Abstract
This study examines the impact of student aspiration factors (expectations, motivation, and enjoyment) on university students’ actual usage of mobile learning systems. The researchers conducted a structured questionnaire survey among 518 randomly selected university students from three universities in Hainan Province to measure their self-reported responses to seven constructs (expectations, motivation, enjoyment, perceived usefulness, perceived ease of use, attitudes, and actual use). A sophisticated two-stage structural equation modelling-artificial neural network approach was applied, and this study revealed non-compensatory and nonlinear relationships between the predictors and the actual usage of university students. The results revealed that all hypotheses were supported except for the direct effect between motivation and perceived usefulness, which was nonsignificant. Furthermore, according to the normalized importance derived from the multilayer perceptron, perceived usefulness (100%), attitude (73.0%), enjoyment (33.5%), motivation (30.8%), perceived ease of use (25.1%), and expectations (12.2%) are significant predictors of the actual usage behaviour of university students. Finally, this study presents theoretical and practical implications for the actual usage behaviour of mobile learning systems among university students.
Introduction
With the rapid advancement of information technology and especially with the widespread adoption of mobile devices and the progress of internet technology, mobile learning is increasingly gaining attention and adoption by educators and learners globally (Adel Bessadok, 2022; Humida et al., 2022). Mobile learning—which is defined as learning activities conducted via mobile devices such as smartphones, tablets, and laptops—breaks the constraints of time and location (Alajlan et al., 2022; Humida et al., 2022), supports personalized learning paths, and fosters active learning and interactive collaboration while also providing a wealth of educational resources and immediate feedback (Liu & Pu, 2020; Nguyen et al., 2018). However, despite the increased efficiency in learning and innovative teaching methods that mobile learning systems offer (Ibrahim et al., 2018; Mailizar et al., 2021; Salloum et al., 2019), factors such as a lack of experience with these systems among students or limitations inherent in mobile devices such as small screen sizes and poor interactivity quality can influence the decision to accept and use mobile learning systems (Agyei & Razi, 2021; Humida et al., 2022). In this study, we specifically focus on university students, considering the widespread application of mobile learning in higher education and its heightened relevance during the COVID-19 pandemic. This group is particularly significant in the adoption and use of mobile learning technologies, demonstrating unique characteristics and potential (Alshurideh et al., 2023). University students have high levels of technological proficiency and frequently engage with technology in their daily studies, which not only enables them to adapt quickly to new learning tools but also provides direct feedback on the effectiveness of mobile learning systems (Chen, 2022). Exploring their perceptions of and motivations towards mobile learning applications is crucial for a deeper understanding and optimization of educational technologies. Thus, to realize the full potential of mobile learning, a collaborative effort among educators, technology developers, and policymakers is needed to overcome the challenges encountered during implementation (Abbad, 2021).
The COVID-19 pandemic accelerated the application of mobile learning systems within higher education institutions (Alshurideh et al., 2023). In this context, researchers have employed various theoretical models to assess and understand the acceptance of mobile learning systems (Niu & Wu, 2022; Shen et al., 2022). While frameworks such as the Technology Acceptance Model (TAM) (Adel Bessadok, 2022; Humida et al., 2022), the Unified Theory of Acceptance and Use of Technology (UTAUT2) (Abbad, 2021; Nguyen et al., 2018)), and the Expectation Confirmation Model (ECM) (Ashfaq et al., 2019) have been used to effectively identify vital factors such as perceived usefulness (PU), self-efficacy (SEL), social influence (SI), attitude (ATT), motivation (MOT), and facilitating conditions (FC) (Agyei & Razi, 2021; Alyoussef, 2023; Azizi et al., 2020), most studies still employ static and linear approaches, such as partial least squares-structural equation modelling (PLS-SEM) and SPSS, which may not capture the dynamics and complexities of student behaviour (Abdou & Jasimuddin, 2020; Adel Bessadok, 2022; Liu & Pu, 2020). Furthermore, although the application of artificial neural networks (ANNs) offers new avenues for analysing nonlinear patterns of student behaviour and predicting actions (Alshurideh et al., 2023; Wu & Wang, 2018), their use in mobile learning research is not yet widespread, particularly in real-time data processing and personalized learning path recommendations (Niu et al., 2022; Wu & Tian, 2022). Thus, this study aims to address this gap by thoroughly analysing existing models and integrating ANN technology to better understand and predict students’ actual usage behaviour towards mobile learning systems.
While an increasing number of studies related to the actual use (AU) of mobile learning have been conducted, some gaps in the literature can still be observed. First, from a theoretical perspective, previous research on the AU of mobile learning systems has often overlooked the factor of student aspirations. These aspiration factors, which include expectation (EXP), MOT, and enjoyment (ENJ)(Adel Bessadok, 2022), greatly influence their decision-making process for adopting and using technology. Understanding these aspiration factors can provide deeper insights into student interactions with mobile learning technologies, thereby enhancing the design and implementation efficiency of mobile learning systems. Second, in terms of methodology, SEM has been widely used for exploratory relationships. However, SEM can identify only linear relationships, which do not apply to the nonlinear and non-compensatory relationships between exogenous and endogenous variables. Methods that handle nonlinear and non-compensatory relationships, such as ANNs, can provide more accurate and comprehensive insights. Such methods can be used to capture and explain more complex interactive effects between constructs, thus offering deeper understanding and predictions that help precisely guide practice and policy formulation.
In summary, this study has two objectives: (1) In the first phase, the ways in which EXP, MOT, and ENJ influence university students’ AU through PU, perceived ease of use (PEoU), and ATT are revealed through SEM. (2) In the second phase, a high-performance prediction model of university students’ AU of mobile learning systems based on an ANN is constructed. These findings provide theoretical and methodological support for university students’ AU of mobile learning systems while promoting the sustainability of mobile learning systems and academic achievement. Thus, the research questions of this study are as follows:
(1)
What are the factors influencing the AU of mobile learning systems among university students?
(2)
To what extent do these factors explain the total variance in the AU of mobile learning systems among university students?
(3)
What is the normalized importance of the factors affecting the AU of university students’ mobile learning systems?
Compared with previous research, this study contributes to the literature in three ways. First, this study integrates the TAM and UTAUT2 models, providing a more comprehensive perspective on how student aspiration factors influence the AU of university mobile learning systems. Second, this study uses the SEM-ANN approach to capture the linear-nonlinear and noncompensatory relationships between exogenous and endogenous variables. Through this approach, the actual usage behaviour of mobile learning systems by university students can be better explained. Finally, the results of this study offer scientifically grounded theoretical guidance that is useful for designers and users of mobile learning systems.
The rest of this paper is organized as follows: the second part is a literature review; the third part introduces the theoretical model of this study and the hypotheses of the relationships between various variables; the fourth part describes the methods and results adopted by this study; the fifth part discusses the results; and finally, the limitations of this study and future research directions are discussed, concluding with the research findings.
Theoretical framework and model
UTAUT2
The UTAUT model is one of the most significant models in technology adoption and was developed by Venkatesh et al. (2003). It comprises four core constructs: performance expectancy, effort expectancy, SI, and FC(Venkatesh et al., 2003). Later, Venkatesh et al. (2012) expanded the UTAUT by adding three additional constructs—hedonic motivation, price value, and habit—to form UTAUT2. Performance expectancy refers to the degree to which an individual believes that using the system will help them achieve academic success (Azizi et al., 2020; DeLone & McLean, 2003). Effort expectancy refers to the ease of use and simplicity of a system (Azizi et al., 2020). SI is the degree to which an individual perceives that significant others (such as peers or teachers) believe that they should use the system for learning (Venkatesh et al., 2003). FC refers to the learner’s insight into the technological and organizational infrastructure and facilities that support the use of the system (Moorthy et al., 2019). The price value is defined as the learner’s understanding of the trade-off between the perceived benefits of the system and the monetary cost of adopting it (Chao, 2019). Hedonic motivation refers to the fun or pleasure of using the system (Chao, 2019). Habit is the extent of habitual behaviour exhibited by students in the learning process (Chao, 2019).
With the rapid development of information and communication technologies and smartphones, understanding user acceptance of information technology has become a crucial research direction in information systems (Adel Bessadok, 2022; Humida et al., 2022). The scope of UTAUT2 includes new technologies, new user groups, and cultural contexts, playing a central role in studies of consumer technology acceptance (Agyei & Razi, 2021). Previous studies have indicated that UTAUT2 can explain up to 70% of the variance in users’ behavioural intention to adopt a technology (Venkatesh et al., 2012). In recent years, UTAUT2 has been increasingly applied in various educational settings (Table 1).
Table 1 Research on the Application of UTAUT.
