Keywords

1 Introduction

Virtual reality (VR) is a technology that generates a three-dimensional (3-D) virtual environment in which users interact with virtual objects through sensing devices [1]. Immersive VR allows users to fully immerse themselves in a virtual environment through stereoscopic displays, such as the head-mounted display (HMD) and cave automatic virtual environment (CAVE) [2]. Immersive VR is currently applied in many fields, including education [3], aerospace [4], and entertainment [2]. Educational institutions are inclined to progressively adopt educational immersive VR applications and displays to facilitate students’ learning due to the potential benefits of immersive VR (e.g., immersion, interaction, imagination) [3, 5]. Therefore, research on the impact of immersive VR on learning outcomes has become increasingly important [5, 6].

In recent decades, researchers have empirically investigated how immersive VR enhances learning outcomes. While some have proven that immersive VR has positive effects on various dimensions of learning outcomes [7,8,9], others have failed to find significant effects [10, 11]. These conflicting results have puzzled educators and learners alike, leaving them wondering whether they should use immersive VR, and if so, in which situations immersive VR enhances learning outcomes. Although some studies have provided a few perspectives, proposing that immersive VR’s impact will be influenced by cognitive load and summary habit [3, 10], there is still a lack of research systematically exploring the influencing process of immersive VR to provide possible explanations for the conflicting results. Therefore, this research intends to fill this gap.

This research builds on the two-path model proposed by Makransky and Petersen [12] to explore how immersive VR influences learning outcomes through both affective and cognitive paths. By extending our understanding of the underlying mechanisms by which immersive VR influences learning outcomes, this study provides useful insights into design strategies that will harness the benefits of immersive VR in education.

2 Literature Review

2.1 Immersive VR

Immersive VR is a system with special hardware that provides users with the psychophysical experiences of being completely isolated from the physical world outside by providing a sense of being fully immersed in a virtual environment [13]. It allows the users to interact with a virtual environment via an HMD or CAVE, which blocks the users’ visual access to the physical world and projects a stereoscopic virtual environment [14,15,16]. Immersive VR techniques, such as 3-D images, high-quality audiovisual effects, and accurate motion tracking functions, are used to enhance the users’ feelings of immersion [17].

To gain a deeper understanding of immersive VR, it is necessary to explore its technological features. According to Lee et al. [18] and Whitelock et al. [19], the most prominent dimensions of VR features are vision and movement. Regarding immersive VR, the vision features include representational fidelity and aesthetic quality. Representational fidelity refers to the realistic degree of the virtual objects and virtual world, which is enhanced by 3-D images, scene content, and smooth motion of the virtual objects [20]. The aesthetic quality, which describes the visual aspects of the virtual environment [21], is also essential because it influences the users’ motivations and interests [22, 23]. Due to its representational fidelity and aesthetic quality, immersive VR has great potential to attract and retain users.

The movement features of immersive VR include immediacy of control and interactivity. Immediacy of control describes the ability to smoothly change the viewpoint and to manipulate objects with expected continuity in the virtual world [18]. In addition, immersive VR has another important feature: interactivity, which is defined as the degree to which users influence the form or content of the immersive virtual environment [24]. Highly interactive VR not only allows users to navigate the virtual world but also to explore, control, and even modify this virtual environment [24, 25]. Interactivity, a core advantage of immersive VR, distinguishes immersive VR from other technologies because it provides its users with effectiveness and motivation through multiple interaction possibilities, which most other technologies fail to offer [24]. For instance, immersive VR attracts users by detecting their head and finger movements, while a great number of technologies only respond to users’ clicking a mouse.

2.2 Immersive VR in Education

Immersive VR has been adopted in the education and training industry, including science and mathematics education [26,27,28], firefighting training [29], and mine-rescuing training [30]. Researchers have examined the effects of immersive VR on learning outcomes [6, 9, 31]; however, there is still much doubt about whether it truly increases learning outcomes [14, 16].

