Exploring Mobile Learning Adoption in Higher Education: A UTAUT2-Based Study with Technostress and Exhaustion as Mediators in Student Engagement

Abstract

The primary objective of this study is to examine factors that influence Mobile Learning adoption and effectiveness in the case of higher education, underpinned by the Unified Theory of Acceptance and Use of Technology 2 (UTAUT2) model. The study was conducted at some of the top Saudi universities. A survey method was used in this study, and the questionnaire includes 12 factors, which include enjoyment, hedonic motivation, and performance expectancy, among others, whose relationships with behavioral intention towards using m-learning and in turn, its actual use, were investigated. Additionally, two mediating factors, namely technostress and exhaustion, were included to examine the effects of students’ engagement with m-learning. The questionnaire was distributed to 500 undergraduate students, and 264 replied. Based on the findings, the traditional UTAUT2 factors had significant effects on students’ intention and use behaviors, which were mediated by technostress and exhaustion. The findings indicated that the effective management of mediators is important to improve student engagement in m-learning and that m-learning strategies can be developed through the information provided. This study validated the UTAUT2 model in the case of m-learning, with a 70% success rate in predicting IT adoption, which exceeded other models. Through the study findings, educators and policymakers can find ways to optimize m-learning platforms. Furthermore, it is recommended that focus is placed on the long-term effects of technostress and exhaustion on the students’ adoption of and engagement with m-learning.

1. Introduction

The integration of technology in higher learning institutions (HLIs) has become essential for enhancing teaching and learning processes, improving accessibility, and supporting sustainable academic development. Universities around the world increasingly rely on digital tools to deliver flexible, interactive, and student-centered learning experiences [1,2]. Technology-enhanced learning environments provide opportunities for improved engagement and more efficient communication between students and instructors, although challenges such as digital disparities, technological readiness, and cognitive overload persist [3,4].

Within this broader digital transformation, m-learning has gained prominence as an innovative extension of technology-supported education. M-learning enables students to access course materials at any time and from any location, supporting autonomous, collaborative, and flexible learning practices [5,6]. The advantages of m-learning include increased convenience, improved learner motivation, and opportunities for personalized learning [7]. Nevertheless, institutions still encounter challenges such as the risk of student distraction, variations in the quality of mobile learning content, and the need for reliable technological and pedagogical support [8,9].

The widespread ownership and daily use of mobile devices—such as smartphones and tablets—have significantly strengthened the potential of m-learning in higher education. Advances in device capabilities and mobile applications have improved the accessibility and effectiveness of technology-mediated learning [10]. Students frequently use mobile devices to communicate, access digital resources, engage in class activities, and manage academic tasks, making such devices central to modern educational practices [11]. This widespread adoption of mobile technologies offers higher education institutions a strong foundation for implementing comprehensive m-learning strategies.

In Saudi Arabia, major national initiatives under Vision 2030 have accelerated the adoption of digital technologies in higher education. Universities have expanded their use of online platforms, mobile applications, and digital learning systems to enhance instructional quality and learning outcomes [11,12]. Saudi university students demonstrate high levels of smartphone usage, presenting significant opportunities for m-learning to enhance engagement and support flexible learning pathways. However, despite this progress, limited research has examined the determinants of m-learning adoption in the Saudi context, particularly concerning psychological factors such as technostress and exhaustion that may influence students’ sustained engagement [13,14].

Technology integration has become a common strategy in improving the students’ learning experience in the dynamic higher education field. M-learning, a type of technological innovation in education, has been the focus of considerable attention due to its contribution to the transformation of traditional classroom environments. This study aims to explore the effectiveness of m-learning adoption in a Saudi university among undergraduate students, using the UTAUT2 as the underpinning theoretical foundation. In the current era riddled with dynamic technological developments, exploring the factors influencing the acceptance of m-learning platforms among students is necessary. This study further examines new mediating effects of technostress and exhaustion, providing an enriching view of the factors’ mediation on the students’ engagement with m-learning platforms. By examining UTAUT2 factors and their influence over behavioral intentions and actual use of m-learning platforms, this study is expected to provide insights into and implications for educators, educational institutions and policymakers concerned with forming future m-learning strategies towards better student experiences.

Although numerous studies have examined mobile learning adoption in higher education, most have focused primarily on traditional UTAUT or UTAUT2 predictors without accounting for psychological strains that emerge in digitally intensive environments. This study extends the UTAUT2 framework by integrating technostress and exhaustion as mediating factors, offering a more holistic explanation of students’ behavioral intention and engagement with m-learning. Unlike prior models that emphasize technological or pedagogical determinants, this study highlights the psychological mechanisms that influence sustainable m-learning use, particularly within the Saudi higher education context. Furthermore, the proposed model explains 66.2% of the variance in behavioral intention—exceeding many previous studies—and provides empirical evidence on how psychological stressors partially mediate engagement outcomes. These contributions position the study as an advancement over existing m-learning approaches, especially by linking technology adoption with students’ well-being and long-term engagement.

The main contributions of this study are as follows. First, the study extends the UTAUT2 model by integrating technostress and exhaustion as mediating psychological variables, offering a more comprehensive perspective on m-learning adoption than prior work focused primarily on technological and behavioral determinants. Second, the study provides empirical insights into mobile learning adoption among undergraduate students in Saudi higher education, a context where research remains limited despite growing national initiatives in digital transformation. Third, the proposed model demonstrates strong explanatory power by accounting for 66.2% of the variance in behavioral intention, thereby advancing current knowledge on factors that influence sustained engagement with m-learning platforms. Together, these contributions enrich theoretical understanding and offer practical guidance for institutions seeking to enhance the effectiveness and sustainability of m-learning initiatives.

The remainder of this paper is organized as follows. Section 2 presents a comprehensive review of the literature on mobile learning and the technological factors influencing its adoption in higher education. Section 3 describes the research methodology, including the conceptual model, constructs, and data collection procedures. Section 4 reports the results of the measurement and structural model analyses. Section 5 provides a detailed discussion and interpretation of the findings. Section 6 outlines the theoretical and practical contributions of the study, while Section 7 offers recommendations and directions for future research. Finally, Section 8 concludes the paper by summarizing the key insights and implications of the study.

2. Literature Overview

This section provides an overview of the literature relevant to the adoption and use of mobile learning (m-learning) in higher education. To introduce the broader technological context of m-learning, Section 2.1 outlines the opportunities and challenges associated with mobile learning technologies. This establishes the technological landscape within which mobile-based instructional practices have emerged. Following this, Section 2.2 presents prior empirical studies on m-learning adoption in higher education, with particular attention to research that has employed TAM, UTAUT, and UTAUT2 models. These studies form the core related work for the present research and highlight the determinants commonly examined in m-learning adoption literature.

2.1. Technologies in Mobile Learning: Challenges and Opportunities

Technology integration in m-learning indicates the dynamic and ever-changing landscape in education, paving the way for further opportunities to enhance the pedagogical experience of students [15,16]. This provides flexibility, tailor-made learning experiences and new engagement and collaboration modalities. Nevertheless, these benefits are not without their related challenges, including the digital divide, potential distractions, issues relating to content credibility, and data security. These challenges must be addressed to fully reap the benefits of mobile learning technologies and ensure an equitable and efficient learner experience [3,8,17].

The present educational landscape is riddled with technological advancement, heralding a paradigm shift towards mobile learning technologies that provide countless opportunities with their related challenges [18,19]. This pedagogical method, which depends on ubiquitous mobile technologies, creates a transformative learning experience while guiding accessibility, engagement, and digital equity in the learning process [9].

One of the foremost opportunities that m-learning presents is its unparalleled accessibility and flexibility. With the widespread use of smartphones and tablets, barriers to geography have been brought down, leading the way for democratized access to education. The saliency of this situation relates to individuals who are limited by logistics and physical disabilities in their access of traditional educational environments. Moreover, the introduction of m-learning technologies marks the era of personalized education, through which adaptive applications and platforms have made it possible to tailor-make learning and styles to each individual, thus promoting experiences of learning efficaciousness [4,20].

Concerning the above, multimedia integration into m-learning platforms in the form of videos, podcasts and interactive quizzes significantly contributes to enhanced engagement and knowledge retention of learners [21,22]. Similarly, the inclusion of gamification aspects like badges and leaderboards motivates the learners through a facilitated competitive and motivational interaction activities, thus enhancing engagement and enjoyment in learning. Mobile technologies are essentially the bridge that promotes collaboration and social learning, learners–educators interaction and a vibrant learning community that goes beyond the boundaries of traditional classroom environments [23,24,25].

