Introduction

Virtual reality (VR) has been increasingly used as an innovative technology to facilitate learning and motivation in various settings (Kim et al., 2020). It is an interactive multimedia environment involving multisensory experiences through the use of computer technology (Bryson, 1998). One key feature defining VR is immersion - a state of “shutting out physical reality” when one is engaged in another reality in a virtual space (Cummings & Bailenson, 2015, p. 3). Immersion is closely related, yet remains distinctive, to the concept of presence (Ryan, 2015; Slater & Wilbur, 1997). More specifically, immersion concerns the technological quality and the capabilities of the delivery systems in creating an illusion of one being transported to a virtual world that resembles another reality (Slater & Wilbur, 1997). Thus, immersion is considered “an objective measure” of how faithfully a virtual environment replicates the real-world environment; while presence emphasizes “the psychological experience” of being in a virtual environment, i.e., a subjective measure referring to a sense of “being there” (Cummings & Bailenson, 2015, p. 2-3). It is argued that immersion is required to elicit a sense of presence (Schott & Marshall, 2018; Witmer & Singer, 1998). The technological quality of a VR environment, i.e., immersive features, affects the level of psychological presence experienced by a user; in other words, higher immersion should lead to higher presence (Cummings & Bailenson, 2015), which, subsequently, influences various cognitive and psychological outcomes, such as learning, enjoyment, cognitive load (the mental effort exerted on working memory at a given time;Sweller et al., 2011), or self-efficacy, which commonly refers to confidence (Makransky & Mayer, 2022; Makransky & Petersen, 2021).

One factor influencing immersion is the type of VR delivery mode. For example, the immersion level of desktop VR (dVR) experienced through a 2-D computer screen is considered low, while the immersion level of 3-D immersive VR (iVR) experienced through a head-mounted display (HMD) is considered high (Slater & Wilbur, 1997; Zhao [Jian]et al., 2020). Another important, yet less studied, factor influencing the level of immersion is auditory information in VR. For example, research indicates that environmental sounds provide essential information about the surroundings (Bosman et al., 2023), or that white noise is effective in isolating users from the real-world sounds (Slater & Wilbur, 1997), both of which can help immerse users in a virtual environment by minimizing the distractions from the physical reality.

The purpose of this study is to investigate the effects of immersion in a virtual science learning environment on the intended cognitive and psychological outcomes. As indicated previously, immersion is an essential technical feature of iVR environments, which can affect the perceived presence and, subsequently various learning and psychological outcomes. However, limited research has examined the effects of immersion from multiple mechanisms in learning environments, such as delivery mode and sound stimuli. Additionally, previous research has produced conflicting findings on the effect of immersion (see Review of Literature), which highlights the need for more empirical studies to better understand how iVR can be used to facilitate learning and psychological outcomes. The present study aims to address this research gap. The key research questions are: In a VR science learning environment, what is the impact of the level of immersion on (a) learning, and (b) psychological outcomes, including cognitive load, self-efficacy, presence, and enjoyment? Immersion is operationalized through two different mechanisms in this study, including the delivery mode (iVR versus dVR) and sound stimuli (ambient environment sound versus no sound).

Review of Literature

Immersion Through Head-Mounted Display (HMD)

Delivery mode is one main mechanism of immersion researched in the literature. Many studies have compared the effects of iVR with dVR on various cognitive and psychological outcomes. As far as presence is concerned, empirical evidence in science learning topics tends to support that iVR provides a heightened sense of presence as compared to dVR, either in K-12 (Lai et al., 2021) or university learning settings (Makransky et al., 2019; Zhao [Jiayan]et al., 2022). Similar results have also been noted in other learning and performing contexts, such as earthquake preparedness (Shu et al., 2019) and virtual wheelchair simulation tasks (Salgado et al., 2022). Only limited research found no significant difference between iVR and dVR on presence (e.g., Alrehaili & Al Osman, 2022). Collectively, research evidence tends to support that iVR is associated with higher perceived presence when compared to dVR.

It is also argued that presence, as a defining feature of iVR, should subsequently impact multiple cognitive and psychological outcomes, such as learning, cognitive load, self-efficacy, and enjoyment (X. Huang et al., 2024; Makransky & Mayer, 2022; Makransky & Petersen, 2021). However, empirical studies do not always support this argument. For example, increased presence in iVR environments does not necessarily lead to better learning in previous studies. Although there is evidence that iVR resulted in better spatial learning outcomes (Srivastava et al., 2019) or superior performance in wheelchair simulator training (Salgado et al., 2022), benefits of dVR over iVR in both science learning (Makransky et al., 2019) and wayfinding tasks (Feng, 2021) have also been reported. Additionally, there is research indicating that dVR led to better knowledge retention in science learning, despite that the iVR group demonstrated enhanced performance in practical lab tests. Still other research reported no significant differences in learning outcomes (Hawes & Arya, 2022).

