1 Introduction

Immersive Virtual Reality (IVR) is considered a potentially cost-effective means for providing students with compelling learning experiences (Hite, 2022; Makransky et al., 2020). IVR employs head-mounted displays to immerse students in an interactive environment, incorporating intuitive features, such as head tracking and body movement (Meyer et al., 2019). Furthermore, IVR can offer a high level of environment representational fidelity, including consistent object behavior, seamless perspective changes, and realistic displays (Dalgarno & Lee, 2010). Recent research has demonstrated that IVR significantly enhances the facilitation of various scientific activities that are otherwise unrealistic or impossible-to-perform for K-12 students (Di Natale et al., 2020; Wu et al., 2020). For instance, IVR can simulate a physical scientific laboratory through a computer-generated, immersive environment that allows for intimate and intuitive engagement. This allows students to interact with the displayed artifacts in virtual laboratory activities (Chan et al., 2021; Naz et al., 2024), avoiding constraints imposed by expensive equipment and materials, time limitations, or safety concerns in the real world (Potkonjak et al., 2016). Evidence indicates that the use of IVR in the laboratory or in science learning influences student outcomes in a way that is either superior to or comparable to traditional laboratory education (Reeves & Crippen, 2021). This may be attributed to the ability of IVR to immerse students in laboratory learning, creating a strong “sense of presence” that leads to increased engagement and involvement, ultimately enhancing learning outcomes (Klingenberg et al., 2020; Makransky & Lilleholt, 2018; Tsirulnikov et al., 2023). Presence can be understood as the psychological sensation of “being there” (Ijsselsteijn & Riva, 2003). When students are fully immersed in a virtual world, they may experience an authentic sense of being present, which can be described from three dimensions—physical, social and self-presence—in accordance with human experience in real life (Lee, 2004; Makransky et al., 2017). Some researchers highlight that IVR applications enhance the sense of physical presence, thereby facilitating scientific learning by increasing situational interest and enjoyment, which subsequently improves their retention of scientific knowledge (Makransky & Mayer, 2022; Petersen et al., 2022).

However, it is worth noting that most students in prior studies were new to IVR technology, and thus potentially influenced by its novelty (Makransky & Lilleholt, 2018; Reeves & Crippen, 2021). The unique advantage of IVR for offering presence, with its ability to stimulate situational interest, enjoyment, and satisfaction, may diminish over time as students become more accustomed to it (Makransky & Petersen, 2021). Therefore, there is a critical need to investigate the sustained affordance of presence in long-term IVR applications. As Matovu et al. (2022) highlighted in their systematic review of IVR for science learning, most learners in reviewed studies were first-time IVR users. Furthermore, Reeves et al. (2021) contended that most studies on laboratory learning have primarily focused on comparing students’ knowledge acquisition in VR Labs with that in traditional physical laboratories, often overlooking their perceptions of learning in an IVR environment. This underscores a significant research gap in understanding students' perceptions regarding the sustained use of IVR for laboratory learning and the evolving role of presence in their experiences, especially over an extended period. Hence, our primary objective is to investigate students’ perceptions of learning in IVR-supported laboratory environments and the evolving impact of presence as IVR becomes a regular fixture in secondary science instruction.

To achieve this, we adopt the Expectation Confirmation Model (ECM), chosen for its suitability in elucidating student perceptions post-IVR experience. Our focus centres on the concept of expectation confirmation, examining how students align their initial expectations about the impact IVRs on laboratory learning with their actual experiences. Drawing from established research (Chang et al., 2018; Makransky & Lilleholt, 2018; Pedram et al., 2020), we also consider factors such as perceived usefulness, satisfaction, and the intention to persist in using IVR for laboratory learning. Furthermore, we delve into the pivotal role of presence, an antecedent variable known to influence students' perceptions of IVR-supported learning within the context of science education (Andersen et al., 2023; Makransky & Lilleholt, 2018; Petersen et al., 2022). In summary, our study aims to address the following key questions:

  1. (1)

    How do students perceive learning in VR Labs as IVR becomes integrated into routine science instruction?

  2. (2)

    What impact does the sense of presence have on student perceptions of learning in VR Labs as IVR becomes standard in daily science instruction?

2 Literature review and hypotheses

2.1 The affordance of IVR on laboratory learning and the role of presence

Virtual experiments delivered through IVR enhance traditional physical experiments by enabling students to explore unobservable phenomena and correlate them with observable phenomena. This facilitates the conducting of multiple experiments within a condensed timeframe, while also highlighting pertinent information and offering immediate feedback and adaptive guidance (de Jong, et al., 2013). A recent systematic review (Reeves & Crippen, 2021) demonstrated that IVR yields positive outcomes in laboratory learning, such as promoting affective outcomes (e.g., intrinsic motivation, interest, and self-efficacy), enhancing students’ understanding of complex scientific knowledge and their ability to transfer scientific knowledge to real-world situations. Regarding virtual hands-on laboratory activities, Makransky et al. (2020) conducted a study in which a total of ninety-nine students between the ages of 13 and 16 participated in IVR laboratory experiences. The findings of the research demonstrated a significant enhancement in students' interest, self-efficacy, and safety perceptions with regard to laboratory work. In another study, Naz et al. (2024) discovered that interactive laboratory activities utilizing IVR significantly improved senior high school students’ practical skills and their comprehension of chemical processes. Moreover, Gao et al. (2023) found that Chinese junior secondary students using IVR for hands-on physical experiments demonstrated significantly improved knowledge transfer to real-world problem-solving tasks compared to video-based learning. The aforementioned studies demonstrate compelling evidence that leveraging IVR in laboratory learning significantly enhances student learning outcomes, thus fostering advancements in both cognitive understanding and affective engagement, while surpassing outcomes achieved through conventional methods.

