Understanding Behavioral Intention to Use Moodle in Higher Education: The Role of Technology Acceptance, Cognitive Factors, and Motivation

Abstract

The adoption of Learning Management Systems (LMSs) such as Moodle in higher education is influenced by a complex interplay of technical, cognitive, and motivational factors. This study extends the Technology Acceptance Model (TAM) by integrating technical system quality (TSQ), educational system quality (ESQ), satisfaction (SAT), anxiety (ANX), and autonomous motivation (AUTO) to provide a more comprehensive understanding of Moodle adoption. A quantitative, cross-sectional research design was employed utilizing Structural Equation Modeling (SEM) to analyze 487 responses from university students with varying levels of Moodle experience. The findings confirm that perceived usefulness (PU), educational system quality (ESQ), and satisfaction (SAT) significantly influence behavioral intention (BI) to use Moodle, whereas Perceived ease of use (PE) was not a significant predictor. Mediation analysis revealed that SAT plays a dominant role in mediating the effects of system and cognitive factors on BI, while ANX exhibited selective and partial mediation effects. Additionally, autonomous motivation (AUTO) moderated the impact of SAT on BI, with results indicating that satisfaction is more critical for low-motivation users. Multi-group analysis further highlighted demographic and usage-based differences, with younger and novice users being more sensitive to technical and cognitive barriers. These findings present both theoretical contributions and inform educational and policy imperatives by extending the TAM in the context of affective and system-level determinants. Institutions should invest in efforts to improve content quality, reduce technology-related anxiety, and establish personalized onboarding and motivation-driven learning strategies.

1. Introduction

Learning Management Systems (LMSs), such as Moodle, are now at the heart of modern higher education, delivering flexible learning and teaching platforms [1,2,3]. Their widespread availability does not always equate to usefulness, since adoption, implementation, and usage difficulties remain [1,2,3]. Moodle remains one of the most widely adopted Learning Management System (LMS) solutions and has been investigated with the Technology Acceptance Model (TAM). The TAM offers a strong theoretical underpinning, but a number of research studies have employed its extended forms to better describe the joint impact of individual, technological, and organizational variables [4,5]. Empirical evidence confirms the effectiveness of Moodle in motivating and improving students’ performance, especially in STEM disciplines. Nevertheless, ample gaps still exist in its application to non-STEM contexts and deployment of interactive and adaptive functionalities with a preference for collaborative learning [4,6]. Apart from individual adoption, infrastructural challenges like inadequate ICT infrastructure, inadequate user training, and limitations in quality assurance still repel widespread Learning Management System (LMS) adoption [7,8,9]. Although perceived usefulness (PU) and ease of use (PE) are the focal points of the TAM, institutional support and system quality are also crucial in determining user participation. Limited pedagogical expertise of teachers and opposition to online instruction lead to underutilization of Moodle’s pedagogic capabilities [10,11,12]. This restricts its usage as a content repository, but not as an interactive learning system. Sabeh et al. [1] and Badia et al. [13] acknowledged the need to expand the TAM with the addition of satisfaction, motivation, and system quality constructs in order to capture the full range of variables influencing adoption. Despite such constraints, Moodle remains a promising platform for blended learning, especially if pedagogical and institutional practices evolve to support interactive engagement [5,10,14]. Filling existing research gaps, the present research uses the TAM to examine the influence of system quality, satisfaction, anxiety, and motivation on students’ behavioral intention to utilize Moodle in higher education.

Most of the existing literature focusing on Learning Management Systems (LMSs), like Moodle, implements core constructs of the Technology Acceptance Model (TAM), namely, perceived ease of use (PE) and perceived usefulness (PU) [2,5]. The crucial cognitive and motivational factors that contribute to behavioral intention are often ignored. The investigation into the mediating variables of satisfaction and anxiety, with the moderating role of autonomous motivation according to Self-Determination Theory (SDT), is very limited, thus leaving gaps in the understanding of user behavior [2,3,15]. Apart from that, system-level factors such as technical system quality and educational system quality are underrepresented in most TAM-based models despite their practical importance. This study therefore addresses these gaps by incorporating cognitive, motivational, and system-level constructs into an extended TAM framework that furthers theoretical knowledge and provides actionable insights for enhancing the adoption of Moodle through tailored interventions, improving system quality and motivation-focused strategies in higher education.

The results showed that perceived usefulness (PU), technical system quality (TSQ), and educational system quality (ESQ) had significant influences on the behavioral intention of Moodle usage (BI), while perceived ease of use (PE) remained insignificant. Satisfaction (SAT) emerged as a key partial mediator of the effects of PU, TSQ, and ESQ on BI, while anxiety (ANX) played a more selective mediating role. Motivation (AUTO) moderated the relationship between SAT and BI by showing that its effect decreases at higher levels of motivation. Multi-group analysis showed that key adoption paths were significantly different across gender, age, prior Moodle usage experience, and frequency of Moodle usage. Younger and inexperienced users were more sensitive to anxiety, whereas experienced users gave greater importance to educational aspects than technical ones.

The rest of the paper is organized as follows: Section 2 provides a literature review of Moodle adoption in higher education, with a focus on the extensions of the TAM by including system quality, cognitive, and motivational factors. Section 3 describes the conceptual framework and research model developed for studying the key determinants of adoption. Section 4 describes the methodology, including the collection, sampling strategy, and data analysis techniques. Section 5 discusses the results of direct, mediating, moderating effects, and multi-group analyses. Section 6 presents practical implications for educators, policymakers, and institutions to improve Learning Management System (LMS) adoption. Finally, Section 7 concludes and outlines the limitations of the study and directions for future research.

2. Literature Review

2.1. Research on Students’ LMS Use and Acceptance Factors

In recent years, the adoption of Learning Management Systems (LMSs)—an example of which is Moodle—has gathered considerable attention with the belief that these will transform higher education through pedagogic renewal and improvement in learner engagement [5,6,16]. Several studies on Learning Management System (LMS) use in the literature highlight some key themes on opportunities and challenges. The research on LMS platforms indicates a uniform transformative potential of students in higher education. On the contrary, significant variabilities have come up with regard to issues about adoption and effectiveness.

LMSs, such as Moodle, have widely been identified as capable of supporting both blended and online learning environments. For example, Ashraf et al. [16] commented on the increasing rise of blended learning in higher education within China, enabled through LMS platforms. The study identifies several key benefits in terms of academic achievement, cognitive engagement, and technical adaptability, but also points out many challenges around technical inefficiencies and resource constraints. Similarly, Ziraba et al. [8] established that Moodle-based e-learning environments (ELEMs) have an effective application in nursing education and bring significant improvement in collaborative learning, satisfaction, and academic performance.

The potential of Moodle as an LMS has also been explored in specific contexts, such as flipped classrooms by Wang [17] and hybrid learning models by Xu et al. [18]. These studies have underlined its utility in facilitating self-regulated learning, engagement, and seamless access to course materials. However, Costa et al. [19] state that Moodle is often underutilized, serving merely as a repository rather than a comprehensive learning tool, thus limiting its potential to enhance teaching and learning processes. Althunibat et al.’s [20] assessment of Moodle during the COVID-19 pandemic further confirms its ease of use, yet also pinpoints certain focused areas for improvement to be fully integrated, and thus supports Zheng et al.’s [3] finding that the habit formation of and tailored strategies notably enhance behavioral intention toward e-learning adoption.

The question of system design is seen as crucial according to studies by Xu et al. [18] and Mella-Norambuena et al. [9], who identify course structure and technological factors as important for satisfaction and use. Both studies point out shortcomings in existing frameworks; for instance, the fact that the TAM only moderately correlates with actual usage illustrates the need to include cognitive and contextual variables. This agrees with Shard et al.’s [2] bibliometric analysis, indicating gaps in understanding teaching styles and perceived learning outcomes. Certain challenges remain regarding the sustainability issue, as can be evidenced in Xie et al. [21], wherein over a period of time, the initial benefits derived from the quality of feedback on Moodle started diminishing and hence call for a more iterative approach towards design. Similarly, Ashraf et al. [16] highlighted the technical and pedagogical challenges of adopting blended learning, with an emphasis on how institutional resources and frameworks are pivotal. In general, LMS integration is considered a positive development; for instance, Ziraba et al. [8] and Xu et al. [18] highlighted increased engagement and learning outcomes. On the other hand, other studies, such as Costa et al. [19] and Ashraf et al. [16], also indicated that there are still several problems in LMS integration, including the underutilization of advanced LMS functionalities and technical inefficiencies. Furthermore, while TAM-based studies are theoretically sound, they often fail to consider the wider educational context, as Shard et al. [2] point out. Other factors, such as system quality, facilitation conditions, and institutional support, are underresearched, though they are crucial in determining adoption and satisfaction [3,9].

Against the backdrop of such challenges, the TAM has been a very reliable starting point for understanding the issue of LMS adoption. However, due to the individualistic level of constructs that it embodies, the TAM can hardly address broader systemic and contextual issues [22,23,24]. Extensions of the TAM model, with the inclusion of technical system quality and education system quality, are proposed to allow its application for studying the adoptive use of platforms like Moodle. The integration of cognitive and motivational factors further explains user engagement and satisfaction in-depth [24,25].