Notably, m-learning-related challenges have multiple facets and levels of importance, with the top one being the digital divide—students in less affluent regions often lack access to technological resources and internet connectivity, and educational inequities continue. Added to this, mobile devices can sometimes be more of a distraction considering the social media and games within, and this could direct attention to something else other than the educational curriculum and content [26,27,28].

Online content quality and credibility are also issues in m-learning related to the amount of content available. This calls for strict measures to guarantee that learners are accessing material and content that are accurate and reliable. Concerning this, online learning activities have magnified data privacy and security issues, and as such, protecting sensitive information from being accessed and cyber-threatened without authority is a continuous challenge that begs for vigilance and the establishment of stringent measures [29,30,31,32].

Table 1. Shows the related works outcomes.

Study	Country	Model/Framework	Key Variables/Factors	Main Findings	Opportunities Identified	Challenges Identified
[33]	Saudi Arabia	UTAUT	PE, EE, SI, FC	PE, EE, SI significant; FC not significant	M-learning improves perceived usefulness and efficiency	Institutional support (FC) insufficient to drive adoption
[34]	Vietnam	UTAUT + Attitudes	PE, EE, SI, FC, Attitude	All factors significantly predicted intention	Mobile-assisted language learning increases flexibility	Requires strong facilitating conditions for sustainability
[12]	Saudi Arabia	Extended TAM (PI, TTF)	PU, PEOU, PI, TTF	All variables positively influenced satisfaction and intention	Useful in remote/online learning contexts	User readiness and content quality vary
[35]	Jordan	Extended ISSM	System quality, content quality, service quality, institutional support	Quality and institutional factors strongly predicted usage	Quality systems enhance m-learning acceptance	Low system/content quality limits engagement
[36]	Saudi Arabia	Hybrid TAM + ISSM	SQ, SEQ, PU, PEOU	Info quality not significant; others significant	Tailored mobile platforms can enhance engagement	Students sensitive to poor information quality
[37]	Iraq	UTAUT	SI, EE, PE	Gender moderated key relationships	Social norms shape mobile learning use	Gender differences require differentiated strategies
[38]	Spain	UTAUT2	HM, Habit, PE	Strong predictors of satisfaction and intention	Hedonic motivation increases long-term use	Need to sustain continuous engagement
[39]	Greece	UTAUT2	Habit, PE, HM	Habit and HM strongly predicted usage	Frequent mobile use supports learning continuity	Risk of distraction from non-learning apps
[10]	Meta-analysis	Device integration	Mobile learning tools, device use	Mobile use improved learning outcomes	Multimodal learning improves performance	Device limitations (battery, screen size)
[4]	Saudi Arabia	M-learning acceptance	PE, EE, SI	Strong acceptance among students	Flexible and adaptive learning opportunities	Digital divide persists
[9]	Egypt	Mobile learning + TAM	PU, PEOU	PU and PEOU influenced adoption	Helpful in emergencies; supports continuity	Institutional support gaps
[24]	Hong Kong	Mobile chatbot learning	Interaction, autonomy, feedback	Positive perceptions; increased self-directed learning	Chatbots enrich learning interaction	Dependence on reliable technology
[20]	Systematic review	M-learning methodologies	Engagement, active learning	Supports experiential learning	Enhances hands-on, active learning	Requires high-quality instructional design
[32]	India	AR-based m-learning	AR engagement, cognitive involvement	AR boosts motivation and engagement	Immersive learning enhances outcomes	Privacy and data concerns remain significant
[26]	Greece	UTAUT2	Learning value, PE, EE	Value and expectancy predicted intention	M-learning increases autonomy and participation	High distraction risk
[24]	Hong Kong	Mobile learning tools	Gamification, learner interaction	Gamification improves engagement	Increased learner motivation	Needs careful instructional alignment

shows some of the related works on the topic.

2.2. Mobile Learning in Higher Education

Education plays a central role in supporting human development, and the integration of technologies such as m-learning has become increasingly important for enhancing teaching and learning processes [19,40]. These digital tools support learners and instructors by improving access, flexibility, and instructional efficiency [41]. Nevertheless, the factors influencing such use are still under-examined. Based on Alfalah [33] findings, m-learning management systems (m-LMSs) use among university students is affected by several factors, namely performance expectancy, effort expectancy, and lecturer’s influence, but facilitating conditions had no significant impact. The author used the UTAUT model, and the study focused on the effect of the COVID-19 pandemic on education. Further findings of the same study revealed that perceived mobile value and academic relevance significantly influenced performance expectancy, while university management support significantly impacted effort expectancy. The author’s findings extended the literature on technology adoption in education, specifically the adoption of m-learning during and after the COVID-19 pandemic.

In a related study, Romero-Rodríguez and Aznar-Díaz [42] conducted an extensive analysis of the m-learning adoption and influence in higher education in Spain, using a survey distributed to 1544 university professors throughout 59 universities in Spain. The research examined the level to which mobile tools are used in university teaching practices. It employed SEM to test ten hypothetical relationships considering the socio-demographic factors influencing m-learning effectiveness. Based on the study’s significant findings, there was a high rate of mobile device use in teaching (73%), which stressed the increasing importance of this type of technology in education. The study also highlighted the critical factors contributing to the development of best practices in teaching using m-learning, having implications for educators, policymakers and researchers in the educational technology field.

Also, the factors that influence the acceptance and use of mobile devices in Vietnam were determined in a Hoi [35] study. The author used the UTAUT as the study framework and added some modifications concerning the learners’ attitudes. A survey was used to collect data from 293 higher educational learners in Vietnam, after which data was exposed to rigorous analysis using the Rasch-based path model. The findings revealed the way performance expectancy, effort expectancy, social influence and facilitating conditions, and learners’ attitudes influence the students’ intentions and behaviors towards using mobile-assisted language learning (MALL). The findings significantly provided information on mobile technology adoption in language education, specifically in Vietnam and developing countries.

Moving on to another related study, the effect of the COVID-19 pandemic on adopting m-learning in higher education was the focus of Alturki and Aldraiweesh’s [12] study focused on the students enrolled in King Saud University in Saudi Arabia. In addition, the authors used extended TAM (which included additional variables of personal innovativeness and task-technology fit) and the original constructs. A survey questionnaire was developed, and copies were distributed to 300 students; based on the findings, the factors had positive correlations with the satisfaction and behavioral intentions of students towards m-learning. The study provided knowledge on the effectiveness and acceptance of m-learning in the context of remote teaching and learning, highlighting its crucial role in education sustainability during a pandemic.

Another study in the topic with an extended model is Althunibat, Almaiah [35], which looked into the crucial factors influencing the use of m-learning applications in higher education institutions during the COVID-19 pandemic. The study specifically extended the Information System Success Model (ISS) through additional factors, including top management support, institutional policy, organizational structure and change management, along with system quality, service quality, content quality, functionality, design quality, usability, and their effect on the satisfaction and use of m-learning systems among students in five Jordanian universities. The quantitative study findings revealed how educational environmental factors and quality factors significantly affected the quality and actual use of m-learning applications.

Furthermore, Drwish and Al-Dokhny [36] analyzed the influencing factors of m-learning’s actual use among higher education students following the pandemic period through a novel hybrid model which combined ISSM elements with those of TAM. The authors examined the effect of quality factors and perceived elements on students’ satisfaction and actual use of ML. The survey questionnaire was distributed to 400 undergraduate and postgraduate students in four universities in Saudi Arabia. Data was exposed to SEM analysis, and based on the major findings, information quality had no significant impact on m-learning use, but other factors had positive effects on it, and these included system quality, service quality, satisfaction, perceived usefulness, and perceived ease of use.

In the same study caliber, Izkair and Lakulu [37] examined the moderating role of gender in the relationship between several factors and m-learning use among higher education students in Iraq. The authors employed the UTAUT model, with the relevant factors being social influence, effort expectancy, performance expectancy, and intention towards m-learning use. Three hundred and three participating students from Iraqi universities were surveyed, and based on the obtained findings, gender had a moderating role in the relationships of UTAUT factors and intention towards m-learning use.

Past studies have recommended determining the factors influencing m-learning use and engagement among higher education students [43,44,45,46].

