The Effects of Social and Spatial Presence on Learning Engagement in Sustainable E-Learning

Abstract

E-learning offers sustainable opportunities for university students by improving educational accessibility and inclusion; however, research exploring learning engagement to ensure continuous learning quality from the perspective of presence remains limited, especially concerning spatial presence and its combined effects with social presence. The purpose of this study is to investigate how social presence and spatial presence interact to predict learning engagement in sustainable e-learning, as well as to determine whether perceived enjoyment mediates these relationships. A total of 442 Chinese university students participated in a self-report survey. Partial least squares structural equation modeling (PLS-SEM) was used to test the conceptual model, indicating that both social presence and spatial presence play a significant positive role in e-learning engagement, with perceived enjoyment confirming its positive role as an important partial mediator. The findings make contributions to empirical research on learning engagement by optimizing presence-related factors and positive emotions in sustainable e-learning. The implications suggest that instructors and platform providers can enhance engagement by improving social presence, spatial presence, and perceived enjoyment, as outlined in the proposed solution model “social presence and spatial presence → perceived enjoyment → learning engagement”. However, due to limitations such as self-reported data, a cross-sectional study design, and cultural context, future research could incorporate behavioral data, longitudinal studies, and explore responses from diverse cultural backgrounds.

1. Introduction

The rise of technology-driven education has reshaped learning environments, making e-platforms indispensable for delivering accessible and inclusive education. By 2022, the global e-learning market had exceeded USD 300 billion and will expand rapidly at a 14% compound annual growth rate (CAGR) from 2023 through 2032 [1]. This shift has promoted the development of e-learning as a sustainable practice, helping reduce energy consumption and lower carbon emissions. Therefore, sustainable e-learning entails leveraging technology to ensure high-quality teaching while minimizing costs by resource reuse [2]. Despite its rapid expansion and the availability of extensive resources, sustainable e-learning faces persistent challenges, particularly in enhancing high-quality, engaging, and interactive content. Learning engagement is directly linked to sustainability: increasing engagement can reduce dropout rates, thereby enhancing the efficient use of educational resources and improving learning equity [3].

Therefore, to ensure sustainable e-learning, characterized by continuity, durability, and proactivity, it is vital to explore strategies that enhance learner engagement. The growth of this concern in understanding user behaviors within e-learning settings has been demonstrated by researchers, particularly in information systems. This includes investigating the essential mechanisms prompting action and approaches to enhance engagement or grades of learners [4,5,6]. Rooted in this body of studies, interactive behaviors exhibited by e-learners are studied to intensify insights. Specifically, the current study explores the present situation of presence and examines how different types of presence influence learner behaviors. Furthermore, the study offers insights into how providers build sustainable e-learning formulated upon presence frameworks to maximize learners’ experience and engagement. This is closely aligned with the United Nations SDG4, which focuses on providing inclusive, high-quality education for everyone and fostering opportunities for lifelong learning [7].

In opposition to traditional face-to-face learning, an e-learning platform could create new virtual and social spaces sustainably and effectively [8]. These spaces offer learners available and personalized learning opportunities, enabling avenues for interaction and experience through LMS-controlled systems. Interactions in e-learning extend beyond the interaction between learners and instructors or learners’ relationships to include interacting with content [9]. Within this context, social presence and spatial presence as two core aspects are grouped into the presence conception [10].

Social presence refers to a sensation of “real” interaction with others in a medium [11], shown to strengthen learners’ involvement. Similarly, spatial presence is defined as a person feeling “being there” in mediated environments [12]. Both social presence and spatial presence play a part in students’ sense of connection and belonging, which are critical to improving learning engagement. The two senses of presence have been recognized as a critical psychological phenomenon within virtual and other media settings [13]. This study also investigates the underlying mechanisms of the relationships. Psychological factors, such as emotional responses (e.g., enjoyment) and cognitive processes, are recognized to significantly affect individual behavior [14]. Social and spatial presence is stated to enhance user engagement by fostering communication with others, bringing responses to computer-mediated stimuli, and promoting a perception of enjoyment within these environments [15]. Building on this foundation, the study empirically examines perceived enjoyment as a mediator in the association of presence with learning engagement.

Further analysis of the literature on engagement reveals that exploring social presence, spatial presence, and perceived enjoyment is crucial for sustainable e-learning. However, the role of spatial presence in e-learning platforms remains less explored. Most studies on presence concern individual structures, such as the Community of Inquiry model [16], neglecting the interaction with social presence and spatial presence, as well as their reciprocal impacts on learning behavior. Moreover, perceived enjoyment as a mechanism linking social presence, spatial presence, and e-learning engagement has received limited attention in prior research. Hence, the main purpose of the research is to tackle gaps in the studies by exploring how social presence interacts with spatial presence and perceived enjoyment to influence learning engagement within e-learning platforms. To this end, the guidance of the research questions is followed by this research: (1) How are the current situations regarding social presence, spatial presence, perceived enjoyment, and learning engagement within sustainable e-learning? (2) How do social presence and spatial presence interact to influence learning engagement within sustainable e-learning? (3) How is perceived enjoyment played as a mediator in the links between social presence together with spatial presence on learning engagement?

