Modelling the predictive effects of social media on the relationship between students’ engagement and academic achievement in higher education institutions

Abstract Social media is becoming more popular, and its resourcefulness has given students new opportunities to learn as well as broaden their academic horizons in various ways assumed if social media is handled appropriately. The main purpose of this study was positioned to estimate the predictive mediating effects of social media on the relationship between students’ engagement and academic achievement of undergraduate university students. In doing so, a correlation study design was employed. Data was collected from 395 participants who were selected through a simple random sampling technique. Structural equation modelling was used to analyze the study data. As a result, the study findings showed both social media and student engagement has a significant direct effect on academic achievement. However, when social media entered the structural relationship between students’ engagement and academic achievement as a mediator variable, the mediating effect was no longer significant, indicating social media played a direct-only non-mediation role. This may be due to social media allowing for students’ minimal engagement. The result demonstrates the need for critical consideration when social media and instruction incorporated into educational activities for students’ benefit. This suggests more research is required to confirm the all-round effects of how social media influence students’ engagement in learning, even though in some literature the results are encouraging. More initiatives are essential because they may reveal efficient intervention tactics for integrating social media into higher education systems.


Introduction
The educational landscape is currently undergoing a tremendous transformation due to the digital revolution (Dwamena et al., 2016). Social networking media and websites were created as a result of significant advancements in communication brought about by the internet's arrival in the 1990s. Therefore, socializing through the Internet has developed to become the norm for people all around the globe (Raza et al., 2017). Social media and networking websites transformed communication, and we now celebrate their advancements in fields like education and entertainment (Kolan & Dzandza, 2018). The once freely available public good of education is now going to be available to those who can afford to pay for specialized services and computer programs that require one's interaction with and exposure to technical tools and websites. Therefore, the designers of the new educational system must be aware of how technology is influencing changes in education through the incorporation of digital and social media websites for both instructors and students in higher education institutions (Rossini et al., 2021).
Social media is now playing a crucial part in how university students learn to provide a sustainable education. Social media's influence on sustainable education is becoming important and compelling (Abbas et al., 2019). The use of social media among the younger generations worldwide is steadily expanding. Social media is commonly used by youngsters of school age, and thus, it will have an impact on students' personal and academic lives (Sivakumar, 2020). Many social media and networking services, including Twitter, Facebook, LinkedIn, Telegram, Google Plus, Orkut, and Google Plus, and other use dynamic social contexts to make it simple for users to establish and maintain online communities through communication and social relationships (Raza et al., 2020). These networking possibilities encourage associations among groups, communities, and individuals with like-minded interests (Gupta & Bashir, 2018).
In addition to becoming the most important communication tool, social media became widely accepted and was employed at higher education institutions since the websites served as an instruction communication channel (Ekpe et al., 2018). The primary question, however, is whether college students and even teachers can adapt to and incorporate the new power of social mediadriven learning and educational technology that is being used in these institutions. As mentioned by Müller and Mildenberger (2021), to meet the demands of a digital society, higher education institutions must provide students with more flexibility and individualized education through the use of new technology and new social media platforms. It is said that because college students frequently have accounts on multiple social media platforms, integrating social media into the educational setting is made simpler. Higher education students' increased use of social media may be partially explained by their tendency to be highly connected when they first enrol (Alshuaibi et al., 2018;Kapila, 2021;Müller & Mildenberger, 2021).
Existing literature show the world has become a global village and social media use has made it a smaller world through social media and how this is changing instruction (Abbas et al., 2019). While social media usage is claimed to enhance instruction, active learning, and academic performance in dominant studies (Gupta & Bashir, 2018;Hassan et al., 2020;Kolan & Dzandza, 2018;Maroco et al., 2016), determining the effects of social media usage on the relationship between students' engagement and academic performance are still unstable. In the previous studies, it is becoming more obvious that social media provides university students with both huge advantages and threats (Abbas et al., 2019). This is because social media may influence students' academic performance either positively or negatively (Kolan & Dzandza, 2018). For instance, it was found that adopting social media as a learning tool can significantly improve students' engagement and academic performance because it enables them to be more involved and effectively incorporated into classroom activities for their benefit (Cassidy, 2016;Kolan & Dzandza, 2018) and encourage students' creativity as well as their learning interest and engagement (Delello et al., 2015).
More importantly, social media use provides the opportunity to continue learning during education affected by emergencies for instance during COVID-19  and armed violence . Furthermore, Raza et al. (2020) concluded that social benefits accrue as a result of students' use of social media as they can talk to their students, keep in touch with their friends, assist create new connections, and provide support. Because using social media satisfies their demands and helps them at every stage of life, students see it as a full platform. Social media has undoubtedly had a significant impact on how to engage and communicate with students since its utilization can improve students' learning experiences in academic settings.
On the other hand, many previous studies reported that even though social media platforms are becoming more and more popular, they are also getting riskier since they offer ways to put off doing students' studies (Dwamena et al., 2016). Many students spend numerous hours online (Dwamena et al., 2016;Ekpe et al., 2018) despite one of the most crucial instruments for developing teaching and learning methods that can adapt to the changing pedagogical and educational landscape in higher education institutions is the use of learning environments supported by social media (Cassidy, 2016). This implies social media's multitasking capabilities probably enhance distraction and occasionally hurt students' performance. Even it may lead to severe health problems unless the educational value of social media usage depends on its appropriate use (Azadi et al., 2021).
Considering the positive and negative sides of social media as stated in the aforementioned studies, the current study was positioned to confirm where the role of social media usage on the relationship between students' engagement and academic achievement resides in light of the inconsistent findings. As seen next, an overview of existing literature was made to give context about the current study variables because it is vital and helps to establish the study hypothesises.