The comparison of iVR and dVR also generated inconsistent findings on other psychological outcomes related to presence, such as cognitive load, self-efficacy, and enjoyment. In Makransky et al. (2019), the university students using iVR experienced significantly higher cognitive load compared to those using dVR during a science simulation. This disparity could be attributed to the possibility that 3-D environments imposed a greater cognitive load than 2-D, which is consistent with the findings of Richards and Taylor (2015). Similarly, Salgado et al. (2022) reported that the iVR group exhibited higher cognitive load during the wheelchair simulator task. However, in a shopping experience study, Ricci et al. (2023) observed a similar cognitive load between the dVR group and the iVR group. Therefore, the relationship between cognitive load and the level of immersion in VR remains unclear and necessitates further investigation. One hypothesis regarding why higher immersion, such as iVR compared to dVR, may not always be beneficial is due to the complex nature of immersive learning environments. For example, immersion-related elements like 360-degree visuals may potentially introduce unnecessary cognitive load that interferes with learning (Makransky et al., 2019). Additionally, learner characteristics such as prior experience with VR may impact the effect of immersion (X. Huang et al., 2024), contributing to the mixed findings in the literature. At the same time, limited research has been conducted on the differential impacts of iVR and dVR on self-efficacy and enjoyment. Lai et al. (2021) revealed that iVR users demonstrated greater computer self-efficacy than their dVR counterparts, who, conversely, achieved higher scores in flow and affective domains. In contrast, Shu et al. (2019) found no significant disparity in self-efficacy in earthquake preparedness between their iVR and dVR participants. As far as enjoyment is concerned, similarly, the research findings are not conclusive, with some studies reporting higher enjoyment in iVR (Zhao [Jiayan]et al., 2022), while others suggesting no difference between the two modes of immersion (Alrehaili & Al Osman, 2022).

In summary, these findings suggest that while iVR can enhance certain aspects of the learning experience, it does not consistently outperform dVR in educational settings. This suggests a more complex relationship between immersion through the delivery mode mechanism and the intended cognitive and psychological outcomes in learning.

Immersion Through Auditory Stimuli

Sound provides another possible mechanism for inducing a more immersive experience in virtual environments (Lindquist et al., 2020). However, to date, only a limited number of studies have explored how sound information might influence perceived presence or other related outcomes. Hendrix and Barfield (1996) was one of the first studies to investigate how auditory information could influence users’ perceptions in a virtual environment. According to the authors, including auditory cues to a virtual environment adds a new sensory channel of information, which should boost perceived presence due to “a richer display medium” (p. 293). The authors validated the hypothesis by recruiting university students to complete a simple virtual room navigation task. Both spatialized and non-spatialized sound stimuli associated with a soda vending machine (e.g., inserting a coin, delivering a soda can) were used in the research. They found that the inclusion of spatialized sound led to a significantly higher sense of presence and realism as compared to no sound or non-spatialized sound. The benefit of sound was also observed in more interactive virtual environments. In Sanders and Scorgie (2002), participants played a computer shooter game in the context of World War II, with either no sound or one of the three sound conditions via three delivery methods (speakers, headphones, and headphones with a subwoofer). The researchers found that sound significantly enhanced the perceived presence as compared to no sound, regardless of the sound delivery method. There was also more recent research focusing on the influence of ambisonic sound versus monophonic sound on perceived presence (Ferdig et al., 2020). Forty-six preservice teachers watched a 360-degree video lesson on elementary math in an online self-paced setting using dVR, accompanied by either monophonic sound or ambisonic sound. Results indicated that the different sound types did not significantly affect presence.