In IVR, presence is deemed a crucial factor that can significantly impact student learning, as per the cognitive affective model of immersive learning (CAMIL) created by Makransky and Petersen (2021). The research conducted by Makransky and Lilleholt (2018) confirms that presence in a virtual laboratory simulation serves as a moderator when predicting students’ perceived learning outcomes, satisfaction levels, and behavioral intentions through motivation and enjoyment. Petersen et al. (2022) discovered that students’ perceived presence in an immersive virtual museum exhibition favorably influences their scientific knowledge retention via intrinsic motivation and situational interest. Pedram et al. (2020) revealed that students’ feelings of presence boost their actual and perceived learning using IVR in the context of safety instruction. Thus, prior studies indicate that the sense of presence may impact students' perceptions of IVR learning during their initial experience. However, it remains unclear how presence affects their perceptions of engagement and intention to persist in IVR-supported learning as they become more accustomed to it. Furthermore, the aforementioned studies (Makransky & Lilleholt, 2018; Pedram et al., 2020; Petersen et al., 2022) did not consider presence as a multifaceted construct and overlooked the aspect of self-presence, which may be perceived by students in VR Labs with hands-on activities, thus failing to provide a comprehensive understanding of the relationship between presence and student perceptions of IVR-supported laboratory learning.

2.2 Student perceptions toward IVR based on the expectation confirmation model

Unlike Davis’ (1989) technology acceptance model (TAM), which focuses on users’ initial behaviour intention to use technology, the expectation confirmation model (ECM) is a post-acceptance model that explains why users continue to use an information system (IS) after accumulating some degree of usage experience (Bhattacherjee, 2001). The ECM includes four post-consumption variables: continuance intention, satisfaction, expectation confirmation, and perceived usefulness. Continuance intention refers to users’ intention to continue using an IS, satisfaction reflects users’ feelings about past IS usage, expectation confirmation gauges the match between users’ expectations and IS performance, and perceived usefulness assesses users’ perception of IS benefits. Bhattacherjee (2001) proposed hypotheses stating that continuance intention is influenced by perceived usefulness and satisfaction, perceived usefulness and expectation confirmation positively affect satisfaction, and expectation confirmation influences perceived usefulness.

ECM may also be used to explain student behaviour towards a product or service associated with virtual reality. For example, Jung (2011) used ECM to examine the factors that influence users’ continued use of virtual social worlds in the context of the platform Second Life. Chang et al. (2018) employed ECM to explore college students’ acceptance of VR-based mental rotation training systems. Zhang et al. (2020) examined college students’ continuance intention towards virtual and remote laboratories by reference to ECM and flow theory. The results of previous research have demonstrated the capability of ECM to explore VR-related topics. Thus, the following hypotheses were proposed:

  • H1. Students’ confirmation of expectations has positive effects on their perceived usefulness of IVR.

  • H2. Students’ confirmation of expectations has positive effects on their satisfaction with IVR.

  • H3. Students’ perceived usefulness has positive effects on their satisfaction with IVR.

  • H4. Students’ perceived usefulness has positive effects on their continuance intention towards IVR.

  • H5. Students’ satisfaction with IVR has positive effects on their continuance intention.

In addition, Alhumaid et al. (2021) discovered that expectation confirmation positively affected students’ intention to continue using mobile learning platforms during the COVID-19 epidemic. Pasha et al. (2021) also investigated the impact of virtual reality with respect to enhancing students’ behavioural intention towards learning management systems and found a direct relationship between expectation confirmation and behavioural intention. The following hypothesis was therefore proposed:

H6. Students’ confirmation of expectations has positive effects on their continuance intention of IVR.

2.3 The impact of presence on user perception and intention toward IVR

Presence in the context of virtual environments refers to users’ subjective feelings of being in an environment generated by computers instead of a real physical location (Witmer & Singer, 1998). Following a review of previous research on presence, Lee (2004, p.27) redefined presence as “a psychological state in which virtual objects are experienced as actual objects in either sensory or nonsensory ways” from the perspective of human experience. Felton and Jackson (2022) adopted a different viewpoint and argued that presence is not merely a technology-induced phenomenon. They defined presence as “the extent to which something (environment, person, object, or any other stimulus) appears to exist in the same physical world as the observer” (Felton & Jackson, 2022, p.1). Regardless of these different definitions of presence, it is prudent to understand presence in a virtual world from several perspectives. By reference to Lee (2004) and Makransky et al. (2017), Table 1 indicates three dimensions of presence in a virtual environment.

Table 1 The three dimensions of presence in a virtual environment (Lee, 2004; Makransky et al., 2017)
Full size table

Previous research has thoroughly examined the importance of physical presence with respect to user perception and behavioural intention. For instance, Pedram et al. (2020) discovered that learners’ sense of physical presence in IVR influences their perceived usefulness of the VR safety training system. Melo et al. (2022) discovered that the physical presence experienced by consumers in the context of virtual tourism increases their happiness and behavioural intention towards tourism products. Ammann et al. (2020) used a VR intervention to validate the causal association between physical presence and participants’ behavioural intention to consume a foodstuff. Shin et al. (2013) observed that presence has a favourable influence on the confirmation of students’ expectations regarding a 3D virtual learning system.