Academic achievement
Academic achievement is the culmination of all of the knowledge that students have acquired through the teaching and learning process (Palominos et al., 2017;Triastuti, 2016). It indicates the student's ability to learn successfully, quickly, and deeply; and how well the educational process, which includes both indoor and classroom activities, is carried out in educational institutions.
Success in learning is the result of students' engagement capacity to learn for a predetermined period. For instance, Palmer et al. (2011) remarked that personal and dispositional engagement attributes including achievement motivation, conscientiousness, academic discipline, and dedication to study are influenced to varied degrees by the students' performance at university. According to earlier research, students' level of learning engagement was statistically shown to be a significant predictor of their satisfaction with instruction (Alshuaibi et al., 2018;Gaceri, 2021). Therfore, it is an essential part of supporting the teaching-learning process because it gets students ready for activities, and the greater students' learning engagement fosters their academic achievement (Kim & Choi, 2018). That is why learning achievement is well understood in the context of the relationship between learners' engagement and the support of other learning platforms including social media (Gupta & Bashir, 2018).

The relationship between students' engagement and academic achievement
Taking into account several definitions of the term "engagement" provided by scholars in prior studies, there is no consensus on what the word means (Alnuaimi, 2022). A synthesis of the literature done by Bailey et al. (2017) showed that definitions could be categorized under six basic categories, including self-engagement with performance, personal role engagement, work task or job engagement, multidimensional engagement, engagement as a composite attitudinal and behavioural construct, and engagement as a management practice. Despite varying definitions, engagement has emerged as one of the key ideas in organizational behaviour and students' learning, it has attracted a lot of attention in academic studies (Hanaysha, 2016).
Even though various expressions were used to describe the term engagement, what about student engagement in learning is the fundamental position of this review section. According to Carmona-Halty et al. (2019), the concept of engagement in learning relates to the learning activities that students carry out and is conceptualized as a pleasant, rewarding, work-related state of mind marked by vigour, dedication, and absorption. More specifically, different dimensions of engagement are mentioned in the literature. For instance, Schaufeli et al (2006) in the work engagement scale students' version, reflected that learning engagement can be understood in three dimentions such as vigour, dedication, and absorption. Vigour refers to high levels of energy and mental resilience while studying, the willingness to invest effort in studying, and persistence even when encountering difficulties; dedication refers to being strongly involved in one's studies and experiencing a sense of significance, enthusiasm, inspiration, pride, and challenge; and absorption refers to being fully concentrated and happily engrossed in what one is studying, where time passes quickly and one finds it difficult to detach him/herself from studying.
Additionally, Hassan et al. (2020) broke down the idea of learning engagement into behavioural and emotional components. Students who are engaged in learning consistently choose assignments that take the initiative to apply newly gained skills to new tasks beyond their level of competence. This involvement makes students positive thinkers, ecstatic and self-assured, and show a deep interest in curiosity about learning activities. Furthermore, Maroco et al. (2016), Fredericks and McCloskey (2012) assumed students' engagement as the cumulative perspective of behavioural, emotional, and cognitive dimensions. Besides, Alshuaibi et al. (2018) demonstrated theFredericks and McCloskey thought that the concept of behavioural engagement is based on involvement in extracurricular, social, or academic activities. When a student adheres to the rules, upholds the classroom standards, and refrains from engaging in disruptive behaviour, this is referred to as being behaviorally engaged; emotional engagement focuses on the extent of positive (and negative) reactions to teachers, classmates, academics, or school. When a student, for instance, identifies with the school or thinks he/she matters to the school, this is referred to as being emotionally involved; and cognitive engagement is defined as the student's level of investment in learning. It involves exercising the necessary effort to understand difficult concepts or master challenging abilities, as well as being deliberate, strategic, and eager to do so. When all the above-mentioned dimensions are taken into account, therefore, students' engagement can be seen as a gauge of institutional quality. As a result, it is the concern of institutions to deliberately create students' educationally engage through improving learning environments for students to achieve enough in their academics. This idea has become more popular in higher education institutions to promote excellent education, supporting students' engagement through face-to-face instruction, online learning, and using social media platforms in the blended form, which is the best input for students' academic success (Hassan et al., 2020). Based on this perspective, the following hypothesis was stated.
Hypothesis 1: Students' engagement has a significant direct effect on academic achievement.
When educators attempt to foster a climate of students' engagement, they do so by establishing specific circumstances that encourage participation (Alnuaimi, 2022). The engaged learner is conscious of their learning environment and collaborates with their peers to enhance academic success. It is anticipated that student's learning and development will improve as a result of their engagement, which causes them to have a deep dedication to or a strong connection to their learning (Wazir et al., 2021). Therefore, as mentioned by Dube et al. (2022), enabling students' behavioural, emotional, and cognitive engagement is crucial if we want them to become technology-infused learners and achieve sufficiently throughout their academic careers.