2. Literature Review

2.1. Learning Engagement

Learning engagement involves the investment of time and endeavor aimed at achieving desired outcomes in goal-oriented educational tasks [17]. This concept is crucial, as engagement is a key driver of learning success [18]. For instance, previous studies attested that learning engagement was a significant antecedent of academic performance, like the grade point average [19,20]. This means that learning engagement meets the demands of the first criterion of knowledge growth. Moreover, learner engagement can be regarded as a catalyst for sustainability, given that it signifies a more persistent affective-cognitive state rather than a momentary state in the learning process [21]. However, engagement in virtual learning environments is especially challenging because learning outside the classroom often exacerbates learning loss [5]. Therefore, learning engagement in such contexts becomes even more critical in such e-learning contexts, and greater engagement is required to compensate for learning loss caused by dropout.

Furthermore, learning engagement can be conceptualized in a theoretical framework. One model embodies a continuous emotional and cognitive condition, stating that the learning engagement framework comprises vigor, dedication, and absorption from Utrecht Work Engagement Scale-Student Version (UWES-S) [21]. This is a classical framework that treats learning engagement as a single construct, aiming to reflect the overall engagement experience of learners in a holistic manner. Additionally, despite its multidimensional nature, most research has focused on isolated variables or singular dimensions of engagement. Such simplified approaches often overlook the intricate details and causal relationships inherent in the engagement process. By adopting a second-order method, this study offers insights into what factors affect the comprehensive engagement that integrates these dimensions.

Both internal and external factors affect learning engagement. Externally, elements such as interactive learning designs and the convenience of e-platforms enhance learning engagement [22]. Emotional support has also been recognized as a critical determinant in higher education E-Systems [23]. Internally, social presence improves engagement, as it is associated with task value and expectancy [24]. Spatial presence is recognized to bring a better user experience, which engages students in virtual environments [25,26]. Perceived positive emotions like enjoyment can also play a critical role in engaging students. However, limited research has explored perceived enjoyment as a mediator and how psychological factors stemming from spatial and social activities jointly contribute to learning engagement, even though these aspects frequently co-occur in educational settings. To address the limitation, the research explores perceived enjoyment and the simultaneous effects of social and spatial presence influencing learning engagement.

2.2. Social Presence

Social presence theory conceptually indicates social presence as the perceived salience of others during interactions, as well as the prominence of interpersonal relationships that arise within such contexts [11,27]. This construct highlights an individual’s capacity to interact with others in communication media. Gunawardena and Zittle [28] expanded this conceptualization in computer-mediated communication (CMC) environments, proposing that social presence reflects the extent to which users perceive the presence of others, whether “real” or “pseudo”, through digital networks. Notably, this perception underscores two aspects: acknowledging the existence of others as well as involving the projection and recognition of identities to maintain meaningful social interactions. Thus, social presence can refer to feeling others as “real” and “present” within mediated communication environments [29]. In sustainable e-learning platforms, social presence signifies the degree students sense a social connection with peers and instructors, thereby strengthening relations and interactions in the learning process and promoting efficient use of resources.

As a pivotal construct in e-learning, social presence contributes to shaping and facilitating meaningful academic experiences, including engagement and enjoyment [30]. For instance, a strong social presence plays a key role in fostering active engagement, collaboration, and emotional support among learners [31]. It is consistently associated with enhanced levels of greater emotional involvement, interaction, and the promotion of collaborative learning [32]. In other words, when learners perceive that others are genuinely “there”, they could more possibly delve into learning tasks and generate enriched cognitive and affective effects [33]. Some studies have posited the pivotal role of social presence on perceived enjoyment [34,35]. In addition, given that engagement is closely and positively correlated with performance, social presence helps to learn in essence, which has been validated by various studies [36]. Additionally, social presence and socialization activities are closely linked to intensified feelings of enjoyment, relatedness, and intrinsic motivation [37]. Peer social presence, in particular, has been shown to significantly predict class enjoyment, perceived learning, and perceived competence [35]. Building on this foundation, hypotheses are put forward as follows:

H1: 

Social presence affects learning engagement positively in sustainable e-learning.

H2: 

Social presence affects perceived enjoyment positively in sustainable e-learning.

2.3. Spatial Presence

Spatial presence is a concept originating from communication studies and cognitive psychology. As a unique dimension of presence, spatial presence is particularly relevant to how individuals cognitively and effectively engage, making it a pivotal factor in immersive media contexts. In sustainable e-learning, fostering spatial presence can reduce distractions and enhance the task completion efficiency of learners, thereby decreasing device energy consumption and optimizing resource use to directly support sustainable practices by minimizing digital waste.

Spatial presence is from spatial presence theory and is defined as the feeling of “being there” within the mediated environments [12]. Specifically, it is conceptualized as a qualia of having a feeling of being on the actual scene, emphasizing its subjective and immersive nature [38]. In other words, spatial presence refers to the cognitive experience of feeling as if one is in a place, or more explicitly, the perceptual illusion of being in a virtual place [39]. As a cognitive construct, spatial presence is understood as a twofold phenomenon, involving perceived self-location and possible actions, both of which are generated through using mediated environments [40]. Mediated environments can be understood as digitally constructed spaces where participants engage and represent themselves to have a presence through various means (e.g., text, audio, images, and avatars), being either remote but real or entirely computer-generated virtual spaces [41]. Falling into the latter virtual category, e-learning platforms can evoke spatial presence more effectively through their interactivity and persistence while supporting instructors, learners, and administrators in content delivery, personalized learning, and interactive communication [13,42]. Accordingly, this study contends that e-learning as computer-generated mediated environments (i.e., virtual environments) may evoke a strong spatial presence of users. And some studies indicate that high-level spatial presence can promote better performance, which means that it exerts an essential role in learning [43].