Full size table
As shown in Table 1, UTAUT2 has been extensively applied to various educational scenarios, especially mobile learning systems (Adel Bessadok, 2022; Humida et al., 2022), flipped classrooms (Agyei & Razi, 2021), learning management systems (Raza et al., 2020), and blended learning (Azizi et al., 2020), among others. In addition to the existing constructs within the UTAUT2 model, previous studies have incorporated additional constructs, such as SN (Humida et al., 2022), social isolation (Raza et al., 2020), PU (Adel Bessadok, 2022), and SEL (Agyei & Razi, 2021), to increase the predictive accuracy of UTAUT2. This model is used primarily in conjunction with the TAM, mainly because, within the same context of researching information technology usage, the TAM is the most widely used and effective model for predicting users’ ATT and intentions toward technology adoption (Adel Bessadok, 2022; Humida et al., 2022).
TAM
The TAM was developed in the 1980s to predict and explain users’ acceptance of information technology (Davis, 1989; Mailizar et al., 2021). The TAM primarily comprises four key variables: PU, PEoU, ATT, and behavioural intention (Davis, 1989). PU is the degree to which an individual believes that using a specific system will enhance their job performance. In contrast, PEoU is the degree to which an individual believes that using a particular system will be free of effort, both physical and mental (Castiblanco Jimenez et al., 2021; Davis, 1989; Wu & Zhang, 2014). ATT refers to an individual’s subjective positive or negative feelings when using the system (Alyoussef, 2023; Mailizar et al., 2021). Behavioural intention is the cognitive process of an individual when preparing to perform a specific action, serving as the direct antecedent to the initiation of the action (Abbasi et al., 2011; Davis, 1989). PU and PEoU determine the user’s ATT and BI, predicting technology acceptance.
In recent years, educational technology has also been continuously refined with the development of the internet and digital technologies. The TAM has been applied in the academic field to predict the acceptance and extent of students’ use of educational technologies (Alyoussef, 2023). In the mobile learning domain, the TAM can be utilized to understand the MOT behind user choices, further optimizing the design of learning tools or applications to meet user needs. A growing amount of research on replicating, modifying, and extending the TAM has been conducted, with some studies incorporating new constructs to improve the accuracy of the TAM’s predictions (Table 2).
Table 2 Research on the Application of the TAM.
Full size table
The literature cited above indicates that the TAM has been applied in the field of education and has been effective in various educational contexts, such as MOOCs (Alyoussef, 2023; Wang et al., 2020), mobile learning systems (Humida et al., 2022), mobile learning (Alajlan et al., 2022; Mailizar et al., 2021), online learning (Liu & Pu, 2020), and learning management systems (Ashrafi et al., 2020). In addition to the primary constructs of the TAM, facets influencing users’ acceptance and use of technology include SN (Humida et al., 2022), IM and EM (Alajlan et al., 2022), and SQ (Mailizar et al., 2021). Furthermore, integrating the TAM with other theoretical research models can effectively compensate for its shortcomings (Wang et al., 2020), such as the UTAUT (Humida et al., 2022; Liu & Pu, 2020), the theory of planned behaviour (TPB) (Wang et al., 2020), and the ECM (Ashrafi et al., 2020). Overall, choosing the TAM as a model to predict and explain the AU of mobile learning systems by students is highly important (Alyoussef, 2023; Humida et al., 2022).
Student aspirations
Students’ EXP is a crucial concept in education and refers to students’ visions and pursuits for their future academic, professional, and personal development (Quaglia & Cobb, 1996). This concept extends beyond mere EXP or hopes, embodying students’ most profound conceptions of what they might achieve at their fullest potential and reflecting their true interests, passions, and goal orientations (Adel Bessadok, 2022). Students’ educational EXP has been proven to be a significant factor affecting their academic performance (Hazel et al., 2013; Tani et al., 2021), as aspirations guide students to recognize the value and benefits of education for their future development, fostering sustained enthusiasm for learning (Wu & Wang, 2020). Student aspirations are related to EXP and manifest in students’ autonomous, internal drive, encompassing their ENJ and satisfaction with the learning process (Moorthy et al., 2019). When students possess a robust intrinsic drive, it is often guided by their MOT to learn and love for learning (Ahmed & Mudrey, 2019; Fu et al., 2022). When these EXPs are met, students can learn more effectively and derive pleasure from the process (Adel Bessadok, 2022).
First, EXP refers to what students believe will happen (Muti Altalhi, 2021). General EXP may include only transient or superficial desires, whereas student EXP is deep-seated, long-term, and sustained goal-setting. Student EXP is more rooted in individual values, interests, and life planning than the simple EXP. For example, in their study on the factors affecting English learners’ adaptation to the flipped classroom, Agyei and Razi (2021) reported that there is a positive correlation between students’ EXP and academic achievements; the higher the students’ EXP is, the better their academic performance.
Second, MOT is the internal force that drives student actions and can be summarized as “the desire to learn.” Student aspirations are a significant source of this MOT (Quaglia & Cobb, 1996). When students have clear aspirations, their MOT tends to be more substantial and more enduring. Adel Bessadok (2022) analysed the success factors of mobile learning systems on the basis of UTAUT2 and TAM theories and revealed that students’ MOT significantly positively affects their intention to use.
Finally, ENJ typically refers to the pursuit of transient, immediate pleasure and satisfaction (Domene et al., 2011). However, an excessive pursuit of perceived expectancy may lead students to make shortsighted choices and abandon long-term, more valuable goals (Smith et al., 2016). Azizi et al. (2020) reported in their study on medical students’ acceptance of blended learning that ENJ significantly impacts students’ behavioural intentions. Moorthy et al. (2019) investigated the factors affecting university students’ usage behaviour of mobile learning systems and reported that perceived expectancy is significantly related to student behaviour; that is, a pleasant learning experience is an essential factor for students using mobile learning systems (Adel Bessadok, 2022).
Research hypotheses
EXP and PU
In the UTAUT2 model, perceived expectancy and effort expectancy are employed to explain users’ anticipation of better usability and ease of use, respectively; thus, the EXP factor is defined as students’ anticipation of the mobile learning experience (Adel Bessadok, 2022; Venkatesh et al., 2002). Previous studies across various fields have discussed the significant positive impact of EXP on PU (Liu & Pu, 2020; Raza et al., 2020). On the basis of the TAM and UTAUT2, Adel Bessadok (2022) investigated the factors of student aspirations that influence the success of mobile learning systems, and the findings indicated a significant positive impact of EXP on PU. In this study, the higher the university students’ EXP of the mobile learning system is, the greater their PU. On the basis of these results, the following hypothesis is proposed for this study:
H1: College students’ EXP with mobile learning systems positively affects PU.
EXP and PEoU
The relationship between EXP and PEoU has received widespread attention in mobile learning systems (Alajlan et al., 2022; Alyoussef, 2023; Humida et al., 2022; Mailizar et al., 2021). First, users with a high EXP of a system or technology tend to tackle difficulties more positively during its use, making the technology seem more accessible to use (Venkatesh et al., 2002). Second, the greater an individual’s EXP of the system is, the more likely they are to devote more time and energy to learning knowledge and skills (Cheng & Yuen, 2020), thereby perceiving the system as more straightforward to use (Davis, 1989). In this study, the higher the university students’ EXP of the mobile learning system is, the higher their PEoU. On the basis of these findings, the following hypothesis is proposed for this study:
H2: College students’ EXP with mobile learning systems positively influences PEoU.
MOT and PU
MOT refers to the internal force within an individual that determines the level, direction, and persistence of effort that they expend at work (Alajlan et al., 2022; Deci et al., 2001). It is important for predicting and adopting mobile learning systems (Azizi et al., 2020; Moorthy et al., 2019). Using the UTAUT2 model, Azizi et al. (2020) explored the factors influencing medical students’ acceptance of blended learning. They reported that learners’ MOT significantly impacts their PU and can even affect their intention to use. In this study, the stronger the university students’ MOT towards the mobile learning system was, the greater their PU. On the basis of these results, the following hypothesis is proposed for this study:
H3: College students’ MOT with mobile learning systems positively influences PU.
MOT and PEoU
As an intrinsic driving force of human behaviour, MOT has been confirmed in numerous fields as a critical factor that positively affects people’s cognition, emotions, and behaviours (Alajlan et al., 2022; Deci et al., 2001). Learners who have solid MOT or a high sense of SEL are more willing to invest time and effort in mastering and using a technology or tool than those who do not (Azizi et al., 2020; Ibrahim et al., 2018). This commitment can make it easier for them to adapt to and understand the tool’s functions, thereby increasing its PEoU (Agyei & Razi, 2021; Salloum et al., 2019). Similarly, highly motivated users seek solutions when encountering usage problems rather than easily giving up (El-Masri & Tarhini, 2017; Raza et al., 2020). Under this persistent MOT, PEoU is significantly increased (Alajlan et al., 2022). In this study, the stronger the MOT of college students towards mobile learning systems is, the greater the PEoU. On the basis of the above findings, the following hypothesis is proposed for this study:
H4: College students’ MOT with mobile learning systems positively influences PEoU.