Social media and its relationship to students engagement and academic achievement
The term "social media" is made up of the words "social" and "media," which together indicate a person's intrinsic urge to connect with others through various venues (Delello et al., 2015). Social media usage is also referred to as online spaces that students use to connect, share, communicate, create or maintain connections with others for academic, entertainment, socializing and other purpose (Gupta & Bashir, 2018). Authors listed some examples of social media platforms and technologies such as blogs, wikis, media-sharing tools (audio, photo, video, text), networking platforms (including Facebook), and virtual worlds (Raza et al., 2020;Yu et al., 2022). Social networking as a communication tool has rapidly increased as a result of the success of mobile applications (Dwamena et al., 2016;Gupta & Bashir, 2018;Khan et al., 2022;Kolan & Dzandza, 2018).
Literature supported the idea that the learning environment is changing in education, moving away from the teacher and the institution toward one where students manage their education, gather their knowledge, and generate it by participating in networks outside of the classroom that is conducive to social media learning (Delello et al., 2015). A study conducted by Raza et al. (2017) provides fresh perceptions about university students' experience of using social media along with their experience of getting connected to the world in that the information can be transferred and utilized. Some of the dominant platforms such as Facebook, Yahoo, Twitter, WhatsApp, Meeting, and Telegram (Raza et al., 2017(Raza et al., , 2020 are just some of the social networking sites that have a lot of userswho employed it for forming relationships, enhancing cooperation, and facilitating the learning process (Qazi et al., 2022). This shows that social media are not just a network of people but also a developed network of varied people through which people and organizations may create profiles, share, and exchange information on a variety of interests and activities (Dwamena et al., 2016).
Social media can help students in many ways because of its capacity to strengthen bonds by making them more accessible. These advantages include giving them a virtual space to discuss their interests or problems with others who share them, supporting them academically, and enhancing their online communication skills and knowledge (Dwamena et al., 2016). Students can devote more time and effort to the content by sharing the ideas offered to them on social media, which furthers educational concepts and connects with them in fresh and relevant ways that traditional media cannot (Alshuaibi et al., 2018;Kapila, 2021).
More than facilitating the current educational setting, well-established social media usage helps to sustain education during uncertainties Assefa et al., 2023;Geng et al., 2022;Micah et al., 2023). Many educators are enthusiastic about the inclusion of social media into higher education institutions as a new educational tool that may be used to promote and improve students' learning experiences (Alshuaibi et al., 2018). It is believed that social media differs from more conventional forms of media because it enables students to engage with the course material both inside and outside of the classroom and to interact more intimately with their peers and teachers (Greenan, 2021) thereby their academic achievement become successfully (Gupta & Bashir, 2018). Considering this premise, therefore, the current study hypothesized that: Hypothesis 2: Students' social media usage has a direct effect on academic achievement. Previous research confirmed that using social media and networking sites for educational purposes has several advantages. For instance, Gupta and Bashir (2018) mentioned four importance of social media to education: (i) students can view educationally relevant films or share information about what they have seen and learned, this empowers students to engage in online discussions with teachers to continue the conversation; (ii) through exchanges on social networking sites, even teachers might pick up new skills from the pupils; (iii) a teacher can also keep an eye on students as they study, reflect, share, interact, and summarize debates; and (iv) social media offer a platform for connecting with classmates and teachers from anywhere and give the flexibility of long duty hours and finally, lead the student to achieve better in their academic.
Additionally, prior research tended to presumptively predicted that when students' engagement is raised as a result of using social media for learning, improved academic achievement will probably follow (Kolan & Dzandza, 2018). Since the advent of these social media platforms, students' academic lives have taken on a new dimension. Several studies have confirmed the significant impact that social media has on students learning engagement and academic success in higher education institutions, including how it can improve relationships, motivate students to learn, provide individualized course materials, and foster teamwork. Therefore, the current study hypothesized that: Hypothesis 3: Students' social media usage has a significant mediation effect on the relationship between students' engagement and academic achievement.
Likewise, Ekpe et al. (2018) pronounced that social media offers an opportunity for effective learning as well as for engaged contact and participation. Teachers and students can use the same platform to exchange and post fresh ideas and information, such as course materials, and receive feedback. For a teacher to find creative ways to use technology in the classroom, therefore, they must first find value in the technology itself, moving from "what social media is" to strategies on "how to use it" in the education system (Delello et al., 2015). Many students now use online and social networking sites as a primary platform for carrying out their everyday interpersonal contacts (Dwamena et al., 2016). Therefore, it's crucial to comprehend how social media use in higher education institutions affects students' learning. There is mounting evidence that multimedia content, tools, and streaming capabilities are successfully engaging students in new types of learning as a result of the social media revolution and its implications on academic success.