Ambient sound is another type of sound information researched in literature, which refers to “the site-specific background sound component providing locational atmospheres and spatial information of public places'', such as wind, rain, birds, rustling leaves, and traffic noises (Chattopadhyay, 2017, p. 352). Even though ambient sounds “carry the primary spatial information” for constructing a sense of presence, their effects have been largely understudied (Chattopadhyay, 2017, p. 354). One empirical study was found investigating how environment-congruent sound would affect university students’ perceptions of virtual greenspaces (Lindquist et al., 2020). Through a 2 (display type: dVR vs. iVR) x 3 (Sound: ambient vs. detailed vs. none) x 5 (environment type: vacant lot vs. community garden vs. habitat vs. woodlot vs. playground) design, the authors found that ambient sound and detailed sound, as compared to no sound, improved perceptions of biodiversity, preference, and realism. It is worth noting that in this study, ambient sound was operationalized as “a generic background sound of city” that was the same for each environmental condition, while detailed sound was operationalized as “sounds from local wooded areas, fields and parks”, in combination with “other sounds” sourced from an institutional library (Lindquist et al., 2020, Methods section). Regardless, the study indicated the role of environmental sound in enhancing VR experiences. This finding was supported by Kern and Ellermeier (2020), where benefits of nature sound in inducing perceived presence and realism were reported.

In sum, these limited studies focused on sound from different perspectives in VR environments, such as sound versus no sound, different sound delivery methods, or different types of sound. They tend to support the idea that rich auditory information can strengthen the psychological presence experienced in virtual environments. However, how sound impacts learning and psychological outcomes such as enjoyment, cognitive load, and self-efficacy is rarely studied, especially in science learning environments. If sound leads to an increased sense of presence, and presence is expected to influence enjoyment, cognitive load, and self-efficacy (Makransky & Mayer, 2022; Makransky & Petersen, 2021), then more empirical studies are called for to verify the hypothetical theory.

Methodology

Participants and Research Design

Participants included 120 adult learners who experienced one of four versions of a condition-dependent VR tour on nature-based science learning. The study involved a 2 (delivery format: desktop VR vs. immersive VR) x 2 (nature ambient sound: yes vs. no) factorial design. Specifically, the immersion level of the tour was implemented through two mechanisms in this study: (1) the delivery format, i.e., desktop VR viewed through a computer screen versus immersive VR viewed through an Oculus Quest; and (2) inclusion or exclusion of nature ambient sound (e.g., birds chirping and water flowing) during the virtual tour. Sixty participants used Oculus Quest to view the VR tour either with ambient sound (iVR+Sound; n = 30) or without ambient sound (iVR-Sound; n = 30). Another 60 participants used a computer to view the virtual tour, and similarly, either with (dVR+Sound; n = 31) or without ambient sound (dVR-Sound; n = 29). A between-subjects design was chosen instead of a within-subject design to avoid order effects caused by participants completing all the conditions in a particular order, leading to effects such as practice and carry over that may confound the findings (Thomas & Hersen, 2011).

Participants were recruited from an online crowdsourcing platform for research, where they were told to explore a nature park in a virtual environment. Participant eligibility criteria included (1) 18 years old or above, (2) English as first language, (3) highest education included high school diploma/A-levels, technical/community college, or undergraduate degree, and (4) a 90% + approval rate for their previous research participation on the platform. Participants in the iVR group needed to meet an additional eligibility criterion – access to an Oculus Quest headset and a swivel chair to take a seated position while navigating the virtual environment. In this sense, the iVR participants were not novice users of immersive HMD devices.

To verify the quality of data and ensure we included only participants who paid attention during the study, we used two attention check questions asking participants to select a specific answer choice or write a specific number in response to the attention prompts embedded in the assessment instruments (X. Huang et al., 2019;Oppenheimer et al., 2009). One participant was removed during the data analysis phase as a result of failing both attention check questions, resulting in a final sample size of 119.The sample included 67 males, 49 females, and 3 non-binary individuals, with 70.6% white, 10.1% black, 9.2% Latinx, 5.9% Asian/Pacific Islander, 3.4% Middle Eastern, and 0.8% of two or more races. Most of the participants (59.7%) identified 25-40 as their age category, with 24.4% aged between 18-24, 13.4% aged between 41-56, and 2.5% aged between 57-66.

Task and Instruments

The VR tour featured a nature trail in an urban park in Southeastern United States. The virtual tour was created using a collection of 360-degree pictures of the trail, with intentionally built-in action buttons (e.g., the park map, information icons, forward/backward buttons) and textual information throughout the tour. Participants could navigate the virtual environment through the embedded action buttons and learn various science topics during the VR tour, such as karst landscape, native and invasive plants and trees, and types of butterflies. The VR environment design followed cognitive load theory (e.g., Sweller et al., 2011) and cognitive theory of multimedia learning (CTML; e.g., Mayer & Fiorella, 2021) to facilitate participant learning while mitigating the potential adverse effects of high immersion associated with iVR environments. More detailed information about the design of the VR environment can be found in X. Huang et al. (2023).

Instruments included measures on participants’ actual learning, perceived learning, self-efficacy, cognitive load, presence, enjoyment, and perceived usefulness. Detailed information about each study measure is described below.