ENJ and PU
ENJ refers to the degree of pleasure that learners experience when they are using mobile learning systems (Venkatesh et al., 2003). Empirical studies have revealed a direct and close relationship between ENJ and PU; if students find more pleasure in using mobile learning systems, they will perceive them as more useful (Abdullah et al., 2016; Humida et al., 2022). For example, on the basis of the TAM and UTAUT models, Humida et al. (2022) explored the behavioural intentions of college students using e-learning systems, and the results revealed that pleasure during use significantly enhanced the PU of the systems. In this study, the more pleasure college students derive from mobile learning systems, the higher the PU. On the basis of the above results, the following hypothesis is proposed for this study:
H5: College students’ ENJ with mobile learning systems positively influences PU.
ENJ and PEoU
Previous research has confirmed that learners’ ENJ significantly positively affects their PEoU (Alyoussef, 2023; Humida et al., 2022; Rekha et al., 2022), among other factors. On the basis of the TAM and UTAUT theoretical frameworks, Humida et al. (2022) predicted the behavioural intentions of mobile learning systems, with the results indicating that ENJ is an essential predictor of PEoU. Using the TAM framework, Li et al. (2021) investigated the determinants of Chinese higher education students’ willingness to engage in e-learning English, and the results indicated a significant positive effect of hedonic motivation on the ease of use of e-learning English for college students. In this study, the more college students enjoy using mobile learning systems, the greater their PEoU. On the basis of the above results, the following hypothesis is proposed for this study:
H6: College students’ ENJ with mobile learning systems positively influences PEoU.
PEoU, PU, and ATT
The TAM is the most widely used and effective model for predicting user ATT and intentions to adopt technology or systems (Adel Bessadok, 2022; Humida et al., 2022). Previous studies in various fields have confirmed the predictive role of the TAM, such as in MOOCs (Alyoussef, 2023; Wang et al., 2020), mobile learning systems (Humida et al., 2022), mobile learning (Alajlan et al., 2022; Mailizar et al., 2021), online learning (Liu & Pu, 2020), and learning management systems (Ashrafi et al., 2020), among others. Drawing on the diffusion of innovation theory and the TAM, Alyoussef (2023) examined the impact of MOOCs on knowledge management. The study revealed that PEoU has a significant positive effect on PU and that both PEoU and PU significantly influence learners’ ATT. Using the TAM, Mailizar et al. (2021) investigated the behavioural intentions of university students to use e-learning during the COVID-19 pandemic. The results indicated that PEoU is an essential predictor of PU, and both significantly affect students’ ATT towards e-learning, with PU being the strongest predictor of ATT. In this study, the higher the students’ PEoU of the mobile learning system is, the higher their PU and the more positive their ATT; similarly, the higher the PU is, the more positive the students’ ATT. On the basis of these findings, the following hypotheses are proposed in this study:
H7: College students’ PEoU of mobile learning systems positively influences PU.
H8: College students’ PEoU of mobile learning systems positively influences ATT.
H9: College students’ PU of mobile learning systems positively influences ATT.
PU and AU
In mobile learning systems, PU is an essential predictor of AU. Previous research has explored the significant positive impact of PU on AU (Albelbisi, 2020; Ashrafi et al., 2020; Dai et al., 2020; Yang et al., 2023), among other factors. On the basis of the TAM and UTAUT models, Humida et al. (2022) conducted a predictive study on university students’ behavioural intentions to use mobile learning. Their study revealed that when learners perceive the usefulness of mobile learning, they are likely to increase the frequency of their AU. That is, PU has a significant positive effect on AU. Ashrafi et al. (2020) reported in their study on factors affecting students’ continued use of learning management systems that PUs can effectively predict AU. In this study, the higher the university students’ PU of the mobile learning system was, the greater the extent of its AU. On the basis of these results, the following hypothesis is proposed in this study:
H10: College students’ PU of mobile learning systems positively influences AU.
ATT and AU
Previous studies in various domains, such as e-learning (Castiblanco Jimenez et al., 2021; Taat & Francis, 2020), online learning (Liu & Pu, 2020; Safsouf et al., 2020), and MOOCs (Al-Mekhlafi et al., 2022; Muti Altalhi, 2021), have confirmed the significant positive impact of learners’ ATT on AU. Employing the TAM and UTAUT models, Abdou and Jasimuddin (2020) examined the use of mobile learning systems, and the results revealed that ATT has a significant positive effect on AU. Liu and Pu (2020) investigated factors affecting learners’ one-on-one online learning and reported that the more positive the learner’s ATT is, the more willing the learner is to use online learning. In this study, the more positive the college students’ ATT towards the mobile learning system is, the more willing they are actually to use it. On the basis of these results, the following hypothesis is proposed in this study:
H11: College students’ ATT towards mobile learning systems positively influences AU.
This study analyses the impact of students’ aspirations on the AU of mobile learning systems. The article integrates the TAM and UTAUT2 to form a new model (Fig. 1). Aspirations are the first part of the new model, manifested as EXP, MOT, and ENJ, which UTAUT2 influences. The second part retains the critical constructs of the TAM—the PU, PEoU, and ATT.
Fig. 1: Research model.
figure 1
The hypothesized model contains ten research hypotheses of this study.
Full size image
Methods
Sample and data collection
The data for this study were collected from May to September 2023 through the online survey platform Sojump (www.Sojump.com), and the proposed hypothesis model was tested. We used a stratified random sampling method, randomly selecting students from three universities in Hainan Province—Hainan Science and Technology Vocational College, Haikou College of Economics, and Qiongtai Normal College. First, we stratified the students by academic year (freshmen, sophomores, juniors, seniors) and then randomly selected students from each stratum to participate in the survey. We used this method to ensure that students from each academic year were fairly represented, thereby enhancing the representativeness and reliability of the sample.
To ensure the representativeness of the sample and the validity of the research results, we set the following selection criteria: first, participants had to be full-time university students, covering all academic years to ensure the diversity and representativeness of the data. Second, we chose universities within Hainan Province to control potential geographic and cultural factors that could affect the results. Third, all of the participants were volunteers and fully understood the research objectives and requirements. Fourth, all of the participants signed an informed consent form before completing the questionnaire, ensuring the ethical and legal compliance of the study. A total of 518 valid questionnaires were collected. The specific demographic characteristics are shown in Table 3. There were 341 males (65.8%) and 177 females (34.2%), and the majority of the participants were aged 18–20 years (65.1%). Freshmen accounted for 98 respondents (18.9%), sophomores accounted for 259 (50%), and juniors accounted for 130 (25.1%). The chi-square test for gender (males = 65.8%, females = 34.2%; P = 0.514) was used to assess the representativeness of the sample. The sample distribution was not significantly different from the population distribution, indicating strong representativeness.
Table 3 Demographic Information of the Sample.
Full size table
Measurement tools
The questionnaire consisted of two parts: the first part comprised demographic information of the participants, whereas the second part involved the constructs relevant to this study. We used established scales, which we appropriately modified to suit the background and objectives of the research, to assess the constructs in this study. In addition to demographic information, we employed a five-point Likert scale ranging from (1) “Strongly Disagree” to (5) “Strongly Agree” for each construct. Appendix A provides detailed listings of the scales adopted in this study.
The UTAUT questionnaire was adapted from the research of A. Bessadok (2022) and was based on the UTAUT2 scale (McLean, 2003), which is used to assess students’ acceptance and use of e-learning systems. To suit the specific context of this study, we carefully modified some questions, mainly changing third-person statements to first-person statements. These adjustments were aimed at enhancing the personal engagement of participants and the directness of the questionnaire. The questionnaire includes 14 items that comprehensively measure students’ acceptance and usage behaviour of mobile learning systems in four dimensions: EXP, MOT, ENJ, and AU.
Regarding the TAM questionnaire revisions, this scale combines the studies of Mailizar et al. (2021) and Ashrafi et al. (2020), which are based on Davis (1989). To adapt to the mobile learning environment, specific wording adjustments were made in the study; we changed the wording of the PU and PeoU questions from passive to active voice to directly link the questions to the participants themselves. Additionally, for the ATT questions, we shifted the focus to the e-learning system. The TAM questionnaire consists of 10 items aimed at comprehensively assessing students’ acceptance and use of mobile learning systems by evaluating PU (4 items), PEoU (3 items), and ATT (3 items).
Data analysis
We employed two-stage approach to validate the hypotheses and establish predictive models for the following reasons. First, SEM, a theory-driven approach, can detect only linear relationships between exogenous and endogenous variables through a compensatory model, where the increase in another offsets the decrease in one variable (Leong et al., 2018). However, the constructs of EXP, MOT, ENJ, PU, PEoU, ATT, and university students’ AU of mobile learning systems are not merely linear or compensatory by nature. Compared with SEM, ANN can capture linear and nonlinear relationships via non-compensatory models (Sharma et al., 2019), leading to higher prediction accuracy. Moreover, owing to its “black box” nature, the ANN is more suitable for prediction than for hypothesis testing. Finally, ANN analysis can further validate the results obtained from SEM. Thus, this study combines the advantages of both methods, integrating a hybrid approach for hypothesis validation and prediction.
