M-Learning Improves Knowledge Sharing Over e-Learning Platforms to Build Higher Education Students’ Social Capital

The education sector is transforming from e-learning to m-learning due to its motivation capabilities that drive students toward self-motivated learning. However, this requires online participation through quality knowledge sharing (KS) of the social capital (SC) of resources. Past-reported low student participation calls for confirming whether social capital affects KS in m-learning. Scholars assessed this effect—however, not in the education sector—and revealed scant clarification of whether m-learning regulates SC through KS. Hence, this is the motivated aim reflected through a conceptual framework, that is, tested using multiple-regression analysis after collecting data from 334 Ahlia University undergraduate business students. The results confirm that m-learning moderates students’ SC through KS in e-/m-learning on Moodle. Implications for theory, practice, and society are revealed in this article.


Introduction
Knowledge is defined as well-managed, structured, and categorized information that is accessible anytime, with the basic systemic code to knowledge management (KM), including explicit, implicit, and tacit knowledge. Explicit, tangible knowledge is published in documents such as evidence-based medicine. Tacit or intangible and inexpressible knowledge is gained through individuals' experiences, values, and instincts (Bose, 2003;Frappaolo, 2008;Killen et al., 2008;Lusignan & Robinson, 2007). Students share knowledge, a KM process, in exchange of information and skills in the form of knowledge between people (Dubé et al., 2006), while learning in classrooms, through interactions with instructors and peers to create social capital (SC) of resources, as stated by the SC theory (SCT) by C.-W. . The SC of resources is summed up resources which are accumulated by people in communities of practice (CoP) or communities of interest (CoI): a group of people joined by a common purpose for sharing wisdom and knowledge (H. H. Chang & Chuang, 2011). A virtually established CoP/CoI is referred to as a virtual community (VC; Razzaque & Eldabi, 2014). Thousands of such social networks exist online as self-support groups, where members discuss concerns, problems, or joint ventures in detail and develop relationships via email groups, discussion forums, and chat rooms (H. H. Chang & Chuang, 2011;Dubé et al., 2006;Eysenbach et al., 2004;Paul, 2006;Robertson, 2011).
The KM concept and its KS process blend with VCs, appreciating its value in the education sector, where scholars endorse e-learning. The growth of information and communication technologies allows learning to take place beyond classrooms, virtually, hereafter referred to as electronic learning (e-learning). Moreover, the widespread use of smartphones provides students unique opportunities to better their learning using mobile devices, such as tablets or personal digital assistants (PDAs), henceforth referred to as mobile learning (m-learning). Notably, scholars are cautioning from e-learning due to its restricted interest as students shy away from interacting with instructors/peers while e-/mlearning. A drop-in interaction causes the death of a VC. There is little evidence suggesting why this is the case for e-learning environments. It is no wonder that instructors struggle to enrich interaction, so students e-learn. Higher education institutions (HEIs) host millennials who are shy of interacting even though they are addicted to social media, and it is unclear whether students interact to make friends or to share knowledge. Scholars suspect the former (Arain et al., 2018;Camarda et al., 2017;Hwang, 2014;Kow et al., 2017;Kunthi et al., 2018;Sohrabi & Iraj, 2016;Yilmaz, 2016Yilmaz, , 2017. The m-learning concept, the current approach to access learning using mobile devices such as laptops, phones, and tablets, is utilized by most of the students to access the internet (Prema, 2012). M-learning promotes online self-study by accessing resources while exchanging knowledge with instructors/peers. M-learning occurs through online tools such as Moodle, WebCT, and Blackboard, along with traditional in-class teaching and learning (Abu-Al-Aish et al., 2012;V. Chang, 2016). For instance, as observed by the authors of this article, the HE instructors use Moodle to deliver the teaching material by sharing knowledge through structured resources such as videos or documents. The instructors attain students' feedback on Moodle as well as record their classroom attendance. Moodle is also used for accepting assignments that students upload and delivering students' online quizzes.
Scholars assessed the effect of VC SC on KS behavior, generalizing findings to contexts other than the HE sectors. For example, Chiu et al. (2006) assessed the relationship between SCT and Knowledge Sharing Quality (KSQ), along with other variables, by collecting data for quantitative analysis from IT-professional VC members, who shared their knowledge about database and programming. H. H. Chang and Chuang (2011), additionally, assessed the same relations, using randomly selected VC members as their study's target population. These include, among others, Harjanti and Noerchoidah (2017), Razzaque and Eldabi (2014), Wu and Hsu (2012), and Widén (2011). There is little research in the education sector assessing the role of students' KS behavior and what role SC (resources for accessing information/ knowledge) plays in this participatory behavior, in addition to how it is achieved when students virtually learn during m-learning. This is not surprising, considering that KS behavior is substantially understood in organizations but is still a lingering challenge faced by students too shy to participate online, despite learning being a collaborative activity. In HEIs, there is astounding evidence that students dislike engaging while learning, which presents a global academic challenge (e.g., Moylan & Razzaque, 2015;. Thus, HEIs face challenges as they strive to improve curriculum designs that could help motivate students to participate online by sharing quality knowledge as they e-/mlearn (Boateng et al., 2007;Liu & Li, 2012).
Accordingly, this study assesses the role of HE students' SC on KSQ while m-learning, that is, the moderation of m-learning on the relation between SCT and KSQ. A conceptual model (Figure 1) was quantitatively assessed on Ahlia University's undergraduate business students (the target population). Section "Literature Review" critiques the reviewed literature and proposes two hypotheses, following the conceptual model shown in Figure 1. The research methodology detailed in section "Research Methodology" describes the design and administration of a survey instrument for data collection, as well as how the collected data were analyzed. Section "Findings" details multiple-regression analysis results. Section "Discussion and Conclusion" discusses and presents a conclusion, while expressing the implications for theory, practice, and society, in addition to the limitations of this study.