Inconsistent Results.

Considerable research has proven the positive influence of immersive VR on learning outcomes, including learning motivation [8, 32, 33], perceived learning effectiveness [6, 34, 35], and objective learning performance [36, 37]. It was reported that immersive VR might increase learners’ motivation through providing interactive experiences [3], immersion [38], and realism [39]. As for perceived learning effectiveness, it would be stimulated by the learners’ perceived presence [32] and their openness to immersive VR [35]. Evidence has also shown that immersive VR increased objective learning performance, which was even better than other traditional teaching methods [36, 37].

However, there are several studies showing that the impact of immersive VR on learning outcomes was insignificant [12, 40, 41]. Immersive VR did not effectively improve the performance of knowledge transfer, although the students felt a strong sense of immersion [11] or presence [42]. One possible explanation was that immersive VR devices were so sophisticated that the learners felt cognitively overloaded in the learning process [10, 42]. In conclusion, although immersive VR provides learners with a strong sense of immersion, it cannot guarantee the improvement of learning outcomes.

The Mediating Role of Learning Processes.

To investigate the effects of immersive VR on learning outcomes, it is important to further identify when and how to use VR’s features to support the learning process, which ultimately enhance learning outcomes. Lee et al. [18] proposed that desktop VR features affected learning through interaction experience (i.e., usability) and psychological factors (i.e., presence, motivation, cognitive benefits, control and active learning, and reflective thinking). Based on the discussion above, this study focuses on analyzing immersive VR features and explaining the process of how immersive VR affects learning outcomes.

3 Theoretical Background and Model

3.1 Technology-Mediated Learning

Technology-mediated learning (TML) describes a situation in which learners interact with learning materials, peers, or lecturers through the intermediary role of advanced information technologies [43]. TML theory suggests that information technology and institutional strategy ultimately influence the learning outcomes through psychological learning process [43]. Studies focusing on how learning is mediated by the intervention of VR have gradually developed TML theory in the VR context [12, 18]. In the work of Makransky and Petersen [12], two paths (i.e., the affective and cognitive paths) are proposed to explain how desktop VR influences learning (see Fig. 1). The affective learning path is concerned with the internalization of learners’ positive emotions and attitudes toward learning [44, 45]. With positive emotions, learners are driven to learn more and apply their knowledge after learning [46]; therefore, affective factors play a crucial role in influencing learning motivation [47] and in the quantity as well as quality of knowledge learned [48]. The cognitive learning path is concerned with the process of improving cognitive abilities and acquiring cognitive objectives [49]. According to Bloom [50], cognitive learning includes memorizing, comprehending, applying, analyzing, synthesizing, and evaluating knowledge, which eventually enhances learners’ acquisition of knowledge, skills, and abilities [49].

Fig. 1.
figure 1

A TML framework in the VR context

Full size image

Based on Makransky and Petersen’s work [12], this study proposes an adapted model of TML theory in the immersive VR context, as shown in Fig. 2 and Fig. 3. There are several differences between our model and Makransky and Petersen’s model [12]. First, this model updates immersive VR features (i.e., representational fidelity, immediacy of control, interactivity, and aesthetic quality) to describe immersive VR.

Fig. 2.
figure 2

Immersive VR features

Full size image
Fig. 3.
figure 3

Research model

Full size image

Second, this study proposes immersion and enjoyment as the affective factors, and control and active learning, cognitive benefits, and reflective thinking as the cognitive factors. Immersion describes the sensation in which learners are fully involved in the artificial world and get confused about the virtual and real world [44, 51]. Enjoyment pertains to the extent to which people find the learning activities enjoyable [52]. Immersion and enjoyment as affective factors indicate the positive emotional responses during the learning process. Three cognitive factors are included in the model. Control and active learning pertains to the instructional design, which allows learners to actively make their own decisions and eventually feel competent as well as self-determined in learning [18]. As for cognitive benefits, it indicates that the learners have improved memorization, understanding, and application of knowledge [18]. Reflective thinking is defined as a state of mind whereby learners inquire about their doubts [12].