Specifically, in the first stage, SEM was used to reveal the effects of EXP, MOT, ENJ, PU, PEoU, and ATT on the AU of mobile learning by university students and to identify significant predictors. In the second stage, the important variables were taken as input neurons in the ANN to predict the actual use behaviour of mobile learning by university students, ultimately obtaining the accuracy of the predictions and the ranking of critical variables.
Results
Measurement model
To evaluate the measurement model, we had to test the reliability and validity of the questionnaire and data. Reliability can be assessed by examining the factor loadings of each variable. According to (Byrne, 2010), factor loading scores exceeding 0.70 indicate high reliability. Table 4 shows that the factor loadings for each item exceeded 0.70, suggesting that they possess high reliability.
Table 4 Model reliability and validity.
Full size table
Internal consistency reliability can be assessed through composite reliability and Cronbach’s alpha coefficients. In this study, the Cronbach’s alpha coefficients ranged from 0.891 to 0.957, which meets the acceptable standards (0.7 is considered an acceptable threshold). Furthermore, following Hair (2021), composite reliability values between 0.60 and 0.70 are deemed acceptable, and those between 0.70 and 0.90 are generally considered satisfactory. In this study, the composite reliability values for all the items ranged from 0.932 to 0.968, which comply with earlier standards.
Convergent validity can be measured by the average variance extracted (AVE). In this study, the AVE values ranged between 0.818 and 0.910, exceeding the threshold of 0.5. According to Henseler et al. (2015), this value is within an acceptable range, indicating that the results have passed the verification of convergent validity.
Discriminant validity refers to the extent to which a construct is distinct from other constructs(Zaiţ & Bertea, 2011). The Fornell–Larcker criterion is a method used to test discriminant validity (Fornell & Larcker, 1981). According to this criterion, the condition for judging discriminant validity is that the square root of the AVE for each construct should be greater than the correlation coefficient between that construct and any other construct (Henseler et al., 2015). Table 5 shows that the square root of the average variance extracted for each construct is greater than its highest correlation with any other construct. Thus, the results meet the requirements for discriminant validity.
Table 5 Fornell–Larcker criterion.
Full size table
According to the cross-loading criteria, an indicator’s outer loading on the associated construct should be higher than any of its cross-loadings (i.e., its correlations) with other constructs (Hair et al., (2021). Table 6 shows that all the constructs’ outer loadings are greater than the correlation coefficients of the indicators with other constructs. Thus, all the constructs in this study exhibit discriminant validity.
Table 6 Discriminant Validity–Cross Loadings.
Full size table
Structural model
Multiple indicators can be utilized to analyse the structural model in this research, including tests for multicollinearity, significance testing of path coefficient estimates, and the coefficient of determination (R2), among others. These indicators aid in assessing the reliability and explanatory power of the model.
Multicollinearity test
According to Hair et al., (2021) recommendation, a multicollinearity test can be conducted to determine if there are any issues with multicollinearity in the model. According to the rule of thumb, a VIF of less than 3.3 indicates an excellent value (Diamantopoulos & Siguaw, 2006), and a VIF of less than ten is commonly accepted to denote the absence of collinearity (Hair et al., (2021)). Table 7 shows that the VIF values for all the variables range from 3.881 to 5.656, suggesting that multicollinearity is not a concern for this study.
Table 7 Multicollinearity test of the structural model.
Full size table
Hypothesis tests
In the structural model, the goal of significance is to determine the impact of exogenous variables on endogenous variables. Table 8 shows that PU (β = 0.591; t = 8.049; p = 0.000) and ATT (β = 0.386; t = 5.228; p = 0.000) significantly correlate with AU. Thus, hypotheses H10 and H11 are supported. PEoU (β = 0.594; t = 7.754; p = 0.000) has a significant positive effect on ATT, as does PU (β = 0.361; t = 4.637; p = 0.000), confirming hypotheses H8 and H9. Additionally, EXP (β = 0.213; t = 2.702; p = 0.007), MOT (β = 0.173; t = 2.746; p = 0.006), and ENJ (β = 0.519; t = 6.492; p = 0.000) are significant predictors of PEoU, so hypotheses H2, H4, and H6 are supported. There is also a significant positive correlation between EXP (β = 0.224; t = 3.037; p = 0.002), ENJ (β = 0.519; t = 8.531; p = 0.000), PEoU (β = 0.222; t = 3.836; p = 0.000), and PU, which supports hypotheses H1, H5, and H7. However, MOT (β = 0.033; t = 0.622; p = 0.534) does not have a significant positive effect on PU; hence, Hypothesis H3 is not supported.
Table 8 Results of the hypothesis tests.
Full size table
Coefficient of determination (R2)
The coefficient of determination (R2) measures the degree to which independent variables explain the variation in the dependent variable. According to Chin (1998), R2 values can be interpreted as strong (0.67), moderate (0.33), or weak (0.19). Table 9 shows that the R2 for the actual usage behaviour of mobile learning systems is 0.896, which is considered strong, indicating that 89.6% of the variance in this endogenous latent variable can be explained.
Table 9 Explanatory power of the model.
Full size table
Common method bias (CMB)
CMB refers to a noncausal association among sample data in research that arises from the use of the same method, timing, survey tools, or subjective judgement of the researchers, potentially interfering with the accuracy of the results and making the observed associations not necessarily accurate (MacKenzie & Podsakoff, 2012). CMB is assessed through two methods. Harman’s single-factor test indicates that no single factor accounts for most of the variance (Podsakoff et al., 2003). The results show that the largest single factor accounts for 26.237% of the variance, which is well below the critical threshold of 50% (Podsakoff et al., 2003). Second, the marker variable technique, which involves adding a theoretically unrelated marker variable to the research model to test for CMB (Lindell & Whitney, 2001), estimates the largest shared variance with other factors at 0.0215 (2.15%), which is very low (Johnson et al., 2011). Thus, on the basis of these two test results, it can be inferred that there is no significant CMB.
ANN analysis
In the subsequent phase, in parallel with studies such as Liébana-Cabanillas et al. (2017), we incorporated significant factors from the PLS‒SEM path analysis as input neurons for the ANN model (Fig. 2). The rationale for applying ANNs includes nonnormal data distributions and nonlinear relationships between exogenous and endogenous variables. Additionally, the ANN demonstrates robustness against noise, outliers, and smaller sample sizes. They are also adaptable to non-compensatory models where a decrease in one factor does not necessitate an increase in another to compensate. The ANN analysis was conducted using IBM’s SPSS Neural Network Module. ANN algorithms can be used to capture linear and nonlinear relationships without a normal distribution (Teo et al., 2015). This algorithm can be learned through training, employing the feedforward-backpropagation (FFBP) algorithm to predict the analysis outcomes (Taneja & Arora, 2019). Multilayer perceptrons and sigmoid activation functions were utilized for the input and hidden layers (Sharma et al., 2019). The error can be minimized through multiple iterations of learning, further enhancing the accuracy of predictions (El Idrissi et al., 2019). Like Leong et al. (2018), we utilized 70% of the samples for the training process and the remainder for testing. To avoid the potential for overfitting, a tenfold cross-validation procedure was conducted, yielding the root mean square error (RMSE) (Ooi & Tan, 2016). Table 10 indicates that the average RMSE values for the training and testing processes were 0.0547 and 0.0572, respectively, confirming an excellent fit for the model.
Fig. 2
figure 2
ANN diagram.
Full size image
Table 10 Root mean square error values.
Full size table
To measure the predictive power of each input neuron, we conducted a sensitivity analysis (see Table 11) in which the relative importance of these neurons was obtained by dividing their importance by the maximum importance to yield their normalized importance, presented in the form of percentages (Karaca et al., 2019). The results show that PU is the most significant predictive factor, with a normalized importance of 100%. The ATT follows this trend with a normalized importance of 73.0%, followed by ENJ at 33.5%, MOT at 30.8%, PEoU at 25.1%, and EXP at 12.2%.
Table 11 Sensitivity analysis.
Full size table
Discussion
In this study, we aimed to analyse the impact of desire factors (EXP, MOT, and ENJ) on university students’ actual usage behaviour in mobile learning systems. On the basis of a systematic review of the literature, we hypothesized that EXP, MOT, ENJ, PU, PEoU, and ATT play significant roles in the actual usage behaviour of university students in mobile learning systems. We employed Smart-PLS to assess the research model. The results indicated that most of the hypotheses received empirical support, with all these factors collectively explaining 89.6% of the variance in the actual usage of mobile learning systems. A detailed discussion of the research findings initially posited in this study is presented in the subsequent paragraphs.