Social Capital
Within VCs, SC is a sum of the resources available in network relations, for example, intellectual capital. SC refers to internal/inter-firm network-focused relations between customers and suppliers, attained through shared knowledge when acquired to stimulate innovation, for example, information exchange effective at an organizational and an individual level. It is the trust and reciprocity harvested in communal connections within network ties barring mutual goals (Gallego, 2010;Materne et al., 2017;Noprisson et al.,  2016 ;Petrou & Daskalopoulou, 2013;Pérez-Luño et al., 2011;Razzaque et al., 2013;Wu & Hsu, 2012). The SC of network-embedded resources facilitates intellectual capital to improve relations attained from KS; when tacit knowledge embraces personal qualities, though difficultly communicated in an organizational/academic institution, SC is quantified from the structural, relational, and cognitive dimension. The structural dimension reflects participants' links that imitate ties/connections. The relational dimension reflects the internal characteristics of relations built on interaction, trust, and reciprocity. The cognitive dimension defines how common language and vision benefits the sharing of resources within a network (Boateng et al., 2007;Gallego, 2010;Pérez-Luño et al., 2011;Razzaque, 2019;Razzaque et al., 2013;Wu & Hsu, 2012).

Social Capital as an Originator of KSQ
Knowledge sharing (KS) is a major concern in KM research, where KS gets hindered by insufficient communication or a lack of personal relations. The current research needs evidence and practical wisdom to generate and transfer knowledge through informal communication. Leaders use stories to share knowledge. Storytelling can be based on a problem statement (Bate & Robert, 2002;Girard & Lambert, 2007;Liu & Li, 2012). KS is possible when perceived personal benefits outweigh valuable knowledge loss. Two KM strategies enable knowledge transfer: codification, which is the storage and distribution of explicit knowledge in information systems (ISs), and personalization, the transfer of tacit knowledge through best practices. Technology facilitates explicit KS, while tacit KS is interpersonal through VC interactions, as it is invisible and intangible for its dynamic and actionable nature (Antonio & Lemos, 2010;Bate & Robert, 2002;H. H. Chang & Chuang, 2011;Razzaque & Eldabi, 2018).
KM and its KS process play a substantial role in HEIs, where KS enhances the SC of resources. However, e-learning research remains vague regarding how virtual teaching-learning cultivates such resources. Although empirical explanation expresses how awareness motivates KS, little evidence remains of cultivating factors of SC through KS while e-learning and now m-learning (referred to as e-/m-learning). This is when social networks harvest knowledge economies. It is vital to stimulate students' interest in KS with instructors/peers, as this research area has not been thoroughly explored, particularly for developing countries, such as in the Middle East. It is also vital to review HEI policies to enable innovative learning outcomes through creativity harnessed from e-/m-learning tools (Aubert & Reiffers, 2003;Forgeard & Kaufman, 2016;Hung et al., 2010;Liang & Chang, 2014;Liang et al., 2013;Nurunnabi, 2017;Sergis et al., 2018). Furthermore, there is not enough research that addresses the challenges that learners face while e-/m-learning through social media participation. There seems to be no evidence of what students learn online and whether they acquire knowledge. Instructors struggle to apply e-learning along with interactions; considering the insufficient quantitative evidence in HE e-/m-learning research, particularly on how online participations influence different circumstances (Camarda et al., 2017;Hwang, 2014;Kow et al., 2017;Kunthi et al., 2018;Sohrabi & Iraj, 2016;Yilmaz, 2016Yilmaz, , 2017. Tacit knowledge hinders KS due to its intangible flora. Knowledge holders require incentives to motivate them toward KS with those they trust while problem solving, which becomes the organizational culture. Student interactions evolve KS, which is vital for learning and academic performance and this is how SC is generated (Cumberland & Githens, 2010;Kim et