In the realm of IVR-supported science education, presence has been recognized as a critical factor influencing students' positive perceptions and behavioral intentions regarding immersive scientific learning. This is because physical presence can create a more immersive and engaging experience, making the technology more appealing and enjoyable to use, as well as more practical and applicable for learning. For instance, Makransky and Lilleholt (2018) observed an unexpected role of presence in predicting students' satisfaction with and willingness to use IVR for science education. Petersen et al. (2020) demonstrated that presence can particularly enhance students' self-efficacy, STEM-related intentions, and outcome expectations in IVR-supported inquiry-based science learning. Furthermore, Andersen et al. (2023) found that immersive virtual field trips, which elicit a higher level of presence, have a more pronounced impact on students' self-efficacy and outcome expectations in science compared to video-based virtual field trips. As a result, the following hypotheses were developed to expand upon the original ECM based on the preceding discussion:

  • H7. Physical presence positively affects students’ expectation confirmation of IVR.

  • H8. Physical presence positively affects students’ perceived usefulness of IVR.

  • H9. Physical presence positively affects students’ satisfaction with IVR.

  • H10. Physical presence positively affects students’ intention to continue using IVR.

Previous research has explored self-presence as a crucial component influencing user perception and behavioural intention in virtual environments. For example, Behm-Morawitz (2013) found that users’ self-presence is positively associated with their satisfaction with their connection with the virtual world. Kim (2020) verified that users’ experiences of self-presence significantly impact their satisfaction with a virtual sports experience. According to the findings of Sun et al. (2015), participants’ perceived presence in the context of VR boxing training influences their perceived usability of the VR training system. Previous studies utilizing ECM have demonstrated that users’ experiences of self-presence directly influence their behavioural intention, such as their continuance intention regarding virtual worlds (Hooi & Cho, 2017) and to engage in healthy eating (Wang et al., 2020). When students interact with different virtual apparatuses through virtual embodiment (i.e., a pair of virtual hands) in the context of immersive laboratory learning, they may feel a projection of their actual arms into the virtual world, resulting in a perceived sensation of self-presence. The ease with which students engage with the virtual world might strengthen their sensation of self-presence and their perceptions/impression of learning. Consequently, we formulated the following hypotheses:

  • H11. Self-presence positively affects students’ expectation confirmation of IVR.

  • H12. Self-presence positively affects students’ perceived usefulness of IVR.

  • H13. Self-presence positively affects students’ satisfaction with IVR.

  • H14. Self-presence positively affects students’ intention to continue using IVR.

Because social presence is concerned with the user's perceptions of interacting with other avatars in a virtual world, a phenomenon which is not considered in this work, this topic was not explored in this study. As a result, this research regarded physical presence and self-presence as antecedent influencing factors relevant to the tasks of explaining and predicting students’ perceptions toward IVR-supported laboratory learning. The research model is presented in Fig. 1.

Fig. 1
figure 1

The research model

Full size image

3 Methods

3.1 Participants and contexts

To examine the incorporation of IVR technology into science curricula, the Chinese government carefully selected a cohort of over 60 primary and secondary schools across six cities in 2020 to serve as the inaugural group of pilot schools. The selection process took into consideration various factors, including students' academic performance, the geographic location of the schools, and the adequacy of their network infrastructure for supporting IVR-assisted science learning. First and foremost, these pilot schools were required to possess the necessary information technology infrastructure to facilitate immersive scientific instruction. This encompassed having well-equipped computer laboratories, interactive whiteboards, VR devices, and high-speed internet access with a total bandwidth exceeding one gigabit. Additionally, science teachers in these pilot schools needed to be proficient in utilizing IVR teaching applications as an integral part of their daily instruction. Meanwhile, pilot schools needed to provide equipment training before IVR-supported science classes are conducted in order to ensure students can use the head-mounted displays and controllers effectively. Moreover, the school principals in these pilot institutions were tasked with providing clear directives to encourage teachers to seamlessly integrate IVR resources into their science courses and instructional methods. For example, teachers were instructed to initially utilize IVR resources to assist students in conducting experiments, allowing them to test hypotheses or theories from their textbooks. In subsequent sharing and discussion sessions, teachers were expected to invite students to share their observations and conclusions from the experiments. At the same time, teachers were encouraged to guide students toward a deeper understanding of scientific concepts and principles. Consequently, these pilot schools were obligated to routinely offer students a comprehensive scientific education that seamlessly integrated IVR technology through a shared virtual laboratory platform.

Following communication with the directors of these pilot schools, four institutions expressed interest in participating in our study. Ultimately, we recruited a total of 167 students aged 13 to 18 (87 boys and 80 girls) from these four pilot schools to participate in our study. These participants used the same equipment to complete IVR science-related programs, and each had completed at least three to four separate IVR science-related programs while wearing a head-mounted display in prior science courses. Furthermore, the IVR resources were employed in these pilot schools to aid students in conducting their scientific experiments and were integrated into the science lessons in a similar manner. As a consequence, they all grew acquainted with the technology and had comparable IVR experiences. Detailed demographic information and their IVR usage experiences regarding the sample of students are presented in Table 2.

Table 2 The students’ demographic profile and IVR usage experiences
Full size table

In this investigation, a sensitivity power analysis was conducted with G*Power 3.0 software to determine an appropriate sample size. According to Faul et al. (2009), sensitivity analyses are employed to afford information about the effect size that research can detect with a certain power given the chosen alpha level and its sample size. Ultimately, the minimum sample size was determined to be 138 with a power level of 0.95, α = 0.05, and f2 = 0.15 (medium effect size). As a result, the obtained sample exceeded the specified sample size and had sufficient statistical power to determine the study model.