University students’ EXP significantly positively influenced their PU of mobile learning systems. This finding is consistent with that of Adel Bessadok (2022), who reported that students’ EXP significantly influenced their PU in their study on student desire factors affecting the success of mobile learning systems. On the one hand, modern university students’ needs and EXP for learning tools are becoming increasingly personalized. Mobile learning systems often provide personalized learning experiences and content recommendations, making students more likely to perceive them as useful. On the other hand, if university students have a high EXP for a mobile learning system, it implies that they believe that the system can bring them actual value and benefits. Once they use it and find that it meets their EXP, they are more likely to perceive the system as useful.
University students’ EXP also significantly influenced their PEoU of mobile learning systems. This result aligns with the findings of Humida et al. (2022), who considered SEL as a form of EXP and studied university students’ behavioural intentions via mobile learning systems and reported that students’ SEL was positively correlated with PEoU. In this study, modern mobile learning systems usually have personalized learning functions that offer a tailored learning experience according to the student’s needs and learning styles. By meeting students’ EXP for a personalized learning experience, the PEoU of the systems is greatly enhanced. Second, university students often have high EXP for social interaction; they wish to interact and collaborate with teachers and peers. Mobile learning systems typically provide platforms for online discussions and collaborative learning, ensuring that students’ EXP for interactive learning is met, thereby strengthening their perception of the system’s ease of use.
However, university students’ MOT did not significantly impact the PU of mobile learning systems. This finding is inconsistent with that of Adel Bessadok (2022), who reported that MOT was an essential predictor of PU during the use of mobile learning systems. This inconsistency may have arisen because the content provided by mobile learning systems often does not precisely align with students’ actual learning needs, course objectives, or learning styles. This misalignment can result in students, despite having strong motivation, finding it difficult to use the PU of the system. Furthermore, even if students are highly motivated, issues such as complex interface design or delayed content updates in the mobile learning system can prevent this MOT from translating into a positive evaluation of the learning system. Thus, even with high MOT, adverse conditions related to system functionality may diminish students’ PU of the system.
College students’ MOT significantly positively affects their PEoU of mobile learning systems. This result is consistent with the findings of Azizi et al. (2020), who reported that MOT has a significant positive effect on PEoU when the factors affecting medical students’ acceptance of blended learning are studied. In this study, students with strong MOT tended to explore new learning tools more actively. Even if they encounter operational difficulties, they may feel that these problems are surmountable, and this proactivity may enable them to adapt to and understand the operation of mobile learning systems more quickly, thus making the system more user-friendly. On the other hand, students with higher MOT are often willing to invest more time and effort when they use new tools. This makes them more likely to master the skills and strategies needed to use mobile learning systems, increasing their PEoU (Alajlan et al., 2022).
College students’ ENJ significantly positively affects their PU of mobile learning systems. This outcome aligns with the conclusions of Humida et al. (2022), who reported that the pleasure derived from the use of e-learning systems by college students is positively correlated with PU. In this study, on the one hand, positive emotional feedback from ENJ can enhance individuals’ positive evaluations of the system. When college students experience fun when they use a mobile learning system, they may more easily believe it is useful, as their emotional experiences are interconnected with their cognitive evaluations. On the other hand, when students find the use of a mobile learning system enjoyable, they may be more willing to engage with it. Such high engagement could lead to a deeper exploration of the system’s features, thereby increasing awareness of its practical value.
College students’ ENJ also significantly positively affects their PEoU of mobile learning systems. This result is in line with the findings of Li et al. (2021), who suggested that ENJ significantly positively affects PEoU when college students use e-learning systems to learn English. In this study, when students felt delighted when using a mobile learning system, they may have more easily overlooked minor usability obstacles. An interface that is fun and engaging could make users more tolerant of potential difficulties or challenges, enhancing their perception of the system’s ease of use. Second, if students find a mobile learning system interesting and enjoyable, they may be more inclined to conduct in-depth exploration and learning, which could increase their confidence and familiarity with the system, thus improving their PEoU.
College students’ PEoU of mobile learning systems significantly positively affects their PU. This result is consistent with the findings of Alyoussef (2023), who considered PEoU to be an essential predictor of the PU. In this study, when students find the system easy to use, they may be more willing to explore its various features and characteristics. Such exploratory behaviour could lead them to discover more functions or tools that are helpful for learning, further strengthening their perception of the system’s usefulness. On the other hand, students do not need to invest too much time and effort in learning how to use simpler systems. This means that they can enter the learning content more quickly, thus increasing their perceived value and usefulness.
College students’ PEoU of mobile learning systems significantly positively affects their ATT. This result aligns with the findings of Mailizar et al. (2021), who argued that college students’ PEoU of e-learning systems significantly positively impacts their ATT. In this study, ease of use implies that students can master the system quickly without investing excessive time and effort to learn its usage. This efficiency creates a sense of satisfaction, reinforcing their positive ATT towards the system. On the other hand, when the system is easy to use, students will use it more confidently because they do not have to worry about making mistakes or encountering technical barriers. This confidence enhances their positive ATT towards the system.
College students’ PU of mobile learning systems significantly positively affects their ATT. This result is consistent with the findings of Liu and Pu (2020), who emphasized that college students’ PU significantly influences their ATT in online learning. In this study, mobile learning systems are designed to enhance and facilitate learning. When college students perceive such systems as useful for their academic success, they are more likely to have a positive ATT towards them. Second, college students believe that using mobile learning systems is beneficial for their current learning and preparing them for future careers. In that case, they are likely to have an even more positive ATT towards the system.
College students’ PU of mobile learning systems significantly positively affects AU. This finding aligns with the research of Humida et al. (2022), who argued that college students’ use of mobile learning systems significantly and positively impacts AU. In this study, PU refers to students’ belief that the system provides the knowledge and resources they need, making it easier to achieve learning objectives. Thus, if students believe that the mobile learning system effectively aids their learning, they are more likely to use it in practice. On the other hand, when students find the mobile learning system useful, they may experience more feelings of achievement, satisfaction, and confidence during its use. These positive emotional experiences encourage them to use the mobile learning system more in practice.
College students’ ATT towards mobile learning systems significantly positively affects AU. This outcome corroborates the findings of Liu and Pu (2020), who assert that college students’ ATT significantly influences AU during online learning processes. In this study, when students have a positive ATT towards the mobile learning system, they may feel more capable of using it effectively, thereby being more likely to use it in practice. Moreover, when students have a positive ATT towards the mobile learning system, they may be more emotionally invested, meaning that they are more likely to find using the system enjoyable and worthwhile and, hence, more likely to use it in reality.
Implications
In the context of mobile learning, we integrate the TAM and UTAUT2 models in this study, utilizing Smart-PLS and ANN to process the collected data. Then, we explore the influence of aspiration factors on college students’ actual usage behaviour of mobile learning systems. This model accounts for 89.6% of the total variance in the actual usage behaviour of mobile learning systems. Thus, the results of this study have specific theoretical and practical impacts.
Theoretical implications
First, this research proposes a new model based on the TAM and UTAUT2 models. This model includes traditional variables such as PEoU and PU and particularly emphasizes the impact of aspiration factors (EXP, MOT, and ENJ) on college students’ usage behaviour in mobile learning systems. This model provides a more comprehensive perspective for understanding the relationship between students’ aspiration factors and the actual usage behaviour of mobile learning systems.
Second, unlike existing linear model research, this study adopts a two-stage SEM-PLS-ANN approach, which includes a linear and compensatory PLS model and a nonlinear, non-compensatory ANN model. This is a novel approach because, in linear compensatory models, the decline of one predictor could be offset by the increase in another. However, this may not always be correct, especially in the context of college students’ mobile learning. Thus, by utilizing a non-compensatory neural network model, we successfully address the shortcomings of linear models, thus providing new theoretical contributions to the literature. Moreover, the ANN highlights the influence of the most critical independent variables (PU, ATT, ENJ, MOT, PEoU, EXP). This confirms the high predictive accuracy of the data through analysis via the neural network.
Finally, the findings provide direct theoretical guidance for designing and optimizing mobile learning systems. This study clarifies the importance of enhancing students’ aspiration factors in the design of mobile learning systems, offering theoretical support for designers to consider incorporating more interactivity, gamification elements, and personalized learning pathways when developing applications.
Practical implications
For developers of mobile learning systems, we recommend first enhancing the system’s PU and PEoU. For example, the relevance of educational content can be increased, personalized learning recommendations can be provided, and the user interface can be optimized. Furthermore, considering that MOT and ENJ have a significant positive effect on PEoU, developers should design incentive mechanisms, such as effective system feedback, to enhance students’ MOT. Finally, developers should utilize artificial intelligence technology to customize personalized learning paths on the basis of students’ learning paces and styles, ensuring the efficient use of educational resources.
For university students, we note that the effective utilization of mobile learning systems is crucial. Given the significant impact of the PU and PEoU on the actual AU, students should explore system features such as personalized recommendations and interface simplicity, which can help enhance learning MOT and satisfaction. Additionally, students should engage in more interactive and entertaining learning activities to enhance the learning experience and improve the system’s PEoU. At the same time, students should regularly set and adjust specific learning objectives on the basis of their personal learning experiences and use system feedback to optimize their learning strategies, ensuring the effectiveness and adaptability of their learning methods.