Learning was assessed using a knowledge retention test (KR20 = .66) where participants were asked to complete 15 multiple-choice questions related to various learning topics covered in the virtual tour.

In addition, open-ended questions were included to obtain qualitative data to better understand participants’ experience with the condition-dependent virtual tour, in particular, what they enjoyed about this experience, what they wished to be part of the VR experience, and whether they would prefer a virtual tour or a walking tour if given a choice. It was expected that participants’ responses to these questions would provide further insights in addition to the quantitative data gathered from the instruments discussed above.

Procedure

Before their condition-dependent VR tour, participants completed a pre-survey on baseline information such as pre-tour self-efficacy, comfort level with VR, and knowledge level of environmental education through the study link. Subsequently, participants in the dVR conditions were directed to another link to access the virtual tour through their computer browser, while participants in the iVR conditions accessed the virtual tour link through their WIFI-enabled Oculus Quest headset and the accompanying controllers. iVR participants were instructed to remain seated in the swivel chair during their virtual experience and that they could use the swivel chair to turn around to view the 360-degree virtual tour. All participants were told to set up the audio volume appropriately on their Oculus Quest headset (iVR conditions) or computers (dVR conditions) as they might receive audio information. They could explore the virtual tour at their own pace without any time restriction.

After the VR tour, participants completed a learning test and a post-survey on the intended outcomes such as post-tour self-efficacy, cognitive load, presence, enjoyment, and perceived usefulness. After participants completed the outcome instruments, several manipulation check questions were asked, such as whether they were able to view the virtual tour without any problem, and whether they heard background ambient sounds during the virtual tour. Participants were instructed that it was possible that the tour did not include such sounds, and they were asked to answer this question honestly.

Finally, participants completed the open-ended questions regarding their experiences and perceptions of the condition-dependent virtual tour. The average time participants spent on the entire study was 43.93 minutes.

Data Analysis

Analyses of pre-existing group differences: To test if there were pre-existing group differences in pre-tour self-efficacy, comfort level with VR, or knowledge level of environmental education, initial MANOVA was performed with delivery format (desktop vs. Oculus) and ambient sound (yes vs. no) as the factors. A Chi-square test was also done to assess pre-existing group differences in prior iVR experience (yes vs. no). Results found significant differences in comfort level with VR and their prior iVR experience (see Results).

Main MANCOVA analysis: While controlling participants’ comfort level with VR and their prior iVR experience as covariates, MANCOVA was performed to test how delivery formats (desktop vs. Oculus) and ambient sound (yes vs. no) impact the intended learning and motivation outcomes: knowledge retention, perceived learning, cognitive load, post-tour self-efficacy, presence, enjoyment, and usefulness.

Testing of MANCOVA assumptions: We first tested if the quantitative data violated any assumptions for performing MANCOVA such as multivariate normality. Shapiro-Wilk tests showed that the data for cognitive load, post-tour self-efficacy, and presence were approximately normally distributed, ps > .05. Shapiro-Wilk tests indicated non-normality in knowledge test score and in the data for perceived learning, enjoyment, and usefulness, ps <.001. However, MANCOVA is considered robust to moderate deviations from normality. Box’s M test was not significant, M = 114.62, F(84, 25668.41) = 1.20, p = .10, indicating equality of covariance matrices. So, the assumption of homogeneity of covariance matrices was not violated. Regarding multicollinearity, the correlations among the dependent variables were all below .90, ranging from r = .06 to .79 (see Table 2), indicating no concerns with multicollinearity.

Results

Descriptive Statistics

Table 1 and Table 2below present the descriptive statistics of the outcome measures and their correlations, respectively. Overall, participants achieved near mastery performance (average 75% accuracy on knowledge retention) and provided largely positive ratings on their virtual experiences.

Table 1. Means & SDs for the Outcome Measures (by Condition & Overall)

Table 2. Correlation Matrix

**: significant at the .01 level (2-tailed). *: significant at the .05 level (2-tailed).

Results of Pre-Existing Group Differences

The results of pre-existing group differences indicated a significant difference in comfort level with VR between the desktop and Oculus groups: the desktop groups had greater comfort level with VR (M = 2.69, SD = .99) than the Oculus groups (M = 1.66, SD = .81), F(1, 113) = 38.42, p < .001, ηp2 = .25. No group differences in pre-tour self-efficacy or knowledge level of environmental education were observed, ps > .05.