3.2 IVR materials and apparatus

This study makes use of IVR learning materials provided by the virtual lab platform (https://vlab.eduyun.cn/portal/home) which are approved for K-12 scientific instruction. These IVR materials on the platform cover a wide spectrum, ranging from elementary school science to secondary school subjects such as physics, chemistry, and biology. They can be experienced by wearing a head-mounted display (e.g., HTC Vive) or other interactive devices. Within the pilot schools we surveyed, several noteworthy IVR learning materials were incorporated into standard science classes; these are detailed in Table 3. The selection process for these representative learning materials in our study adhered to specific criteria, requiring that these materials had been adopted and integrated into science instruction in at least two pilot schools. The instructors strategically included these IVR materials in science courses across different grade levels to align with student’s learning progress.

Table 3 IVR material blended in science lessons in junior or senior secondary schools
Full size table

To set up teaching materials on the IVR device, users need to install the STEAM VR and VIVE PORT applications. VIVE PORT guides users to install drivers for IVR hardware. The head-mounted display can then be connected to a desktop computer via cables and a streaming box. Interactive controllers are paired with the display by pressing buttons. The STEAM VR window displays the connection status of the headset and controllers. A steady green light indicates they are operational. Once connected, students can access virtual experiments on the virtual lab platform. Figure 2 illustrates how students explore the virtual scene using the head-mounted display and manipulate the virtual apparatus with interactive controllers.

Fig. 2
figure 2

The interfaces of the virtual experiment and the equipment used by students

Full size image

In the virtual science experiment “Observing chloroplast and cytoplasmic flow using high-power microscopes” (see Fig. 2), Fig. 2(a) depicts the virtual environment and apparatus, such as the lab workbench, beakers, and microscopes; Fig. 2(b) shows an operation involving the grasping of a beaker with virtual hands by pressing the trigger button on the controller, while the name of whatever is placed in the captured beaker is automatically displayed; Fig. 2(c) depicts an operation involving manipulating the fine focus knob of microscopes with virtual hands by pressing the menu button on the controller to receive action prompts, with the parameter information of the manipulation automatically displayed; and Fig. 2(d) shows the devices used by the students during the virtual experiment, including a head-mounted display and a pair of interactive controllers.

The IVR examinations on the virtual lab platform share a common scenario but feature different instruments depending on numerous scientific themes. For example, the experiment “Observing chloroplast and cytoplasmic flow using high-power microscopes” includes a virtual microscope, while “Measuring the resistance of a conductor according to Ohm's law” incorporates a virtual slide rheostat, ammeter, and voltmeter. Figure 3 depicts the consistent setting with different instruments for these experiments. Elements like the lab workbench, surrounding walls, and the decorations are fixed and visible through the head-mounted display. Apparatus on the lab workbench, such as beaker and instruments, can be operated using controllers. Consequently, interactive equipment and technology enable students to have hands-on experiences that closely resemble those seen in the actual world. As the result, students may have comparable engaging experiences when conducting virtual experiments on a variety of scientific topics. Moreover, these virtual lab hands-on activities were designed based on the guided activity principle (Moreno & Mayer, 2007), enabling students to complete the science experiments with step-by-step procedural guidance.

Fig. 3
figure 3

The screen shots of two different virtual experiments

Full size image

3.3 Instruments

A three-part questionnaire was developed to assess students’ presence in and perceptions of immersive laboratory learning. The first portion of the questionnaire gathered the demographic information of participants, such as their gender, age, and grade. The second portion investigated students’ perceived presence in IVR. Because the IVR contents did not present students with social actors, only physical presence and self-presence were examined in this study. Finally, 9 items were adopted from Makransky et al. (2017) to measure students’ perceived presence in IVR. The following 14 items were modified from the ECM scale developed by Bhattacherjee (2001) to measure students’ perceptions of immersive laboratory learning. All items were measured on a five-point Likert scale with 5 indicating ‘‘strongly agree” and 1 indicating ‘‘strongly disagree”. The items were translated into Chinese, and slight modifications were made to suit the purposes of the present study. To maintain the original meanings, we had an educational technology specialist assess the revised items. A middle school teacher then reviewed them to ensure clarity for students. Finally, five to eight students from a pilot school, but not involved in the study, tested the revised items and found them clear and understandable. Table 4 shows all the items included in the questionnaire.

Table 4 Constructs and items included in the questionnaire
Full size table

3.4 Procedure

The survey was conducted with the consent and approval of school administrators and teachers, in close collaboration with relevant government agencies overseeing experimentation. All procedures involving student participants in this study adhered to the ethical requirements set forth by the funding agency concerning research with human subjects. Given that the majority of participants were minors, parental forms were distributed to their parents through the assistance of teachers prior to the survey. Additionally, we ensured that participants were informed about the voluntary nature of their participation, affirming their right to withdraw from the study at any point. We emphasized that their responses would be kept anonymous and confidential, thereby mitigating potential bias arising from nonresponse.

To efficiently collect data from diverse regions, online survey questionnaires were distributed. The questionnaires were distributed to students by their teachers immediately after completing a virtual experiment, in alignment with their ongoing learning progress. In order to ensure data quality, students were required to complete the questionnaire on the same day they received it if they wished to participate. The survey took place from March 2022 and spanned a period of three months, in line with the teaching schedule of each pilot school.

4 Results

In this work, we employed partial least squares structural equation modelling (PLS-SEM) (Hair et al., 2014) to examine the structural relationships among different variables due to its advantages in dealing with small sample sizes or nonnormal data with a focus on prediction. Furthermore, PLS-SEM has been used extensively in e-learning research (Lin et al., 2020). The reliability and construct validity of the measurement model were first evaluated by confirmatory factor analysis (CFA). Subsequently, the structural model was evaluated to test the hypothesized links among variables.