However, research on the impact of spatial presence on learner engagement in general e-learning remains insufficient. While little research has attempted to model spatial presence as a predictor of learning, findings are mixed and often context-dependent. For example, some studies suggest that spatial presence enhances user engagement in virtual education and training environments [44]. Similarly, Berki [45] observed that spatial presence fosters involvement within educational desktop virtual reality platforms. However, opposite evidence indicates that spatial presence may also lead to lower engagement under 3D virtual reality conditions [46]. These discrepancies underscore the need for further empirical investigation into the nuanced effects of spatial presence on e-learner outcomes.

Beyond engagement, spatial presence has also been found to predict perceived enjoyment in virtual settings [47]. For instance, research has demonstrated that spatial presence enhances enjoyment in mobile gaming experiences [48]. Similarly, research on news media shows that spatial presence induced by 360-degree photography and video significantly enhances enjoyment [49]. Likewise, research suggests that sensing physical presence through multisensory VR setups has been shown to contribute positively to users’ overall enjoyment of virtual experience [50]. These discoveries put forward that spatial presence fosters the illusion of immersion that amplifies enjoyable responses.

Overall, spatial presence has been studied in some mediated environments, whereas its specific effects within e-learning environments remain insufficiently explored. There is a notable limitation in understanding the specific mechanisms by which spatial presence influences learning engagement and perceived enjoyment in such contexts. Existing research provides limited and sometimes conflicting insights, leaving the broader implications of spatial presence in e-learning ambiguous. Addressing these potential relationships, the hypotheses are given:

H3: 

Spatial presence affects perceived enjoyment positively in sustainable e-learning.

H4: 

Spatial presence affects learning engagement positively in sustainable e-learning.

2.4. Perceived Enjoyment

Perceived enjoyment, sometimes referred to simply as enjoyment, can be conceptually described as the degree to which the perceived use of a computer system is inherently enjoyable in itself [51]. Within the broader context of emotions, enjoyment is a subset of achievement emotions, as defined in the three-dimensional taxonomy [15]. Specifically, enjoyment is characterized by positive valence, activating emotion, and activity focus, centering on the learning process and immediate task-related experiences, varying from outcome emotions. This means that enjoyment is tied to deep engagement and active participation, focusing on experiencing ongoing achievement activities. Lacking fun in learning can lead to failure in learning engagement [52]. Therefore, perceived enjoyment, as an activity-focused achievement emotion, significantly influences key aspects of learning. These aspects include effort investment, engagement, the utilization of flexible learning strategies, satisfaction, and final performance [4,53,54]. Considering that perceived enjoyment offers the foundation for academic performance, it can be regarded as an essential role in learning growth [55].

Additionally, enjoyment as the task-related achievement emotion is induced by contextual factors such as social presence and spatial presence [37]. Perceived enjoyment is the fundamental mediating mechanism by which social presence affects usage intentions and behaviors within social media [56]. Specifically, social presence emphasizes feelings regarding meaningful interactions and connections with peers or instructors, improving strong bonds and collaboration. Thus, optimistic feelings like satisfaction can be induced, thereby enhancing active learning engagement [57]. Moreover, spatial presence reflects learners’ immersion in the virtual learning contexts, enhancing attention and decreasing distractions [58]. Building on this, spatial presence in a mediated world fosters deeper enjoyment, absorption, and task immersion, reducing disengagement [40]. Therefore, perceived enjoyment may function as a mediator linking social presence and spatial presence to learning engagement. This is consistent with the established “motivation–emotions–engagement” framework [59], which underscores the relationship between emotional experience and psychological state. In addition, the four core variables have been widely examined in previous empirical studies as distinct constructs [37].

Nevertheless, the current studies are limited in exploring the effectiveness of the reciprocal effect of social presence and spatial presence and the underlying mechanisms. The findings could offer valuable insights into how intrinsic motivation and contextual factors drive engagement. Accordingly, the mediation model is depicted in Figure 1, and the mediating hypotheses are formulated:

H5: 

Perceived enjoyment positively affects learning engagement in sustainable e-learning.

H6: 

Social presence plays a significant role in learning engagement through perceived enjoyment in sustainable e-learning.

H7: 

Spatial presence plays a significant role in learning engagement through perceived enjoyment in sustainable e-learning.

3. Methods

3.1. Participants

To ensure robust data analysis, the inverse square root method was used to conduct the a priori sample size calculation [60]. This method determined the lowest significant path coefficients (β_min = 0.197) to detect a minimum sample size of 160, thereby reducing the risk of Type II errors (false negatives) due to insufficient sample size. The inverse square root approach can be especially robust for non-normal data distributions and is suitable for various research designs. To account for potential data loss from incomplete surveys or entry errors, this study aimed to collect a slightly larger sample to maintain analytical integrity and flexibility.