Study design and participants
The purpose of this study was to estimate the mediating role played by social media use on the relationship between students' engagement and academic achievement of undergraduate university students. A correlation study design was used in this process. The magnitude of the study variables' association and effect were assessed following the conceptual framework and previous theoretical investigations. Undergraduate students from three universities such as Bahir Dar, Woldia, and Wollo Universities, Ethiopia, were included as the study participants. The study sample size was determined by using a statistical method, considering Yamane's formula, n = N 1þN e ð Þ 2 where n = sample size, N = population size, and e = precision rate. Considering this into account at a probability of 95% confidence interval, a total of 395 participants selected through simple random sampling technique, were involved in the study.

Instrumentation and measurement
Based on the purpose of the current study, social media usage and students' engagement were considered as an independent variables, and academic achievement was considered as the dependent variable. The study's data collection instrument, which is the questionnaire, was adapted from existing scales to measure all these variables. Below, the details are presented about how the study variables were operationalized and measured.

Students' social media usage
Social media and networking sites are becoming more and more popular, and their versatility has given students new opportunities to learn as well as broaden their horizons and quality of life in various ways. It has been assumed that, if social media is handled appropriately, it can, despite some drawbacks, connect and mediate students' learning engagement and their academic accomplishment in several ways (Dwamena et al., 2016;Khan et al., 2022). Therefore, a social networking usage questionnaire developed and validated by Gupta and Bashir (2018) was used to measure students' social media usage and the role it plays on the relationship between their learning engagement and academic outcome.
Although several instruments were developed to measure social media usage across different contexts, the current study chose to use the Gupta and Basheer's scale because it was developed in the context of higher education and the majority of the constructs were found to be appropriate for measuring this variable. As a result, the social media usage measure, which has five dimensions was employed. These aspects include academics (seven items), socialization (six items), entertainment (four items), informativeness (three items), and constraints (four items). All of the items were created with a five-Likert scale measurement system, ranging from "1" representing "never" to "5" for "always". The measurement of social media usage was based on self-reported data, which can be biased because we assumed participants may not accurately report their social media use due to social desirability or memory biases. To minimise this dread, we employed the methods suggested in Latkin et al. (2017) work through pleasant wording and prefacing of questions, clearly defining the role of "study participant," and assessing and addressing motivations for socially desirable responses.