4.1 Measurement model

To evaluate the measurement model, we examined the reliability and construct validity, which consisted of convergent validity and discriminant validity.

As shown in Table 5, the Cronbach’s α values (0.861 ~ 0.963) and composite reliability (CR) values of each construct (0.905 ~ 0.976) were larger than 0.7, thus indicating good construct reliability (Hair et al., 1998). Regarding convergent validity, the value of the average variance extracted (AVE) of each construct ranged from 0.706 to 0.930, exceeding the minimum required level of 0.7. The value of the factor loading of each measured item exceeded the cut-off value of 0.5, with loading values ranging from 0.763 to 0.972. These results indicate that the convergent validity of each construct was satisfactory (Fornell & Larcker, 1981). We also evaluated the discriminant validity of the measurement model to determine whether the latent variables were independent of each other (Tsai et al., 2011).

Table 5 Construct reliability and convergent validity
Full size table

As presented in Table 6, all the square roots of the AVE of the constructs were larger than the correlation coefficients among each construct, exceeding the cut-off value of 0.5. According to the Fornell-Larcker criterion (Fornell & Larcker, 1981), these results indicate good discriminant validity. Overall, the results reported above indicate that the measurement model used for our study of each of the constructs exhibited good reliability and validity. The model fit is also favorable with a Standardized Root Mean Square Residual (SRMR) value of 0.050, below the recommended threshold of 0.08 according to Hu and Bentler (1999). Additionally, considering distance measures like d_ULS and d_G for bootstrap-based model fit assessment, we found values of 0.390 and 0.812, respectively, both below the recommended threshold of 0.95 (Henseler et al., 2016).

Table 6 Discriminant validity (Fornell-Larcker criterion)
Full size table

4.2 Structural model and hypothesis testing

The statistical significance of the PLS path model was estimated using the bootstrapping resampling approach. According to previous studies (Lin et al., 2020; Streukens & Leroi-Werelds, 2016), bootstrapping with 5000 subsamples is recommended as a minimum requirement for PLS-SEM analysis. Thus, this study examined the structural relationships among the different latent variables proposed in the research model using 5000 bootstrap subsamples. As displayed in Fig. 4, eight significant predictive relations were identified in the model.

Fig. 4
figure 4

The structural model

Full size image

As shown in Fig. 4, expectation confirmation positively affected perceived usefulness (β = 0.499, p < 0.001), satisfaction (β = 0.759, p < 0.001), and continuance intention (β = 0.583, p < 0.001), thereby supporting H1, H2, and H6. Perceived usefulness had a positive impact on satisfaction (β = 0.207, p < 0.01) but no impact on continuance intention, thus supporting H3 and rejecting H4. In addition, satisfaction positively affected continuance intention (β = 0.391, p < 0.01), supporting H5. Physical presence not only positively affected expectation confirmation (β = 0.410, p < 0.001) but also positively affected perceived usefulness (β = 0.396, p < 0.001), supporting H7 and H8. There were no significant relationships between physical presence and satisfaction or between physical presence and continuance intention, thereby rejecting H9 and H10. Self-presence had a positive impact on expectation confirmation (β = 0.397, p < 0.001) but no direct effect on perceived usefulness, satisfaction, or continuance intention; thus, H11 was supported, while H12, H13, and H14 were rejected.

The results of our study show that students’ perceived presence can positively predict their confirmation of expectations regarding IVR-supported science learning, explaining 58.4% of the total variance (R2 = 0.584). In addition, students’ perceived physical presence in the IVR environment and confirmation of their expectations regarding its actual performance jointly predicted students’ perceived usefulness of IVR-supported science learning, explaining 69.1% of the total variance (R2 = 0.691). Both perceived usefulness and expectation confirmation had positive impacts on satisfaction, explaining 86.5% of the total variance (R2 = 0.865). In addition, students’ satisfaction with the learning experience in the IVR environment and confirmation of their expectations of the actual performance of IVR jointly predicted their continuance intention towards IVR-supported science learning, explaining 91.4% of the total variance (R2 = 0.914).

In addition to the use of R2 to assess the model’s explanatory power, Stone-Geisser’s Q2 statistic was also employed to examine the predictive relevance of the endogenous constructs. The Q2 values of expectation confirmation, perceived usefulness, satisfaction, and continuance intention were 0.527, 0.582, 0.773, and 0.843, respectively. According to Hair et al. (2019), Q2 values greater than 0.50, 0.25, and 0 indicate large, medium, and small predictive relevance of the PLS-path model, respectively. Overall, the proposed model indicated good predictive accuracy according to the results of the Q2 value (Hair et al., 2019).

Table 7 shows the outcomes of the hypothesis testing. Our results verified 8 of 14 total hypotheses. According to Cohen (1988), f2 values greater than 0.02, 0.15, and 0.35 indicate small, medium, and large effect sizes, respectively. Thus, the effects of physical presence on expectation confirmation (f2 = 0.142) and perceived usefulness (f2 = 0.239) were close to medium and exceeded medium, respectively. The effect of self-presence on expectation confirmation (f2 = 0.132) was close to medium. Furthermore, expectation confirmation had substantial effects on perceived usefulness (f2 = 0.379), satisfaction (f2 = 1.642), and continuance intention (f2 = 0.595). In addition, the effects of perceived usefulness on satisfaction (f2 = 0.122) were close to medium. According to these results, students’ perceived physical and self-presence in the virtual environment had a relatively modest influence on their confirmation of expectations regarding IVR. Likewise, expectation confirmation had a significant influence on students’ perceived usefulness of IVR learning, satisfaction with the IVR learning experience, and intention to continue using IVR.