The Technology Acceptance Model (TAM) has been one of the widely used foundational frameworks for the study of LMS adoption. Core constructs, in particular perceived ease of use (PE) and perceived usefulness (PU), have emerged as significant predictors of behavioral intention. Confirmations of the TAM’s predictive power in explaining LMS usage, such as the studies by Zheng et al. [3] and Salloum et al. [25], confirm this, with [3] adding particular emphasis on Habit and Hedonic Motivation based on the Unified Theory of Acceptance and Use of Technology 2 (UTAUT2). Likewise, Shard et al. [2] point out that PE, PU, and user attitudes are significant factors in e-learning acceptance. However, several gaps are observed in TAM-based research regarding how cognitive factors like satisfaction, anxiety, and attitude interact with technology acceptance constructs.

Though the literature reveals areas raising a series of challenges, it shows a potentially powerful Moodle LMS once well designed, supported with effective user training and motivation strategies. The results of the previous studies form a strong basis, but important gaps still exist that this research aims to fill. While embedding system-level constructs such as TSQ and ESQ together with TAM can enrich conventional models in this research study, our study more importantly tries to combine cognitive and motivational factors like satisfaction and autonomous motivation to increase the comprehensiveness of Moodle adoption. These contribute to bridging the theoretical and practical gaps related to Learning Management System (LMS) research, which can be actionable for educators and institutions.

TAM-Based Research on Moodle Adoption

The Technology Acceptance Model (TAM) integrated into the study of Moodle adoption has provided useful insights regarding user acceptance and behavioral intention. In many studies, the importance of TAM constructs, such as PE and PU, was underlined relating to the use of Learning Management Systems (LMS) such as Moodle by students. While some results expose shortcomings of the classical version of the TAM, several scholars have developed various versions and extensions because of the omissions in the traditional TAM [26,27,28].

The basic constructs of TAM, PE, and PU have been widely validated across contexts. For example, Yeou [23] revalidated the robustness of the TAM on assessing Moodle adoption in blended learning settings by noting significant relationships between PU, PE, and students’ attitudes toward Moodle. Also, Escobar-Rodriguez et al. [29] mentioned that knowledge about the factors influencing the intention to use Moodle can help students increase their skills and improve their academic performance by reaffirming the predictive validity of the TAM. However, subjective norms and technological complexity again come into play, and as Lavidas et al. [27] asserted, the TAM does not capture the intricacies of user behavior at higher levels of education.

Researchers have attempted to extend the TAM with additional constructs. For instance, Govender et al. [30] established that some external factors such as technical support, system quality, and information quality are significant antecedents of PE and PU. Surprisingly, their study showed a weak relationship of PE with PU, against the assumptions considered by the TAM. In a similar vein, Salloum et al. [25] added the constructs of computer self-efficacy, perceived enjoyment, and accessibility, which proved effective in explaining the impact on both PE and PU for enriching the TAM’s explanatory power. These findings were in alignment with Sharma et al. [28], who pointed out that system quality and service quality are two prime predictors for the adoption of LMSs. Despite these extensions, there are still significant gaps in the understanding of how cognitive and motivational factors interact with TAM constructs. Barz et al. [31] stressed the role of self-regulated learning and affinity for technology in affecting PE and PU, but stated that prior experience and self-efficacy with digital media did not have a significant effect. In a similar vein, Al-Adwan et al. [26] added constructs such as self-directed learning and learning tradition to the Unified Theory of Acceptance and Use of Technology (UTAUT), with mixed impacts on BI. These studies emphasize the need for an integrative approach that considers individual and contextual variables.

Besides, institutional and system-level factors are often not considered in traditional TAM applications. AL-Nuaimi et al. [24] proved that the quality of the technical system had a crucial effect on PU, while the quality of the system and information influenced PE. These results confirm that institutional support and resources play an important role in creating conditions for LMS adoption. Similarly, Sharma et al. [28] found service quality to be the main predictor of continued LMS use by instructors, pointing out the interaction of technical and personal factors. Furthermore, the complexity of LMS adoption is underlined by contradictions in the literature. For example, Govender et al. [30] indicated that a positive attitude toward Moodle did not ensure greater usage, against the assumptions laid down by the traditional TAM. In a similar vein, Jiang et al. [22] questioned the suitability of the Unified Theory of Acceptance and Use of Technology (UTAUT) for higher education and, hence, suggested Jiang et al. [22], modifying existing models to capture the dynamics of LMS use in academic settings.

Collectively, these findings show that the TAM has a sound basis in research on Moodle adoption despite the challenges and inadequacies of traditional constructs in capturing complexities in user behavior [26,27,28]. In addition, the TAM is able to accommodate cognitive factors such as satisfaction and anxiety along with system and educational quality contextual variables to provide a holistic framework. This study tries to fill some of the gaps in the literature by incorporating factors that enhance the explanatory and predictive power of the TAM, making theoretical contributions and informative practices toward the advancement of technology acceptance studies and optimization of the diffusion process in a higher learning institution context of LMSs.

2.2. Expanding TAM with the Cognitive Factors of Satisfaction and Anxiety

Cognitive factors, integrated with the Technology Acceptance Model (TAM), such as attitude, satisfaction, and anxiety, enriched our understanding of e-learning adoption and provided a better explanation of the users’ behaviors [32]. These factors act as either mediators or amplifiers of the traditional TAM construct relationships, such as PEOU and PU, with BI, pointing out crucial dimensions usually not considered in basic TAM frameworks [33,34,35,36].

Overall, satisfaction has been widely recognized as a critical determinant of e-learning success. For example, Ifinedo et al. [33] discovered that satisfaction acted as the central mediator between system usability variables, such as PEOU and PU, and other student-related outcomes, including academic performance and perceived learning impacts. Satisfaction, based on information quality, was also found by Du et al. [37] to lie at the core of perceived learning outcomes, which in turn are driven principally by the structured and up-to-date content of the learning materials. Other studies have identified contextual factors affecting satisfaction. While Yawson et al. [38] identified generational differences in satisfaction and varied responses to the design and delivery methods of courses among age cohorts, Nguyen-Viet et al. [39] related it to gamified learning environments in which engagement and enjoyment acted as mediators. Such findings stress that satisfaction is not a unitary construct but is instead contingent upon demographic and contextual variables, and adaptive strategies need to be devised for LMS implementation.

Anxiety has already been well established as a negative predictor in the adoption of e-learning. Chen et al. [34] indicated that computer anxiety negatively influences BI mediated by PEOU, thus indicating that reducing anxiety can enhance adoption intentions. Soria-Barreto et al. [40] also pointed out that anxiety is one of the strongest determinants of continuance intention across various cultural settings and is thus universally influential. However, contradictions arise with respect to how anxiety interacts with other constructs. For example, Sabeh et al. [1] found that anxiety negatively influences PEOU, but does not have a significant effect on PU. In contrast, Bossman et al. [41] found that anxiety indirectly decreases satisfaction and performance through perceived learner satisfaction. The aforementioned results suggest the need to further elaborate on how anxiety works in different e-learning environments and user groups.

Attitude has also emerged as the most consistent mediator within the realm of TAM-based research. Azizi [42] explained that attitude towards the acceptance of Moodle is important and how positive attitudes, as influenced by elements like computer self-efficacy and PU, will influence usage intentions. In another related work, Escobar-Rodriguez et al. [29] established that attitudes toward Moodle adoption do enhance the students’ academic abilities and grades, thus highlighting the need to ensure positive perceptions. On the other hand, research also underlines certain challenges. Rivers [43] showed that personality traits of agreeableness and conscientiousness affect attitudes indirectly; hence, individual differences need to be considered. This view was further supported by Barz et al. [31], who established that while attitude is influenced by self-regulated learning and affinity for technology, prior experience does not have a strong impact—a contrast that reveals a gap in understanding how pre-existing technological familiarity interacts with attitude formation.

While satisfaction and attitude are well researched as mediators, the interaction of these with broader systemic variables, such as technical support and information quality, has not been well explored [44]. Furthermore, anxiety as a moderator for relationships between TAM constructs and BI is not fully explored, which also calls for an extension into contexts with differing levels of technological infrastructure and user readiness [35]. Moreover, these findings take into consideration cultural and demographic variables. Works like Al-Adwan et al. [26] and Soria-Barreto et al. [40] underline how different socioeconomic and cultural contexts shape satisfaction and anxiety, underlining an ongoing need to conduct research including more diverse profiles of learners.

Based on these insights, this study attempts to extend TAM-based research by systematically incorporating attitude, satisfaction, and anxiety into its framework [35,39,43,44]. To what extent the mediating effects between constructs such as PU, PEOU, and BI, together with variables like technical and educational system quality, are investigated is a proposed way to address the identified gaps. More precisely, the study attempts to explain how such cognitive factors might interact with the technological and contextual dimensions of a more complete explanation of Moodle adoption within higher education settings, thus adding both to theoretical enrichment and practical strategies aimed at enhancing e-learning adoption.

2.3. The Moderating Role of Motivation

While motivation is increasingly recognized as a critical factor in higher education for e-learning adoption, its moderating role in behavioral intention and learning outcomes has been underexplored. Consistently, motivation has been cited as playing a bridging role between technology acceptance and effective system use. Motivation, intrinsic or extrinsic, influences the key constructs of technology acceptance models and shapes user engagement with e-learning systems like Moodle [45,46,47].