The Chi-square test result for prior iVR experience (yes vs. no) showed that the proportion of participants who reported prior iVR experience significantly differed by group, X2(3, N = 119) = 30.38, p < .001. All participants in the Oculus groups (n = 59) reported having used an Oculus or similar iVR device before, whereas 60% of the participants in the desktop groups (n = 36 out of 60) reported having used an Oculus or similar iVR device before.

Results of Main MANCOVA Analyses

The MANCOVA results (participants’ comfort level with VR and their prior iVR experience as covariates) showed main effects of delivery format on four of the outcome measures. Specifically, desktop format (vs. Oculus) resulted in higher perceived learning (Ms= 5.61 vs. 5.34 , SDs = 1.03 vs. .98), F(1, 107) = 9.64, p =.002, ηp2 = .08; higher perceived usefulness (Ms= 6.13 vs. 5.99 , SDs = 1.02 vs. .93), F(1, 107) = 5.22, p = .024, ηp2 = .05; higher perceived presence (Ms= 5.52 vs. 5.25 , SDs = .85 vs. 1.06), F(1, 107) = 4.51, p =. 036, ηp2 = .04; as well as higher post-tour self-efficacy (Ms= 5.58 vs. 5.43, SDs = 2.29 vs. 1.97), F(1, 107) = 4.72, p =.032, ηp2 = .04. That is, after controlling for the effects of participants’ comfort level with VR and their prior iVR experience, the desktop groups consistently reported higher perceived learning, usefulness, presence, and post-tour self-efficacy than the Oculus groups.

Regarding the covariates, comfort level with VR was significantly and negatively associated with perceived learning, B = -.33, p =.002; perceived usefulness, B = -.28, p =.007; perceived presence, B = -.25, p = .014; as well as post-tour self-efficacy, B = -.57, p = .01.

No other effects reached significance, ps >.05.

Supplemental Analysis

Given that participants reported self-efficacy twice (pre- and post-tour), we further analyzed if there was a significant change in self-efficacy level before and after the tour and whether that change depended on the group. The GLM repeated measures test indicated a significant increase in self-efficacy from the pre-survey to the post-survey, F(1, 106) = 18.58, p < .001, ηp2 = .15. There were no significant interactions between time of efficacy measurement (pre-post) and group (delivery format or sound factor), ps > .05, indicating that all groups reported higher self-efficacy perceptions from before to after the tour.

Qualitative and Other Findings

Participant responses to the three open-ended questions (what they enjoyed about this experience, what they wished to be part of the VR experience, and whether/why they would prefer a virtual tour or a walking tour if given a choice) were coded and analyzed to identify emerging themes (see Data Analysis). Themed results are presented below by questions.

Preference of a walking tour or a VR tour:When asked whether participants would prefer a walking tour or a VR tour if given a choice, their responses revealed some interesting patterns. Figure 1 summarizes participants’ frequency of responses by groups.

Figure 1. Participants’ Preferences of a VR vs. Walking Tour: Themes and Frequency

As indicated in Figure 1, 16% more participants in the iVR groups (n = 17; 28%; nine iVR+Sound participants and eight iVR-Sound participants) indicated that they would prefer a VR tour as opposed to the dVR groups (n = 7; 12%; five dVR+Sound participants and two dVR-Sound participants). Consistent with these findings, the two iVR groups had less participants (n = 25; 42%; 13 iVR+Sound participants and 12 iVR-Sound participants) indicating that they would prefer a walking tour as compared to the two dVR groups (n = 38; 63%; 19 participants in each group). In other words, 21% less participants in the iVR group would like a walking tour as compared to the dVR participants. Approximately the same number of participants in the iVR groups (n = 16; 27%) and the dVR groups (n = 15; 25%) mentioned both VR and walking tours depending on specific situations (e.g., walking tour when it is close by and VR tour when it is far away).

When asked to explain their preference, participants choosing a VR experience mostly mentioned its accessibility, flexibility, immersion capabilities, or learning affordances. For example, one participant wrote “I can experience the tour and feel like I am there from the comfort of my own home. I can feel just as immersed. Also, it saves on money and transport issues.” Another participant mentioned “As you could spend as much time as you please within the areas which interest you the most and learn at your own pace.” Participants choosing a walking tour mostly mentioned they would like a full immersion in a natural environment with all senses, e.g., “You can breathe the fresh air, you can grab leaves, feel the trees, and dip your feet in the water. It's a more engaging experience”.