2.3. Innovation of the Present Study

Although numerous studies have investigated mobile learning adoption in higher education using frameworks such as TAM, UTAUT, and UTAUT2 [12,33,34], most have focused primarily on technological and behavioral determinants such as performance expectancy, effort expectancy, social influence, and facilitating conditions. While these factors have contributed substantially to the understanding of m-learning adoption, the existing literature has largely overlooked the psychological strains associated with technology use—particularly technostress and exhaustion—which are increasingly relevant in digitally intensive learning environments.

To address this gap, the present study introduces two key innovations. First, it extends the UTAUT2 framework by incorporating technostress and exhaustion as mediating psychological constructs. Prior work has shown that technostress and related forms of strain can adversely affect users’ well-being, satisfaction, and continuance intentions [47,48,49]. More recent studies in educational settings similarly highlight the need to consider emotional and cognitive overload when evaluating technology adoption [14,50]. By integrating these constructs into UTAUT2, this study provides a more comprehensive explanation of how psychological mechanisms influence students’ intention to use and engage with m-learning platforms.

Second, this research contributes context-specific insights by examining m-learning adoption within Saudi Arabian higher education, where empirical studies remain limited despite rapid national digital transformation efforts [11,12,13]. The inclusion of technostress and exhaustion within this context provides a novel lens for understanding the sustainability of m-learning practices among undergraduate students.

Although UTAUT2 has been widely applied to explain technology adoption behaviors, the model primarily focuses on cognitive and behavioral determinants such as performance expectancy, effort expectancy, and hedonic motivation. However, extensive research on technostress and digital strain argues that technology use also generates psychological demands that influence user engagement, satisfaction, and long-term adoption [48,49]. These studies indicate that emotional exhaustion, cognitive overload, and technology-induced stress can reduce intention to use and impede performance, yet such mechanisms are not incorporated into UTAUT2. This reveals a clear theoretical gap; the dominant acceptance models overlook the affective and strain-related processes that accompany mobile learning environments. Integrating technostress and exhaustion into UTAUT2 therefore allows for a more comprehensive explanation of m-learning adoption by capturing both the motivational factors emphasized in acceptance theory and the psychological constraints emphasized in technostress theory. This synthesis provides a theoretically grounded rationale for the proposed model rather than a purely additive extension.

Together, these contributions position the current study as an advancement over prior m-learning approaches by linking technology acceptance with learner well-being and long-term engagement.

On the other hand, research on human–technology interaction increasingly acknowledges that emotional, motivational, and strain-based constructs often coexist and influence one another dynamically rather than in isolation. Theoretical frameworks such as technostress theory [49], cognitive load theory [51], and technology acceptance models [52] concur that feelings of overload and strain frequently co-occur with changes in intrinsic motivation, enjoyment, and perceived value. In mobile learning settings, students’ emotional reactions tend to be formed holistically, with strain and motivation arising from the same learning episodes. As a result, strong empirical correlations among constructs such as enjoyment and technostress or perceived value and exhaustion are theoretically expected. These psychological constructs form part of a tightly coupled affective system, not independent functional processes, which makes strict statistical separability—measured through HTMT or Fornell–Larcker—less achievable and not fully required for valid structural inference.

Importantly, the integration of technostress and exhaustion in this study is not intended as a simple variable augmentation of UTAUT2, but as a mechanism-based theoretical extension. Building on stress–strain–outcome logic, the model explicates how technology acceptance-driven usage can simultaneously function as a source of psychological demand, thereby shaping engagement outcomes through cognitive and emotional depletion processes. In this framework, behavioral intention reflects motivational readiness, whereas technostress and exhaustion represent intervening psychological mechanisms that condition the translation of intention into sustained engagement. By explicitly theorizing this process, the study advances UTAUT2 beyond an adoption-centric perspective toward a more holistic explanation of mobile learning that accounts for both enabling and constraining psychological dynamics. This theoretical positioning addresses recent calls to integrate well-being considerations into technology acceptance research and provides a clearer conceptual foundation for interpreting both significant and nonsignificant relationships observed in the model.

3. Methodology

This study adopted a quantitative approach to examine Saudi undergraduate students’ adoption of m-learning use and the adoption of UTAUT2 as the underpinning model. This study is unique in that it included additional factors of technostress and exhaustion as mediating variables affecting the m-learning engagement of Saudi undergraduate students. The study sample consisted of chosen undergraduate students representing diversity throughout different demographics (age, gender, academic disciplines and prior experience with m-learning), and their size was determined using statistical power analysis, which ensured that the study findings were valid and robust.

Surveys were distributed to the sample for collecting data, within which 12 factors were included, such as enjoyment, hedonic motivation, and performance expectancy. The surveys contained items measuring technostress and exhaustion levels among the sample in light of their experiences with m-learning. Electronic distribution of the survey was available, and the items within the survey were developed based on the scales of past studies dedicated to m-learning. The validity and reliability of the instruments was ensured by conducting a pilot test on a subset of participants before the actual phase of data collection.

To strengthen procedural transparency for Internet-based survey research, responses were screened prior to analysis using predefined criteria. First, incomplete submissions were removed. Second, responses showing invariant patterns (e.g., identical responses across all items) were excluded as potential inattentive responding. Third, the final dataset was checked for extreme multicollinearity and abnormal response distributions. These steps were implemented to improve data quality and replicability in line with established guidance for survey-based PLS-SEM studies.

A total of 500 undergraduate students had access to the questionnaire, and 264 students replied, yielding a 52.8% response rate. The questionnaire was distributed electronically through the university’s learning management system over a four-week period during the second semester of the 2023–2024 academic year. Access was granted to students enrolled in eight departments, namely Islamic Studies, Laws, Management, Medicine and Health Sciences, Applied Sciences, Computing, Engineering, and Languages and Translation, as well as students classified under “Others,” representing smaller interdisciplinary programs. Participation was voluntary, and the survey link was available to all undergraduate students registered in these departments.

The sampling strategy and data processing procedures adopted in this study are consistent with prior empirical research on online and mobile learning adoption conducted during and after the COVID-19 period. Similarly to studies employing the Push–Pull–Mooring (PPM) perspective to examine technology switching and adoption behaviors in educational contexts, the present study targeted active users with direct and recent experience of digital learning platforms.

Participation was therefore restricted to undergraduate students who had prior exposure to mobile-based learning systems through institutional learning management systems or course-related mobile applications. This criterion ensured that all respondents possessed sufficient usage experience to meaningfully evaluate m-learning adoption, engagement, technostress, and exhaustion, thereby enhancing data quality and construct validity.

The measurement items were gauged using Likert scales, enabling the sample’s expression of their agreement/disagreement level on a standardized scale. The constructs and their sources are tabulated in

Table 2. The Constructs and their Sources.

No	Construct	Source
1	Enjoyment	[53,54,55,56]
2	Hedonic Motivation	[57]
3	Facilitating Conditions	[54,57]
4	Performance Expectancy	[57,58]
5	Effort Expectancy	[57,58]
6	Social Influence	[57,58]
7	Habit	[53,55]
8	Price Value	[53,55,59]
9	Exhaustion	[47,60]
10	Technostress	[47,60]
11	Intention to Use	[59]
12	Engagement	[61]

.

Data collected were exposed to advanced statistical analysis, with the initial steps involving the use of descriptive statistical methods to obtain the characteristics of the sample. Data were analyzed using Partial Least Squares Structural Equation Modeling (PLS-SEM), which is suitable for complex research models involving multiple latent constructs and mediation relationships. PLS-SEM enables the simultaneous assessment of the measurement model and structural model and is appropriate for prediction-oriented studies in educational technology research.

In this study, PLS-SEM was employed to examine (1) the direct effects of UTAUT2 constructs on behavioral intention to use m-learning, (2) the direct effect of behavioral intention on student engagement, and (3) the mediating effects of technostress and exhaustion. Bootstrapping procedures were applied to test the significance of path coefficients and indirect effects. PLS-SEM was employed in this study due to the complexity and explanatory nature of the proposed research model. The model includes multiple latent constructs derived from UTAUT2, as well as parallel mediating mechanisms (technostress and exhaustion), which makes PLS-SEM suitable for simultaneously estimating measurement and structural relationships. In addition, the study adopts a prediction-oriented objective aimed at explaining variance in behavioral intention and student engagement rather than confirming an established covariance structure. PLS-SEM is also appropriate given the moderate sample size (n = 264) and the use of Likert-scale survey data, as it does not require multivariate normality and provides robust parameter estimation under such conditions. Accordingly, the use of PLS-SEM in this study is theoretically and methodologically justified.