Students' engagement
Though there is much debate about the operationalization and dimensionality of students' engagement, measuring this construct is a key factor in academic achievement. The term engagement was operationally defined as the amount of time and energy put forth in educationally effective practices measured in three dimensions such as emotional, cognitive, and behavioural engagement. Each dimension has its indicators to be measured. Even though there are many existing students' learning engagement measurement scales developed and validated by scholars (Doğan, 2014;Schaufeli et al. 2006), the current study preferred to use the University Student Engagement Inventory (USEI) scale developed by Maroco et al. (2016) since its context is certainly related to university students' context as this study too. The USEI contained three dimensions: emotional, cognitive, and behavioural engagement. A final set of fifteen items, five items for each of the three dimensions was booked to measure students' learning engagement. All the items were rated in a five-scale measurement ranging from "1" representing "strongly disagree" to "5" for "strongly agree".

Academic achievement
At the higher education institutions where this study was focused, students' learning achievement is seen as the result of a learning process as a whole that can be completed within a specific time frame and can be quantified in terms of cumulative grade point average (CGPA) scores ranges from 0.00 until 4.00. The higher the CGPA, the better the academic performance a student has. The choice of using the CGPA to measure academic achievement seemed acceptable because it is the most often used objective indicator of academic performance/ability in many higher education institutions. Correspondingly, interested by the Alshuaibi et al. (2018) methods of subjective assessment, on a five-point scale, students were asked to rate how satisfied they were with their current CGPA, ranging from "1" representing "Strongly unsatisfied" to "5" for "Strongly satisfied." The subjective evaluation was used in addition to the CGPA's objective measurement because it was assumed that many students wouldn't want to reveal accurately such information.

Procedures
The entire data collection procedure was lasted for almost two months (from November 28, 2022, to December 29, 2022. In the process, the researchers' Institutional Review Board (IRB) granted permission to use the students' response. The related literature review was carried out to adapt the appropriate scale for data collection instruments, which was then reviewed several times before the main instrumentation procedure phase. To ensure inter-item reliability, a pilot test was done with 33 students who weren't part of the actual study sample. The questionnaire was then prepared and distributed to the selected students from each university. After confirming each participant's willingness to participate in the study, the researchers invited them to complete the printed survey at a time. To guarantee that the technical filling is understood by the students and data is obtained as logically as possible, a briefing was given to them. When participants had questions and troubles about any of the items, the necessary explanations were provided as well.

Data analysis
Responses from participants were noted following the response category. The missing data was handled by using the mean substitution technique (Schumacker & Lomax, 2016) since it works best when only a small number of missing values are present in the data. Structural equation modelling (SEM) supported by its variety of tools such as confirmatory factor analysis (CFA) and path analysis was used to analyze the data because it enables the estimation of multiple layers of the relationships between constructs (structural model) and relationships between indicators and their corresponding latent constructs (measurement model) simultaneously (Collier, 2020).
Since the study was done to estimate the direct and indirect (mediation) effects of the independent variables on the dependent one, a bootstrapping technique was used in the analysis. This is especially for testing the indirect effect (Awang, 2012). In addition to reporting the significant effects of one construct on the other from the structure model, the relative contribution or the effect size (R 2 ) of each path in the model (Collier, 2020;Awang, 2012) was estimated. Although the judgment of the level of R 2 depends on the specific research discipline, the current study regarded Cohen's d effect size which is categorized into three ranges such as the effect size values of below 0.13, between 0.13-0.26, and above 0.26 as having small, medium, and large effects, respectively (Collier, 2020;Hair et al., 2021). In doing all the statistical issues, statistical analysis software such as IBM SPSS 27 and AMOS 23 was used to calculate the data.

Results
Each of the 395 participants of the study received a copy of the questionnaire. Out of them, 374 returned their response and were taken into account for further analysis. However, the remaining 21 (5%) of the submissions were incomplete, and therefore, were discarded from the analysis because they were insufficient. Inspired by the methods suggested by Awang (2012), Collier (2020) and Hair et al. (2021) , the current study's result of the assessment of measurement and structural models were reported as follows.