What participants enjoyed:When asked what they enjoyed most about the virtual experience, several themes emerged from participants’ responses, as summarized in Figure 2. The most frequently mentioned theme related to the educational and informative aspect of the tour, i.e., their learning experience in the virtual environment. For example, “It was really cool learning about all the different species of plant and butterfly, and how the ecosystem worked. It was something I was very naïve about before, but I feel that I learnt a lot”. Similarly, another participant commented they enjoyed “learning about topics I previously had no idea about...I thought the information provided was really concise, interesting and memorable.” One interesting observation was that more participants in the two iVR groups (n = 38; 63%; 19 participants in each group) highlighted the learning aspect of the tour than their peers in the two dVR groups (n = 24; 40%; 12 participants in each group).

Figure 2. What Participants Enjoyed About the Experience: Themes and Frequency

Another major theme emerged about enjoyment related to the design environment of the VR tour, with participants referring to the tour as being self-paced, interactive, intuitive, and with 360-degree views and smooth transitions. For example, one participant noted “I love this type of interactive experience…Navigation feels more or less natural and there's a lot to look at, with all the points along the way being quite engaging.” Another commented, “I enjoyed the fact that I could zoom into certain objects, and as I zoom, more details regarding the object would appear.” Similarly, one participant liked that “there were different icons pointing to specific plants to explain about them.” Notably, more participants in the two dVR groups (n = 18; 30%; 10 dVR-Sound participantsand eight dVR+Sound participants) mentioned that they enjoyed the design aspect of the tour than those in the two iVR groups (n = 9; 15%; five iVR+Sound participants and four iVR-Sound participants).

Sound is another theme of what participants enjoyed about the virtual tour, particularly, for the two groups that received the ambient sound (n = 13; 22%; five iVR+Sound participants and eight dVR+Sound participants). This theme is also closely related to the theme of “sense of being there” in that 67% of these 13 participants mentioned that they enjoyed the sound effect because it helped create a feeling of being there in the park. For example, one participant commented “The audio was evocative and conjured up the atmosphere of a sunny, warm afternoon outdoors and the feeling that I was actually walking on the trail in live time in person.” Similarly, another participant wrote “I enjoyed the sound of birds and sometimes even the crunching of leaves, it made me feel I was there”. It is worth noting that the ambient sound of the virtual tour did not include that of “the crunching of leaves”. This brings up another interesting observation of the study. That is, for the manipulation check question whether they heard background ambient sounds during the virtual tour, 14 participants (23%), including three in dVR (10%) and 11 in iVR (37%), who did not receive such auditory cues reported “yes”. Implications of this observation will be addressed in the Discussion section.

In addition to sound, there was also mention of visuals being helpful in creating a sense of being there, e.g., “The visuals also did its justice, there were times where I felt I was truly involved”. Some other themes emerged included enjoyment of the nature environment (e.g., “I enjoyed being surrounded by nature”), exploration of a new place (e.g., “...to explore a beautiful place that I will likely never visit in person”), and the relaxing experience (e.g., “it was a very calming experience”).

What participants wished to be part of the VR experience: When participants were asked what they wished to be part of their VR experience, similar themes emerged from their responses, as shown in Figure 3, including (more/better) sound, (more) learning, (more interactive) design environment or technology, and (higher resolution) visuals. It is worth noting that 42% of the participants (n = 25) who did not receive the sound cues indicated that they would like to have such sound effects included in their virtual experiences, e.g., “I wish there was sound to immerse myself more in the experience”. Interestingly, more dVR-Sound participants (n = 17) than iVR-Sound participants (n = 8) mentioned sound as a missing element in their experience.

Figure 3. What Participants Wished to be Part of the VR Experience: Themes and Frequency

Conclusion

As immersion is an essential feature of VR technology, it is critical to investigate how the level of immersion impacts various learning and psychological outcomes. However, little previous research has explored immersion through different mechanisms in a VR environment. In this study, the level of immersion was operationalized at two levels, i.e., the delivery format (dVR vs. iVR) and auditory information as environment ambient sound (yes or no). It is one of the first studies that explored the level of immersion through two different mechanisms in virtual learning environments.

Main Findings

Our findings indicate that although immersion did not affect participants’ knowledge retention, nor their cognitive load and perceived enjoyment, it impacted other cognitive and psychological outcomes in science learning. Particularly, regardless of whether auditory information was included or not, a less immersive learning environment in the format of dVR led to higher perceived learning, presence, self-efficacy, and perceived usefulness, as compared to a more immersive iVR version of the learning environment. In other words, the less immersive 2D virtual experience resulted in more beneficial effects in the intended outcomes relative to the more immersive 3D virtual experience. These findings do not support previous research showing the favorable results stemming from iVR such as perceived presence or self-efficacy (Lai et al., 2021; Makransky et al., 2019).