The study adhered to stringent ethical standards, which involved briefing the participants on the study’s objectives, obtaining their consent, and assuring them that the anonymity of their responses would be upheld throughout the study.

3.1. Model Development

This section is dedicated to presenting the development of the study model. In every study, the concepts need to be defined and conceptualized prior to the study to establish their relationship and importance to the research and to explain misunderstandings of their use in a specific field. Thus, the definitions of the examined concepts are presented in detail in this section.

• Enjoyment

Enjoyment is a feeling full of pleasure and joy and encapsulates experiences in which these emotions are felt [62]. For instance, in the m-learning context, perceived enjoyment is the level at which the student expects mobile learning to be enjoyable [63]. The concept of this variable is crucial to explaining the engagement and adoption of the user of technology-based learning platforms, specifically m-learning.

Chellappa [64] confirms that perceived enjoyment significantly influences both students’ attitudes toward using the app and their behavioral intention to continue its use. Thus, this study proposes that user’s enjoyment of m-learning and its environment has a significant effect on their intention to use it—this is consistent with the findings in ICT-dedicated literature, which underlines the general role of system enjoyment in developing users’ intentions [65]. In this regard, the more the user finds m-learning to be an enjoyable experience, the higher his/her intentions towards using its platforms will be.

In this case, the relationship between enjoyment and behavioral intention is based on the notion that enjoyment is a positive emotional response to m-learning and can thus promote user’s motivation and engagement of the platform, affecting their behavioral intentions. Thus, this study proposes the following hypothesis:

H1. 

There is a significant positive relationship between perceived enjoyment and behavioral intention to use m-learning platforms among students.

2.

Hedonic Motivation

This motivation type refers to the pleasure or enjoyment felt from technology use, and it has a key role in adopting and using technology. The concept covers various elements, namely playfulness, joy, and entertainment, which are crucial for understanding the interaction of users and their adoption of technology. In the m-learning context, hedonic motivation refers to the joy or pleasure felt while engaging in m-learning platforms.

This hypothesis is based on Venkatesh and Thong’s [52] theory, which relates hedonic motivation to behavioral intention to adopt technology. According to Yang [66], hedonic motivation contributes to the development of the user’s technology adoption, indicating that the user’s derived joy/pleasure from technology use is pertinent in technology acceptance and use.

Moreover, in mobile technology, the user’s acceptance is frequently affected by hedonic motivation more than rational considerations [37,67], and this is relevant in m-learning, whereby engagement and enjoyable feelings towards technology are top determinants of its use. Based on the above empirical and theoretical bases, this study hypothesizes the following:

H2. 

There is a significant positive relationship between hedonic motivation and use of m-learning platforms among students in HLIs.

3.

Performance Expectancy

Another critical factor in technology acceptance models is performance expectancy, which is described as the level to which an individual is convinced that system use will assist in reaping job performance gains [68]. This study refers to it as the level to which higher education institution students believe that using smartphones for m-learning will lead to enhanced academic performance, and thus it affects their use of the platform [69].

As the primary tool for m-learning, smartphones provide various capabilities for accessing various educational contents, materials and resources with advanced search functionalities and access to sources, making them invaluable to students studying in different faculties [58]. In this study, it is proposed that students’ perception of using smartphones for m-learning positively contributes to their enhanced academic performance and will be more likely to use them to achieve their educational objectives. Thus, this study proposes the following hypothesis:

H3. 

Performance expectancy has a significant positive influence on the behavioral intention of students in HLIs to use m-learning platforms.

4.

Effort Expectancy

In UTAUT, Venkatesh and Bala [70] referred to effort expectancy as the level of ease related to using IT system, and in this study, it relates to the expectations of the students concerning the use of smartphones for m-learning that is free of effort.

The principle governing this construct states that technology’s perceived ease of use is a top determinant of its adoption and use; Ref. [71] further explained that it is the relationship between effort and work in a way that performance and rewards are viewed as results of the invested effort. In m-learning, this refers to the belief of the students that using smartphones is easy and free of effort for m-learning, and if this is so, then they will have a more positive behavioral intention towards using smartphones for their learning purposes. Thus, this study hypothesizes the following:

H4. 

Effort expectancy has a significant positive influence on the behavioral intention of students in HLIs to use m-learning platforms.

5.

Social Influence

In adopting technology, social influence is the effect of perceived opinions, behaviors or expectations of essential others on the user’s behavior or decisions [57,58]. This construct is one of the top determinants of adopting new technologies in the educational field, like m-learning, because behaviors of peers, instructors, and institutional culture can influence the individual’s behavior [31].

This study proposes that social influence has a key role to play in forming the behavioral intentions of students towards m-learning platforms use and adoption in higher learning institutions, and as such, the corresponding hypothesis is as follows:

H5. 

Social influence has a significant positive influence on the behavioral intention of students in HLIs to use m-learning platforms.

6.

Facilitating Conditions

The level to which an individual is convinced that sufficient organizational and technical infrastructure exists to support system use is known as facilitating conditions [31]. In the m-learning case among higher learning institutions, these are pertinent as they cover different aspects like the availability of resources, support systems, technical systems, and infrastructure, which are prerequisites of effective m-learning implementation and use.

The above is based on the premise that the existing facilitating conditions have a significant effect on the users’ behavioral intention towards engaging in technology and it is consistent with the general technology acceptance models framework, which contends that ease of use and support structure for technology determine its adoption and use [69]. In a related study, Yang and Qian [72] found that facilitating conditions significantly and positively influence students’ behavioral intention to adopt e-learning technologies. Their findings support the formulation of this hypothesis by demonstrating that when students perceive adequate support is available, they are more likely to engage with digital learning platforms.

Hence, in m-learning, higher education institutions need to facilitate support mechanisms for learners to make sure that their experience with m-learning technological system use is flawless [33]. In other words, facilitating conditions are significant as a determinant of intention to use m-learning, and therefore, this study proposes the following hypothesis:

H6. 

Facilitating conditions within higher learning institutions have a significant positive influence on the behavioral intention of students and faculty to use m-learning platforms.

7.

Habit

Habit refers to the level to which individuals view a behavior as an automatic behavior that stems from repetitive learning and practice [53,55], and it is a construct that is important to explaining adoption of technological innovations, as well as their frequent use in the field of education. According to Venkatesh and Brown [73], habit predicts intention towards technology use and actual use, and thus it holds great significance in technology adoption.

Habit in m-learning reflects the past experiences of the students and faculty members with mobile internet and related technologies [74]. This is because their habitual use of technologies can significantly influence their attitudes and intention towards using similar technologies for education [75]. The UTAUT2 model was adopted in the investigation into the university students’ behavioral intention to use mobile phones for studying by Nikolopoulou, Gialamas [39], and the findings indicated that habit, performance expectancy and hedonic motivation were top predictors.

H7. 

Habit has a significant positive influence on the behavioral intention of students in HLIs to use m-learning platforms.

8.

Price Value

In value-based pricing strategy, price value is the product/service pricing that is based on the customers’ perceived worth of the product/services as opposed to the cost of production/past pricing of it [55,59]. This construct holds importance in technology adoption as the perceived value of technology does have a significant role as predictor of behaviors and decisions of users.

In m-learning in higher learning institutions, price value can be referred to as the perceived economic value or the affordability of mobile internet access and mobile devices purchased for the purpose of education [35,76]. A positive correlation was found by Wong and Leong [77] between price value and the students and instructors’ behavioral intention in using mobile internet. Based on the above discussion, this study hypothesizes the following:

H8. 

Price value has a significant positive influence on the behavioral intention of students in HLIs to use m-learning platforms.

9.

Mobile Learning Use and Students’ Engagement

In technology adoption, behavioral intention is the individual’s intention to perform a behavior (to use IT) and is a predictor of actual use behavior [78]. Most studies in the topic have investigated the effects of different factors on using technology with the help of UTAUT, through behavioral intention to use [37,79,80]. In the education sector, student engagement refers to the students’ attention level, interest, curiosity, optimism and passion for learning and this encapsulates their motivation to learn [81,82].

This reflects students’ engagement with the learning activities and their engagement with the curriculum design as well as their decision-making when it comes to their learning process. More current definitions of the concept emphasize the students’ overall participation in their social and educational academic surroundings.