Third, combining the work of Lee et al. [18] and Makransky and Petersen [12], this research adopts motivation, learning effectiveness, and learning performance as learning outcomes. Motivation and learning effectiveness are subjective constructs that reflect learners’ viewpoints toward learning. Motivation refers to an internal psychological state that stimulates learners’ behaviors and indicates a direction [53]. Learning effectiveness measures the extent to which learners believe that they have gained knowledge [54]. As an objective construct, learning performance refers to the learners’ overall knowledge and skills acquisition after the VR intervention [55].

3.2 Research Hypotheses

According to TML theory, immersive VR features enhance both immersion and enjoyment. Through increasing the realistic degree of the virtual environment, immersive VR allows learners to easily feel as though they are in the virtual world, enhancing their feeling of immersion [56, 57]. Moreover, with diverse interaction methods, smooth movements, and instant feedback, immersive VR enables learners to directly interact with the 3-D virtual world as they would in the real world, increasing their feeling of immersion [58, 59]. Furthermore, immersive VR enables learners to navigate the virtual world and control the virtual objects in different ways, bringing them interesting experiences and increasing their enjoyable feelings. There is extensive evidence proving that immersive VR features are significantly relative to learners’ enjoyment [6, 12]. Based on the discussion, we propose the following hypotheses.

  • Hypotheses 1a–b: Immersive VR features positively influence a) immersion and b) enjoyment.

TML theory suggests that immersive VR has a positive impact on cognitive factors, including control and active learning, cognitive benefits, and reflective thinking. With vision and movement features, immersive VR not only visualizes knowledge; it also allows learners to actively control the learning objects, which enhances their understanding, memorization, and reflective thinking of the learning content with less cognitive effort than traditional learning [27, 60,61,62]. Based on the discussion, we propose the following hypotheses.

  • Hypotheses 1c–e: Immersive VR features positively influence c) control and active learning, d) cognitive benefits, and e) reflective thinking.

Based on TML theory, the effects of the technology features on cognitive learning processes are mediated by the interaction experience, which is represented by technology usability, including ease of use and usefulness [12, 18]. In this hypothesized model, ease of use is excluded because immersive VR is not always easy for learners to use. Learners tend to believe that learning with immersive VR is useful because immersive VR provides diverse learning experiences and timely feedback on their behaviors [63, 64]. When learners feel that immersive VR is useful, they might also think it is relevant, important, and valuable to their learning, resulting in their willingness to actively control learning activities [65]. Moreover, because perceived usefulness increases learning initiative, it offers the possibility of promoting conceptual understanding [66], and significantly influence reflective thinking [12, 18]. Based on the discussion, we propose the following hypotheses.

  • Hypothesis 1f: Immersive VR features positively influence usefulness.

  • Hypotheses 2a–c: Usefulness positively influences a) control and active learning, b) cognitive benefits, and c) reflective thinking.

TML theory proposes that all psychological factors have positive effects on learning effectiveness. Immersion and enjoyment are proven to promote the positive impact of immersive VR on learning [18, 67, 68]. When immersed in the virtual environment, learners are not disturbed by the outside world; they are focused on learning the content, making their learning effective. When learning with immersive VR is enjoyable, the learners will gain high interests and improve their learning efficiency, which enhances their perceived learning effectiveness [68, 69].

When learners have high control and perform active learning, they make decisions about their learning path, learning pace, and instruction methods [70]. As a result, the learners are able to adjust the learning process according to their personal situations and discover the most suitable learning methods, which allows them to feel their learning effectiveness [18, 71]. With cognitive benefits, learners gain better understanding, memorization, and application of knowledge in immersive VR [18]. After improving their cognitive abilities, learners can comprehend, recall, and apply knowledge in a shorter time, which leads to their perceived learning effectiveness. Reflecting thinking enhances learning effectiveness by enabling learners to critically reflect on what they have learned and their doubts [61, 72, 73]. By gaining new knowledge after reflection, learners tend to believe that their learning is effective. Based on the discussion, we propose the following hypotheses.