Among such motivational constructs that are related to affecting students’ attitudes towards an e-learning system, enjoyment, expectation, and aspiration come forward. According to Bessadok [48], perceived enjoyment is among the primary predictors of perceived usefulness, which positively impacts user attitude towards an e-learning system. Waheed et al. [47] identified content delivery and communication-related features of intrinsic motivation as aspects that significantly enhance learning and academic outcomes in students. These findings align with Self-Determination Theory (SDT), which emphasizes intrinsic motivation in order to attain higher levels of cognitive engagement [48].

Not only does motivation determine behavioral intention, but it also mediates the relationships between technological factors and learning outcomes. Maldonado et al. [49] integrated motivation into the Unified Theory of Acceptance and Use of Technology (UTAUT) to show how this factor was positively associated with behavioral intention. Furthermore, Pan [45] showed how motivation mediates the influence of technological self-efficacy on attitudes toward self-directed learning, indicating that motivated students are more likely to engage meaningfully with an e-learning system. Conversely, most external motivators, such as facilitating conditions and social influences, often have inconclusive effects. Though Zhang et al. [50] mention hedonic motivation as an important determinant, according to Maldonado et al. [49], facilitating conditions in using e-learning portals show a negligible impact, indicating that external motivators may need some contextual adjustments to influence engagement effectively.

Not all contexts bear equal motivational effects. Maldonado et al. [49] stress that cultural and regional aspects are very important in moderating motivational effects, while Sabah [51] illustrate that individual differences in motivational drivers indicate an important effect on perceptions related to the blended learning environment. These works highlight the need to design adaptive systems that will be able to consider diverse learner profiles and cultural backgrounds. Despite its seeming importance, motivation is a relatively underresearched aspect in most e-learning research. Jiang et al. [52] criticize the limited emphasis that LMS adoption research places on the relationship between motivation and learning success. Furthermore, Law et al. [46] identify that while motivation strengthens social and cognitive presence, its direct influence on learning performance is not clearly established in some contexts.

Current research has placed much emphasis on the integration of motivational aspects into the model of e-learning adoption to enhance the latter’s explanatory power. Considering motivational constructs other than the traditional factors of PE and PU allows studies to present a more complete overview of user behavior. Drawing on these insights, motivation in the current study is treated as a moderating variable that fluctuates the relationships among the constructs of the TAM, particularly between PU and PE and behavioral intention. Therefore, by placing motivation in the context of cognitive and technological factors, this research attempts to advance theoretical and practical interest in the area of e-learning adoption.

Thus, based on the accumulated evidence from the literature review, we highlighted and examined predictive factors of behavioral intention to adopt and use Moodle in higher education, formulating the following hypotheses:

H1.

Perceived usefulness (PU) directly influences behavioral intention (BI) to use Moodle in higher education.

H2.

Perceived ease of use (PE) directly influences behavioral intention (BI) to use Moodle in higher education.

H3.

Technical system quality (TSQ) directly influences behavioral intention (BI) to use Moodle in higher education.

H4.

Educational system quality (ESQ) directly influences behavioral intention (BI) to use Moodle in higher education.

H5a.

Anxiety (ANX) directly influences behavioral intention (BI) to use Moodle in higher education.

H5b.

Satisfaction (SAT) directly influences behavioral intention (BI) to use Moodle in higher education.

H6a.

Anxiety (ANX) mediates the relationship between perceived usefulness (PU) and behavioral intention (BI) to use Moodle in higher education.

H6b.

Satisfaction (SAT) mediates the relationship between perceived usefulness (PU) and behavioral intention (BI) to use Moodle in higher education.

H7a.

Anxiety (ANX) mediates the relationship between perceived ease of use (PE) and behavioral intention (BI) to use Moodle in higher education.

H7b.

Satisfaction (SAT) mediates the relationship between perceived ease of use (PE) and behavioral intention (BI) to use Moodle in higher education.

H8a.

Anxiety (ANX) mediates the relationship between technical system quality (TSQ) and behavioral intention (BI) to use Moodle in higher education.

H8b.

Satisfaction (SAT) mediates the relationship between technical system quality (TSQ) and behavioral intention (BI) to use Moodle in higher education.

H9a.

Anxiety (ANX) mediates the relationship between educational system quality (ESQ) and behavioral intention (BI) to use Moodle in higher education.

H9b.

Satisfaction (SAT) mediates the relationship between educational system quality (ESQ) and behavioral intention (BI) to use Moodle in higher education.

H10a.

Autonomous motivation (AUTO) moderates the relationship between anxiety (ANX) and behavioral intention to use Moodle (BI).

H10b.

Autonomous motivation (AUTO) moderates the relationship between satisfaction (SAT) and behavioral intention to use Moodle (BI).

3. Research Methodology

3.1. Conceptual Model and Rationale

This research focuses on the adoption of Moodle in higher education institutions with the incorporation of perceived ease of use (PE), perceived usefulness (PU), technical system quality (TSQ), educational system quality (ESQ), satisfaction (SAT), anxiety (ANX), and autonomous motivation (AUTO). Anchored in the TAM but extended with additional constructs, our framework allows us to comprehensively analyze factors affecting BI to use Moodle, addressing critical gaps in the literature concerning the interplay of cognitive, system, and motivational factors amid the bourgeoning body of research concerned with LMS adoption in higher education [4,6,44,45,46].

Technical and educational system quality (TSQ and ESQ) are crucial factors in determining the usability and perceived value of Moodle. TSQ addresses technical reliability, stability, security, and personalized functions of the system, while educational system quality (ESQ) examines the platform in terms of the support it offers for different learning styles, interactivity, and assessment features. Both contribute to the perceptions of Moodle’s usefulness, PU, through increased efficiency and productivity while learning, as noted in several earlier studies [24,26]. At the TAM’s core, PE and PU are fundamental constructs that impact BI, while PE describes the ease of use with Moodle and PU represents an improvement in academic performance and learning effectiveness. Both the constructs have a direct and indirect impact on BI through cognitive mediators such as satisfaction (SAT) and anxiety (ANX) [44,53].

SAT represents users’ post-use evaluations, reflecting their fulfillment and enjoyment of Moodle’s performance, while attitude represents their evaluative judgments influenced by PE and PU. These mediators thus reflect the translation of system features and usability perceptions into sustained adoption intentions, in line with prior findings [43,54]. On the other hand, anxiety (ANX) is a psychological barrier that includes fear of failure, error, and technological intimidation. Thus, it has a negative effect on PE and BI [34,43]. ANX can be reduced through user training and supportive system design, thus enhancing acceptance. In addition to the cognitive and technical variables typically explored in TAM-based models, motivational constructs have been increasingly recognized as important drivers of technology acceptance. Furthermore, the autonomous motivation (AUTO) from the Self-Determination Theory (SDT) adds a motivational dimension, since it differentiates between intrinsic and extrinsic motivational drivers of Moodle use [51]. AUTO is also assumed to moderate relationships between TAM constructs and BI because it amplifies positive perceptions and thereby nurtures sustained engagement, as earlier research has shown for instance [46,51]. This comprehensive framework integrates technical, cognitive, affective, and motivational factors that provide a holistic understanding of Moodle adoption in higher education.

The conceptual model, illustrated in Figure 1, visualizes these hypothesized relationships, capturing the interplay of system features, cognitive mediators, motivational factors, and behavioral intentions in Moodle adoption.

3.2. Data Collection and Sampling

This study employs a quantitative cross-sectional research design, which is appropriate for testing and examining the relationships among variables that influence Moodle adoption in higher education [55,56]. These variables include technical system quality (TSQ), educational system quality (ESQ), perceived ease of use (PE), perceived usefulness (PU), satisfaction (SAT), autonomous motivation (AUTO), anxiety (ANX), and behavioral intention (BI). The cross-sectional approach allows data collection at one point in time, thus providing a snapshot of the drivers of Moodle adoption without the need for any longitudinal follow-up [57,58]. This design fits the purpose of the study concerning the exploration of technology adoption behaviors within the educational context.

Therefore, a stratified random sampling procedure is realized to achieve a representative dataset. Indeed, stratification allows balanced representativeness from different groups, such as educational level, academic discipline, and frequency of Moodle use, for detailed insights into subgroup comparisons, enabling enhancements in the generalizability of the results [59,60,61]. Complementarily, snowball-sampling techniques can be employed when targeting responses that are hard to attain by surveys, such as doctoral candidates or students from less specialized faculties [62,63]. While snowball sampling is not a probability-based method, it compensates for specific drawbacks of stratified sampling by enhancing sample diversity and providing broader representation [63,64].

The data collection was performed via an online structured questionnaire, whereas the items were adapted from previously validated scales in the TAM and its related frameworks to ensure content validity with respect to Moodle. The online structured questionnaires were completed on platforms like Google Forms. A questionnaire was sent out and distributed using institutional mailing lists, social media networks, and professional contacts, which provided the highest level of coverage and response rates. This approach allowed for wide demographic coverage, as well as access to the substantial population comprising university students and faculty members. The survey spanned a period of three months (December 2024–February 2025), adequate for the respondents to participate.

Given the exploratory nature of this study’s design, a self-report questionnaire was adopted as the primary tool for data collection to systematically study the interrelationships among the identified determinants. Items in the questionnaire were adapted from previously validated scales to ensure relevance to the context of the study and consisted of a total of 32 items (see Appendix A,

Table A1. Measurements used for data analysis.