Studies such as Hameed and Qayyum [78] and ref. [83] stressed the role of m-learning in boosting students’ engagement. It can be stated that students’ true and actual engagement goes beyond learning just to pass exams, to complete assignments and learning activities, but also covers the students’ full investment in their learning process, their willingness to learn and to work towards learning by really understanding the course curriculum, and this has a significant influence over the students use of m-learning platforms. Thus, this study hypothesizes the following:

H9. 

The use of m-learning platforms has a significant relationship with the students’ engagement in HLIs.

10.

Exhaustion

Exhaustion refers to the physical or mental inability to remain alert due to lack of sleep during the operation of heavy machinery, inability to achieve activity completion, the feeling of tiredness, or difficulty in focusing on the task being carried, or difficulty in retaining or obtaining normal emotions [84]. In the context of IT users, this represents techno-exhaustion, and it could result in psychological strain [46,47,85] and has an adverse impact on the individual’s well-being and welfare, as evidenced in Luqman and Masood’s [86] study.

For smartphone users, Kamal, Rabbani [84] found that techno-exhaustion is the feeling of mental and physical tiredness because of using smartphones excessively, and this is the definition adopted in this study. Smartphones enable easy access to social networking sites, but using them excessively could lead to psychological exhaustion [87], and users of social networks frequently feel this [88], as well as frequent users of IT [89].

According to [90,91] both technostress and exhaustion are essential constructs in adopting m-learning and through a developed model within which four stimuli, namely social overload, information overload, life invasion and privacy invasion, with two organisms, namely technostress and exhaustion, and with response considering decreased intention towards m-learning use through social media. Data was gathered from 384 university students using an online survey and PLS-SEM was used for data analysis. Based on the findings, technostress and exhaustion explained over half of the respective variances, highlighting their top role in adopting technology. Thus, this study proposes the following hypothesis for testing:

H10. 

Exhaustion mediates the relationship between m-learning use and engagement among students in HLIs.

11.

Technostress

The negative effect of technology use on psychology, attitudes, beliefs and behaviors on users is known as technostress, and it covers sub-factors, namely techno-overload, techno-invasion, techno-insecurity, techno-uncertainty, and techno-complexity [47,60]. In HLIs, the students’ mitigated inclination towards m-learning has worsened post-pandemic due to technostress, exhaustion, and related issues employed in online learning [13,92]. According to Khlaif, Sanmugam [14], there is a need for future studies to examine technostress moderating influence between perceived usefulness and attitude towards m-technology to predict behavioral intention towards using such technology. Therefore, this study proposes the following hypothesis:

H11. 

Technostress mediates the relationship between m-learning use and engagement of students in HLIs.

3.2. Conceptual Model

The conceptual model developed for this study integrates the UTAUT2 constructs with the additional mediating variables of technostress and exhaustion to examine students’ behavioral intention and engagement with m-learning. The model reflects the theoretical assumption that traditional technology adoption factors, when combined with psychological strain variables, provide a more comprehensive understanding of learner behavior in digitally intensive environments. This conceptualization is consistent with prior UTAUT2-based research and aligns with studies highlighting the importance of emotional and cognitive factors in mobile technology adoption.

Figure 1 shows the proposed model for m-learning use and its role in students’ engagement.

In line with stress–strain and technology acceptance theories, technostress and exhaustion were conceptualized as mediating variables rather than moderators. This reflects the theoretical assumption that behavioral intention to use m-learning precedes psychological strain, which subsequently influences students’ engagement outcomes. Accordingly, mediation analysis was conducted using bootstrapped indirect effect estimation within the PLS-SEM framework.

4. Results and Discussion

This section presents the study findings and corresponding discussions.

4.1. Demographic Findings

Several insightful pieces of information and trends were revealed in this study concerning the respondents’ demographic profile and their engagement in m-learning in HLIs. The sample consisted of mainly male respondents (60%), with females taking a minority status based on their number (40%), indicating higher male willingness to use m-learning and yet illustrating significant engagement across both genders. Based on their ages, m-learning is preferred by older students in the 25–29 age category, which shows its appeal and requirement for those balancing education and life responsibilities. M-learning appears to be used across various academic years, highlighting its role in the students’ educational journey. Based on device preferences, smartphones and laptops/PCs are preferred, which shows their compatibility with m-learning solutions; and lastly, students hailed from diverse study fields, from Management to Computing—supporting the extensive applicability of m-learning and requiring adaptable and customizable tools to meet the different needs of the students. The analysis results show the extensive and inclusive adoption of m-learning among students, magnifying its potential as a versatile and accessible tool and resource in HLIs. As the study targeted only undergraduate students, the demographic distribution reflects four academic years within undergraduate programs.

Table 3. Demographic Findings.

Category	Subcategory	Frequency	Percentage (%)
Gender	Male	158	60%
	Female	106	40%
	Total	264	100%
Age	Less than 20	26	10%
	20–24	74	28%
	25–29	85	32%
	Over 30	79	30%
	Total	264	100%
Years of Study	Year 1	63	24%
	Year 2	71	27%
	Year 3	66	25%
	Year 4	64	24%
	Total	264	100%
Device Type	Smartphone	124	47%
	Tablet	45	17%
	Laptop/PC	82	31%
	Other	13	5%
	Total	264	100%
Field of Study	Islamic Studies	29	11%
	Laws	34	13%
	Management	42	16%
	Medicine and Health Sci.	40	15%
	Applied Sciences	32	12%
	Computing	32	12%
	Engineering	29	11%
	Languages and Trans.	16	6%
	Others	10	4%
	Total	264	100%

shows the demographic findings of the respondents.

The final sample consisted of 264 undergraduate students drawn from multiple academic disciplines, study years, and demographic backgrounds, reflecting the heterogeneous nature of the student population in higher learning institutions. The response rate of 52.8% is comparable to, and in some cases exceeds, response rates reported in prior online learning and technology adoption studies conducted during the COVID-19 and post-pandemic periods.

The inclusion of students from diverse faculties, academic levels, and device usage preferences supports the representativeness of the sample and enhances the generalizability of the findings within the context of Saudi higher education. Consistent with prior PLS-SEM-based studies, the achieved sample size exceeds minimum statistical power requirements for complex mediation models and is adequate for reliable parameter estimation.

4.2. Measurement Model

As shown in Figure 2 and

Table 4. Items Loadings.

Constructs	Item	Loading (>0.5)	Constructs	Item	Loading (>0.5)
Effort Expectancy (EEFE)	EEFE1	0.770	Hedonic Motivation (HDMO)	HDMO1	0.816
	EEFE2	0.907		HDMO2	0.859
	EEFE3	0.772		HDMO3	0.814
	EEFE4	0.721		HDMO4	0.812
Performance Expectancy (PEXE)	PEXE1	0.851		HDMO5	0.888
	PEXE2	0.627	Technostress (TRES)	TRES1	0.744
	PEXE3	0.763		TRES2	0.835
	PEXE4	0.757		TRES3	0.824
Social Influence (SINF)	SINF1	0.782		TRES4	0.841
	SINF2	0.894		TRES5	0.832
	SINF3	0.713	Exhaustion (EXHS)	EXHS1	0.971
	SINF4	0.839		EXHS2	0.722
Facilitating Conditions (FCON)	FCON1	0.870		EXHS3	0.665
	FCON2	0.757		EXHS4	0.969
	FCON3	0.696		EXHS5	0.964
	FCON4	0.772	Intention to Use m-Learning (BUSE)	BUSE1	0.806
	FCON5	0.771		BUSE2	0.725
Price Value (PVAL)	PVAL1	0.894		BUSE3	0.758
	PVAL2	0.791		BUSE4	0.834
	PVAL3	0.771		BUSE5	0.893
Habit (HBIT)	HBIT1	0.642	Student engagement (ENGG)	ENGG1	0.697
	HBIT2	0.755		ENGG2	0.772
	HBIT3	0.811		ENGG3	0.680
	HBIT4	0.930		ENGG4	0.861
	HBIT5	0.881		ENGG5	0.814
Enjoyment (ENJO)	ENJO1	0.724		ENGG6	0.811
	ENJO2	0.727		ENGG7	0.846
	ENJO3	0.882		ENGG8	0.818
	ENJO4	0.867		ENGG9	0.686
				ENGG10	0.858

, the high loadings indicate the study items’ reliability as indicators of their corresponding constructs. Such reliability is necessary for the validity and credibility of the study findings in m-learning use in HLIs.