For educators, understanding and applying the data analysis results of mobile learning tools is crucial for improving teaching effectiveness. Educators should adjust teaching methods and content on the basis of students’ feedback on the PU and PEoU. Additionally, given the importance of experience and ENJ in enhancing learning outcomes, educators should adopt more engaging teaching methods and increase the frequency of teacher‒student interactions to increase student engagement and satisfaction. Moreover, educators should actively participate in professional training and seminars to continuously improve their teaching and technical skills and better meet students’ learning needs.
Limitations and future work
Although this study has provided valuable insights into mobile learning systems, it has several significant limitations. First, this study explored the factors that college students focus on when accepting mobile learning systems but did not address other important factors that may affect their AU behaviour, such as personal skills, social influence, technical support, and system quality. These omitted variables could significantly impact the acceptance of mobile learning systems by college students. Additionally, this study did not fully consider specific teaching and learning contexts, such as different learning environments (e.g., blended learning, online learning) and the nature of learning tasks (e.g., collaborative learning, autonomous learning), which may significantly affect the use and effectiveness of mobile learning. Additionally, since this study is limited to a specific group of university students, the generality and breadth of the research results may be restricted; thus, future research should consider including a broader audience and ensure a balanced gender ratio in the sample to increase the applicability of the research findings. Moreover, the study did not fully consider the individual differences in technology use among student groups, such as learning styles, technological proficiency, and SEL, which could affect students’ acceptance and use of mobile learning systems. Finally, although SEM was used for statistical analysis, future research could employ more complex statistical methods, such as multilevel modelling analysis and time series analysis, to increase the accuracy and depth of the research results, which would help to more comprehensively understand the dynamic relationships and trends in data over time.
Conclusion
In this study, we proposed and tested a new research model based on the TAM and UTAUT2 models to predict the desired factors of students affecting the AU of mobile learning systems. A significant innovation of this study over traditional models is the use of ANN technology to analyse and understand complex patterns and nonlinear relationships in the data. The use of an ANN enhances the model’s predictive ability and reveals subtle yet important variable relationships that traditional statistical methods may overlook. For example, in this study, MOT did not significantly impact PU, which might suggest the role of other potential moderating or mediating factors. The introduction of ANN provided a powerful tool for deep learning and pattern recognition, enabling the identification of more precise data variations from complex data, which is crucial for understanding and improving the AU behaviour of students in mobile learning. On the basis of these results, this study offers scientifically sound and feasible suggestions for designers of mobile learning systems, learners themselves, and educational practitioners, suggesting improvements in the functionality of mobile learning systems, identifying the desired factors affecting the AU of mobile learning systems, and ultimately encouraging more people to use mobile learning systems in practice.
Data availability
The raw data supporting the results of this study are in the supplementary file.
References
Abbad MMM (2021) Using the UTAUT model to understand students’ usage of e-learning systems in developing countries. Educ Inf Technol 26(6):7205–7224. https://doi.org/10.1007/s10639-021-10573-5
MATHGoogle Scholar
Abbasi MS, Chandio FH, Soomro AF, Shah F (2011) Social influence, voluntariness, experience and the internet acceptance: An extension of technology acceptance model within a South‐Asian country context. J Enterp Inf Manag 24(1):30–52
MATHGoogle Scholar
Abdou D, Jasimuddin SM (2020) The use of the UTAUT model in the adoption of E-learning technologies. J Glob Inf Manag 28(4):38–51. https://doi.org/10.4018/jgim.2020100103
MATHGoogle Scholar
Abdullah F, Ward R, Ahmed E (2016) Investigating the influence of the most commonly used external variables of TAM on students’ perceived ease of use (PEOU) and perceived usefulness (PU) of e-portfolios. Comput Hum Behav 63:75–90
Google Scholar
Agyei C, Razi Ö (2021) The effect of extended UTAUT model on EFLs’ adaptation to flipped classroom. Educ Inf Technol 27(2):1865–1882. https://doi.org/10.1007/s10639-021-10657-2
MATHGoogle Scholar
Ahmed W, Mudrey RR (2019) The role of motivational factors in predicting STEM career aspirations. Int J Sch Educ Psychol 7(3):201–214
MATHGoogle Scholar
Al-Mekhlafi A-BA, Othman I, Kineber AF, Mousa AA, & Zamil AMA (2022) Modeling the impact of massive open online courses (MOOC) implementation factors on continuance intention of students: PLS-SEM approach. Sustainability 14(9). https://doi.org/10.3390/su14095342
Alajlan N, Hadwan M, Ibrahim DM (2022) Effectiveness of adopting cloud-based E-learning at Qassim University. Comput J 65(5):1098–1106. https://doi.org/10.1093/comjnl/bxaa146
MATHGoogle Scholar
Albelbisi NA (2020) Development and validation of the MOOC success scale (MOOC-SS). Educ Inf Technol 25(5):4535–4555. https://doi.org/10.1007/s10639-020-10186-4
MATHGoogle Scholar
Alshurideh M, Al Kurdi B, Salloum SA, Arpaci I, Al-Emran M (2023) Predicting the actual use of m-learning systems: a comparative approach using PLS-SEM and machine learning algorithms. Interact Learn Environ 31(3):1214–1228
MATHGoogle Scholar
Alyoussef, IY (2023) The impact of massive open online courses (MOOCs) on knowledge management using integrated innovation diffusion theory and the technology acceptance model. Educ Sci 13(6). https://doi.org/10.3390/educsci13060531
Ashfaq M, Yun J, Waheed A, Khan MS, & Farrukh M (2019) Customers’ expectation, satisfaction, and repurchase intention of used products online: empirical evidence From China. SAGE Open 9(2). https://doi.org/10.1177/2158244019846212
Ashrafi A, Zareravasan A, Rabiee Savoji S, Amani M (2020) Exploring factors influencing students’ continuance intention to use the learning management system (LMS): a multi-perspective framework. Interact Learn Environ 30(8):1475–1497. https://doi.org/10.1080/10494820.2020.1734028
Google Scholar
Azizi SM, Roozbahani N, Khatony A (2020) Factors affecting the acceptance of blended learning in medical education: application of UTAUT2 model. BMC Med Educ 20(1):367. https://doi.org/10.1186/s12909-020-02302-2
PubMedPubMed CentralMATHGoogle Scholar
Bessadok A (2022) Analyzing student aspirations factors affecting e-learning system success using a structural equation model. Educ Inf Technol 27(7):9205–9230. https://doi.org/10.1007/s10639-022-11015-6
MATHGoogle Scholar
Byrn, BM (2010) Structural equation modeling with AMOS: basic concepts, applications, and programming (multivariate applications series). New York: Taylor & Francis Group, 396, 7384
Castiblanco Jimenez IA, Cepeda García LC, Marcolin F, Violante MG, & Vezzetti E (2021) Validation of a TAM extension in agriculture: exploring the determinants of acceptance of an e-learning platform. Appl Sci 11(10). https://doi.org/10.3390/app11104672
Chao C-M (2019) Factors determining the behavioral intention to use mobile learning: an application and extension of the UTAUT model. Front Psychol 10:1652
PubMedPubMed CentralMATHGoogle Scholar
Chen J(2022) Application of mobile learning in higher English education systems using cognitive web services. Int J e-Collab 19(2):123
ADSMATHGoogle Scholar
Cheng M, Yuen AHK (2020) Junior secondary students’ acceptance and continuance of e-learning system use: a multi-group analysis across social backgrounds. Behav Inf Technol 41(2):324–347. https://doi.org/10.1080/0144929x.2020.1811378
MATHGoogle Scholar
Chin WW (1998) The partial least squares approach to structural equation modeling. Mod methods Bus Res 295(2):295–336
MATHGoogle Scholar
Dai HM, Teo T, Rappa NA, & Huang F (2020) Explaining Chinese university students’ continuance learning intention in the MOOC setting: a modified expectation confirmation model perspective. Comput Educ 150. https://doi.org/10.1016/j.compedu.2020.103850
Davis FD (1989) Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 319-340
Deci EL, Koestner R, Ryan RM (2001) Extrinsic rewards and intrinsic motivation in education: reconsidered once again. Rev Educ Res 71(1):1–27
MATHGoogle Scholar
DeLone WH, McLean ER (2003) The DeLone and McLean model of information systems success: a ten-year update. J Manag Inf Syst 19(4):9–30
MATHGoogle Scholar
Diamantopoulos A, Siguaw JA (2006) Formative versus reflective indicators in organizational measure development: A comparison and empirical illustration. Br J Manag 17(4):263–282
MATHGoogle Scholar
Domene JF, Socholotiuk KD, Woitowicz LA (2011) Academic motivation in post-secondary students: effects of career outcome expectations and type of aspiration. Can J Educ /Rev Can de l'éduc 34(1):99–127
Google Scholar
El-Masri M, Tarhini A (2017) Factors affecting the adoption of e-learning systems in Qatar and USA: extending the unified theory of acceptance and use of technology 2 (UTAUT2). Educ Technol Res Dev 65(3):743–763. https://doi.org/10.1007/s11423-016-9508-8
MATHGoogle Scholar