Technical System Quality (TSQ)
TSQ1	Moodle includes the necessary features and functions I need.	Adapted from AL-Nuaimi et al. [24]
TSQ2	Moodle does not crash frequently.
TSQ3	Moodle protects my information from unauthorized access by logging only with my account and password.
TSQ4	Moodle provides me with a personalized entry page (e.g., showing my modules, recommending additional modules and courses)
TSQ5	Moodle launches and runs right away. (deleted)
Perceived Ease of Use (PE)
PE1	Learning to operate Moodle is easy for me.	Adapted from AL-Nuaimi et al. [24]
PE2	My interaction with Moodle is clear and understandable.
PE3	It is easy for me to become skillful at using Moodle.
PE4	Overall, I believe that Moodle is easy to use.
Anxiety (ANX)
ANX1	I feel anxious about using Moodle.	Adapted from Sabah [51]
ANX2	I hesitate to use Moodle for fear of making mistakes I cannot correct.
ANX3	Moodle is somewhat intimidating to me.
ANX4	I hesitate to use Moodle for fear of failure and self-doubt. (deleted)
Autonomous Motivation (AUTO)
AUTO1	It is important for me to use Moodle in my learning.	Adapted from Sabah [51]
AUTO2	I value the benefits of using Moodle.
AUTO3	I think it is important to make the effort to use Moodle.
AUTO4	I study using Moodle because it is meaningful to me.
Perceived Usefulness (PU)
PU1	Using Moodle enables me to learn more efficiently.	Adapted from Teo et al. [54]
PU2	Using Moodle improves my academic performance or productivity.
PU3	Using Moodle enhances the effectiveness of my learning.
PU4	Overall, I find Moodle to be useful for my learning.
Behavioral Intention (BI)
BI1	I intend to continue using Moodle in my studies.	Adapted from Teo et al. [54]
BI2	I will use Moodle in the future for learning.
BI3	I plan to use Moodle in my studies as often as needed.
Educational System Quality (ESQ)
ESQ1	I believe that communication facilities have been effective learning components in my study.	Adapted from Al-Adwan et al. [68]
ESQ2	Moodle provides evaluation components and assessment materials (e.g., quizzes, assignments).
ESQ3	Moodle provides me with different learning styles (e.g., flash animation, video, audio, text, simulation, etc.) and they are interesting and appropriate in my study.
ESQ4	Moodle provides interactivity and communication facilities such as chat, forums, and announcements.
Satisfaction (SAT)
SAT1	I am satisfied with the performance of Moodle.	Adapted from Al-Adwan et al. [68]
SAT2	Moodle satisfies my educational needs.
SAT3	Overall, I am pleased with the experience of using Moodle.

). It consisted of two main parts: one regarding the demographic information and the other for the scale-based items relating to the constructs of interest in this study. The survey items were measured using a five-point Likert scale, ranging from 1, indicating strongly disagree, to 5, indicating strongly agree.

The questionnaire was first pretested on a small sample representative of the target population to ensure that the questions were clear and culturally relevant. Based on the results from this pilot, minor adjustments were made to refine the question phrasing for construct validity. Reliability was further ensured using Cronbach’s alpha during data analysis to assess internal consistency.

The sample size estimation was based on the recommendations for SEM, considering at least 10 participants per estimate of a parameter [65,66,67]. For the model in this investigation, there are 32 observed variables; therefore, the sample should be composed of at least 320 participants to guarantee adequate statistical power. To increase the robustness of the model and allow for subgroup analyses, the expected sample size exceeds the minimum requirement, or 320 participants. In all, 487 responses were gathered, which provided adequate power for SEM and allowed for generalization across a wide range of higher education contexts. Diversity in the sample was ensured through stratified random sampling and snowball methods so that that any findings could be generalized in the context of higher education.

3.3. Measurement Scales

Adapted scales were employed to measure and validate the constructs in this study to ensure reliability and context in relation to the adoption of Moodle in higher education settings. Technical system quality (TSQ) was measured by using a 5-item scale that measures functionality, stability, security, and personalization of the technical system [24]. Perceived ease of use (PE) was measured on a 4-item scale reflecting how simple and clear it was to interact with Moodle [24]. Anxiety (ANX) was assessed using a 4-item scale that tapped apprehensions associated with the use of Moodle [51]. Autonomous motivation (AUTO) utilized a 4-item scale to capture intrinsic motivation for engagement [51]. Perceive usefulness (PU) was assessed using a 4-item scale representing the impact of Moodle on enhancing learning efficiency, effectiveness, and productivity [54]. Similarly, behavioral intention (BI) was measured using a 3-item scale that assessed the intention to continue using Moodle [54]. Lastly, educational system quality (ESQ) was measured using a 4-item scale that gauged interactivity [68], learning style support, and assessment. Satisfaction (SAT) was measured on a 4-item scale to reflect overall contentment and enjoyment [68].

3.4. Sample Profile

The sample included 487 participants, with a nearly gender-balanced distribution of 239 (49.1%) female and 248 (50.9%) male participants. Age-wise, most of the respondents fell within the range of 18–25 years old (246, 50.5%), followed by 26–30 years old (144, 29.6%), and then 31–40 years old (97, 19.9%). With regard to educational background, 189 participants (38.8%) had a bachelor’s degree, 267 (54.8%) were pursuing or had completed their master’s, 16 (3.3%) were PhD candidates, and 15 (3.1%) held a doctoral degree. Prior experience with Moodle varied, with 82 respondents (16.8%) having no experience, 131 (26.9%) reporting minimal experience, 168 (34.5%) having moderate experience, and 106 (21.8%) possessing extensive experience. The frequency of Moodle use was also diverse, with 129 participants (26.5%) using Moodle less than once per week, 151 (31.0%) using it 1–2 times per week, 109 (22.4%) using it 2–5 times per week, and 98 (20.1%) using it more than five times per week. The primary motivation for using Moodle included preference over other methods (170, 34.9%), convenience for assignments and submissions (92, 18.9%), easy access to materials (129, 26.5%), and course requirement (96, 19.7%). The results are summarized in

Table 1. Sample profile.

		Frequency	Percentage
Gender	Female	239	49.1%
	Male	248	50.9%
Age	18–25	246	50.5%
	26–30	144	29.6%
	31–40	97	19.9%
Education	Bachelor’s degree	189	38.8%
	Master’s degree	267	54.8%
	PhD candidate	16	3.3%
	Doctoral	15	3.1%
Prior Experience with Moodle	No experience	82	16.8%
	Minimal experience	131	26.9%
	Moderate	168	34.5%
	Extensive	106	21.8%
Frequency of Moodle Use	<1 times per week	129	26.5%
	1–2 times per week	151	31.0%
	2–5 times per week	109	22.4%
	>5 times per week	98	20.1%
Primary Motivation for Using Moodle	Preferred over other methods	170	34.9%
	Convenience for assignments and submissions	92	18.9%
	Easy access to materials	129	26.5%
	Required for course completion	96	19.7%

.

4. Data Analysis and Results

The Structural Equation Modelling (SEM) methodological framework was used to carry out the analysis in this study using SmartPLS 4, Version 4.1.0.0. As structured by Nitzl et al. [69], SEM is one of the variance-based analyses recognized to be extremely effective and thus appropriate for research in management and social sciences. PLS-SEM was chosen because it allows verification of causal models based on the maximization of the variance explained in dependent latent constructs [70,71,72]. Multi-group analysis (MGA) was used further to test subgroup differences; thus, contextual variations in relationships can be found, which are usually missed when traditional regressions are performed [73,74,75]. The analytical procedure followed the recommendations provided by Wong [76] for the accurate estimation of beta coefficients, standard errors, and reliability indicators. Indicators in the reflective measurement model needed to show sufficient relatedness to their respective latent constructs; that is, outer loadings greater than 0.7 were considered acceptable.

4.1. Common Method Bias (CMB)

To examine the results for validity and reliability, a systematic check for common method bias (CMB) was conducted by following the methodological framework provided by Podsakoff et al. [77]. Specifically, Harman’s single-factor test was used to observe whether a single factor accounted for the majority of variance in the model. The unrotated principal factor analysis indicated that the single largest factor explained 24.58% of the total variance, well below the critical threshold of 50%. Although CMB was not a critical concern in this study, its addressing reinforces the validity of the variable relationships and strengthens confidence in the findings through a reduction in potential biases [77,78].

4.2. Measurement Model

The first step of PLS-SEM is to analyze the measurement model in detail, for which the constructs have been measured through reflective indicators. The key properties concerning this measurement assessment would include composite reliability, indicator reliability, convergent validity, and discriminant validity, according to the suggestion by Hair et al. [70].

Indicator reliability, according to Vinzi et al. [71], is another important issue of measurement model assessment that refers to the proportion of a variable’s variance that is accounted for by its respective construct. This assessment is drawn based on outer loading magnitude, which needs to have a value over 0.70, following the guidelines of Wong [76] and Chin [79]. However, Vinzi et al. [71] observed that whereas loadings above this threshold are preferable, social science studies often have lower loadings for some of their indicators. Removal of low-loading items should be done on the basis of their contribution to the composite reliability and convergent validity in order to avoid early exclusion of such items. Hair et al. [80] recommend that indicators with loadings between 0.40 and 0.70 should only be eliminated if their removal leads to a significant improvement in either the composite reliability or the AVE for the respective construct.