Table 7 Summary of hypothesis tests
Full size table

Table 8 shows the indirect effects of the latent variables alongside several significant mediated paths. A general guideline in this case is that if the value of VAF (the ratio of the indirect-to-total effect) is less than 0.2, no mediating effect is evident; if the VAF value is between 0.2 and 0.8, a partial mediating effect is in play; and if the value of VAF is more than 0.8, a full mediating effect is detected (Hair et al., 2016; Nitzl et al., 2016). The results of the mediation analysis are displayed in the last column of Table 8.

Table 8 Specific indirect mediating effects
Full size table

According to these findings, expectation confirmation plays an essential mediating role in the associations between students’ perceptions of presence and their perceived immersive laboratory learning, including perceived usefulness of IVR, satisfaction with the IVR learning experience, and their intentions to continue using IVR to learn science in the future. Specifically, expectation confirmation partially mediates the influence of physical presence on continuance intention. Additionally, expectation confirmation mediates the relationship between self-presence and continuance intention to some degree. In terms of satisfaction, expectation confirmation partially mediates between physical presence and satisfaction. Expectation confirmation fully mediates the effect of self-presence on satisfaction. Regarding perceived usefulness, expectation confirmation partially mediates the influence of physical presence on perceived usefulness. Expectation confirmation also fully mediates the impact of self-presence on perceived usefulness. In addition, satisfaction partially mediates the influence of expectation confirmation on continuance intention.

5 Discussion

The current study investigated the impact of the ECM, encompassing expectation confirmation, perceived usefulness, satisfaction, and continued intention—along with the two additional variables physical presence and self-presence—on the perceptual experiences of immersive laboratory learning among secondary students. According to our results, students’ confirmation of IVR expectations influences their perceived usefulness of IVR learning as well as their satisfaction with the IVR learning experience. Students’ satisfaction with the IVR learning experience can be positively predicted by their perceived usefulness of IVR, and their degree of satisfaction impacts their intention to continue using IVR to learn science. These outcomes align with previous studies on 3D virtual learning environments (Shin et al., 2013) and virtual communities (Feng et al., 2019), providing further validation for the idea that students' intention to continue using IVR for laboratory education is indirectly affected by both physical presence and self-presence. Thus, to promote the sustained use of IVR for laboratory education, it is sensible to maintain students’ experiences with physical presence and self-presence via IVR learning.

The aforementioned results show that it is essential to improve the confirmation of students’ expectations of IVR-supported laboratory learning. Students are more inclined to continue using IVR to carry out laboratory learning if they experience a greater expectation confirmation, since this experience strengthens their perceived benefits of IVR learning and their levels of satisfaction with this approach to learning. Furthermore, our study found that expectation confirmation has a direct effect on students’ continuance intention towards IVR-supported laboratory learning, which is consistent with the results reported by Alhumaid et al. (2021), who showed that expectation confirmation is a significant factor in predicting students’ intentions regarding the usage of mobile learning throughout the COVID-19 period. In contrast to previous research (Alhumaid et al., 2021; Feng et al., 2019), our analysis did not find any association between perceived usefulness and continuance intention. A reason for this finding might be that students discontinue using IVR to learn science due to cyber sickness (Munafo et al., 2017), even if they recognize the benefits of its use. For example, Dehghani et al. (2022) found that users’ perceived health hazards have a detrimental influence on their continuance intention regarding VR gadgets. Therefore, addressing cybersickness in IVR education is crucial to enhance students' perception of the alignment between their expectations and experiences in IVR learning, thereby promoting its sustained use for laboratory education.

Students’ experience of physical presence experience in an IVR context has a strong direct impact on their perceived usefulness of IVR learning. This finding is in line with the results reported by Pedram et al. (2020) who showed that students’ experience of presence in an IVR environment favourably affects their perceived usefulness of IVR training. Fokides and Atsikpasi (2018) further revealed that participants’ perceived presence within a three-dimensional virtual museum positively affects their perceived usefulness of learning. However, Makransky and Lilleholt (2018) found that usability measured in terms of ease of use and perceived usefulness predicts students’ experience of presence when students are provided with both desktop VR contents and IVR contents. Furthermore, no correlation was established between perceived usefulness and presence in the context of the desktop VR experience (Ai-Lim Lee et al., 2010; Makransky & Petersen, 2019). This aforementioned inconsistency in the relationship between perceived usefulness and presence may be attributed in part to different understandings of presence and different ways of measuring it. In previous studies (Ai-Lim Lee et al., 2010; Makransky & Petersen, 2019; Pedram et al., 2020), presence has been regarded as a single construct and assessed using a self-developed scale featuring only one item (Ai-Lim Lee et al., 2010) or using measures developed by Witmer and Singer (1998) or Sutcliffe et al. (2005). In contrast, our research considered presence to be a multidimensional sense construct and assessed it using two subscales—namely, physical presence and self-presence. On the other hand, the diverse degrees of presence that may be achieved by different research environments, such as desktop VR and IVR, may result in inconsistent outcomes. It is apparent that students’ experiences of presence are substantially stronger in IVR than in desktop VR settings (Makransky & Lilleholt, 2018; Zhao et al., 2020).