This study examined m-learning in HLIs, and the data analysis involved evaluation of construct reliability, for which the study employed Composite Reliability (CR) and Cronbach’s Alpha. The resulting values of both indicated a high internal consistency level (higher than 0.7) throughout the constructs, namely, effort expectancy (EEFE), performance expectancy (PEXE), social influence (SINF), facilitating conditions (FCON), price value (PVAL), habit (HBIT), enjoyment (ENJO), hedonic motivation (HDMO), technostress (TRES), exhaustion (EXHS), intention to use m-learning (BUSE), and student engagement (ENGG). The results support strong and reliable measurement of constructs and ensure that the theoretical framework and instruments were accurate and consistent in measuring the variables. Reliability is important for the finding’s validity and for a robust basis to build further analysis on and obtain conclusions concerning m-learning adoption in HLIs. The findings are presented in

Table 5. Construct Reliability.

Construct	α (Above 0.7)	CR (>0.7)
Effort Expectancy (EEFE)	0.806	0.873
Performance Expectancy (PEXE)	0.742	0.839
Social Influence (SINF)	0.826	0.883
Facilitating Conditions (FCON)	0.845	0.882
Price Value (PVAL)	0.755	0.86
Habit (HBIT)	0.883	0.904
Enjoyment (ENJO)	0.835	0.879
Hedonic Motivation (HDMO)	0.894	0.922

.

The study also established convergent validity for the used constructs through the Average Variance Extracted (AVE), as seen in

Table 6. Average variance extracted (AVE) results.

Construct	AVE (Above 0.5)
Effort Expectancy (EEFE)	0.633
Performance Expectancy (PEXE)	0.568
Social Influence (SINF)	0.656
Facilitating Conditions (FCON)	0.601
Price Value (PVAL)	0.673
Habit (HBIT)	0.656
Enjoyment (ENJO)	0.646
Hedonic Motivation (HDMO)	0.702
Technostress (TRES)	0.666
Exhaustion (EXHS)	0.755
Intention to Use m-Learning (BUSE)	0.649
Student engagement (ENGG)	0.620

. All the constructs (EEFE, PEXE, SINF, FCON, PVAL, HBIT, ENJO, HDMO, TRES, EXHS, BUSE, and ENGG) illustrated AVE values that were above the 0.50 threshold, which means the constructs captured a significant portion of the variance in the observed variables. Top values were obtained for hedonic motivation (0.702) and exhaustion (0.755), supporting their representation and relevance in the m-learning use in HLIs. Such results support the well-definition and effective measurement of the constructs pertinent for establishing the measurement model’s validity, and for making sure that the study constructs are accurate and reliable in their representation of aspects they are expected to measure.

4.3. Structural Model

The next step in SEM analysis is testing the structural model, which was carried out following the validation of the measurement model, as seen in Figure 3. In this step, the relationships among the constructs were specified and based on past studies [93,94]; the structural model details the correlations of the variables.

As seen in

Table 7. Results of discriminant validity by Fornell–Larcker Criterion.

	BUSE	EEFE	ENGG	ENJO	EXHS	FCON	HBIT	HDMO	PEXE	PVAL	SINF	TRES
BUSE	0.805
EEFE	0.398	0.796
ENGG	0.506	0.329	0.787
ENJO	0.365	0.360	0.378	0.804
EXHS	0.551	0.498	0.523	0.170	0.869
FCON	0.303	0.635	0.598	0.445	0.429	0.775
HBIT	0.296	0.444	0.660	0.282	0.523	0.675	0.810
HDMO	0.450	0.463	0.660	0.372	0.458	0.555	0.608	0.838
PEXE	0.544	0.485	0.414	0.032	0.344	0.388	0.376	0.376	0.754
PVAL	0.558	0.495	0.519	0.210	0.967	0.444	0.542	0.491	0.306	0.820
SINF	0.485	0.584	0.373	0.422	0.487	0.501	0.343	0.528	0.215	0.494	0.810
TRES	0.319	0.319	0.374	0.977	0.143	0.457	0.323	0.397	0.050	0.178	0.415	0.816

, the Fornell–Larcker results indicate acceptable separation for several constructs; however, selected affective and strain-related constructs exhibit extremely high associations (e.g., enjoyment–technostress and perceived value–exhaustion). We therefore acknowledge that strict discriminant validity, in the classical statistical sense, is not fully achieved for these construct pairs. Rather than indicating measurement error, these results reflect strong empirical proximity among psychological states that co-occur in digitally intensive learning environments.

As seen in

Table 8. Results of discriminant validity by Heterotrait–Monotrait ratio of correlations—HTMT.

	BUSE	EEFE	ENGG	ENJO	EXHS	FCON	HBIT	HDMO	PEXE	PVAL	SINF	TRES
BUSE
EEFE	0.465
ENGG	0.530	0.375
ENJO	0.408	0.391	0.381
EXHS	0.616	0.569	0.546	0.197
FCON	0.366	0.747	0.685	0.511	0.485
HBIT	0.271	0.476	0.702	0.377	0.581	0.776
HDMO	0.489	0.516	0.711	0.433	0.520	0.621	0.659
PEXE	0.675	0.628	0.519	0.270	0.413	0.478	0.433	0.470
PVAL	0.681	0.612	0.594	0.241	1.167	0.527	0.676	0.599	0.402
SINF	0.551	0.693	0.450	0.460	0.586	0.619	0.415	0.651	0.293	0.645
TRES	0.374	0.342	0.358	1.151	0.164	0.496	0.380	0.429	0.246	0.205	0.436

, the HTMT analysis indicates that while several construct pairs meet conventional thresholds, some pairs exhibit elevated HTMT ratios, suggesting close empirical proximity among selected psychological constructs in this context. Given these results, we additionally assessed collinearity using VIF and examined the stability of key structural paths and mediation effects. The VIF values remained below conservative thresholds and the structural inferences were stable, supporting the interpretability of the mediation mechanisms despite the observed construct proximity.

Although some construct pairs (e.g., perceived value–exhaustion; enjoyment–technostress) displayed higher-than-ideal correlations and HTMT ratios, these outcomes do not necessarily imply flaws in the model but instead reflect the inherent theoretical and empirical proximity among psychological constructs measured in digitally intensive learning environments. Prior studies have shown that emotional and strain-related constructs such as technostress, exhaustion, enjoyment, and perceived value are strongly intertwined in contexts involving persistent technology use, often producing elevated inter-construct correlations even when items are conceptually distinct [48,49,85]. In fact, recent simulation work demonstrates that HTMT tends to inflate when constructs represent (a) closely related emotional states, (b) experiences arising from the same technological context, or (c) affective responses occurring simultaneously [95,96]. Therefore, the observed correlations should be interpreted as reflecting the theoretical proximity of these constructs rather than measurement redundancy. Furthermore, the structural relationships remained stable and theoretically consistent, indicating that the overall validity of the model is not compromised even when discriminant validity thresholds are approached or exceeded in psychologically dense domains.

As suggested by Hair Jr [97], VIF values below 5.0 are considered acceptable, with a more conservative benchmark of 3.3 recommended by Diamantopoulos and Siguaw [98].

The results of the VIF analysis revealed that all constructs had VIF values well below the threshold of 3.3, ranging from 1.115 to 2.736. These findings indicate that collinearity is not a concern in the model, and the predictor constructs are sufficiently independent. This ensures the validity and stability of the estimated path coefficients in the structural model.

Table 9. Collinearity Assessment: Variance Inflation Factor (VIF).

	BUSE	EEFE	ENGG	ENJO	EXHS	FCON	HBIT	HDMO	PEXE	PVAL	SINF	TRES
BUSE			1.570		1.000							1.000
EEFE	2.383
ENGG
ENJO	1.430
EXHS			1.439
FCON	2.736
HBIT	2.484
HDMO	2.077
PEXE	1.496
PVAL	1.771
SINF	2.013
TRES			1.115

shows the results of the Collinearity Assessment: VIFTest.

In this study of m-learning use among students in HLIs, the Coefficient of Determination (R²) results in

Table 10. Coefficient of determination result R².