Measurement model assessment
Assessments of unidimensionality, validity, and reliability were calculated to assess the measurement model used in this study (see Table 1). Since all items for each latent construct had excellent factor loading (the cutoff is > 0.5), the model's unidimensionality was attained. Regarding validity, Table 1 shows that all constructs met the necessary criteria, which is the Average Variance Extraction (AVE) > 0.5, indicating the reflecting measurement model's convergent validity was determined to be acceptable. The discriminant validity assessment of the correlation between exogenous constructs was also determined to be below the threshold point of 0.85, signifying the correlation between constructs is lower than 0.85. Additionally, as seen in Table 2, discriminant validity is obtained for all constructs when a diagonal value (in bold) is higher than the correlations of the corresponding construct values in its row and column (Awang, 2012). The square roots of the AVE, as shown in Table 2, were all higher than the correlations of the corresponding constructs, indicating sufficient discriminant validity was achieved.
Regarding the reliability assessment, both the internal consistency value of Cronbach's alpha and the composite reliability of each construct wereobtained. As with conventional internal consistency interpretation, the value of Cronbach's alpha should be > 0.7 (Collier, 2020;Awang, 2012) although above 0.6 was also acceptable (Hair et al., 2021). As observed in Table 1, all dimensions of the reflective constructs Cronbach's alpha value and composite reliability coefficients were above the recommended threshold value of 0.7, suggesting acceptable internal consistency and composite reliability. These results confirmed the items of the scales indeed measured the same underlying construct.

Model fitness measures
Based on the threshold suggested by Hair et al. (2021), Collier (2020), and Awang (2012) the current study model fitness assessment was estimated. The results are presented in Table 3, indicating that the required level of all fitness index categories was achieved.

Structural model assessment
Hypothesis 1: Students' engagement has a significant direct effect on academic achievement.
As Table 4 shows, the standard coefficient between students' engagement and academic achievement was found 0.256. According to the absolute critical value of 2.912, it can be settled that in a probability of 95% confidence interval, students' engagement had a significant direct effect on academic achievement (β = 0.256, p < 0.01) with an R 2 value of 0.066. This means due to the direct (unmediated) effect of students' engagement on academic achievement, when engagement goes up by one standard deviation, academic achievement goes up by 0.256 standard deviations. This is in addition to any indirect (mediated) effect that engagement may have on academic achievement. In addition, it was calculated that the coefficient of determination shows 6.6% of the variance in academic achievement could be estimated using the students' engagement which is in the small range.
Hypothesis 2: Students' social media usage has a significant direct effect on academic achievement.
As Table 5 shows, the standard coefficient between students' social media usage and academic achievement was 0.179, and according to the absolute critical value of 2.021, social media usage had a significant effect on academic achievement (β = 0.179, p < 0.05) with an R 2 value of 0.075. That is to say, the standardized direct (unmediated) effect of social media usage on academic achievement was .179. This indicates, due to the direct (unmediated) effect of social media usage on academic achievement, when social media usage goes up by one standard deviation, academic achievement goes up by 0.179 standard deviations. This is in addition to any indirect (mediated) effect that social media may have on academic achievement. Furthermore, the coefficient of determination showed 7.5% of the variance in academic achievement could be estimated using the students' social media usage and the effect size was found in the small range.
Hypothesis 3: Students' social media usage has a significant mediation effect on the relationship between students' engagement and academic achievement.
Depending on the data presented in Table 6, social media was entered into the students' engagement and academic achievement relationship model to see whether students' engagement has a significant indirect effect on academic achievement through the mediating role of social media played. It was confirmed in the literature that when the mediator variable enters the model, the value of beta coefficient for the first entered exogenous construct is expected to reduce, or in other words, the direct  (Awang, 2012;Collier, 2020). Hence, as shown in Tables 6, and 7, the standard coefficient of social media usage, which is the mediator in the relationship between engagement and academic achievement was 0.136. This mean, due to the indirect (mediated) effect of students' engagement on academic achievement, when engagement goes up by one standard deviation, academic achievement goes up by 0.136 standard deviations. Before social media usage entered the model, the direct effect of the first construct which is students' engagement on academic achievement was 0.256. However, when the mediator entered, the coefficient of determination of the indirect effect of students' engagement on academic achievement through social media usage was reduced to 0.136, indicating the indirect effect of engagement on academic achievement is no longer significant (β = .136, p > 0.05) with an R 2 value of 0.17. Since the coefficient of determination of the indirect effect is lower than the direct effect and the direct effect is significant but not the indirect effect, the mediation role of social media usage in the relationship is said to be what Hair et al. (2021) call a direct-only non-mediation role in which due to no significant effects of the mediation process.
Moreover, as shown in Table 6, the mediation path in the model had an R 2 value of 0.17, which means that the coefficient of determination shows 17% of the variance in academic achievement could be estimated using the combined effects of both students' engagement and social media usage, demonstrating the combined effect size is in the medium range.