E-learning helps people worldwide obtain extensive knowledge, providing sustainable development, especially for the developing world [61]. A total of 442 e-survey responses were gathered from Chinese university learners. All respondents voluntarily and anonymously participated in the survey, with assurances that their privacy would be protected. All participants gave their approval to complete the questionnaire. Due to the shift to online learning prompted by the COVID-19 pandemic, all Chinese university students were required to engage in e-learning to ensure continuity of learning. Participants from Chinese universities experienced credit-bearing e-courses through online learning platforms such as CHAOXING and XuetangX, which are similar to Coursera at the international level.

To ensure relevance, participants were confirmed in a screening question as to whether they had prior experience with e-learning, defined as having completed at least one e-course. Among the participants, 419 reported having prior e-learning experience, indicating that 94.8% had engaged in online learning. Only 23 reported having no prior experience with e-learning, accounting for 5.2%. The proportion aligns with the actual situation in China, where the online course offering rate in universities has reached 91% [62]. Furthermore, data reports show that 92.41% of university students in China have used online platforms for learning [63]. This realistic scenario offers evidence for the ecological validity of findings in e-learning environments.

Finally, a total of 362 valid samples were obtained and showed an overall response rate of 81.90%, following the exclusion of 80 invalid responses due to no e-learning experience, excessively short completion times, and overly uniform response patterns. Hence, the result met the a priori minimum sample size requirement. In the final collection, 72.7% of females and 27.3% of males were comprised. Participants’ educational backgrounds included junior college students (8.6%), undergraduates (65.7%), master’s students (24.6%), and doctoral candidates (1.1%). Most respondents were in the 18–26 age range (88.9%). The sample closely aligns with the typical demographic profile of university learners in China, as reported by the National Bureau of Statistics, with undergraduates comprising the majority and doctoral candidates representing the smallest proportion [64].

Furthermore, this study controls for demographic factors that can predict perceptions of use practices in e-learning [65]: gender, education level, and age, to guarantee robustness and minimize the risk of omitted variable bias. This approach isolates the unique link between the predictors and learning engagement.

3.2. Instruments

The questionnaire includes items on demographics, social presence, spatial presence, perceived enjoyment, and learning engagement. All items were translated into Chinese, and a pretest with 15 learners made sure of clearness and comprehensibility. A survey on the 5-point Likert scale is used often in learning [66], and it was used to evaluate the responses from “absolutely disagree” to “absolutely agree” in this study. Assessing the internal consistency and strong reliability of scales was shown for e-learning engagement (α = 0.921), perceived enjoyment (α = 0.859), social presence (α = 0.820), and spatial presence (α = 0.817).

The social presence scale, consisting of 5 items, is based on the work of Hostetter and Busch [67] and measures students’ extent to feel and interact with other characters online. The spatial presence scale, developed by Rodríguez-Ardura and Meseguer-Artola [68], also contains 5 items. The perceived enjoyment scale, derived from Lee [69], includes 3 items.

Lastly, the UWES-S was adopted for learning engagement measurement, which includes 9 items from Schaufeli, Bakker, and Salanova [21]. Specifically, learning engagement is multidimensional and modeled as the reflective second-order factor, with three 3-item first-order factors: vigor, dedication, and absorption. All the items of the four variables are present in

Table 1. Variables and items.

Variables	Items
Social presence	1. I feel comfortable communicating via the platform.
	2. I feel comfortable participating in course discussions of e-learning.
	3. I feel comfortable interacting with other participants in the course of e-learning.
	4. I feel that my point of view was acknowledged by other participants in the course of e-learning.
	5. I am able to form distinct individual impressions of some course participants in e-learning.
Spatial presence	1. I forget about my immediate surroundings when I use e-learning.
	2. Using e-learning makes me forget where I am.
	3. After using e-learning, I feel like I come back to the ‘real world’ after a journey.
	4. When I use e-learning, I feel I am in a world created by technology.
	5. When I use e-learning, the world generated by technology is more real for me than the ‘real world’.
Perceived enjoyment	1. Using e-learning is pleasurable.
	2. Using e-learning is exciting.
	3. Using e-learning is interesting.
Learning engagement	Vigor 1: During my study in e-learning, I feel bursting with energy.
	Vigor 2: When I get up in the morning, I feel like going to study in e-learning.
	Vigor 3: At my study in e-learning, I feel strong and vigorous.
	Dedication 1: I am enthusiastic about my study in e-learning.
	Dedication 2: My study in e-learning inspires me.
	Dedication 3: I pride myself on my study in e-learning.
	Absorption 1: I feel happy when I am studying in e-learning intensely.
	Absorption 2: I am immersed in my e-learning study.
	Absorption 3: I get carried away when I am studying in e-learning.

.