  • Hypotheses 3a–e: a) Immersion, b) enjoyment, c) control and active learning, d) cognitive benefits, and e) reflective thinking positively influence learning effectiveness.

Affective factors are proven to have a positive impact on learning motivation [12]. Immersion provided by immersive VR allows learners to fully engage in virtual learning activities, providing them with a different learning environment and unique learning experiences. These novel learning experiences enhanced by immersion stimulate the learners’ motivation and curiosity to explore the virtual world and learn more [74]. As for enjoyment, it is believed to positively affect the learners’ motivation because enjoyment is one of the internal needs that enhances intrinsic motivation in learning [75]. When the learners’ inherent needs for happiness are fulfilled, they will be motivated to learn. Based on the discussion, we propose the following hypotheses.

  • Hypotheses 4a–b: a) Immersion and b) enjoyment positively influence motivation.

Motivated learners are likely to make a positive evaluation of their learning effectiveness [43, 64, 76]. If learners are motivated, they will actively devote more time and energy to learn and gain more knowledge; therefore, they tend to feel that the learning is effective [77]. The relationship between learning effectiveness and learning performance has also been widely discussed. Learners usually think learning is effective because they believe that they gain more new knowledge, which will simultaneously improve their learning performance. Moreover, with positive perception of the learning effectiveness, the learners are more willing to learn, resulting in the improvement of learning performance [76, 77]. Based on the discussion, we propose the following hypotheses.

  • Hypothesis 5a: Motivation positively influences learning effectiveness.

  • Hypothesis 5b: Learning effectiveness positively influences learning performance.

4 Methods

4.1 Research Subjects and Procedures

To test the hypotheses we propose, this study employed a survey method. Sixty undergraduate and postgraduate students were recruited from a university in Hong Kong. First, the participants were asked to answer demographic questions and biology questions related to the immersive VR learning content, which took about 10 min. Next, the participants played an immersive VR game with HMD for about 10 min. Last, the participants were asked to answer a list of biology questions in the first session again and fill in a questionnaire containing all constructs in the research model.

4.2 Software and Hardware

Participants were asked to play an educational VR application, The Body VR. This VR application applied immersive animations and narration to instructing learners about the knowledge of cells and the human body. A simulative and enlarged human body system was designed to allow the learners to shuttle through the human body. The learners could look around the virtual environment at 360 degrees. They were also allowed to touch, rotate, and send out virtual objects, such as red blood cells and white blood cells. Figure 4 shows a screenshot of this game. A 2017 HTC Vive® as the HMD was used to offer the fully immersive VR experiences. Two handheld motion controllers were provided for interaction during the experiences.

Fig. 4.
figure 4

The screenshot of the Body VR

Full size image

4.3 Questionnaires

The questionnaires consisted of demographic questions, a biology test, and measurement items. The demographic questions included age, gender, nationality, major, and educational background. The biology test questions were multiple-choice questions about the information introduced in the learning application. These questions were used to test the prior knowledge as well as the learning performance of the participants. Two pilot tests with 10 students were conducted to ensure the test performances were differentiated. The sequence of the questions used before and after playing was counterbalanced. The measurement items of the other constructs were adapted based on the previous research, and they were measured with a 5-point Likert scale, as shown in the appendix.

5 Results

Among the sixty participants, 41.7% (n = 25) of them were male, and 58.3% (n = 35) were female. Their average age was 20. SmartPLS was used to test the model. The next session shows the results of the measurement model and the structural model.