In this research, we discovered that physical presence and self-presence play distinct roles in influencing students’ perceived laboratory learning in an IVR environment. Physical presence has a direct influence on perceived usefulness and an indirect impact on satisfaction with and intention to continue using IVR. These results were also discovered by Makransky and Lilleholt (2018), who found that presence makes an indirect impact on students’ perceived IVR learning (e.g., satisfaction, behavioural intention). A recent study also confirmed that physical presence favourably affects student learning outcomes in IVR via situational interest and intrinsic motivation (Petersen et al., 2022). The aforementioned findings highlighted the fact that physical presence is a significant component influencing student learning perception in an IVR environment, particularly in the context of a VR Lab. A strong experience of physical presence may be created throughout the design process by offering users fluidly displayed object movements and changes of perspective, a realistic display of the object and surroundings, a high degree of freedom regarding controlling views and manipulating objects, and 3-D audio technology as well as force feedback upon touching virtual objects (Chan et al., 2021; Dalgarno & Lee, 2010; Makransky & Petersen, 2019). Namely, these sophisticated techniques create a highly immersive and engaging virtual world, and both immersion in and the interactivity of the virtual environment have been verified to predict students’ experience of physical presence in IVR learning (Petersen et al., 2022).

Unlike physical presence, perceived self-presence in the IVR environment has an indirect influence on students’ satisfaction, perceived usefulness, and continuance intention towards IVR use. This result is in line with the findings of previous research that examined self-presence in persuasive virtual environments, such as virtual sports (Kim, 2020; Sun et al., 2015), VR tourism (Adachi et al., 2022; Tussyadiah et al., 2018), and VR advertising (Jung et al., 2022; Martínez-Molés et al., 2022; Song et al., 2021). For example, users’ experience of self-presence has a substantial impact on their satisfaction with and the perceived usability of VR sports (Kim, 2020; Sun et al., 2015). The experience of self-presence in VR tourism influences users' favourable destination image, resulting in a greater level of travel intention (Adachi et al., 2022; Tussyadiah et al., 2018). The sense of self-presence in VR advertising has a favourable influence on consumers’ enjoyment, favourable brand attitude, product knowledge, and purchase intention (Jung et al., 2022; Martínez-Molés et al., 2022; Song et al., 2021). These findings highlight the significance of self-presence and its impact on users in VR-related environments, such as IVR environments featuring virtual hands-on activities. To preserve students’ self-presence in virtual environments, a well-designed virtual avatar and synchronous visuomotor (or tactile) stimulation has been proven to be effective. For example, Jahn et al. (2020) noted that preserving the congruence of visual stimuli that come into contact with the body and subsequent touch feedback can preserve students’ strong sense of self-presence in a virtual environment. In addition, avatar-self similarity in the virtual world has been found to positively affect users’ self-presence experience in a virtual environment via self-awareness (Hooi & Cho, 2017), whereas disruption of visuo-tactile synchrony has been found to break the illusion of users’ body ownership and can decrease users’ feeling of self-presence in virtual environments (Kokkinara & Slater, 2014).

According to our findings, expectation confirmation is a crucial mediator between students’ perceptions of presence and their perceived learning. This finding is in accordance with the results delivered by Makransky and Petersen (2021), and Makransky and Lilleholt (2018). In these studies, the mentioned researchers discovered that students’ felt presence in an IVR environment can improve their learning and satisfaction via affective mediators such as motivation, enjoyment, and curiosity; however, they caution that these mediators (e.g., curiosity, interest, or enjoyment) may diminish as students become more familiar with IVR. From these findings, it can be inferred that examining the affordance of presence in the long-term application of IVR is essential (Makransky & Lilleholt, 2018; Radianti et al., 2020). By using expectation confirmation as a crucial mediator, our study endeavoured to explore the affordance of presence in laboratory learning when IVR was incorporated into conventional scientific classes.

Contrary to our initial expectations, the findings of our study indicate that neither physical presence nor self-presence have a direct impact on satisfaction or the intention to continue using IVR for laboratory learning. However, they do exert indirect effects on satisfaction and continuance intention. This observation distinguishes our results from prior research on persuasive virtual environments such as virtual sports, VR tourism, and VR advertising. In the context of IVR laboratory learning, students' satisfaction and their intent to continue using this medium may not be directly contingent on their presence experiences in virtual environments. As highlighted by Han (2020), even when students perceive a strong sense of physical presence during virtual field trips, they might still hold high expectations which, if unmet, could lead to dissatisfaction with the immersive learning experience. Additionally, concerns related to social isolation and technology addiction may further complicate students' satisfaction. This observation aligns with the ECM proposed by Bhattacherjee (2001) which suggests that when students' experiences with IVR learning fall short of their expectations, satisfaction is unlikely, and the intention to continue using it diminishes. Our study also underscores the critical role of expectation confirmation as a mediating factor between students' perceptions of presence and their perceived IVR learning, including perceived usefulness, satisfaction, and continuance intention. In summary, our study expands the application of the ECM to immersive laboratory learning and underscores the significance of both physical presence and self-presence as essential antecedent variables that can impact students' favorable perceptions of immersive laboratory learning. Moreover, our findings respond to the call made by Lee (2010) for further investigation into the factors influencing expectation confirmation and how they can be manipulated in future research on ECM to enhance students' overall e-learning experiences.