3.3. Data Analysis

Descriptive statistics and correlations were examined employing SPSS 26.0. In addition, to assess the relationships of the model, SmartPLS 3.2.9 was used for partial least squares structural equation modeling (PLS-SEM). Several reasons justify the reasons why this method [70]. First, it is particularly suitable for testing complex models involving higher-order constructs. The model’s complexity, which includes multidimensional constructs, guides that PLS-SEM can handle such relationships effectively [71]. Additionally, PLS-SEM is appropriate for non-normally distributed data. The Kolmogorov–Smirnov normality test, conducted using SPSS, indicated non-normal data, making PLS-SEM an ideal choice. Moreover, this approach has been empirically validated in cross-disciplinary studies [72,73]. Hence, given this study’s goal to test specific hypotheses regarding how social presence, spatial presence, and perceived enjoyment predict second-order learning engagement, PLS-SEM was deemed appropriate for hypothesis testing and the assessment of these theoretical relationships.

The mediating role was estimated by employing bias-corrected bootstrapping with 95% confidence intervals. Non-zero confidence intervals indicate significant indirect effects. Additionally, several indices were used to assess model quality, including model fit, predictive relevance (Q²), and effect size (f²) [70]. To evaluate model fit, the following indicators were considered: SRMR (standardized root mean square residual), NFI (normed fit index), and the coefficient of determination (R²). The values of NFI (0.092 > 0.09) and SRMR (0.077 < 0.08) suggest a good model fit. Estimating the predictive relevance of the model, Q² was calculated through the blindfolding method.

4. Results

4.1. Descriptive Analysis

The means, standard deviations, and correlation matrix are presented (see

Table 2. Description of statistics and correlation between variables.

Variables	(1)	(2)	(3)	(4)	(4a)	(4b)	(4c)
(1) Social Presence	1
(2) Spatial Presence	0.641 ***	1
(3) Perceived Enjoyment	0.659 ***	0.638 ***	1
(4) Learning Engagement	0.741 ***	0.721 ***	0.727 ***	1
(4a) Vigor	0.687 ***	0.606 ***	0.633 ***	-	1
(4b) Dedication	0.634 ***	0.601 ***	0.644 ***	-	0.667 ***	1
(4c) Absorption	0.671 ***	0.720 ***	0.672 ***	-	0.696 ***	0.729 ***	1
Mean	3.640	3.492	3.749	3.576	3.344	3.698	3.609
Standard Deviation	0.710	0.771	0.837	0.770	0.992	0.806	0.827

Note: *** Significant at the 0.001 level (2-tailed).

). Observably, all the correlations of learning engagement (and its first-order construct: vigor, absorption, and dedication), perceived enjoyment, and social presence, as well as spatial presence, are positive and significant (p < 0.001), indicating that stronger social and spatial presence is connected with greater enjoyment and engagement. Moreover, after using SPSS 26.0 to perform Harman’s single-factor analysis, the first extracted factor accounted for 39.09% of the variance, which did not exceed the critical threshold of 50%. Therefore, there is no severe common method bias [74]. In addition, the mean scores for social presence, spatial presence, perceived enjoyment, and learning engagement exceed the theoretical median of 3. The result suggests that learners’ psycho-emotional experiences and engagement are generally at a medium-high level on the e-learning platforms. These results address limitations in prior studies, particularly concerning that e-platforms can provide relatively high levels of spatial presence. The high mean values imply that learners’ psychological needs and user experience expectations are mostly met within e-learning platforms. Although standard deviations exceed 0.7, indicating some individual variability, the overall consistency suggests stable data with most learners sharing similar experiences.

4.2. Measurement Model

In terms of reliability, first-order and second-order measurement models were revised and assessed (see

Table 3. Reliability and validity of the measurement model.

Variables	Factor Loading	Cronbach’α	CR	AVE
The first-order reflective construct
Perceived Enjoyment1	0.836	0.755	0.859	0.671
Perceived Enjoyment2	0.847
Perceived Enjoyment3	0.771
Social Presence1	0.759	0.723	0.828	0.547
Social Presence2	0.790
Social Presence3	0.722
Social Presence4	0.683
Spatial Presence1	0.767	0.699	0.815	0.525
Spatial Presence2	0.696
Spatial Presence3	0.710
Spatial Presence4	0.723
Learning engagement: The second-order reflective construct
Vigor	0.907	0.874	0.922	0.798
Dedication	0.893
Absorption	0.880

). The results showed good reliability for each construct, with values of Cronbach’s alpha exceeding 0.7. All factor loadings start at 0.683 and go up to 0.869 (>0.5), and values of composite reliability (CR) were over 0.8, confirming overall reliability in the measurement model [70].

Regarding validity, the model shows satisfactory convergent validity by evaluating AVE (average variance extracted, >0.5) in Table 3 and discriminant validity in

Table 4. Discriminant validity.

Variables	(1)	(2)	(3)	(4)
(1) Learning Engagement	0.899	0.879	0.874	0.825
(2) Perceived Enjoyment	0.731	0.831	0.828	0.848
(3) Social Presence	0.713	0.635	0.756	0.765
(4) Spatial Presence	0.657	0.633	0.566	0.731

Note: In bold, the main diagonal is the square root of the AVE; below the main diagonal are displayed the Fornell-Larcker criterion values, while above the main diagonal is displayed the HTMT values.