El Idrissi T, Idri A, Bakkoury Z (2019) Systematic map and review of predictive techniques in diabetes self-management. Int J Inf Manag 46:263–277. https://doi.org/10.1016/j.ijinfomgt.2018.09.011
Google Scholar
Fornell C, Larcker DF (1981) Evaluating structural equation models with unobservable variables and measurement error. J Mark Res 18(1):39–50. https://doi.org/10.1177/002224378101800104
Google Scholar
Fu X, Yan T, Tian Y, Niu X, Xu X, Wei Y, Hu Q, Ouyang Z, Wu X (2022) Exploring factors influencing students’ entrepreneurial intention in vocational colleges based on structural equation modeling: evidence from China. Front Psychol 13:898319
PubMedPubMed CentralGoogle Scholar
Hair J, Jr., G Tomas M Hult, Christian M Ringle, M Sarstedt (2021) A primer on partial least squares structural equation modeling (PLS-SEM) (3rd ed) SAGE Publications, Inc
Hazel CE, Vazirabadi GE, Gallagher J (2013) Measuring aspirations, belonging, and productivity in secondary students: validation of the student school engagement measure. Psychol Sch 50(7):689–704
MATHGoogle Scholar
Henseler J, Ringle CM, Sarstedt M (2015) A new criterion for assessing discriminant validity in variance-based structural equation modeling. J Acad Mark Sci 43(1):115–135. https://doi.org/10.1007/s11747-014-0403-8
MATHGoogle Scholar
Humida T, Al Mamun MH, Keikhosrokiani P (2022) Predicting behavioral intention to use e-learning system: a case-study in Begum Rokeya University, Rangpur, Bangladesh. Educ Inf Technol 27(2):2241–2265. https://doi.org/10.1007/s10639-021-10707-9
Google Scholar
Ibrahim R, Leng NS, Yusoff RCM, Samy GN, Masrom S, & Rizman ZI (2018) E-learning acceptance based on technology acceptance model (TAM). J Fundam Appl Sci 9(4S). https://doi.org/10.4314/jfas.v9i4S.50
Johnson RE, Rosen CC, Djurdjevic E (2011) Assessing the impact of common method variance on higher order multidimensional constructs. J Appl Psychol 96(4):744–761. https://doi.org/10.1037/a0021504
PubMedMATHGoogle Scholar
Karaca Y, Moonis M, Zhang Y-D, Gezgez C (2019) Mobile cloud computing based stroke healthcare system. Int J Inf Manag 45:250–261. https://doi.org/10.1016/j.ijinfomgt.2018.09.012
MATHGoogle Scholar
Leong L-Y, Jaafar NI, & Ainin S (2018) Understanding Facebook commerce (f-commerce) actual purchase from an artificial neural network perspective. J Electron Commer Res 19(1)
Li C, He L, Wong IA (2021) Determinants predicting undergraduates’ intention to adopt e-learning for studying English in Chinese higher education context: a structural equation modelling approach. Educ Inf Technol 26(4):4221–4239. https://doi.org/10.1007/s10639-021-10462-x
MATHGoogle Scholar
Liébana-Cabanillas F, Marinković V, Kalinić Z (2017) A SEM-neural network approach for predicting antecedents of m-commerce acceptance. Int J Inf Manag 37(2):14–24. https://doi.org/10.1016/j.ijinfomgt.2016.10.008
MATHGoogle Scholar
Lindell MK, Whitney DJ (2001) Accounting for common method variance in cross-sectional research designs. J Appl Psychol 86(1):114–121. https://doi.org/10.1037/0021-9010.86.1.114
CASPubMedMATHGoogle Scholar
Liu N, Pu Q (2020) Factors influencing learners’ continuance intention toward one-to-one online learning. Interact Learn Environ 31(3):1742–1763. https://doi.org/10.1080/10494820.2020.1857785
MATHGoogle Scholar
MacKenzie SB, Podsakoff PM (2012) Common method bias in marketing: causes, mechanisms, and procedural remedies. J Retail 88(4):542–555
MATHGoogle Scholar
Mailizar M, Burg D, Maulina S (2021) Examining university students’ behavioural intention to use e-learning during the COVID-19 pandemic: an extended TAM model. Educ Inf Technol 26(6):7057–7077. https://doi.org/10.1007/s10639-021-10557-5
MATHGoogle Scholar
McLean WHDaER (2003) The DeLone and McLean model of information systems success: a ten-year update
Moorthy K, Yee TT, T’ing LC, & Kumaran VV (2019) Habit and hedonic motivation are the strongest influences in mobile learning behaviours among higher education students in Malaysia. Aust J Educ Technol 35
Muti Altalhi M (2021) Towards understanding the students’ acceptance of MOOCs: a unified theory of acceptance and use of technology (UTAUT). Int J Emerg Technol Learn 16. https://doi.org/10.3991/ijet.v16i02.13639
Nguyen TD, Nguyen DT, & Cao TH (2018) Erratum to: acceptance and use of information system: E-learning based on cloud computing in Vietnam. In Information and Communication Technology (pp. E1-E1). https://doi.org/10.1007/978-3-642-55032-4_72
Niu X, Niu Z, Wang M, Wu X (2022) What are the key drivers to promote entrepreneurial intention of vocational college students? An empirical study based on structural equation modeling. Front Psychol 13:1021969. https://doi.org/10.3389/fpsyg.2022.1021969
PubMedPubMed CentralMATHGoogle Scholar
Niu X, Wu X (2022) Factors influencing vocational college students’ creativity in online learning during the COVID-19 pandemic: The group comparison between male and female. Front Psychol 13:967890
PubMedPubMed CentralMATHGoogle Scholar
Ooi K-B, Tan GW-H (2016) Mobile technology acceptance model: an investigation using mobile users to explore smartphone credit card. Expert Syst Appl 59:33–46. https://doi.org/10.1016/j.eswa.2016.04.015
Google Scholar
Podsakoff PM, MacKenzie SB, Lee JY, Podsakoff NP (2003) Common method biases in behavioral research: a critical review of the literature and recommended remedies. J Appl Psychol 88(5):879–903. https://doi.org/10.1037/0021-9010.88.5.879
PubMedGoogle Scholar
Quaglia RJ, Cobb CD (1996) Toward a theory of student aspirations. J Res rural Educ 12(3):127–132
MATHGoogle Scholar
Raza SA, Qazi W, Khan KA, Salam J (2020) Social isolation and acceptance of the learning management system (LMS) in the time of COVID-19 Pandemic: an expansion of the UTAUT model. J Educ Comput Res 59(2):183–208. https://doi.org/10.1177/0735633120960421
MATHGoogle Scholar
Rekha IS, Shetty J, Basri S (2022) Students’ continuance intention to use MOOCs: empirical evidence from India. Educ Inf Technol 28(4):4265–4286. https://doi.org/10.1007/s10639-022-11308-w
MATHGoogle Scholar
Safsouf Y, Mansouri K, Poirier F (2020) An analysis to understand the online learners’ success in public higher education in Morocco. J Inf Technol Educ: Res 19:087–112. https://doi.org/10.28945/4518
Google Scholar
Salloum SA, Qasim Mohammad Alhamad A, Al-Emran M, Abdel Monem A, Shaalan K (2019) Exploring students’ acceptance of E-learning through the development of a comprehensive technology acceptance model. IEEE Access 7:128445–128462. https://doi.org/10.1109/access.2019.2939467
Google Scholar
Sharma SK, Sharma H, Dwivedi YK (2019) A hybrid SEM-neural network model for predicting determinants of mobile payment services. Inf Syst Manag 36(3):243–261. https://doi.org/10.1080/10580530.2019.1620504
MATHGoogle Scholar
Shen Y, Huang L, Wu X (2022) Visualization analysis on the research topic and hotspot of online learning by using CiteSpace—based on the Web of Science core collection (2004–2022). Front Psychol 13:1059858
PubMedPubMed CentralGoogle Scholar
Smith ML, Mann MJ, Georgieva Z, Curtis R, Schimmel CJ (2016) What counts when it comes to school enjoyment and aspiration in the middle grades. RMLE Online 39(8):1–13
PubMedPubMed CentralGoogle Scholar
Taat MS, Francis A (2020) Factors influencing the students’ acceptance of E-learning at teacher education institute: an exploratory study in Malaysia. Int J High Educ 9(1):133–141
Google Scholar
Taneja A, Arora A (2019) Modeling user preferences using neural networks and tensor factorization model. Int J Inf Manag 45:132–148. https://doi.org/10.1016/j.ijinfomgt.2018.10.010
MATHGoogle Scholar
Tani M, Gheith MH, & Papaluca O (2021) Drivers of student engagement in higher education: a behavioral reasoning theory perspective. Higher Educ 1-20
Teo A-C, Tan GW-H, Ooi K-B, Hew T-S, Yew K-T (2015) The effects of convenience and speed in m-payment. Ind Manag Data Syst 115(2):311–331. https://doi.org/10.1108/IMDS-08-2014-0231
MATHGoogle Scholar
Venkatesh V, Morris MG, Davis GB, & Davis FD (2003) User acceptance of information technology: toward a unified view. MIS Q 425-478
Venkatesh V, Speier C, Morris MG (2002) User acceptance enablers in individual decision making about technology: Toward an integrated model. Decis Sci 33(2):297–316
Google Scholar
Venkatesh V, Thong JY, & Xu X (2012) Consumer acceptance and use of information technology: extending the unified theory of acceptance and use of technology. MIS Q 157-178
Wang Y, Dong C, Zhang X (2020) Improving MOOC learning performance in China: an analysis of factors from the TAM and TPB. Comput Appl Eng Educ 28(6):1421–1433. https://doi.org/10.1002/cae.22310
MATHGoogle Scholar
Wu B, Zhang C (2014) Empirical study on continuance intentions towards E-Learning 2.0 systems. Behav Inf Technol 33(10):1027–1038. https://doi.org/10.1080/0144929x.2014.934291
MATHGoogle Scholar
Wu X, Tian Y (2022) Predictors of entrepreneurship intention among students in vocational colleges: a structural equation modeling approach. Front Psychol 12:797790
PubMedPubMed CentralGoogle Scholar
Wu X, & Wang M (2018) Selection of cooperative enterprises in vocational education based on ANP. Educ Sci Theory Pract 18(5)
Wu X, Wang M (2020) Influence of professional identity and core self-evaluation on job satisfaction of vocational education teachers and the mediating effect of work stress. Rev Argent De Clin Psicol 29(2):31
CASMATHGoogle Scholar
Yang H, Cai J, Yang HH, Wang X (2023) Examining key factors of beginner’s continuance intention in blended learning in higher education. J Comput High Educ 35(1):126–143. https://doi.org/10.1007/s12528-022-09322-5
PubMedMATHGoogle Scholar
Zaiţ A, Bertea P (2011) Methods for testing discriminant validity. Manag Mark J 9(2):217–224
MATHGoogle Scholar
Download references
Acknowledgements
This research was funded by Education Department of Hainan Province, grant number: Hnjg2020-126, and Ministry of Education of the People’s Republic of China, grant number: 2021BCI02003.