6 Conclusions and implications

6.1 Conclusions

The present study investigated the ways in which presence affects students’ perception toward immersive laboratory learning after they have participated in several topics of scientific programs using IVR equipment. Our results showed students’ perceptions of the usefulness of IVR are positively influenced by physical presence in the IVR environment. The perception of physical presence also has an indirect impact on student learning satisfaction and continued intention towards IVR in the context of future study. As per the findings of our study, it has become evident that students’ sense of self-presence within the IVR exerts an indirect influence on their satisfaction, perceived usefulness, and their intention to continue using IVR. This influence operates through the mechanism of expectation confirmation. These study outcomes not only build upon the foundations of prior research but also offer empirical support for the significant impact of self-presence on students' perceptions of learning within the IVR context. Moreover, the concept of expectation confirmation plays a pivotal role as a bridge connecting the sense of presence and students' perceptions of immersive laboratory learning. Consequently, when implementing IVR in laboratory education, it is paramount to ensure that students' IVR expectations are met and confirmed.

6.2 Theoretical and practical implications

The implications drawn from findings on physical presence have significant relevance for technology developers involved in the design and development of IVR environments for educational settings. Firstly, it is imperative that the designed virtual environment incorporates a high degree of interactivity and immersion. This is crucial because extant research, such as the work by Petersen et al. (2022), has demonstrated that these factors positively impact the physical presence experienced by learners. To be more specific, the IVR environment should offer students a heightened sense of control within the virtual environments and provide a rich degree of physical realism. Secondly, the chosen interaction techniques integrated into the virtual world should designed with a simplicity that minimizes the perceived extraneous cognitive load stemming from the interaction process. In other words, the interaction methods should be user-friendly and intuitive, as difficulties with the interaction process may detract from students' sense of physical presence, redirecting their attention towards the technical aspects rather than the intended educational content. This observation aligns with the findings of other research, such as a study by Brogni et al. (2003). Furthermore, the findings on self-presence hold practical implications for technology developers. They underscore the critical importance of enhancing students' experience of self-presence, particularly in virtual labs where hands-on activities are integral. To achieve this, it is advisable for developers to consider implementing well-designed virtual avatars (as suggested by Hooi & Cho, 2017) and synchronous visuomotor or tactile stimulation (following the insights from Jahn et al., 2020) in the design of virtual environments. These strategies can effectively preserve and enhance students' self-presence within these educational virtual settings.

According to our findings regarding the significance of presence, educators should prioritize IVR content that offers a high degree of physical realism and instils a strong sense of control among students. Additionally, educators should consider IVR content with immersive design features, as highlighted by Matovu et al. (2022). Such features may include elements like virtual body ownership, character role-playing, a well-defined storyline, challenging tasks, and opportunities for extensive social interactions. Furthermore, to alleviate the extraneous cognitive load associated with interacting with IVR content, educators should ensure that students receive appropriate guidance on how to effectively engage with immersive technologies. For instance, when the learning task involves interactive 3D object observation, students should be well-informed about how to manipulate visuals to derive meaningful insights, as suggested by Lee (2018). Verstege et al. (2021) also proposed several design requirements and corresponding principles for IVR-supported laboratory learning, which include creating a positive learning experience, facilitating students in achieving intended learning outcomes, and enabling students to complete assignments independently.

Given the crucial influence of expectation confirmation on student perception of IVR learning, educators must implement effective instructional strategies to enhance the actual performance of IVR learning and ensure that it aligns with students' expectations. These strategies can include pre-training, peer instruction, and the use of generative techniques like summarizing or enactment, drawing insights from research such as Meyer et al. (2019), Zhong et al. (2022), Zhao et al. (2020), and Makransky et al. (2021). Furthermore, following the suggestions of Makransky and Lilleholt (2018), a proactive approach is preferred over relying solely on the technology itself. This approach involves careful construction of IVR content to foster a more constructive learning experience. One strategy to achieve this is by incorporating elements of gamification within IVR content, aligning with prior studies such as Shin et al. (2013) and Zhang et al. (2020), so as to reinforce students' confirmation of their expectations regarding learning in virtual environments. In addition to the efforts of educators, it is imperative for policymakers to take action by developing training programs to help teachers address technological and pedagogical challenges, as emphasized by Cardullo and Wang (2021), when integrating IVR into science curricula.

7 Limitations and future works

This research offers intriguing insights but has some limitations. The primary one is the small sample size, with its subjects taken from just four pilot schools; this could introduce selection bias and affect generalizability. To address this, future studies should include a more diverse range of pilot schools. Additionally, given the participation of adolescents aged 13–18, considering potential moderating factors is important for understanding students' intentions to continue using IVR. This may involve including a broader age range in the sample and considering relevant moderating variables (such as spatial ability) in further research (Makransky & Petersen, 2021). Moreover, the participants in our study were drawn from different pilot schools, potentially resulting in a nested data structure. Given our limitation of having fewer than the recommended minimum of 50 schools as per Maas and Hox (2005), it is not feasible to precisely measure the nesting effect in this study. Future research should aim to collect data from a larger number of schools to adequately address this issue. Lastly, it is crucial to note that this study focused exclusively on a specific type of virtual learning environment, namely the virtual lab with hands-on activities, utilizing the virtual lab platform (https://vlab.eduyun.cn/portal/home). Consequently, caution should be exercised when extrapolating the results to other contexts. Future studies should encompass a broader range of IVR learning environments, including emerging platforms like the metaverse, to enhance the understanding of their impact. Additionally, the virtual hands-on activities completed by participants in this study varied based on their individual learning progress, potentially affecting the level of perceived presence. Exploring the influence of presence on perceived learning in diverse educational contexts, such as social and emotional learning within IVR environments (as suggested by Tan et al., 2022), would offer a promising avenue for future research. Moreover, while social presence may exist within IVR learning environments, this aspect was not explored in our study. Thus, investigating its effects in the future would contribute to a more comprehensive understanding of the impact of presence on students' perceived IVR learning.