One possible reason may be related to a lack of novelty effect in our study. The novelty effect refers to the hypothesis that the short-term beneficial effects of a new technology could potentially be attributed to the “novelty” of the technology itself (W. Huang et al., 2021;Koch et al., 2018). Previous literature has provided empirical evidence of the novelty effect in different fields, particularly in Computer-supported Cooperative Work (CSCW) and Human–Computer Interaction research (see Koch et al., 2018, for a discussion), although more empirical studies in VR environments are needed to support the novelty effect (X. Huang et al., 2024). Our iVR participants were not novice users of immersive HMD devices, which may explain why the iVR experience did not deliver better intended outcomes. As the novelty effect hypothesis considers participants’ prior experience with the technology used, it leads to another related possible explanation of the study results, although from a different perspective. That is, as perceptions are relative to previous experiences (Southwell et al., 2007), is it possible that the dVR participants were comparing their particular experience with other less immersive technologies such as computer-based video tours without participant interactivity, thus, tended to rate this virtual experience more positively? Similarly, is it likely that the iVR participants were comparing their virtual tour with their previous iVR experiences that were more interactive, such as iVR games including 3D modeling, thus, tended not to rate their virtual experience as more positive? We asked these questions partly because when having a common reference, i.e., participants’ preference of a walking tour or a similar VR tour they experienced, our qualitative data seem to indicate perceived benefits of iVR, in contrast to the quantitative results. More specifically, 16% more participants in the iVR groups preferred a similar virtual tour as opposed to the dVR groups, and 21% more dVR participants preferred a walking tour as compared to their iVR counterparts. These findings seem to indicate that more iVR participants were convinced that their version of the virtual tour was comparable to an actual walking tour, at least in terms of immersion capabilities or learning affordances as reported by some participants. In view of these inconsistent findings, having the same group of participants who experience both versions of the VR tour may allow us to compare the two modes of the virtual experience more directly rather than in relation to their previous similar experiences.

With regard to immersion in terms of auditory information, theoretically, ambient sound should lead to increased presence, which, consequently, may positively impact other cognitive and motivational outcomes (Kern & Ellermeier, 2020; Lindquist et al., 2020). As noted in the qualitative findings, some participants who received the background ambient sound explicitly mentioned that they enjoyed the sound effects as they helped further immerse them in the virtual environment. At the same time, some of their counterparts who did not receive the ambient sound voiced the desire to have such sound effects included as part of their experience. However, this expressed benefit of sound was not reflected in the participants' ratings of perceived presence or other intended perceptions in the study. More specifically, our quantitative results indicated that background ambient sound did not exert a significant effect on any of the intended outcomes, which failed to support the benefits of sound or negative influence of silence as shown in previous research, even in terms of perceived presence (Hendrix & Barfield, 1996; Kern & Ellermeier, 2020).

One possible explanation is that our VR environment consisted of a series of 360-degree pictures accompanied with textual information of the nature-trail, which provided rich and detailed visual information that in itself is sufficient to elicit a sense of presence; thus, the inclusion of the ambient sound served more as background information and did not result in added practical value in that aspect. Another related possible explanation is that when participants were immersed in a virtual environment, the rich visual information might have elicited associated but imaginary auditory information for some participants who received no such auditory cues. For example, 10% of participants in the dVR-Sound group and 37% participants in the IVR-Sound group incorrectly reported that they heard such auditory cues during the virtual tour. Even though these are subjective self-reported data after the VR experience that might not be entirely reliable, it is interesting to note the number of participants in the no-sound groups who claimed receiving such associated auditory cues. This echoes previous iVR research where non-existing sound was reportedly heard by participants (Kern & Ellermeier, 2020). Perhaps more importantly, as the number of such participants in the iVR group is almost 4 times greater than those in the desktop environment, it is possible that the more immersive 3D visual environment more likely elicited the effect of “merged senses” (Bilby, 2015) or “hearing” the visuals (Goller et al., 2009), in the sense of participants automatically filling in the blank of the auditory channel with associated sound in a real-world nature trail. This “merged senses” effect was likely triggered by the rich visual information presented in the virtual nature trail environment, e.g., “hearing” the non-existent sound of leaves crunching, as reported by one participant.

Theoretical Implications

First, this study shows that different mechanisms to operationalize immersion level could impact cognitive and psychological outcomes differently. Therefore, it is important to examine immersion from multiple perspectives. In addition, due to the limited number of studies on the impact of sound as an immersion level in teaching and learning environments, more research is encouraged in this area.