. In terms of discriminant validity, the criterion of [75] was assessed. The results showed that the square root of AVE (the bold diagonal number) is all above the correlation values between variables (below the diagonal), confirming that each variable can be distinguished effectively. Furthermore, the heterotrait–monotrait ratio (HTMT, above the diagonal), the strictest testing method of discriminant validity, was also applied. HTMT values ranged from 0.794 to 0.892, all below the 0.9 threshold. In conclusion, the measurement model demonstrated acceptable validity.

4.3. Structural Model Assessment and Hypothesis Testing

While controlling for gender, age, and education level, links between four main hidden constructs were evaluated. Using SmartPLS 3.2.9 software, the predictive capacity in the structural model was assessed by employing PLS-SEM. The hypothesized model was considered by standard assessment criteria [70], including the explanation ability (R²), effect size (f²), predictive relevance (Q²), and significance level. Before assessing the structural model, to avoid biased regression results, collinearity was checked. The values of the variance inflation factor (VIF) are between 1.244 and 2.566 (<3), indicating that collinearity will not adversely affect the estimation of path coefficients [70].

Subsequently, R² values were assessed and fell within the expected range of 0 to 1. Specifically, R² values were 0.486 for perceived enjoyment and 0.662 for learning engagement, indicating moderate explanatory power. Furthermore, the adjusted R² were 0.489 and 0.667, respectively, both of which are close to the unadjusted R² values, showing that most of the independent variables contribute to the model meaningfully. Furthermore, the thresholds of 0.02, 0.15, and 0.35 in f² values indicate small, moderate, and large effect sizes on dependent variables in the structural relationships. Perceived enjoyment, social presence, and spatial presence can explain moderate, moderate, and small effect sizes on learning engagement. On perceived enjoyment, both social presence and spatial presence can explain moderate effect size.

Additionally, the predictive relevance of the PLS path relationship was evaluated using Q² values. These values not only indicate the model’s predictive performance but also its ability to explain out-of-sample data. For endogenous constructs, Q² values must exceed zero to confirm the model’s satisfactory predictive capacity. Specifically, when Q² is 0, 0.25, and 0.5, predictive relevance is considered small, moderate, and large, respectively. After performing the blindfolding procedure, Q² values of 0.321 for perceived enjoyment and 0.525 for learning engagement indicated moderate to large predictive relevance.

As in Figure 2, hypothesis testing outcomes showed that all path coefficients were significant (p < 0.001). Learning engagement is affected by social presence in a significant and positive way (β = 0.307, p < 0.001) and perceived enjoyment (β = 0.386, p < 0.001); hence, hypotheses H1 and H2 are supported. In terms of spatial presence, a significant and positive impact can be observed on both perceived enjoyment (β = 0.401, p < 0.001) and learning engagement (β = 0.227, p < 0.001); therefore, hypotheses H3 and H4 are supported. Finally, learning engagement can also be influenced by perceived enjoyment (β = 0.391, p < 0.001), confirming hypothesis H5. In the end, within e-learning platforms, the path coefficients suggest that spatial presence interacts with social presence and plays a similar role in learning engagement, effectively complementing existing research on this topic. As a result, all direct effect hypotheses were supported.

A post hoc analysis was conducted using the minimized coefficients of direct effects. Based on the inverse square root equation N > (z_0.95 + z_0.8/|β_min|)², where z_0.95 = 1.645 represents the 95% confidence level, and z_0.8 = 0.842 corresponds to the 80% power level. Using a β_min value of 0.227 from the PLS-SEM outcome, the analysis showed that a sample size of 121 can provide robust and sufficient statistical power. With a final sample size of 362, the analysis met the required statistical efficacy, allowing further exploration of patterns and relationships.

To explore how perceived enjoyment acts as a mediator, bootstrapping was performed with 5000 resamples at the 95% confidence interval.

Table 5. Relationship effect analysis between variables.

Hypotheses	Total Effects	Direct Effects	Indirect Effects
Perceived Enjoyment → Learning Engagement	0.391 ***	0.391 ***	-
Social Presence → Learning Engagement	0.458 ***	0.307 ***	0.151 ***
Social Presence → Perceived Enjoyment	0.386 ***	0.386 ***	-
Spatial Presence → Learning Engagement	0.384 ***	0.227 ***	0.157 ***
Spatial Presence → Perceived Enjoyment	0.401 ***	0.401 ***	-

Note. *** Significant at the 0.001 level.

shows that social presence affects learning engagement by perceived enjoyment (β = 0.151, 95% CI [0.095, 0.226]). Similarly, perceived enjoyment mediates the link between spatial presence and learning engagement indirectly and substantially (β = 0.157, 95% CI [0.105, 0.219]). These mediation effects are confirmed significant because zero is not contained in both confidence intervals. The mediation effect of perceived enjoyment in the social presence-learning engagement relationship was 32.97%, suggesting partial mediation. Similarly, spatial presence played a role in learning engagement with a 40.89% mediating effect of perceived enjoyment. Thus, all hypotheses related to indirect effects were supported, and the proposed model was validated.

5. Discussion

To enhance the sustainability of e-learning, the research sought to explore the relationships among social presence, spatial presence, perceived enjoyment, and second-order learning engagement in e-learning platforms, with control for age, education level, and gender. By examining these variables in Chinese university students, the research offers an important understanding of how sense of presence and emotional experiences influence student engagement in e-learning environments.