Following the recommendations of Gefen et al. [81], the measurement model was optimized by eliminating two indicators, TSQ5 and ANX4, as they exhibited factor loadings below 0.500, as presented in

Table 2. Factor loading reliability and convergent validity.

Constructs	Items	Factor Loadings	Cronbach’s Alpha	rho_A	CR	AVE
Anxiety	ANX1	0.751	0.573	0.639	0.768	0.530
	ANX2	0.843
	ANX3	0.563
Autonomous Motivation	AUTO1	0.616	0.590	0.643	0.784	0.552
	AUTO3	0.864
	AUTO4	0.728
Behavioral Intention	BI1	0.779	0.810	0.824	0.889	0.728
	BI2	0.916
	BI3	0.858
Educational System Quality	ESQ1	0.768	0.795	0.796	0.867	0.619
	ESQ2	0.802
	ESQ3	0.794
	ESQ4	0.783
Perceived Ease of Use	PE1	0.830	0.814	0.577	0.841	0.576
	PE2	0.850
	PE3	0.769
	PE4	0.546
Perceived Usefulness	PU1	0.888	0.891	0.897	0.924	0.752
	PU2	0.856
	PU3	0.876
	PU4	0.850
Satisfaction	SAT1	0.911	0.890	0.891	0.932	0.819
	SAT2	0.902
	SAT3	0.903
Technical System Quality	TSQ1	0.864	0.858	0.920	0.902	0.704
	TSQ2	0.926
	TSQ3	0.930
	TSQ4	0.588

This table presents the outer factor loadings of each item on its associated latent construct, as well as the internal consistency indicators: Cronbach’s alpha, rho_A, composite reliability (CR), and average variance extracted (AVE). Loadings above 0.70 are considered acceptable; items with low loadings (e.g., ANX4, TSQ5) were removed. AVE values above 0.50 confirm convergent validity. All reliability estimates meet or exceed the recommended thresholds.

.

The indicators of reliability in this research included Cronbach’s alpha, rho_A, and composite reliability. Following Wasko et al. [82], the threshold of 0.700 was met on constructs, such as BI, ESQ, PE, PU, SAT, and TSQ, while for the rest of the constructs, the scales were discovered to be moderately to highly reliable, as evidenced in earlier studies also [73,74,75]. The rho_A coefficient, being conceptually situated between Cronbach’s alpha and composite reliability, was above the threshold of 0.7 in most instances and thus supported the results of Sarstedt et al. [74], which in turn supported the results on reliability provided by Henseler et al. [83].

Convergent validity was deemed sufficient as the AVE for most constructs surpassed the 0.50 threshold, as suggested by Fornell et al. [84]. However, Fornell et al. [84] further argued that constructs with AVE less than 0.50 may still be considered to have acceptable convergent validity if the composite reliability is greater than 0.60. Discriminant validity was established using inter-construct correlation analysis, where the values were lower than the square root of AVE, following the approach of Fornell et al. [84]. This was further verified by the heterotrait–monotrait ratio proposed by Henseler et al. [83], with all values below the strict threshold of 0.85, as depicted in

Table 3. HTMT ratio.

	ANX	AUTO	BI	ESQ	PE	PU	SAT	TSQ
ANX
AUTO	0.399
BI	0.591	0.375
ESQ	0.677	0.432	0.730
PE	0.110	0.069	0.067	0.102
PU	0.639	0.308	0.568	0.661	0.053
SAT	0.424	0.273	0.618	0.438	0.092	0.442
TSQ	0.148	0.097	0.215	0.115	0.420	0.041	0.156

Note: This table shows the HTMT ratios between each pair of latent constructs. HTMT values below the threshold of 0.85 indicate acceptable discriminant validity. All values in this analysis meet this requirement, confirming that each construct is empirically distinct.

and

Table 4. Fornell and Larcker criterion.

	ANX	AUTO	BI	ESQ	PE	PU	SAT	TSQ
ANX	0.728
AUTO	−0.249	0.743
BI	0.424	−0.269	0.853
ESQ	0.510	−0.306	0.588	0.787
PE	0.026	−0.046	0.059	0.077	0.759
PU	0.476	−0.223	0.487	0.565	0.035	0.867
SAT	0.286	−0.199	0.525	0.371	−0.104	0.401	0.905
TSQ	0.009	−0.021	0.194	0.095	0.326	0.025	0.146	0.839

Note: The diagonal values (in bold) represent the square roots of the AVE for each construct, which should be greater than the inter-construct correlations in the corresponding rows and columns. This condition is met across all constructs, supporting discriminant validity in the measurement model.

.

4.3. Structural Model

The structural model of the proposed research framework was assessed on the basis of examining the R² and Q² values, considering the path coefficients’ significance [80]. In this study, the R² values were 0.375 for behavioral intention, 0.313 for anxiety, and 0.237 for satisfaction, indicating that the values fell between 0 and 1 and, hence, were acceptable. Besides, the predictive relevance (Q²) scores ranged from moderate to high: behavioral intention scored 0.326, anxiety 0.297, and satisfaction 0.215.

Further validation of the model was achieved by hypothesis testing, which presented the significance of the relationships among constructs. The path coefficients were estimated using the bootstrapping method, as recommended by Hair et al. [80]. The mediation analysis was performed based on the bias-corrected, one-tailed bootstrap approach described by Preacher and Hayes [85] and Streukens and Leroi-Werelds [86] using a 10,000-sample bootstrapping procedure. A summary of these analyses is shown in

Table 5. Hypothesis testing.

Hypothesis	Path	Coefficient (β)	SD	t-Value	p-Value	Results
H1	PU → BI	0.113	0.043	2.657	0.004	Supported
H2	PE → BI	0.019	0.035	0.548	0.292	Not Supp.
H3	TSQ → BI	0.106	0.038	2.807	0.003	Supported
H4	ESQ → BI	0.340	0.043	7.952	0.000	Supported
H5a	ANX → BI	0.285	0.038	7.417	0.000	Supported
H5b	SAT → BI	0.422	0.039	10.942	0.000	Supported

Note: This table summarizes the direct relationships between latent variables and behavioral intention (BI), including standardized path coefficients (β), standard deviations (SD), t-values, and p-values obtained via bootstrapping (10,000 samples). Significant results are marked as “Supported” and indicate acceptance of the corresponding hypotheses (H1–H5b).

.

Perceived usefulness (PU) was positively significant to behavioral intention (BI) to use Moodle (β = 0.113, SD = 0.043, t = 2.657, p = 0.004), supporting H1. However, perceived ease of use (PE) was not a significant predictor of BI (β = 0.019, SD = 0.035, t = 0.548, p = 0.292), and thus H2 was rejected. Technical system quality (TSQ) was positively and significantly related to BI, with β = 0.106, SD = 0.038, t = 2.807, p = 0.003, thus supporting H3. Similarly, educational system quality (ESQ) was found to have a positive and significant effect on BI, with β = 0.340, SD = 0.043, t = 7.952, p < 0.001, thereby supporting H4. Anxiety (ANX) was positively significant to BI for the affective and cognitive mediators, with β = 0.285, SD = 0.038, t = 7.417, p < 0.001, thus confirming H5a. Moreover, satisfaction was the strongest predictor of BI, showing a highly significant and positive relationship (β = 0.422, SD = 0.039, t = 10.942, p < 0.001), confirming H5b. The results indicate that PU, TSQ, ESQ, ANX, and SAT significantly contribute to the adoption of Moodle, but that PE does not directly affect BI.

4.3.1. Mediation Analysis

A mediation analysis was conducted to examine the indirect effects of perceived usefulness (PU), perceived ease of use (PE), technical system quality (TSQ), and educational system quality (ESQ) on behavioral intention (BI) via anxiety (ANX) and satisfaction (SAT). The results are presented below.

The mediation analysis indicated that anxiety (ANX) partially mediates the relationship between perceived usefulness (PU) and BI (β = 0.078, SD = 0.017, t = 4.506, p < 0.001), thus supporting H6a. Also, SAT partially mediated the relationship between PU and BI (β = 0.121, SD = 0.024, t = 5.115, p < 0.001), thus supporting H6b. Since the direct effect of PU on BI remained significant, these findings suggest partial mediation, as both direct and indirect pathways are important in developing BI. For perceived ease of use (PE), anxiety (ANX) did not mediate the relationship between PE and BI (β = −0.000, SD = 0.014, t = 0.019, p = 0.493), thus negating H7a. On the other hand, satisfaction (SAT) fully mediated the relationship between PE and BI (β = −0.079, SD = 0.035, t = 2.287, p = 0.011), thus supporting H7b. Since the direct effect of PE on BI was non-significant, this result indicates full mediation, with the implication that PE influences BI only through SAT. Regarding technical system quality (TSQ), ANX did not significantly mediate its relationship with BI, (β = −0.009, SD = 0.012, t = 0.755, and p = 0.225); thus, H8a was rejected. The effect of TSQ on BI was partially mediated by SAT (β = 0.076, SD = 0.023, t = 3.309, and p < 0.001), thus supporting H8b. Since TSQ had a direct effect on BI, the result shows that partial mediation occurred with TSQ, affecting BI directly and indirectly through SAT. Finally, ANX partially mediated the relationship between ESQ and BI (β = 0.102, SD = 0.022, t = 4.636, p < 0.001), thus supporting H9a. Similarly, SAT partially mediated the relationship between ESQ and BI (β = 0.087, SD = 0.023, t = 3.789, p < 0.001), thus supporting H9b. Given the retained significant direct effect of ESQ on BI, this indicates partial mediation, illustrating the influence of ESQ on BI through both direct and mediated paths. The results are summarized in

Table 6. Mediation analysis.