Second, it is important to consider how learners’ prior experience with VR devices might influence the intended learning and motivational outcomes (X. Huang et al., 2024). As the iVR participants in this study were not novice users of immersive HMD devices, we can safely assume that the “novelty effect” discussed in some previous literature (Crompton, 1979; Koch et al., 2018) should not influence the results of the present study. Our study indicates that when learners are experienced users of HMD devices, a more immersive 3D virtual experience may not lead to better cognitive and motivational perceptions than a less immersive 2D virtual experience.

Third, our findings indicate it is important to evaluate participants’ VR experiences using both quantitative and qualitative methods. Qualitative data could provide an additional perspective to examine participants’ perceptions. When qualitative data are not consistent with the quantitative data, it can prompt researchers to dig deeper into the “why” question in order to come up with a plausible explanation.

Last but not least, as the immersion question relates to the well-known “media debate” - whether media affects the learning process (Clark, 1994; Kozma, 1994), at the supervisory level, we may conclude that media does affect psychological outcomes, and in this particular case, the less immersive 2D VR environment seemed to prove more beneficial. The debate is likely to continue due to the inconclusive results in the field. However, it may be a more complex question than simply comparing different types of media, which cannot be answered without the consideration of its context, such as target audience, their prior experience, and the media environment and task. A more important goal may be to focus on maximizing the capabilities inherent in a media technology that facilitates the learning process, which inevitably leads to the question of how we design such learning environments.

Practical Implications and Recommendations

This study shows the great potential of virtual reality in developing science learning and motivational outcomes. For both the iVR and dVR groups, participants achieved near mastery performance and provided mostly positive feedback regarding their virtual experiences. As some participants reported, a VR tour can be a great alternative, and in some cases even more preferable, to an in-person tour due to its accessibility, flexibility, learning affordances, and immersion capabilities.

The research findings also have implications for instructional designers, which we present as recommendations. As indicated by the qualitative data, how learners perceive the learning value of their experience connects greatly to their perceived enjoyment. It is important that the design should allow the learners to see the relevance of their learning, especially in terms of what they perceive as new information gathered from the learning experience. In addition, designers should optimize theory-based, technology-enabled features such as interactivity design and self-paced navigation to enhance the holistic virtual experience. These may include features like zoom in and zoom out, information icons, and intuitive navigation buttons to aid learning while minimizing the potential negative effects of high immersion in iVR environments (see X. Huang et al., 2023). Finally, even though sound as an immersion level did not exert an impact as shown in the quantitative data, participants' qualitative responses indicated they favored the inclusion of background ambient sounds as a means to enhance a sense of being there. Therefore, when creating nature-based informal VR learning environments, designers should consider adding such sound effects as an immersion factor.

Limitations and Future Directions

We acknowledge several limitations of this study. First, we purposefully did not include a pre-measurement on learning due to the concern of the testing effect, i.e., taking a pre-test could influence subsequent post-test performance (Brown et al., 2014). As a result, it is not possible to tell if participant learning increased from the pre-test to the post-test. Even though we found no difference among the groups on their self-reported knowledge level of the general learning domain prior to the intervention, assessing their prior knowledge on the specific learning topics may be helpful if the goal was to investigate the effectiveness of the intervention on pre- and post- learning difference. Second, participants’ actual learning was measured by a knowledge retention test immediately after the VR tour, so it remains unclear if the same result pattern will be obtained in terms of long-term knowledge retention or transfer of learning in a different context. Third, in this study, we used a between-subjects design where independent groups of participants experienced the dVR version and the iVR version of the science learning virtual tour, respectively. A within-subject design where participants experience both versions of the virtual environment may shed more light on the results of the study. This approach, for example, can further alleviate concerns related to participants’ prior VR experience. In addition, as one researcher coded the qualitative data, no inter-rater reliability was assessed. Although the data presented was straightforward and the coder reviewed the coding multiple times to ensure accuracy, caution is advised when interpreting the qualitative findings. Finally, we recruited the participants online and they completed the study on their own following the instructions provided. Future research may replicate the study using local participants and investigate whether the same pattern of results can be obtained.

Funding

This research was partly supported by a Research & Creative Activities Program grant (RCAP #20-8027) at Western Kentucky University.

Authorship Contribution Statement

Huang: Conceptualization, design, data acquisition, data analysis, drafting manuscript, editing/reviewing, securing funding, material support. Zhao: Data analysis, drafting manuscript, editing/reviewing. Jiang: Drafting manuscript. Kang: Drafting manuscript. Huss: Editing/reviewing, securing funding, material support. North: Editing/reviewing, securing funding, material support.