An important result of the research is that a stronger spatial presence predicts more e-learning engagement. This result extends and aligns with prior work on spatial presence, which has predominantly focused on immersive environments such as virtual reality [44]. This research further demonstrates that underlying spatial presence is pivotal to engaging even in more general e-learning, such as desktop-based courses. A potential explanation for this correlation is that a stronger sense of spatial presence may reinforce learners’ intrinsic motivation. When learners experience a pronounced feeling of “being there” within the e-learning environment, their curiosity and interest in the content may increase, leading to more active learning behaviors. Furthermore, spatial presence could engage deeper vigor, absorption, and dedication by encouraging the use of multiple sensory inputs and greater mental effort to process and assimilate information. Therefore, spatial presence may emerge as a significant facilitator of efficient e-learning, potentially reducing energy consumption and supporting sustainable practices. This further emphasized the significance of complementing spatial presence in sustainable e-learning environments.

Aligning with prior studies, this research found that social presence predicts e-learners’ engagement [76]. When students feel a strong social presence, it becomes more possible to be involved in communication, sharing resources, and engaging in collaborative learning [77]. This result can be further explained by several underlying mechanisms. Specifically, an elevated sense of social presence fosters learners’ feelings of belonging and group affiliation, which, in turn, stimulates their willingness to actively engage in learning tasks. Moreover, it alleviates feelings of isolation and reduces social anxiety, thereby diminishing psychological barriers to meaningful participation. Additionally, the perception of social presence strengthens both intrinsic and extrinsic motivation by promoting meaningful interactions and encouraging knowledge-sharing and collaborative behaviors. It also contributes to lowering cognitive load and fostering trust among learners—factors that collectively sustain and enhance their engagement. Therefore, social presence fosters a sense of belonging in e-learning, mitigates learner burnout, and supports sustainability through optimized use of educational resources. The finding underscores that building socially rich settings in e-education is crucial for enhancing student engagement. Furthermore, this study also finds that the combined effect of social presence and spatial presence influences engagement enhancement. Specifically, it confirmed that social presence and spatial presence generally co-occur, while current studies always ignore their reciprocal effects. Social presence creates connections for learners, while spatial presence offers immersion within the learning space. Together, these two elements interact to enrich the learning experience, making it more engaging and immersive. This provides insights into how these factors interplay with one another to enhance learner engagement.

Perceived enjoyment is found to mediate the link that social presence and spatial presence enhance e-learning engagement. This result is particularly noteworthy because the mediating role of particular perceived enjoyment has received limited attention. Specifically, the findings suggest that when learners experience enjoyment, they will engage in e-learning tasks more actively and sustainably. This finding is consistent with the literature that emphasizes the role of enjoyment in study engagement [78]. In other words, the feeling of enjoyment motivates learners to devote more time, effort, and perseverance to e-learning tasks. Additionally, this study demonstrates that both social presence and spatial presence contribute directly to learners’ perceived enjoyment, which is consistent with earlier findings [35,37]. By clarifying the bridging role of enjoyment between presence and engagement, the present study extends previous literature and offers further insight into the underlying mechanism. Several possible explanations may account for the results. From the perspective of intrinsic motivation, social and spatial presence may stimulate enjoyment by making learners feel socially connected and spatially immersed. In turn, perceived enjoyment can further promote intrinsic motivation and strengthen learners’ willingness to engage in learning tasks. Additionally, the presence experience may alleviate learners’ cognitive burden by creating a supportive and engaging learning environment, enabling them to concentrate their cognitive resources on learning rather than dealing with feelings of isolation or disengagement. From the standpoint of need satisfaction, social and spatial presence may fulfill learners’ basic psychological needs for relatedness and autonomy, thereby enhancing their enjoyment of the learning process. This increased sense of enjoyment encourages sustained engagement and persistent effort in e-learning activities. Therefore, when learners experience higher levels of enjoyment, particularly those who are sensitive to social and spatial cues, they are more inclined to invest additional time, energy, and cognitive effort in e-learning, ultimately contributing to improved engagement levels.

Overall, the research contributes to e-learning research by exploring a novel framework that integrates social presence and spatial presence to improve perceived enjoyment and further learning engagement to maintain sustainable e-learning in virtual learning environments. This framework offers a structured approach for platform providers and instructors, guiding them to leverage these findings and build more immersive, enjoyable, and engaging e-learning environments, bringing conduciveness to deep learning and sustained participation.

6. Conclusions

Given the significance of sustainable e-learning for education development, this study examined how social presence, spatial presence, and perceived enjoyment as the mediator affect learning engagement within e-learning environments. The empirical results support all seven proposed hypotheses, indicating that social presence, combined with spatial presence, positively impacts learning engagement, with perceived enjoyment mediating this effect.