5.1 Measurement Model

We assessed the measurement model through an estimated coefficient or loading, convergent validity, and discriminant validity. All items load significantly on their latent constructs. We assessed the convergent validity through composite reliability (CR) [47], cronbach’s alpha (CA), and average variance extracted (AVE). The lowest value of CR and CA should be 0.7, while the lowest value of AVE should be 0.5 [78]. All constructs meet the requirement (see Table 1).

Table 1. CR, CA, and AVE
Full size table

The correlational method was applied to evaluate the discriminant validity. The correlations between the constructs should be lower than the squared root of AVE. Discriminant validity is achieved in this model (see Table 2).

Table 2. Correlation between the constructs
Full size table

5.2 Structural Model

Figure 5 and Fig. 6 display the results of the structural model. The parameters of the model include the path coefficients and squared multiple correlation (R2). The path coefficients illustrate the effects of a variable as a cause of another variable, indicating the effective connectivity between the constructs. R2 explains to what extent the variance of a construct is explained by the independent constructs. In general, the results show that most hypotheses have been supported, except H1d, H1e, H3a, H3b, and H3e.

Fig. 5.
figure 5

Loadings of immersive VR features

Full size image
Fig. 6.
figure 6

Structural model

Full size image

As shown in Fig. 6, the model explained 49.8% of the varience in motivation, 65.6% of the varience in learning effectiveness, and 42.0% of the varience in learning performance. Moreover, the immersive VR features are significant antecedents to usefulness (beta = 0.536, t = 5.543), immersion (beta = 0.549, t = 7.386), enjoyment (beta = 0.628, t = 8.206), and control and active learning (beta = 0.225, t = 2.022). Usefulness is a significant antecedent to control and active learning (beta = 0.620, t = 7.055), cognitive benefits (beta = 0.638, t = 6.359), and reflective thinking (beta = 0.683, t = 7.380). Immersion (beta = 0.237, t = 2.036) and enjoyment (beta = 0.563, t = 3.606) are significant antecedents to motivation. Control and active learning (beta = 0.366, t = 2.267), cognitive benefits (beta = 0.289, t = 1.992), and motivation (beta = 0.439, t = 2.952) are significant antecedents to learning effectiveness. Learning effectiveness (beta = 0.308, t = 2.813) is a significant antecedent to learning performance. However, immersive VR features are not significant antecedents to cognitive benefits (beta = 0.221, t = 1.832) and reflective thinking (beta = 0.068, t = 0.661). Immersion (beta = −0.051, t = 0.501) and enjoyment (beta = −0.168, t = 1.301) are not significant antecedents to learning effectiveness. The control variables of this model include age (beta = 0.240, t = 1.910), gender (beta = 0.041, t = 0.334), degree (beta = −0.199, t = 1.641), interest in technology (beta = 0.037, t = 0.356), previous VR playing experiences (beta = 0.131, t = 1.214), and prior knowledge (beta = 0.395, t = 3.238).

6 Discussion

This research was set to explore how immersive VR influences learning outcomes. To do so (1) we conceptualized and operationalized new technological features to describe immersive VR and (2) we unpacked the paths from the immersive VR features to the learning outcomes, focusing on the affective and cognitive factors.

First, we proposed four subdimensions of immersive VR features, and we empirically tested the constructs. The convergent and discriminant validity satisfied the criteria, indicating that two new VR features (i.e., interactivity and aesthetic quality) are valid factors that describe immersive VR’s characteristics.

Second, we tested how immersive VR features influenced learning performance through the affective path. The results indicate that immersive VR features have a direct impact on the affective factors (i.e., immersion and enjoyment). We did not find the direct effects of the affective factors on learning effectiveness and performance. Instead, the affective factors significantly increase motivation, which further influences learning effectiveness and performance.

Third, we also tested how the immersive VR features influenced learning performance through the cognitive path. The results indicate that the effects of immersive VR features on the cognitive factors (i.e., control and active learning, cognitive benefits, and reflective thinking) are mediated by usefulness. Except reflective thinking, control and active learning as well as cognitive benefits have a direct impact on learning effectiveness.