Hypothesis	Direct Effects	Coeff. (β)	SD	t-Value	p-Value	Results	Mediation Type
	PU → BI	0.113	0.043	2.657	0.004
	PE → BI	0.019	0.035	0.548	0.292
	TSQ → BI	0.106	0.038	2.807	0.003
	ESQ → BI	0.340	0.043	7.952	0.000
	Total Effects	Coeff. (β)	SD	t-value	p-value
	ESQ → BI	0.189	0.028	6.856	0.000
	PE → BI	−0.080	0.041	1.949	0.026
	PU → BI	0.199	0.028	7.161	0.000
	TSQ → BI	0.067	0.027	2.530	0.006
	Specific Indirect Effects	Coeff. (β)	SD	t-value	p-value
H6a	PU → ANX → BI	0.078	0.017	4.506	0.000	Supp.	Partial mediation
H6b	PU → SAT → BI	0.121	0.024	5.115	0.000	Supp.	Partial mediation
H7a	PE → ANX → BI	−0.000	0.014	0.019	0.493	Not Supp.	No mediation
H7b	PE → SAT → BI	−0.079	0.035	2.287	0.011	Supp.	Full mediation
H8a	TSQ → ANX → BI	−0.009	0.012	0.755	0.225	Not Supp.	No mediation
H8b	TSQ → SAT → BI	0.076	0.023	3.309	0.000	Supp.	Partial Mediation
H9a	ESQ → ANX → BI	0.102	0.022	4.636	0.000	Supp.	Partial Mediation
H9b	ESQ → SAT → BI	0.087	0.023	3.789	0.000	Supp.	Partial Mediation

Note: The table includes direct, total, and specific indirect effects for each hypothesized path, using bootstrapped standard errors and significance testing. The results classify mediation as full, partial, or not supported, following Preacher and Hayes’ bias-corrected bootstrap method.

.

These findings stress the position of satisfaction as a dominant mediator, which explains key relationships between system-related and cognitive constructs with behavioral intention. On the other hand, anxiety was found to partially mediate the relationships between PU and BI as well as ESQ and BI but did not mediate the effects of PE and TSQ on BI. By implication, reducing anxiety alone is not sufficient to drive Moodle’s adoption, while increasing satisfaction enhances the user engagement process.

4.3.2. Moderation Analysis

A moderation analysis was conducted to examine the influence of autonomous motivation (AUTO) on the relationships between satisfaction (SAT), anxiety (ANX), and behavioral intention (BI) to use Moodle.

Regarding the moderating effects, AUTO did not significantly moderate the relationship between ANX and BI (β = 0.004, SD = 0.042, t = 0.104, p = 0.459), leading to the rejection of H10b. This implies that autonomous motivation does not change the influence of anxiety on behavioral intention; hence, the effect of anxiety on BI remains invariant across different magnitudes of motivation. Conversely, AUTO significantly moderated the relationship between SAT and BI (β = −0.084, SD = 0.048, t = 1.760, p = 0.039), supporting H10a. The negative interaction effect indicates that with increased autonomous motivation, the positive effect of satisfaction on BI is reduced (

Table 7. Moderation analysis.

Hypothesis	Path	Coefficient (β)	SD	t-Value	p-Value	Results
	ANX → BI	0.285	0.038	7.417	0.000
	SAT → BI	0.422	0.039	10.942	0.000
	AUTO → BI	−0.116	0.040	2.911	0.002
H10a	Moderating effect (AUTO × SAT → BI)	−0.084	0.048	1.760	0.039	Supported
H10b	Moderating effect (AUTO × ANX → BI)	0.004	0.042	0.104	0.459	Not Supp.

Note: This table reports the moderation effects in the model, focusing on the interaction terms between AUTO and both SAT and ANX in relation to behavioral intention (BI).

). Hence, a diminishing return effect occurs, whereby motivated persons may depend less on their satisfaction to drive their behavioral intentions. That is, satisfaction can play a critical role in BI for users with low motivation, but the satisfaction level of highly autonomous learners is less dependent on influencing their behavioral intention to use Moodle.

A simple slope analysis was conducted to further examine the moderating effect of autonomous motivation (AUTO) on the relationship between satisfaction (SAT) and behavioral intention (BI) to use Moodle. The results revealed that with low AUTO, the positive relationship between SAT and BI was much stronger (β = 1.012), suggesting that satisfaction is important in forming behavioral intention. Conversely, for those with high AUTO, the effect of SAT on BI is weaker (β = 0.676), which indicates that highly motivated users rely less on satisfaction in maintaining their intention to use Moodle.

As mentioned above, satisfaction exerts a stronger influence on BI for users with lower autonomous motivation, the BI of highly motivated individuals remains unaffected by their degree of satisfaction. This pattern is confirmed by a simple slope graph (Figure 2), which for low AUTO features a steeper slope, depicting a stronger dependence on SAT.

4.3.3. Multi Group Analysis

The multi-group analysis (MGA) examined whether the structural relationships in the model differ significantly between gender, age, prior experience, and frequency in using Moodle.

The results show that for ANX, ESQ had a significantly different impact for males and females (Δβ = 0.213, p = 0.016). This suggests that for males, ESQ has a stronger effect on anxiety compared to females, indicating that male participants are more sensitive to variations in the quality of the educational system when experiencing anxiety related to the use of Moodle. In the rest of the paths, the differences between males and females were non-significant (p > 0.05).

The MGA revealed significant age-based differences in structural relationships. Specifically, TSQ was a stronger predictor of ANX for the youngest users, 18–25- vs. 26–30-year-olds (Δβ = 0.267, p = 0.004), thus proving that the technical factors contribute more to anxiety in younger students. ANX had a stronger influence on BI for younger participants than older ones: 18–25 vs. 26–30 (Δβ = 0.160, p = 0.041; 18–25 vs. 31–40: Δβ = −0.225, p = 0.009). Regarding the impact on ANX, it appears that ESQ showed significantly different effects between the 18–25- and 26–30-year age group (Δβ = −0.191, p = 0.042). No other paths were significantly different (p > 0.05), indicating stability across the groups.

The results from the MGA suggest that there is a significant difference in the strength of the satisfaction effect on behavioral intention between novice and expert Moodle users (Δβ = −0.141, p = 0.032). This indicates that among novices, satisfaction plays a greater role in driving the intention to use Moodle, while in the case of experienced users, behavioral intention may be driven by other factors beyond satisfaction. The remaining paths were not significant, t (p > 0.05), indicating that the relationships among the constructs did not differ significantly across groups of individuals with prior experience in using Moodle.

The MGA results indicated significant differences in key relationships based on Moodle usage frequency. For example, the effect of TSQ on satisfaction (SAT) was significantly stronger for the low-frequency users (Δβ = 0.390, p = 0.001), suggesting that system quality plays a more important role in shaping satisfaction among less frequent users. In contrast, ESQ had a stronger effect on SAT for the high-frequency user group (Δβ = −0.165, p = 0.042); that is, for experienced users, the quality of the educational system is more valuable in determining their satisfaction. Moreover, AUTO × ANX on BI showed a significant moderation effect (Δβ = 0.145, p = 0.053), which indicates that the role of motivation in mitigating anxiety’s impact on behavioral intention is different across groups. None of the other relationships showed significant differences, which suggests that most structural relationships are invariant across levels of user frequency. The significant differences are reported in

Table 8. Significant MGA results with group comparisons.

Path	Group Comparison	Difference (Δβ)	p-Value
ESQ → ANX	Male vs. Female	0.213	0.016
TSQ → ANX	18–25 vs. 26–30	0.267	0.004
ANX → BI	18–25 vs. 26–30	0.160	0.041
ANX → BI	18–25 vs. 31–40	−0.225	0.009
ESQ → ANX	18–25 vs. 26–30	−0.191	0.042
SAT → BI	Novice vs. Expert	−0.141	0.032
TSQ → SAT	Low Usage vs. High Usage	0.390	0.001
ESQ → SAT	High Usage vs. Low Usage	−0.165	0.042
AUTO × ANX → BI	Low Usage vs. High Usage	0.145	0.053

Note: This table reports only the statistically significant differences in structural path coefficients (Δβ) between groups based on gender, age, prior Moodle experience, and frequency of Moodle use.

.

5. Discussion

The findings of the present study affirm the key determinants that shape the behavioral intention of BI to utilize Moodle in a higher education framework. Also, PU was predicted to have the most significant influence on BI due to its high importance in the TAM (H1) [27,87]. This finding supports previous research [29,42], which suggests that students tend to use and adopt Moodle more when they perceive it as beneficial for improving their learning efficiency and academic performance. However, the relatively moderate effect size suggests that though PU is important, factors other than perceived usefulness underline the adoption of Moodle.