Discussion
Examining the mediation role of social media usage on the association between students' engagement and academic achievement was the main goal of the current study. The results appeared to suggest that each exogenous construct has a significant direct effect on academic achievement,  but the effect size was small in range. Students' engagement and social media usage were hypothesized to be significant predictors of academic performance of university students under the study. Students' engagement and social media usage were found to significantly account for 6.6 percent and 7.5 percent of the variance in academic performance, respectively. However, the mediation path no longer had a significant effect on academic achievement. This may be due to social media allowing for students' minimal engagement., In previous study, it is suggested that contract and individualized teaching-learning methods are efficient approaches to improving students' learning capacities in higher education institutions. Therefore, students' engagement associated with other factors like social media usage is expected to improve their learning outcomes. Two factors contribute to students' learning engagement: an instructional strategy in which they assume responsibility for their education, and personal characteristic in which they assume responsibility for their motivation-based education (Triastuti, 2016). According to the results of the current study, student engagement significantly predicts academic achievement. In addition to the other fundamental competencies like motivation, general abilities, and innovation-specific competencies, this may be explained by the fact that students have an open mind and a desire to participate in their learning, which may have an impact on the learning outcome depending on the setting in which they can confidently learn their field of study.
The mediating effect of social media use on academic achievement was also investigated in this study. It was estimated that engaged students would benefit from using social media since it enables them to study from a variety of sources, enhancing their comprehension and mastery of the subject. However, the current study's findings indicated that when social media use was included as a mediator variable, the coefficient of determination of students' engagement decreased. This suggests that the relationship between student engagement and academic achievement was not significant through the use of social media. As a result, the current study is consistent with earlier findings, which suggested that social media was blamed for the decline in educational standards and the production of graduates of subpar quality (Ekpe et al., 2018). More research is required to confirm the many effects of student engagement components and how social media influence students' participation in learning, even though in some reports the results are encouraging (Alshuaibi et al., 2018). More initiatives are essential because they may reveal efficient intervention tactics for integrating social media into higher education.
The modest mediation effects of social media use in the relationship between students' engagement and academic achievement in the current study may be influenced by several factors. This may be because engaged students perform well in the course overall, which implies that regardless of their use of social media, students are capable of fundamentals for learning. The disposition for learning is a key concept that cannot be overlooked in any discussion about learner engagement. Individuals should possess this quality or have previously developed it before engaging in any learning activity if they want to reap the greatest rewards for their time and effort. This includes some dominant constructs, namely: a desire to learn; a positive attitude toward the learning situation; a willingness to make the investment of time and effort that is necessary for learning, the ability to persevere, an understanding of the importance and value of learning (Chorrojprasert, 2020). More crucially, Gupta and Bashir (2018), Kolan and Dzandza (2018) presupposed that social media is undeniably an important tool in human life and would continue to be so as far as communication is concerned. Today, humanity is reaping enormous benefits from its existence in both the majority of scholarly pursuits and simple communication.
These social media networks have been used extensively to support and, to some extent, facilitate many forms of education. As opposed to earlier, getting information from friends, lecturers, or specialists both locally and abroad is no longer difficult, and the internet is the driving force behind this achievement. Social media users, especially students, must be aware of its risks and take caution when using it because it is "a good servant but a dangerous master" and can also be "characterized as a two-edged sword." The results of the study show that despite social media having minimal effects to engage students, technology-enabled is productive for educational organizations (Al Halbusi et al., 2022;Li et al., 2022;NeJhaddadgar et al., 2020), students can derive from social media networks, such as information sharing, relationship building, and participating in group discussions from near and far, among others. Parallelly, there is to some extent addiction and distraction of attention caused by the use of social media, which could have serious consequences on academic performance. Hence, students' technology-driven behaviour (Li et al., 2022) can be supported to sustain their learning engagement practices. Because as mentioned in Rahmat et al. (2022) work, technology-infused and social media-supported learning could be a solution to proceed with education not only in normal circumstances but also in the advent of natural and man-made crises may bring a global crisis that resulted in the closure of universities and educational institutions.