Author information
Authors and Affiliations
Hainan Vocational University of Science and Technology, Haikou, China
Henan Wu, Xiaoping Que & Ling Pan
Authors
Henan Wu
View author publications
You can also search for this author inPubMedGoogle Scholar
2. Xiaoping Que
View author publications
You can also search for this author inPubMedGoogle Scholar
3. Ling Pan
View author publications
You can also search for this author inPubMedGoogle Scholar
Corresponding author
Correspondence to Xiaoping Que.
Ethics declarations
Competing interests
The authors declare no competing interests.
Ethical approval
The researchers confirm that all research was performed in accordance with relevant guidelines/regulations applicable when human participants are involved (e.g., Declaration of Helsinki or similar). This study was approved by the Ethics Committee of Hainan Vocational University of Science and Technology, with the approval number: HVUST-2022-01-0016. The approval is valid from January 16, 2022, to January 16, 2025. The scope of approval covers the collection and use of survey data from university students, with particular attention to informed consent and data confidentiality.
Informed consent
Consent was obtained in written form, with participants signing an informed consent form, and the process took place from May to September 2023. All authors of this manuscript obtained consent from all university students participating in the study. Participants agreed to take part in the study, understanding the nature and scope of their involvement, and consented to the collection and use of their data (including survey responses, interview records, etc.) for analysis by the research team. They also consented to the potential use of study results for academic publications, with strict confidentiality of personal information, and their data will not be used for any other purposes.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Data Set 1
Appendix A. Scales of all the constructs
Appendix A. Scales of all the constructs
Construct Item Original content Adjusted content Source
MOT MOT1 He thinks the flexibility in time and space makes using e-learning system very pleasant. I think the flexibility in time and space makes using e-learning system very pleasant. A. Bessadok ([2022](https://www.nature.com/articles/s41599-025-04717-y#ref-CR15 "Bessadok A (2022) Analyzing student aspirations factors affecting e-learning system success using a structural equation model. Educ Inf Technol 27(7):9205–9230.
https://doi.org/10.1007/s10639-022-11015-6")); DeLone and McLean ([2003](https://www.nature.com/articles/s41599-025-04717-y#ref-CR26 "Diamantopoulos A, Siguaw JA (2006) Formative versus reflective indicators in organizational measure development: A comparison and empirical illustration. Br J Manag 17(4):263–282"))|
|MOT2|He uses e-learning system to be similar to other students at prestigious universities.|I use e-learning system to be similar to other students at prestigious universities.|
|MOT3|The use of e-learning system makes him feel that he belongs to a part of the technology revolution|The use of e-learning system makes me feel that he belongs to a part of the technology revolution.|
|EXP|EXP1|He believes that the use of e-learning system will make him more competitive in the local job market.|I believe that the use of e-learning system will make him more competitive in the local job market.|A. Bessadok (2022); DeLone and McLean (2003); Venkatesh et al. (2012)|
|EXP2|He believes that using e-learning system will improve his diploma.|I believe that using e-learning system will improve his diploma.|
|EXP3|He believes that using e-learning system will enhance his skills.|I believe that using e-learning system will enhance his skills.|
|ENJ|ENJ1|For him, e-learning system is a stress-free process due to reduced learning tasks as the task is accomplished immediately.|For me, e-learning system is a stress-free process due to reduced learning tasks as the task is accomplished immediately.|A. Bessadok (2022); Venkatesh et al. (2002); Venkatesh et al. (2012)|
|ENJ2|He prefers to do course tasks through the e-learning system than manually.|I prefer to do course tasks through the e-learning system than manually.|
|ENJ3|He finds it entertaining to learn through e-learning system.|I find it entertaining to learn through e-learning system.|
|ENJ4|He enjoys using the e-learning system to learn.|I enjoy using the e-learning system to learn.|
|PU|PU1|The use of e-learning during university closure due to the COVID 19 pandemic helps me to access learning resources.|Using e-learning allows me to accomplish his tasks more quickly.|Davis (1989); Mailizar et al. (2021); Venkatesh et al. (2002)|
|PU2|Using e-learning will improve learning performance in distance learning during the COVID 19 pandemic.|I believe that using e-learning improves his learning performance.|
|PU3|The use of e-learning will increase my productivity in distance learning during the COVID 19 pandemic.|Using e-learning helps me learn effectively.|
|PU4|The use of e-learning is beneficial for my learning activities during the CoviD19 pandemic.|I believe e-learning in general is useful to him.|
|AU|AU1|He intends to continue using e-learning in the future.|I intend to continue using e-learning in the future.|A. Bessadok (2022)|
|AU2|He will always try to use e-learning in his day life.|I will always try to use e-learning in his day life.|
|AU3|He is intending to visit the e-learning system portal frequently to check news or course information.|I am intending to visit the e-learning system portal frequently to check news or course information.|
|AU4|He plans to continue to use e-learning frequently.|I plan to continue to use e-learning frequently.|
|ATT|ATT1|Using LMS is a good idea.|Using e-learning system is a good idea.|Ashrafi et al. (2020)|
|ATT2|I like using the LMS.|I like using the e-learning system.|
|ATT3|It is desirable to use LMS.|It is desirable to use e-learning system.|
|PEoU|PEoU1|Learning to use e-learning system in distance learning during the COVID 19 pandemic is easy.|I would find e-learning system easy to use.|Mailizar et al. (2021)|
|PEoU2|It is easy to navigate my university’s e-learning system in distance learning during the COVID 19 pandemic.|Learning to operate e-learning system would be easy for me.|
|PEoU3|The use of an e-learning system during the COVID 19 pandemic is flexible.|It would be easy for me to become skilful at using e-learning system.|
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
Reprints and permissions
About this article
Check for updates. Verify currency and authenticity via CrossMark
Cite this article
Wu, H., Que, X. & Pan, L. Analysing students’ aspiration factors that impact actual use of mobile learning systems: a two-stage SEM-ANN approach. Humanit Soc Sci Commun 12, 406 (2025). https://doi.org/10.1057/s41599-025-04717-y
Download citation
Received:29 November 2023
Accepted:10 March 2025
Published:20 March 2025
DOI:https://doi.org/10.1057/s41599-025-04717-y
Share this article
Anyone you share the following link with will be able to read this content:
Get shareable link
Sorry, a shareable link is not currently available for this article.
Copy to clipboard
Provided by the Springer Nature SharedIt content-sharing initiative