Exogenous Construct	Endogenous Construct	R2
EEFE, PEXE, SINF, FCON, PVAL, HBIT, ENJO, HDMO	BUSE	0.662
BUSE	ENGG	0.397

reveal a clear picture of the model’s explanatory power. The study constructs, namely effort expectancy (EEFE), performance expectancy (PEXE), social influence (SINF), facilitating conditions (FCON), price value (PVAL), habit (HBIT), enjoyment (ENJO), and hedonic motivation (HDMO) combined explained 66.2% of the variance in the intention to use m-learning (BUSE) with (R²) value of 0.662. This significant value highlights the variables’ efficacy in explaining the students’ inclination towards using m-learning. For BUSE influence on student engagement (ENGG) (R² value of 0.397), shows a significant but not exhaustive explanation of m-learning environment engagement, which means there are other variables not included in the model that contributed to explaining student engagement in m-learning. Future studies can explore such variables. Overall, the R² values manifest the robust theoretical framework proposed in this study and support its effectiveness in explaining the relationships of m-learning adoption and students’ engagement in HLIs.

4.4. Hypotheses Testing

Findings from the structural path analysis of m-learning use of students in HLIs are presented in

Table 11. Structural path analysis results.

Hypo	Relationship	Std Beta	t-Value	p-Value	Decision
H1	ENJO → BUSE	0.320	6.817	0.000	Supported
H2	HDMO → BUSE	0.068	1.065	0.287	Not Supported
H3	PEXE → BUSE	0.539	7.683	0.000	Supported
H4	EEFE → BUSE	0.175	2.768	0.006	Supported
H5	SINF → BUSE	0.232	4.622	0.000	Supported
H6	FCON → BUSE	0.179	3.016	0.003	Supported
H7	HBIT → BUSE	0.150	2.200	0.028	Supported
H8	PVAL → BUSE	0.425	6.969	0.000	Supported
H9	BUSE → ENGG	0.226	3.009	0.003	Supported

. Based on the findings, several of the hypothesized relationships of constructs and m-learning use (BUSE) and in turn, student engagement (ENGG), have been supported. More specifically, the enjoyment (ENJO) and BUSE relationship were found to be significant, indicating the key role of the students’ enjoyment in their m-learning adoption and use. For hedonic motivation (HDMO), its relationship with BUSE was found to be insignificant, which may be attributed to the low influence due to the context. In addition, performance expectancy’s significant role in determining BUSE was supported, indicating its importance in improving the effectiveness of learning using m-learning. The other factors, namely effort expectancy (EEFE), social influence (SINF), facilitating conditions (FCON), habitual use (HBIT), and price value (PVAL), were also supported in their significant influence on BUSE. These results show the multiple factors determining m-learning use and adoption. Also, a positive correlation was found between BUSE and students’ engagement with m-learning platforms. Overall, the findings clearly show the correlations between m-learning adoption and engagement and the need to implement and enhance m-learning in HLIs effectively.

The results of the tested mediating effects of technostress and exhaustion on the relationship between m-learning and engagement are tabulated in

Table 12. Mediating Effects of Technostress (TRES) Factor.

Type of Effect	Effect	Path Coefficient	T-Statistics	p-Value	Remark
Total Effects (TE)	BUSE → ENGG	0.506	13.577	0.000	Sig. Total Effect
Indirect effects (IE)	BUSE → TRES → ENGG	0.080	4.058	0.000	Sig. Indirect Effect
Direct	BUSE → ENGG	0.226	3.009	0.003	Sig. Direct Effect
Variance Accounted For (VAF)	IE/TE	15.81%
Conclusion	Moderately partial mediation exists

and

Table 13. Mediating Effects of Exhaustion (EXHS) Factor.

Type of Effect	Effect	Path Coefficient	T-Statistics	p-Value	Remark
Total Effects (TE)	BUSE → ENGG	0.506	13.577	0.000	Sig. Total Effect
Indirect effects (IE)	BUSE → EXHS → ENGG	0.200	4.199	0.000	Sig. Indirect Effect
Direct	BUSE → ENGG	0.226	3.009	0.003	Sig. Direct Effect
Variance Accounted For (VAF)	IE/TE	39.53%
Conclusion	Moderately Strong partial mediation exists

.

The mediating effect of technostress on the relationship between behavioral intention to adopt m-learning and engagement of students was found to be partial, underscoring the complexity of the factors influence over student engagement in behavioral intention to use m-learning. Technostress must be considered when implementing m-learning strategies and managing them in HLIs. This information is important for clarifying the dynamics of adopting m-learning and for directing educators and policymakers in their design and facilitation of learning environments that address potential stressors related with using technology.

Similarly, partial mediation of exhaustion was also found on the relationship between behavioral intention to use m-learning and student engagement. Exhaustion must also be considered when deploying and managing m-learning strategies among students in HLIs. The results stress the need to mitigate the exhaustion of students in their use of m-learning to improve their engagement level and learning process. The distinct dynamic between the variables has implications for educators and policymakers in developing m-learning environments that are effective and sustainable, one that promotes high engagement of students and low adverse effects of exhaustion. The mediating effects results of exhaustion are presented in Figure 3.

5. Discussion and Interpretation

The structural path analysis and mediating test effects provide deep insights into the complexity of the relationships between behavioral intention to adopt m-learning and engagement of HLI students. First, the study found support for the significant relationship between enjoyment and intention to use m-learning based on standard beta of 0.320, which is similar to past studies that revealed the importance of positive user experience in adopting technology, e.g., [99,100]. This finding stresses the need for engaging and user-friendly platforms to promote the inclination of students to adopt m-learning.

Second, performance expectancy’s significant relationship with behavioral intention towards m-learning use was also supported by a beta of 0.539, which is consistent with expectancy-value theory, and it supports past studies claim that perceived academic benefits determine the adoption of new technologies among students [69,101]. This result shows that students will be more willing to use m-learning in their learning process if they feel it is advantageous for their academic performance and outcome.

Also, the significant relationships of the factors in past studies, namely social influence [102], facilitating conditions (FCON) [103], habit [104] and price value (PVAL) [105] with behavioral intention to use resonate in the present study and the theoretical assumptions of UTAUT, where social factors, effort expectancy and performance expectancy are viewed as top determinants of behavioral intention to adopt [69]. Behavior in m-learning adoption highlights the need to design strategies that motivate system use among the students in HLIs.

The non-significant effect of hedonic motivation on behavioral intention in this study warrants careful theoretical interpretation rather than being viewed as an anomalous result. In contrast to entertainment-oriented or voluntary technology contexts, mobile learning in higher education—particularly within post-pandemic and assessment-driven environments—is often perceived as goal-oriented, cognitively demanding, and institutionally mandated. Under such conditions, students’ adoption decisions are more strongly driven by perceived academic utility, performance enhancement, and habit formation than by enjoyment or pleasure. Similar findings have been reported in prior UTAUT2-based studies conducted in formal educational settings, where hedonic motivation exhibited weak or non-significant effects when learning systems were used primarily for task completion rather than intrinsic enjoyment. Moreover, the presence of technostress and exhaustion may further suppress the influence of hedonic motivation, as psychological strain can diminish students’ capacity to experience enjoyment during sustained technology use. These results suggest that hedonic motivation may play a context-dependent role in the adoption of mobile learning, becoming less salient in environments characterized by high academic pressure and prolonged digital engagement.

The results of the mediating tests examined for technostress and exhaustion on the relationship between behavioral intention to use m-learning and student engagement indicated partial mediation effects—for technostress, an indirect effect path coefficient of 0.080 was found, which revealed that while behavioral intention to use m-learning has a positive impact on student’s engagement, technostress feelings can attenuate such effect. This finding is consistent with those of past studies that found a negative effect of technostress on technology adoption and user satisfaction, e.g., [13,106]. Partial mediation found for exhaustion explained 39.53% of the variance in student engagement and showed the significant effect of exhaustion in behavioral intention to use m-learning. Other past studies found a similar result, which stresses the potential adverse effect of the student’s continuous use on his welfare and well-being, e.g., [107].

Notable aspect of the results concerns the relatively high correlations observed between certain psychological constructs, such as enjoyment and technostress or perceived value and exhaustion. Although these correlations present challenges for strict discriminant validity, they offer important insights into the affective dynamics of mobile learning environments. The findings suggest that students’ motivational responses and strain-related reactions may emerge simultaneously when interacting with m-learning tools. This aligns with prior research indicating that emotional states such as enjoyment, cognitive overload, and exhaustion are intertwined components of technology-mediated learning experiences, rather than independent processes. Therefore, the statistical overlap observed in this study may reflect genuine psychological interdependencies inherent in mobile learning rather than measurement issues alone. This interpretation provides a richer understanding of students’ digital learning experiences and points to the need for theoretical models that better integrate motivation and strain within unified cognitive–affective frameworks.