Conclusion
The current study confirmed that both social media and students' engagement have a significant direct effect on academic achievement. Since these exogenous variables can improve students' learning experiences, their significance is widely acknowledged in the academic setting. However, when social media entered the structural relationship between students' engagement and academic achievement, the mediating effect was no longer significant, indicating social media played a directonly non-mediation role. This may be due to social media allowing for students' minimal engagement in their learning. This suggests more research is required to confirm the many effects of student engagement and how social media influence students' learning success, even though in some reports the results are encouraging. More initiatives are essential because they may reveal efficient and comprehensive intervention tactics for integrating social media into higher education systems.

Implications
This study could have the following implications for both the academic and scientific community to inform future research, educational practices, and the need for policies related to social media use in education. First, theoretical knowledge about how student learning engagement and academic achievement are associated in the context of higher education institutions can be improved by looking at the mediating impact of social media usage. As to the context of this study, due to the current government's bold demand for "Digital Ethiopia," many universities in Ethiopia consider enhancing technologysupported learning environments. To ascertain students' learning outcomes in this process, it is crucial to look into the potential of students' social media usage and their engagement in learning during the technological age. Therefore, the current study contributes to the body of knowledge that it could aid in igniting further discussions about how to engage students to support their academic success through social media-infused learning as the future of learning and teaching is going to be shaped especially in the context of the Ethiopian higher education system. Second, to ensure that students are willing to advance using the online learning system supported through various social media platforms, the results of this study will serve as a caution and reference that without critical consideration, social media is not always helping to improve student learning performance and preparation for e-learning. It informs the need for future research to fully understand the role that social media usage plays in determining the relationship and impacts between students' learning engagement and academic accomplishment, even if this concept is not new.

Recommendations
The current study makes the following recommendations: Students are easily made addicted to social media if it is not properly controlled. Students may then experience several health problems such as poor sleep, eye tiredness, negative body image, depression, anxiety, and other issues that have a significant negative impact on their ability to learn and perform well academically. These problems are getting worse today since Ethiopian higher education institutions offer free wi-fi Internet access in their compound. Therefore, it is advised that higher education institutions educate students about the harmful impacts of social media and how to reroute them, particularly through offline, mindful activities. To keep students away from harmful social media platforms that might not have a big impact on their learning, some restrictions should be taken into account.
Additionally, it is advised that university lecturers and professors incorporate social media use into educational activities to improve students' learning engagement and academic performance by assisting them in developing appropriate social media usage skills, which prepares them for future e-learning-supported educational activities.

Limitations of the study and future study suggestions
The study has some restrictions: First, because just three public universities were used for the study samples, the respondents in the curent study are not necessarily representative of the entire student found in the country's public universities, and the findings might not be accurate for all at large. Despite these drawbacks, the results of this study could help guide concerned organizations regarding the significance of taking social media usage into account in the process of the education system.
Second, the measurement of study variables, for instance, social media usage was based on selfreported data, which may be biased since participants may not accurately report their social media use due to social desirability or memory biases. Also, the study relied on a single instrument to measure social media usage (Gupta & Bashir, 2018) and student engagement (University Student Engagement Inventory). While these instruments are validated, we suggest further study using multiple measures which could provide a more comprehensive picture of the constructs being studied. Additionally, the current study only measured students' academic achievement using cumulative grade point average (CGPA) and a subjective assessment of satisfaction with CGPA. More importantly, this study suggests further investigation using other measures of academic achievements, such as standardized test scores or completion rates which could provide additional insights.
Third, while the current study discussed the relationships between the study constructs, it did not thoroughly examine the various characteristics influencing the overall learning environment of students' engagement, social media, and the associated factors which may have their impact on explaining academic achievement. Hence, further study is suggested to determine academic achievement through consideration of various factors and constructs. Also, to gather data for the study, the questionnaire was used mostly. As a result, it was insufficient to fully capture students' perspectives on the relationship between study variables. Therefore, a replication of this study using alternative data collection techniques is advised to corroborate the results and further understand the study constructs.