors of the engagement itself.Through the Tinto Integration Model, Handagoon and Varma (2019) explain the relationship between internal and external factors from individuals to involvement.The results show that the more social support and self-efficacy students have, the lower their levels of stress and anxiety will be.However, if it is associated with involvement the results have no effect.Benlahcene's research (2022) uses the Broadenand-build theory.The results explain that when a person has a high level of well-being it will encourage him to be more advanced.Students with high flourishing will be more involved in assignments, feel positive emotions, be able to self-regulate, and provide feedback on learning.Meanwhile, Cahyadi et al. (2021) used the Conservations of Resources Theory, which concluded that students who experience incivility at work will negatively affect their personal lives, including their engagement in the evening class.
Regarding predictors, there are variables related to individual characteristics and environmental factors that can affect student engagement.The researchers conclude that the variables that have been found can predict engagement positively or negatively.However, there are several variables related to demographic terms that would like to be discussed further.The first is about the age range.Generally, undergraduate students are in the range of 18 to 22 years.However, in some literature, the sample used is undergraduate students who are currently working and are over 26 years old (Rahim, 2022;Cahyadi et al., 2021).This causes some different engagement dynamics for each group.According to developmental theory, students between 18 to 25 are in the emerging adult stage; which is learning and exploring career paths, and possibilities in life, and becoming independent (Arnett, 2000).In contrast to the profile of working students, the purpose of their education is to complement their work experience with new knowledge (Rahim, 2022).This distinction leads to different predictors that affect engagement.For working students, the competence of lecturers in operating learning applications does not significantly affect their involvement (Rahim, 2022).Factors that influence their self-efficacy and motivation to study in college.In addition, incivility in the lecture class does not affect engagement.These results were obtained because they did not take this situation too seriously, compared to experiencing incivility at work (Cahyadi et al., 2021).
Meanwhile, for emerging adult students, there are different dynamics that emerge.Higher education is now dominated by Gen Z, which tends to have a shorter attention span but has frequent reliance on technology (Hampton et al., 2019).Baloran et al. (2021) and Hoi (2021) found that students wanted more interactions with lecturers and between students.In addition, integration with technologies such as Facebook and the Learning Management System (LMS) makes the learning process more enjoyable.According to Benlahcene et al. (2020), students who have access to learning many new things will have higher engagement.However, as the academic year progresses, there are changing forms of engagement as well.New students tend to be involved in studying material and doing coursework.Meanwhile, third and final-grade students are more engaged in discussions (Baloran et al., 2021).
Second, differences in the type of university and pre-university education also affect engagement indirectly.Research Lim et al. (2022) explained that students at state universities have higher expectations of lecturer teaching methods than students at private universities.This expectation is formed because the process of entering state universities is more difficult and students want their expectations to be met in the form of studentcentered learning.The same thing was found in the study of Tang et al. (2018), where pre-university education affects students' adaptability.Students who previously attended international schools will have a high level of acculturation.When they enter a global university with a high plurality, they will more easily adapt and be involved in the learning process.
Third, cultural factors also influence predictor mechanisms of engagement.In Handagoon and Varma's research (2019), social support did not affect engagement.This is not in line with previous studies which say that social support predicts engagement (Benlahcene et al., 2022;2020).This result is influenced by Thai culture which values support from the family (microsphere), not from educational institutions (microsphere).Meanwhile, in the study by Benlahcene et al. (2020), it was found that lecturers who are supportive and meet the needs for autonomy raise only agentic involvement, not in other forms.This autonomy is quite sensitive to cultural values that are closely related to individualist societies.While the samples of this study were Malaysian students, in which Malaysia adheres to a collectivist culture so that it upholds concern and harmony between individuals.

Table 1 .
Search Result in Five Database

Database Total Articles identified Total Articles that Do Not Meet Criteria Total Articles That Meet Criteria Article
, Utrecht Work Engagement Scale (Schaufeli et al., 2002), The Online Student Engagement Scale (Dixson, 2015), Cognitive Engagement (Wang et al., 2016), Student Engagement (Carini et al., 2006), and Student Engagement Instrument

Table 2 .
Overview Matrix