al., 2015;Noprisson et al., 2016). Scholars studied KS for its integration, quality, quantity, and contribution and confirmed its benefits. However, there is not enough evidence of the moderating contribution of m-learning for the success of students' KS behavior, and research is yet to confirm the belief that students' SC affects their KS interactions and improves their academic performance (Liu & Li, 2012). Assessing the role of students' SC on KSQ is vital, since students shy away from online interaction, though it is a vital representative of successful learning through a collaborative effort. The curricula of HEIs are being redesigned to encourage KS interactions (a global academic challenge; Boateng et al., 2007). Based on this argument, the first hypothesis (Students' SCT positively and significantly affects their KSQ behavior) and its four subhypotheses (as shown in Figure 1) state the following in the context of virtual learning environments:

M-Learning Moderates Between Social Capital and Knowledge Sharing
M-learning is used from kindergarten to the HE level. Ample research has assessed the rate of acceptance and adoption of m-learning technologies in the education sector (e.g., Dumpit & Fernandez, 2017). M-learning influences both teaching and learning and has recently received abundant academic attention with rising adoption rates among HEIs since the advent of cloud computing. M-learning is popular for facilitating instructor/peer participation without physical networking for KS or collaboration. However, it faces constraints from the prism of mobility/communication, in addition to other restrictions such as battery life, environmental issues such as scalability/availability, and security issues such as privacy. This makes it imperative to improve the capability of mobile devices (Alghabban et al., 2017;Arain et al., 2018;Li & Wang, 2018). Moreover, m-learning is moderated to meet millennial students' demand for blending traditional in-class teaching-learning with e-/m-learning and social media participation for KS, accessible anytime and anywhere, and for cultivating SC while studying whenever and wherever. However, the underpinning theory behind using m-learning for teaching-learning remains underresearched, with a lack of pedagogical understanding (Li & Wang, 2018). There is also unclear evidence of both tangible and intangible benefits behind m-learning's moderation: to enhance learner participation to build SC of resources.
Scholars have assessed the extent to which m-learning is applicable, as mentioned previously, as well as its awareness. For example, Sobri et al. (2010) investigated the awareness, opinions, and requirements of 82 HE students who used m-learning-based learning practices. They revealed that substantial awareness and support was available from HEI students for the incorporation of such tools for teaching-learning strategies. In addition, Mao's (2014) assessment of 300 undergraduates from Southwest University revealed that 76% of learners were satisfied with m-learning. Of these students, 84% indicated that they preferred m-learning for problem solving. However, the literature reports mixed claims regarding educational technology utilization. For example, K. Lee et al. (2014) advise instructors to develop learning pedagogy in classrooms without the moderation of technology-supported learning. Even so, only a few studies have investigated academic perceptions of m-learning, despite them being imperative in achieving learning outcomes where m-learning overcomes the limitations of time and space of traditional teaching-learning, allowing independent learning (Jafari Navimipour & Zareie, 2015;Rowe et al., 2015). KS in m-learning enriches knowledge (Kunthi et al., 2018). Based on this argument, the second hypothesis (depicted in Figure 1) is as follows: Hypothesis 2: Students' m-learning moderates between their SC of resources and KSQ.

Research Methodology
The present study's deductive approach highlights two hypotheses (Figure 1) through two research questions: (a) What is the effect of students' SC on KSQ in online environments and (b) What is the moderating effect of m-learning between SC and KSQ. For hypothesis testing, a 36-item online survey was designed (Supplemental Appendix).