On the other hand, PE did not have a significant effect on BI, which contradicts the findings of the classical TAM (H2). Whereas the role of PE has been validated in several prior studies pertaining to different LMS adoption contexts [25,30], its nonsignificant nature in this study suggests that ease of use alone may not be a decisive factor in adoption, especially among users with prior exposure to e-learning. This supports the prior claim of Lavidas et al. [27], that the assumptions of the TAM may fall short in the case of LMS adoption in higher education, where prior experience with digital platforms alleviates concerns over usability. Instead, the influence of PE seems to be indirect, mainly on user satisfaction (SAT), as will be discussed later using mediation analysis.

BI was significantly influenced by both TSQ and ESQ, with the strongest direct effect being that of ESQ (H3 and H4). These findings confirm the results of previous studies affirming that pedagogical effectiveness, interactivity, and varied learning support are key in LMS adoption [24,28]. The strong effect of ESQ suggests that not only usability but also functionality for supporting different learning styles, structured assessments, and communication with instructors are valued by students when working with Moodle. In line with the perspective of Escobar-Rodriguez et al. [29], LMSs should go beyond usability by implementing educational affordances that increase learning outcomes. The strong effect of TSQ on BI reinforces system reliability, security, and smooth accessibility, complementing and reiterating several related studies that have identified infrastructure as a critical determinant of LMS adoption [30].

More importantly, anxiety (ANX) had a significant effect on BI (H5a), supporting the view of Bossman et al. [41] and Teo et al. [54] that apprehensions about technology may act as deterrents to the adoption of innovations. This aspect is relevant in academic circles where students feel insecure about their prowess with digital tools, thus making them resistant to the use of LMSs. These findings are in alignment with those of Barz et al. [31], who identified that self-regulated learning and digital literacy drive students’ intention to use e-learning environments. Thus, user anxiety needs to be addressed through structured training, intuitive design, and troubleshooting support to guarantee better LMS adoption. This finding supports an emerging literature suggesting that affective outcomes and user experience, operationalized via satisfaction, are more relevant to further technology use than are known cognitive notions, such as perceived ease of use. While the TAM initially placed emphasis on ease of use as a key driver of intention, subsequent LMS research has pointed out that satisfaction increasingly mediates this relationship or supersedes it entirely [41]. This trend is worth highlighting, as it underlines the importance of focusing not only on system usability, but also on rich, engaging learning experiences that promote learner autonomy and enjoyment.

Finally, satisfaction turned out to be the strongest predictor of BI, further supporting its vital role in technology acceptance (H5b). This result also corroborates the findings of Bossman et al. [41], who indicated that users base their continued use of a system on post-use satisfaction. The strong effect of SAT therefore suggests that in Moodle, overall experience, enjoyment, and fulfillment are highly related to continued intention to use the platform. This provides support to other studies that consider user satisfaction a vital mediator between system quality, ease of use, and adoption intention [25,28].

5.1. Mediation Analysis Results

Mediation analysis provides an in-depth understanding of how these factors—cognitive and system-related—influence behavioral intention to use Moodle. In fact, satisfaction emerged as a dominant mediator in explaining the key relationships between technological, cognitive, and affective constructs with the adoption of Moodle. Anxiety, on the other hand, played a more limited role, primarily mediating the influence of perceived usefulness (PU) and educational system quality (ESQ) on BI, while showing no significant mediation for perceived ease of use and technical system quality. Conversely, SAT partially mediated the effect of PU on BI (H6a), showing that while students consider Moodle useful, it is their satisfaction that strengthens the continuance of the intention to continue using it. It has already been argued in the literature that satisfaction is an important determinant in continued e-learning adoption [31,33,38]. Similarly, SAT fully mediated the relationship between PE and BI (H7b), indicating that ease of use alone does not affect BI directly, but through user satisfaction. In this regard, it supports the work of Ifinedo et al. [33], who argued that ease of use often plays a secondary role in the overall learning experience and engagement during LMS adoption. Furthermore, SAT partially mediated the relationship between TSQ and BI (H8b) and that between ESQ and BI (H9b), which implies that both system reliability and educational effectiveness contributed to adoption mainly through satisfaction. This reiterates the role played by improved stability, interactivity in learning, and content for ensuring student satisfaction to make them continue longer [24,28].

ANX partially mediated the effect of PU on BI (H6a), suggesting that students who are less anxious will be in a better position to perceive Moodle as useful and thus more likely to adopt it. This was in agreement with previous research indicating that when apprehensions about technology reduce, perceived usefulness and intentions of adoption can increase among students [34,40]. Similarly, partial mediation of ANX between ESQ and BI (H9b) showed that well-structured and stimulating course materials and instructional support reduce anxiety, which in turn facilitates the adoption of an LMS. However, ANX did not mediate the relationships between PE and BI (H7a) or TSQ and BI (H8a). In other words, while technical reliability or usability can influence adoption, they do not necessarily lead to meaningful alterations in user anxiety. These findings align with studies indicating that anxiety becomes more relevant at the early stage of the adoption curve or for users with low technological proficiency, whereas individuals accustomed to digital learning environments show relatively low anxiety when working with a well-designed LMS platform [1,41].

5.2. Moderation Analysis Results

The moderation analysis follows with the impacts of AUTO on the relationship between satisfaction (SAT), anxiety (ANX), and behavioral intention (BI) in the adoption of Moodle. While it emerged that AUTO significantly moderated the influence of SAT on BI (H10a) but not the relationship between ANX and BI (H10b), this finding consolidates the increasing recognition of motivation as a crucial factor in the adoption of e-learning and also reveals its role within the contextual process of technology acceptance.

Indeed, the analysis showed that AUTO significantly moderated the relationship between SAT and BI, supporting the inverted U-shaped return effect: the positive effect of satisfaction on behavioral intention becomes weaker as users’ autonomous motivation increases. This means that the role of satisfaction in determining BI is more influential in less intrinsically motivated learners than highly motivated learners, who do not depend on their level of satisfaction when developing the intention to use Moodle. This finding was further supported by the simple slope analysis, which indicated that the positive effect of SAT on BI was much stronger among individuals with low AUTO as opposed to those with high AUTO [52,88]. This result is supported by such works as Self-Determination Theory (SDT), which states that highly autonomous learners use e-learning platforms because of intrinsic interest and not for the sake of extrinsic satisfaction [89].

This finding also extends the present literature on motivation in e-learning. Whereas Waheed et al. [47] and Bessadok [48] identified that intrinsic motivation improves engagement with the learning system, the studies by Pan [45] and Maldonado et al. [49] demonstrated that motivation mediates the influence of technological self-efficacy on learning outcomes. However, our study indicates that motivation also moderates key cognitive relationships, especially the degree to which satisfaction leads to behavioral intention. This underscores the need for differentiated engagement strategies: while improving satisfaction is critical for less motivated students, highly autonomous learners may benefit more from features that enhance self-directed learning and personalization, as noted by Law et al. [46].

However, AUTO did not moderate the relationship between ANX and BI, which indicates that autonomous motivation does not change the influence of anxiety on behavioral intention. In other words, regardless of whether students have high or low intrinsic motivation, their level of anxiety continues to influence their intention to adopt Moodle. This contradicts some previous studies that suggested that motivation may buffer negative emotional states in technology adoption [40]. However, this is in agreement with studies suggesting that technology-related anxiety is affected more by factors of external support structures, like training and the usability of systems, than internal motivational states [34,41].

5.3. Multi-Group Analysis (MGA)

The MGA highlighted several demographic and contextual differences in the adoption of Moodle while confirming stability in the majority of the relationships. Gender-based differences emerged for the relationship between ESQ and ANX, as males appeared more sensitive to ESQ when anxiety was high, indicating that instructional design would play a more important role in the reduction of anxiety for male students [7,15]. The age-based differences indicated that TSQ had greater influences on shaping anxiety among the relatively younger participants, aged between 18 and 25 years, probably as they are less accustomed to the use of Moodle [2,5]. Also, ANX impacted BI more in younger users. The findings extend the call for targeted onboarding programs. Also, the impact of ESQ on ANX showed differences between relatively younger and older users, representing sensitivities that may differ among educational content users of different ages. Prior experience with Moodle significantly moderated the role of satisfaction in BI. While novices rely more on SAT, experienced users may depend on habit or familiarity, underscoring the fact increased usability and effective support for new users are crucial to driving adoption [4,6]. In addition, Moodle usage frequency demonstrated divergent paths of adoption. Low-frequency users rely more on TSQ for satisfaction; that is, for infrequent users, the quality of the system is important in terms of engagement. Conversely, for high-frequency users, ESQ is more effective toward SAT, supporting the view that instructional design becomes more critical for sustaining engagement. Also, the moderation of ANX on BI by AUTO proved that motivation may play a somewhat different role in mitigating anxiety across levels of use.

The MGA results also highlight the demographic and contextual factors as very important in shaping Moodle adoption behaviors. The main relationships are moderated by gender, age, experience, and frequency of use, suggesting that one-size-fits-all LMS implementation approaches are likely to be ineffective and that each institution needs to adopt engagement strategies tailored to the unique needs of its different learner segments [4,8,9].

Table 9. Summary of key findings.