6.1. Contributions

This study offers a novel perspective distinct from prior research. Unlike existing studies that primarily examine the isolated effects of social or spatial presence, the model empirically validates their interactive influence on engagement, an area that has remained largely unexplored, particularly in Asian contexts. Moreover, while previous research has investigated spatial presence, it has not been widely examined within general e-learning environments, as in this study. The major contributions of this study lie in both theoretical and practical domains. From a theoretical standpoint, this research expands the current literature on e-learning engagement through empirical evidence on perceived enjoyment as the mediator between social presence and spatial presence. By using a second-order construct, it presents a more comprehensive view of engagement and introduces an innovative mechanism to enhance learner involvement within e-learning environments. In terms of practice, the findings offer valuable insights for instructors and platform providers, enabling them to foster learner engagement through more scientifically grounded strategies. Ultimately, this study contributes to the long-term sustainability of educational quality, aligning with the objectives of SDG4.

6.2. Implications

The findings highlight the importance of fostering interaction and immersive experiences to enhance engagement and enjoyment, providing critical implications to improve sustainable e-learning.

Theoretically, the research clarifies the links in e-learning environments among social presence, spatial presence, and learning engagement, emphasizing the indirect effects of perceived enjoyment. It advances how social presence and spatial presence together drive engagement, verifying the impact of spatial presence that is often overlooked in e-learning settings. This study enriches the presence theory in e-learning and confirms the mechanisms linking the sense of presence and engagement through positive-activating-task emotions (i.e., perceived enjoyment). It also offers a novel perspective on second-order learning engagement, offering more comprehensive insights.

Practically, this study provides specific implications for instructors and platform providers involved in e-learning platforms, contributing to the understanding of sustainable e-learning practices. The positive effect of social presence encourages instructors to actively foster interaction with students. Feasible suggestions for instructors to cultivate social presence include establishing teacher image (such as teacher introduction and updating content regularly) and providing necessary support (such as timely feedback and clarification) for learners. Additionally, instructors could implement co-regulated learning strategies, such as organizing virtual learning discussion groups, role play, and peer review, to elicit learners’ sensation of connection. For platform providers, it is essential to design and refine interaction features that facilitate real-time communication. Encouraging users to personalize their profiles by adding social elements can strengthen self-representation and the feeling of being acknowledged. Additionally, implementing AI-driven recommendation systems for study groups and communities can create a learning environment where users feel a sense of connection with specific individuals, not merely platforms, thereby enhancing engagement.

Concerning spatial presence, integrating location and action cues in operable virtual courses should be considered to help learners immerse and engage themselves in learning spaces. In terms of self-location, platform providers can create well-scaffolded, engaging virtual environments that simulate the experience of a physical classroom. These environments should incorporate multisensory stimuli such as text, voice, and images, aiming to replicate the feeling of being physically present in a real-world setting. Instructors should fully leverage technology to create a dedicated, distraction-free e-learning space for students to engage with course material. This can help students psychologically detach from their physical surroundings, enhancing the sense of immersion. Considering this, improving instructors’ digital literacy is essential to better facilitate the spatial presence of the learning process. Instructors can also integrate scientific knowledge with real-life contexts to create more immersive learning experiences. In terms of possible actions, instructors can employ strategies like implementing on-the-hour time points to activate execution for enhancing spatial presence. By modifying learners’ behavioral states, these interventions can arouse learners to take action in e-learning environments. Platform providers can contribute to spatial presence by optimizing interactive design that supports learners’ diverse learning actions. Providing natural and visually interactive approaches, such as free navigation and first-person perspectives, can encourage users to engage more actively in learning activities. Leveraging both social and spatial presence can boost user engagement by incorporating psychological interventions. These implementations offer cost-effective approaches that demand minimal effort but substantial benefits to learning engagement.

Perceived enjoyment is crucial in enhancing learner engagement in e-learning settings. Given that both social presence and spatial presence enhance perceived enjoyment, instructors should provide strong support and encourage learners to engage with their peers and course materials, thereby stimulating their interest in academic tasks. Platform providers should focus on creating location-aware and active user interfaces that ensure seamless, accessible, and immersive learning experiences, ultimately boosting users’ sense of pleasure and excitement. To further increase learning engagement by heightened perceived enjoyment, instructors should employ engaging pedagogical strategies such as gamification and storytelling-based instruction. Providing growth-oriented feedback and promoting positive emotional guidance can also foster perceived enjoyment and trigger intrinsic motivation. Regarding platform providers, they can design enjoyable and innovative tools, such as micro-interactions (e.g., animated buttons, page-turning effects) and reward mechanisms, while offering personalized learning pathways tailored to individual needs and interests. Collectively, these efforts can significantly enhance the appeal of e-learning, thereby strengthening user enjoyment and fostering greater user engagement.

6.3. Limitations and Future Research Directions

While this study offers valuable insights into sustainable e-learning theories and practices, future research is supposed to deal with some present limitations. Firstly, this research relies on self-reported measures, introducing the possibility of bias in assessing variables. So, incorporating objective instruments (e.g., behavioral tracking) to offer more accurate pictures of engagement is needed. Secondly, the research put cautious explanations regarding causal inferences because of the cross-sectional dataset. To implement this and to better understand how the relationships develop, longitudinal studies are a kind of solution. Third, other variables can be considered to advance the framework of learning engagement. For example, future research can investigate students in different levels of e-learning experience and different disciplines or subjects as key variables. Finally, the object of the investigation is Chinese university learners; thereby, the university’s findings should be verified in diverse educational settings and academic disciplines.