Instrument Design and Development
The four-part survey was introduced by a cover letter expressing the study's aim and ensuring voluntary and confidential participation. Next was a survey of three demographics items-gender, student status, and undergrad education level (Table 2)-which led to the third part: a 5-point Likerttype scale questionnaire, ranging from 1 (strongly agree) to 5 (strongly disagree). This included 17 SCT items adopted from Chiu et al. (2006), 5 KSQ items adopted from H. H. Chang and Chuang (2011), and 11 m-learning items adopted from Hunaiti (2012). These were deemed as suitable items for adoption since they were reported as reliable and valid by Chiu et al. (2006), H. H. Chang andChuang (2011), andHunaiti (2012). SCT was the independent variable, KSQ was the dependent variable, and m-learning was the moderating variable. This ensures the content validity of this study's instrument. The survey (Supplemental Appendix) was carried out via Google forms in November 2018, which took 2 months to collect data from 336 respondents. An online link for this survey was distributed to Ahlia University's Moodleusing business students through WhatsApp.

Sampling and Data Collection
Ethical approval was attained from the Deanship of Graduate Studies and Research Ethics Committee, Ahlia University, before survey distribution. The survey's reliability and validity were re-assessed to ensure a holistically fit model ( Figure  1) in the higher education context of 700 Ahlia University undergraduate business students, 336 of whom responded to the survey. The collected data were screened for missing data and outliers using SPSS v23. The sample of 336 exceeds the minimum required sample size (249) to generalize to the 700 undergraduate business students at Ahlia University with a 5% margin of error, 95% confidence level, and 50% response distribution (Raosoft Inc., 2004). This ensures the instrument's external validity, which refers to its generalizability from a sample to a population. Demographic information from the 336 responses (Table 1) reveals that most of the participants were female and held a full-time senior-level status. This was then followed by juniors, sophomores, and freshmen (Table 2).
The positive relationship between SCT and KSQ was enhanced when m-learning was introduced as a moderator, indicating a strong effect between the two: β = .655, p < .05. Interestingly, further tests revealed a decrease in significance when gender was introduced as a moderator: β = .655 dropped to β = .159, p < .05. Upon further investigation by distinctly assessing each moderating gender, root concern was discovered. When the male gender moderated, a positive relationship was revealed between SCT and KSQ as β increased to .568, p < .05. However, when the female gender was moderated, a negative relationship was expressed between SCT and KSQ, that is, β dropped to .506, p < .05. This indicated that the issue is specific to the female gender.

Discussion
This study was inspired by the research gap that was discovered, combined with the author's experience as an academic, with the aim of understanding the need to assess the role of students' SC in the quality of their KS behavior as regulated by their m-learning activities. At Ahlia University, business students are inclined to use cell phones for m-learning by accessing teaching/learning resources both inside and outside the classroom, via Moodle. Therefore, the research had such an aim because students shy away from participating in teaching/learning activities when classrooms are blended with e-/m-learning. This is also a global HEI concern, as expressed by others (Camarda et al., 2017;Hwang, 2014;Kow et al., 2017;Kunthi et al., 2018;Sohrabi & Iraj, 2016;Yilmaz, 2016Yilmaz, , 2017. Previous studies have assessed the effect of SC on KS for VC members. These scholars generalized findings in contexts other than the HE sectors. For example, Chiu et al. (2006) quantitatively assessed the relationship between SCT and KSQ among VC IT-professionals. Moreover, H. H. Chang and Chuang (2011) assessed this relationship between random VC member participants. Scholars reported that there is insufficient evidence of education sector research assessing how students' SC of resources affect their KS behavior, especially when they need to virtually learn and specifically while m-learning. Since this is the norm at Ahlia University, it is also the context of this study. It is an important research topic because while organizational KS behavior is substantially understood, it is an academic challenge as HEI students show reluctance in virtually participating in collaborative learning. Due to astounding evidence that students dislike engaging while learning, HEIs strive to design curricula that motivate KS  participation among peers (Boateng et al., 2007;Liu & Li, 2012). The model in Figure 1 is specially designed to tackle such an academic challenge-to understand why this is the case. As illustrated in this model, a correlation was observed in the study, that is, the relation between SCT and KSQ, and the moderating role of m-learning to assess its effect on SCT and KSQ when students engage in m-learn. An additional analysis with the gender moderated also confirmed that males are more influential than females in improving the relationship between SCT and KSQ when indulging in m-learning activities on Moodle. This study discovered a positive relationship between m-learning and SCT and KS, as evidenced by β = .601, p < .05; in other words, SC was found to be valuable in HEIs for KS. Therefore, HEIs can confidently set up standard VCs and promote them during lectures, which in turn can be promoted to students through superior curricula that facilitate KS to accelerate students' participation (Harjanti & Noerchoidah, 2017). The SC's social interaction ties, trust, and shared vision are endorsed differently, depending on the context (Widén, 2011). Through the positively significant role of SCT on KSQ, this study confirms the claims of scholars (i.e., Boateng et al., 2007;Kim et al., 2015;Liu & Li, 2012;Noprisson et al., 2016) that students' SC facilitates KSQ. The second quality of the model presented in Figure 1 is the moderation of m-learning between SCT and KSQ, expressing a moderation between SCT and KSQ. As revealed by Pearson (2016), almost 78% of instructors report using digital learning in combination with classroom teaching and learning strategies, as this is beneficial for e-learning flipped with traditional face-to-face learning in the classroom.
Additional tests were conducted on the model given in Figure 1 to assess the effect of gender on SCT → KS, indicating a negative relationship. The β of SCT and KS was .540, which dropped to .159 when both genders were introduced. To understand the reason, data pertaining to males were separated from those of females. With the moderation of the male gender, a positive relationship was found between SCT → KS, β = .568, p < .05. However, when the female gender was moderated, a negative relationship was revealed