Key Relationship	Finding
H1: PU → BI	Perceived usefulness (PU) predicts behavioral intention (BI) strongly, but a moderate effect size suggests that there might be other determinants besides usefulness.
H2: PE → BI	Perceived ease of use (PE) does not significantly affect BI, possibly due to prior LMS experience of students.
H3: TSQ → BI	Technical system quality (TSQ) has a substantial influence on BI, highlighting the importance of platform reliability and accessibility.
H4: ESQ → BI	Educational system quality (ESQ) has the strongest direct impact on BI, validating the importance of pedagogical aspects and interactivity.
H5a: ANX → BI	Anxiety (ANX) significantly impacts BI; higher anxiety reduces the likelihood of adopting Moodle.
H5b: SAT → BI	Satisfaction (SAT) is the strongest overall predictor of BI, validating its dominance in the long-term use of technology.
H6a–H9b: Mediation effects	SAT partially or fully mediates the effect of PU, PE, TSQ, and ESQ on BI. ANX mediates the effect of PU and ESQ, but not PE or TSQ.
H7b: PE → SAT → BI	Full mediation: Ease of use impacts BI only through satisfaction.
H6a: PU → SAT → BI	Partial mediation: Usefulness increases satisfaction, which strengthens BI.
H9a: ESQ → ANX → BI	Partial mediation: Better educational quality reduces anxiety, thereby increasing BI.
H10a: AUTO × SAT → BI	Moderation confirmed: Satisfaction matters more for less autonomously motivated students; the effect weakens with high AUTO.
H10b: AUTO × ANX → BI	Not supported: Autonomous motivation does not moderate the impact of anxiety on BI.
MGA: Gender	Males are more sensitive to ESQ when anxiety is high.
MGA: Age	Younger users (aged 18–25 years) are more affected by TSQ and ANX, requiring more support and system reliability.
MGA: Experience	Novice users rely more on satisfaction for BI; experts depend on familiarity or habit.
MGA: Usage frequency	Low-frequency users depend more on TSQ for satisfaction; high-frequency users prioritize ESQ.
MGA: AUTO × ANX → BI (usage-based)	Moderation approaches significance; motivation may play a nuanced role across usage frequency groups.

shows an organized overview of the principal empirical findings, reporting supported associations, mediation and moderation effects, and group differences revealed through analysis to supplement the narrative analysis.

6. Practical Implications

This study provides valuable insights for educators, policymakers, and LMS developers aiming to improve the use of Moodle and enhance digital learning in higher education. Institutions can create focused strategies that enhance engagement, reduce barriers, and improve user satisfaction by addressing technical, educational, cognitive, and motivational factors. To serve novice or less proficient learners better, institutional policy needs to make formal onboarding programs, accessible digital literacy training, and proactive technical support a staple. Such policies can directly reduce technology-related anxiety (ANX) while enhancing perceptions of system and educational quality. Integrating such support into the student experience early on can instill confidence and ease the transition to the Moodle learning environment.

6.1. Implications for Educators and Higher Education Institutions

To begin with, for institutions of higher learning and educators, ensuring TSQ and ESQ is critical, as both factors considerably influence students’ behavioral intention in using Moodle [5,6]. Institutions should improve system stability, security, and usability while upgrading instructional design, content delivery, and interactive learning tools. Moreover, with satisfaction being the most important mediator, efforts have to be aimed at enhancing users’ experiences in terms of creating personalized learning tracks, gamifying, and thus making course content adaptive. This would increase perceived usefulness among the students and then create long-term usage intentions among them [4,16].

Other key issues involve reducing technology-related anxiety (ANX). Institutions should provide structured onboarding programs, digital literacy training, and responsive technical support to put students at ease, especially younger and inexperienced users who struggle with the complexity of systems. Besides, as frequent users emphasize educational quality, while infrequent users are influenced by system quality, tailored interventions must account for both groups of users. Beginners can appreciate better support in usability and navigation, whereas advanced users might need more functionalities like peer collaboration and assessment functionalities [5,42,50].

In addition, the results could be used to redesign teaching strategies through Moodle to enhance engagement and satisfaction. Some of the redesigns include incorporating greater collaboration, using feedback mechanisms effectively, clearly organizing navigation, and developing an interactive online presence [4,16]. The use of interactive features of Moodle, such as forums, quizzes, and personalized content, can help reduce anxiety among learners and perceived educational value drivers in the utilization and continued use of Moodle [4,16].

6.2. Implications for Policymakers

In addition to institutional actions, the findings of the present study bring into view supportive policy frameworks for education policymakers and decision-makers to adopt LMSs, combining technical infrastructure improvements with faculty training and student support services. That policy should inspire investment in high-quality digital learning environments; LMS platforms should be designed to assist diverse learning preferences and technological competencies. Additionally, budgeting and incentives for faculty development programs will enable instructors to better utilize Moodle’s interactive and pedagogical tools, leading to improved student engagement [32,37,52].

Secondly, institutional policy needs to be in the interests of novice or less proficient learners by including scaffolding features in LMS uptake. These include compulsory orientation modules, peer mentoring schemes, and continued access to help desks or live technical assistance. Not only do these policies help reduce initial anxiety, but they also allow all students to engage constructively with Moodle regardless of their prior experience with digital learning environments. By mandating inclusive digital accessibility guidelines and ongoing feedback cycles with final users, policymakers can assist in keeping system and learning quality responsive to shifting learner demands [32,37,52].

6.3. Implications for LMS Developers and Designers

This study, therefore, provides a basis on which LMS developers and system designers could focus their efforts, developing user-centered design principles that enhance usability and educational functionality. Features that support customization, accessibility, and real-time feedback should be developed with priority for satisfaction and long-term adoption. Besides, the incorporation of an autonomy-supportive feature in the system, such as self-directed learning modules and AI-driven course recommendations, might be helpful for highly motivated learners who rely less on system satisfaction but require greater flexibility and personalization [5,90].

Overall, the findings contribute to the reinforcement of multilayer approaches for LMS adoption, where technical efficiency is balanced with pedagogical efficiency and user experiences to attend to the diverse needs of students. Stakeholders can utilize evidenced-based practices to develop learning environments that are more digital, inclusive, and engaging—thereby delivering the full value that Moodle and similar platforms offer to higher learning institutions.

7. Conclusions, Limitations, and Future Research

The current research enriches the understanding of Learning Management System adoption, integrating technical, cognitive, and motivational factors within the Technology Acceptance Model (TAM). The results underline that high usability of the system in itself will not guarantee effective adoption; what actually matters is educational quality, user satisfaction, and motivation. While perceived usefulness is still a principal driver of behavioral intention, user engagement itself is most powerfully influenced by antecedents such as TSQ, ESQ, satisfaction, and anxiety. In addition to the theoretical contribution, this study emphasizes the practical guidelines for educators, policymakers, and institutions to adopt an inclusive strategy toward improving robustness in the technology infrastructure, pedagogic efficiency, and diversification in the learning pathways.

This research is not without limitations; however, these limitations provide avenues for further research to build upon and extend the findings. One limitation involves the cross-sectional research design, which evaluates participants’ perceptions at a single point in time [56,91]. In this respect, even though the approach offers a snapshot of Moodle adoption factors, it fails to consider the possibility of longitudinal changes in user behavior, satisfaction, or motivation over time. Longitudinal designs for future studies could trace changes in behavioral intention to use an LMS and actual LMS use over time as experience with Moodle grows or when policies at the institutional level change [55]. Another limitation is related to the fact that the data come from self-reports, requiring participants to subjectively assess their experience with Moodle. Although we controlled and checked for common method bias, which, in our case, did not pose an issue, self-reported data may create social desirability or response bias due to overreporting or underreporting of satisfaction, anxiety, or motivation [92,93]. Therefore, usage data can also be based on other objective data besides self-reporting to have a comprehensive overview of user engagement. Moreover, while the sample was diverse, it was limited to higher education students only; this may be a factor to consider in the generalization of other user groups, such as instructors, administrators, or even corporate e-learning users [5,15]. While the research extended the TAM to include other aspects, some additional constructs could be considered. These could include perceived interactivity, social presence, and features for collaborative learning that might increase the power of the model [27,28]. The present limitations offer an opportunity to make improvements to the present work by redefining and refining the LMS adoption model to foster an inclusive environment for effective digital learning. Additionally, the study did not take into account users’ qualitative Moodle experience, such as perceived quality, interactivity, and variety of course materials they might have access to, and also did not address how Moodle usage would be placed against their specific course materials. These experiential factors are likely to significantly affect users’ satisfaction, anxiety, and behavioral intention, but were outside the scope of this study. Also, we did not control for disciplinary origins (e.g., STEM and non-STEM students) directly, which could influence use patterns and LMS expectations. Future research could add these contextual variables to the model and investigate how the field of study moderates adoption relationships. Finally, the cultural and institutional environment of this research should be taken into account when interpreting the results. The data were gathered from Greek universities only, which function within a certain national policy context of education and the digital environment. These contextual elements—like centralized curriculum policies, degrees of digital preparedness, and institutional support mechanisms—might influence students’ perceptions of Moodle in ways that are distinct from other nations. Therefore, future research could explore whether the suggested model is applicable across different educational systems and cultural environments.

As the concept of digital learning environments continues to evolve, understanding the complex interplay between technical, cognitive, and motivational factors in LMS adoption remains very relevant to improving student engagement and educational outcomes. In continuously refining our notion of digital learning environments, institutions can further optimize the use of technology and ensure better experiences for students in shaping the future of technology-enhanced education.