6.1 Theoretical Implications

There are several theoretical implications provided by this study. First, with two new subdimensions, the VR features that are shown in this model comprehensively characterize the technological features of immersive VR. Researchers are advised to adopt these features as a basis for their immersive VR research in the future.

Second, we have identified how immersive VR enhances learning outcomes through affective and cognitive paths. While the results are consistent with the research by Makransky and Petersen [12], this study further develops their findings. Rather than manipulating the learning outcomes as a second-order construct [18] or ignoring the subjective evaluation of learning effectiveness [12], we also have an advanced understanding of the relationships between specific learning outcomes. We find that motivation is proposed to influence learning performance through learning effectiveness. Prior literature either failed to explain why some psychological factors failed to influence learning performance [12] or identify the precise relationships between them [18]. By unpacking these relationships, we have clearly proposed and confirmed how the affective and cognitive factors influence different learning outcomes and how immersive VR effectively influences learning outcomes.

Third, we also provide a new perspective to explain the contradictory research results on the effects of immersive VR to some extent. We have discovered the indirect impact of immersion and enjoyment on learning effectiveness. For example, immersion is considered to be one of the advantages that contributes to the positive effects of immersive VR on learning outcomes, partially because it allows learners to devote themselves to the virtual learning environment and focus on learning [9, 79,80,81]. It was predicted that learners should feel that their learning is effective when they are immersed in the learning environment. However, our research indicates the indirect impact of immersion on learning effectiveness, which is aligned with the research by Hamari et al. [82]. The framework shows that immersion needs to increase motivation before positively affecting learning effectiveness. More specifically, if immersive experiences fail to motivate learners, learners will not feel that their learning is effective. In conclusion, our study suggests that it will be difficult for immersive VR to enhance learning outcomes if it cannot influence these affective and cognitive factors.

6.2 Practical Implications

Our work illuminates crucial immersive VR features that play an important role in the learning process. To effectively enhance the positive effects of immersive VR, designers of immersive VR applications should focus on improving the technological features proposed in this research.

Moreover, this research points out the significant psychological factors (i.e., the affective factors and cognitive factors) of learners that immersive VR designers need to pay attention to. The relationships between the psychological factors and learning outcomes are also revealed from the affective and cognitive paths. Based on these findings, effective strategies to increase immersive VR’s positive effects on learning will be proposed. For instance, designers are advised to increase learners’ motivation through enhancing their immersive and enjoyable feelings. In general, many strategies are concluded from this research to improve learners’ immersive VR learning experiences.

6.3 Limitations and Future Directions

This research has limitations, which should be considered in interpreting the findings. The limitations might become future research opportunities. First, given that this study used a single VR application, the generalizability of our findings is not ensured. Researchers are advised to test the model by using diverse immersive VR applications and subject samples. Researchers might benefit from increasing the sample size or recruiting subjects with different educational backgrounds to test the model. Second, although this research included the major psychological factors associated with the learning process in the model, there might be other factors that may mediate the effects of immersive VR on the learning outcomes, such as cognitive load [10], flow [83], and satisfaction [32]. Therefore, researchers will benefit from taking a more comprehensive set of factors into consideration when developing an understanding of immersive VR use in education.

7 Conclusions

This research builds on a two-path model to explain how immersive VR influences learning outcomes through the affective and cognitive paths. Prior studies have shown that VR technologies may or may not enhance learning performances. By investigating how immersive VR features influence learning processes and outcomes, our research provides explanations for the inconsistent findings. We speculate that if immersive VR features do not stimulate the affective and cognitive factors to enhance motivation and learning effectiveness, it will be difficult to eventually improve the learners’ performance. We also suggest that the interaction experiences and psychological factors should be considered when designing an immersive VR to enhance learning.