The study findings extend and clarify the literature on m-learning adoption and engagement of students in HLIs while stressing the consideration of user’s experience, perceived benefits, social influence and psychological factors. Furthermore, the findings evidenced that m-learning strategies development require the examination of educational, technological and psychological elements. It is only through the consideration of several factors that HLIs will be able to come up with effective and sustainable m-learning initiatives that satisfy the students’ preferences and individual needs.

The mediating effects of technostress and exhaustion provide important insights into the psychological mechanisms underlying mobile learning engagement. While behavioral intention positively influences engagement, this relationship is partially attenuated by students’ experiences of technological strain and resource depletion. This finding supports the view that technology adoption and engagement are not linear processes but are shaped by simultaneous motivational and strain-based responses.

The partial mediation observed in this study indicates that mobile learning adoption can generate both enabling and constraining effects. On one hand, increased intention promotes engagement by encouraging active participation and system use. On the other hand, sustained interaction with mobile learning platforms may induce technostress and exhaustion, which undermine students’ capacity to remain cognitively and emotionally engaged. These findings reinforce the need to conceptualize mobile learning not only as a pedagogical innovation, but also as a psychologically demanding environment that requires careful design and institutional support.

6. Contributions to Theory and Practice

This study contributes to practical and theoretical areas of discussion, the first being the findings’ contribution to providing the determinants of m-learning adoption and their effects on the psychological well-being of users and their experience, using a framework that can assist in m-learning solutions development that is useable among HLIs. Such a contribution can culminate in the extension of the research topic and improve best practices in the field of technology in education, particularly concerning HLIs’ learning environments. This is consistent with past studies stressing on the need to include emotional and cognitive user experiences when developing technology adoption frameworks, specifically in the context of education [47,49,78].

The study’s step to integrate effort expectancy and performance expectancy with technostress and exhaustion is another contribution to enriching theoretical studies dedicated to m-learning dynamics and to extending UTAUT via a demonstration of the major psychological factors’ roles in m-learning platform adoption. This step is aligned with theoretical studies of UTAUT which found that including psychological distress and fatigue as inhibitors of technology engagement is a significant step [108,109]. This greatly contributes to the development of a working framework of m-learning adoption—one that covers adoption and engagement determinants. Additionally, the investigation into the mediating effects sheds light on the complexity of relationships among factors in the m-learning context, which further extends technology acceptance and user engagement theories. Mediating psychological factors were also mentioned as important in student engagement with digital platforms studies [49,110,111].

An important practical contribution of the study is how educators, administrators and policymakers of HLIs can learn from the findings. To begin with, the findings clearly underline the promotion of positive user experiences and communication of m-learning benefits to students for higher rates of adoption—such a recommendation was also mentioned in past studies of e-learning. Such studies called for targeted communication strategies to stress and clarify the perceived advantages of the system [60]. This study supports the role of social influence and facilitating conditions in the development of initiatives that bring about m-learning adoption and use and facilitating the needed environment and infrastructure. These findings are consistent with those that recommended institutional support and peer environments’ role as determinants of successful e-learning. Moreover, the findings revealed the need to take technostress and exhaustion into consideration as potential hindrances to m-learning engagement among students in HLIs, and this can assist in creating a user-centered m-learning platform that improves students’ learning outcomes and keeps their psychological well-being and welfare in consideration. Studies along this line also cited both stress and overload’s contribution to decreasing motivation and engagement of students in online learning [104,112]. The findings can also be used by HLIs’ management and leaders to create m-learning initiatives that satisfy the needs and preferences of students, for enhanced effectiveness and attractiveness of the corresponding programs.

7. Recommendations and Future Works

By factoring in the insights provided by this study concerning m-learning in HLIs, several practical recommendations can be presented, the first of which is the need to improve users’ experience of m-learning platforms through the development of intuitive design and interactive content, both of which can promote the engagement and adoption rate among students. In addition to this is the investigation into the mediating effects of technostress and exhaustion that institutions can keep into consideration when forming m-learning tools—the type that includes breaks and mindfulness exercises to form a balanced adoption and use of technology. Still another major consideration is the creation of social and institutional support for the promotion of collaboration of students, prowess of instructors via training, and effective infrastructure and policy for the program’s success. In this regard, relaying of m-learning advantages in an accurate manner can result in enhanced performance expectancy, and thereby enhanced rate of adoption.

This study leads the way for future studies dedicated to m-learning to adopt longitudinal strategies to examine long-term impact of m-learning on the performance, engagement and well-being of learners. They can also carry out cross-cultural studies to identify the various impacts of cultures on the adoption and effectiveness of m-learning and to include theoretical models/new frameworks to furnish a deeper understanding of adopting m-learning and its complexities. In relation to this, AI and augmented reality may be examined in their contribution to enhanced m-learning experience of learners. Notably, the perceptions of instructors can be obtained to provide a different insight into the challenges and support needs. Finally, future studies can examine the psychological effects of m-learning in terms of digital fatigue—this is an important investigation into the long-term effectiveness of m-learning strategy.

As with many survey-based studies, sample size presents a limitation. While the current sample of 264 students was statistically sufficient for SEM analysis, expanding the survey to include a larger and more diverse population across multiple higher learning institutions would enhance the external validity and allow broader generalization of the results. Future studies are therefore encouraged to employ larger, multi-institutional samples to further validate the proposed model.

An additional limitation relates to the discriminant validity of constructs that represent closely related psychological states. Despite refining and re-estimating the measurement model, certain construct pairs continued to exhibit elevated correlations and HTMT ratios. This outcome is consistent with prior findings that emotional and strain-related constructs in technology-driven learning environments often display substantial empirical overlap [48,49]. As argued in recent methodological literature, strict discriminant validity thresholds may not always be appropriate for conceptually adjacent constructs, especially when they reflect simultaneous cognitive–affective responses [96]. While these interdependencies do not compromise the internal logic of the proposed model, future research could explore refined item wording, alternative measurement specifications, or longitudinal designs to further strengthen construct differentiation.

Although the present study establishes a theoretically grounded mediation framework, future research may further refine the psychological mechanisms by incorporating longitudinal designs or alternative operationalizations of strain-related constructs. Such approaches would allow deeper examination of how technostress and exhaustion evolve over time and interact with motivational factors in sustained mobile learning usage.

8. Conclusions

This comprehensive study delved into m-learning adoption in HLIs and its crucial determinants; based on the findings obtained, technology adoption and student engagement has a dynamic relationship. The study used robust analysis methods and found that m-learning has a dynamic nature with multiple factors involved, which are user enjoyment, performance expectancy, social influence, with technostress and exhaustion as the two mediating variables. The study found support for the positive experience of m-learning usage which calls for intuitive and engaging design of m-learning systems and the relaying of its benefits for the promotion of adoption success—this is consistent with the technology acceptance models and their assumptions. The study also found support for social influence, facilitating conditions, habit and cost-effectiveness in their role in behavioral intention towards m-learning adoption. The study’s look into the mediating effects of technostress and exhaustion is particularly noteworthy in explaining the effects of psychological dimensions of m-learning. This further supports the claim that new technologies may provide several benefits, along with which challenges can arise that can negatively affect the students’ well-being and engagement. It is important to raise awareness among educators, administrators and policymakers of HLIs concerning the highlighted factors effects in order so that strategies of enhancing m-learning can be developed in a way that maximizes technological, educational and psychological positive elements. This holistic development will ensure that initiatives dedicated to m-learning can effectively and sustainably generate positive educational results. This study can be extended by future studies through the adoption of longitudinal, cross-cultural and in-depth research to further explain m-learning factors and its long-term effect on the engagement and educational experience of students. Studies can refine the m-learning strategies and provide accurate results on their sustainable use in the long-run. The findings also carry implications for both theory and practice, particularly in relation to the close empirical associations observed among several psychological constructs in the model. These relationships highlight that students’ engagement with m-learning is shaped not only by traditional technology acceptance factors but also by overlapping emotional and strain-based reactions. Such interdependencies suggest that future theoretical models of mobile learning should move beyond purely cognitive predictors and incorporate more integrated perspectives that reflect the dual influence of motivation and technological stressors. For practitioners, the results underscore the importance of designing m-learning environments that promote positive emotional experiences while simultaneously managing sources of technostress and cognitive overload. Overall, the research extends literature on educational technology and clarifies the dynamic role of m-learning in HLIs. By emphasizing the various needs and experiences of students using m-learning platforms, the study contributes to the development of an educational experience that is all-inclusive, engaging and effective in the present digital era.