Implications
The empirical findings of this study express the need for instructors to design curricula that encourage virtual participation in learning VCs through innovative regulation of m-learning activities. In other words, to appreciate the implications of practicing the model shown in Figure 1, it is vital to design curricula with both teaching and learning strategies that embed the concept of flipped learning with social media. Since students prefer blending social media with traditional classroom learning, the business students of Ahlia University in particular, who are not keen on interacting while e-learning, can enjoy flipped learning in their curricula (Elmasry et al., 2016;Khorsheed, 2015;Kidanu, 2018;Sohrabi & Iraj, 2016). In this scenario, what is traditionally done in classrooms can now be done at home and vice versa (Aubert & Reiffers, 2003;Sohrabi & Iraj, 2016). Flipped e-learningwhen instructors blend social media with e-learning and traditional learning in the curricula design-is appreciated by students, as reflected through appropriate learning outcomes. An example of social media e-learning is Khan Academy (Elmasry et al., 2016;Sohrabi & Iraj, 2016).

Limitation and Future Research
One major limitation faced in this study was that the model ( Figure 1) was empirically fitting when SCT's identity and shared language were excluded. The relations (identity → KSQ and shared language → KSQ) were insignificant and their exclusion allowed the model to reach its fitness: 188.733 at significance .00, as depicted in Table 6. In fact, this limitation can contribute toward the context of this study, as future research could investigate why these relations were insignificant, though important, in the SCT construct, as well as what moderator variable can establish/improve the two relations. Also, future researchers can adopt the instrument and the research design followed by this study to first, assess the present model longitudinally to identify changing traits of the target population at an individual and group level and second, the model could be assessed in other geographical locations, sectors, and other higher education stakeholders involved in e-learning and m-learning, such as instructors and managers.

Conclusion
The findings of this study bare social implications on the Gulf Corporation Council's (GCC's) developing countries, as they can be generalized beyond the business students of Ahlia University, Bahrain's private HEI. The HE sectors in developing countries recognize the importance of KS and KM tools, such as e-/m-learning platforms, for improving teaching-learning quality through the best practices to implement KM tools (J. Lee, 2018). Therefore, this research aimed to study how factors such as m-learning affected students' SCT and KSQ. Since knowledge economies are transforming into creative economies, with education being fundamental for human development, this was an important topic to explore. The Middle East lags in this area, and its current issues such as fluctuating oil prices and high unemployment require revamping of HEIs for quality human capital to fill approximately 40 million jobs in the GCC countries. These include Bahrain, Kuwait, Oman, United Arab Emirates, Saudi Arabia, and Qatar, with Bahrain ranking highest for its investments in higher education quality (Aubert & Reiffers, 2003;Deloitte, 2017;El-Khoury, 2015;Tadros, 2015).

Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.