1. Introduction
With the impact of the COVID-19 pandemic, education systems around the world have encountered huge challenges. Many schools have been forced to cancel in-person classes and shift to online teaching and learning [
1]. An increasing number of schools have adopted distance education and are making use of various platforms for teaching activities, for example, video conferencing, emails, and massive open online courses (MOOCs) [
2]. Zhu and Liu [
2] further indicated that these types of teaching activities mitigate the negative effects of the pandemic to some extent; however, several problems are still found. First, students normally learn at home rather than in a school setting, which leads to the decline of self-regulation for some students. There is even a situation where students open the teacher’s live streaming and play with another mobile device at the same time. As a result, students’ attention levels are weaker than those in face-to-face classes, which is in line with García-Peñalvo, Corell, Abella-García, and Grande-de-Prado’s [
3] finding.
In these circumstances, landscape architecture education faces a greater challenge in addition to the above problems. Since landscape architecture education requires exposure to a substantial number of design cases, landscape designs, landscape architecture, and historical remains, situated learning is critical to this area of education. Case studies are extremely important for landscape architecture education. It can provide students with opportunities to be exposed to landscape designs; however, the contexts it can offer in classroom activities are limited [
4]. It is evident that with the influence of the pandemic, students can only explicate these works through photos or videos. It is truly difficult for students to empathize with the works and enhance their inspiration for creation by only responding to the teacher’s demonstration of pictures or videos at the other end of the computer. In the past, when there was no pandemic, the teacher could take students out for field investigations, or students could even experience these places in person. Nonetheless, this became difficult after the outbreak of COVID-19 [
5]. Therefore, technology or a method is urgently needed to solve the aforementioned problems at this moment.
The development of science and technology has led to changes and progress in teaching methods. Among these educational technologies, the application of virtual reality (VR) technology in education has brought tremendous progress to situated learning [
6,
7]. With the maturity of VR technology, several experts have started to integrate it into classroom teaching. VR technology has been applied to many disciplines, and its application value and feasibility have also been proved by studies, for example, in STEM education [
8], EFL students’ English learning [
9], geoscience [
10], medical education [
11], and physics [
12]. The adoption of VR in education has brought about enormous changes to teaching and learning in many disciplines. With VR technology, situational effects that are usually difficult to show in the classroom can be well presented. With the raging pandemic, campus closures, and shortcomings in online learning, VR technology has shown its great potential. Especially in landscape architecture education, its prospects for development are promising.
In addition, we have detected that online education faces a few problems during COVID-19 as we mentioned above. VR technology is needed since landscape architecture education in particular requires numerous contexts. However, the production of materials and the preparation of equipment for 3D model-based VR technology are complicated and cumbersome. Thus, it will consume more human and material resources during the implementation and will have higher operational requirements for teachers and students [
7,
9,
13,
14]. In order to better present the contexts of landscape architecture education, the present study adopted a spherical video-based virtual reality approach. Compared to 3D model-based VR technology, a spherical video-based immersive virtual reality (SV-IVR) learning system is convenient, with simple and intuitive operation. Only a smartphone and a pair of portable cardboard goggles are required to operate the system. No matter whether it is the preparation and creation of materials or the operation for students, SV-IVR can meet the needs of students and teachers and provide educators with a viable alternative [
15].
Furthermore, the use of SV-IVR in education is a relatively new research field in educational technology [
16]. Currently, there are still relatively few studies using this new SV-IVR in landscape design courses. Therefore, we proposed SV-IVR, and also innovated the instructional design to produce a specifically for landscape architecture education SV-IVR learning system which is a systematic cyclical learning process with level design. This can be considered as an innovation under COVID-19. This is for the field of landscape architecture, because landscape architecture design is different from other proven courses, as it requires more three-dimensional spatial imagination. Among the existing studies on landscape architecture education and VR, most were related to its educational potential [
17]. In addition, a majority of studies on the use of VR technology in landscape architecture education also mentioned the disadvantages of 3D model-based VR technology, such as its complexity, high cost, and time-consuming nature [
18,
19]. It is believed that VR technology should be further improved in order to be better applied in courses. As a result, the current study integrated SV-IVR into a landscape architecture history class, which is a relatively new attempt.
In order to assist landscape architecture students in their learning performance, enhance their learning attitudes, and develop their self-regulation during the COVID-19 era, this study developed a landscape architecture SV-IVR learning system based on the SV-IVR approach. A quasi-experimental study was adopted to compare the effect of the SV-IVR approach and conventional technology-supported learning approach on a landscape architecture history class. The following research questions were proposed:
- (1)
Can the SV-IVR learning system enhance students’ learning achievements in comparison with the conventional technology-supported learning approach?
- (2)
Can the SV-IVR learning system enhance students’ learning attitudes in comparison with the conventional technology-supported learning approach?
- (3)
Can the SV-IVR learning system enhance students’ self-regulation in comparison with the conventional technology-supported learning approach?
- (4)
Can the SV-IVR learning system enhance students’ self-efficacy in comparison with the conventional technology-supported learning approach?
- (5)
Can the SV-IVR learning system reduce students’ cognitive load in comparison with the conventional technology-supported learning approach?
Based on the research questions, we propose the following hypotheses.
- (1)
The SV-IVR learning system can enhance students’ learning achievements in landscape architecture class.
- (2)
The SV-IVR learning system can enhance students’ learning attitudes in landscape architecture class.
- (3)
The SV-IVR learning system can enhance students’ self-regulation in land-scape architecture class.
- (4)
The SV-IVR learning system can enhance students’ self-efficacy in landscape architecture class.
- (5)
The SV-IVR learning system will not increase students’ cognitive load in landscape architecture class.
In order to better understand the proposed hypotheses, the following literature review and other content are presented.
3. Materials and Methods
In order to examine the effects of the landscape architecture SV-IVR learning system on landscape education, a quasi-experimental study was conducted in a landscape architecture history course. The quasi-experimental design is a more scientific approach to research because it explores the causal relationships between independent and dependent variables in a well-controlled context. For some experimental studies that are not easy to conduct, the quasi-experimental research method can be adopted to design some control methods to minimize the potential factors affecting the validity of the study [
41]. As the basic course in landscape education, the importance of a landscape architecture history curriculum is self-evident. The present study aimed to explore the effects of the SV-IVR learning system on students’ learning achievements, learning attitudes, self-regulation, self-efficacy, and cognitive load in landscape education.
The analysis of covariance (ANCOVA) was used in this study to perform the statistical analysis of the data. ANCOVA treated the pre-test score as a predictive variable (or control variable) of the post-test score, and then detected whether the adjusted post-test score had intergroup differences after the adjustment of the pre-test score [
42].
3.1. Participants
A total of 140 first-year students majoring in landscape design at a university in China were recruited for this study, with 70 students (39 females and 31 males) randomly chosen as the experimental group and 70 students (41 females and 29 males) as the control group. Their average age was 20 years. The experimental group adopted the SV-IVR learning system in the learning process while the control group adopted the conventional technology-supported learning approach. All of the students were taught by the same professor who had more than eight years of teaching experience in landscape architecture history. Due to the pandemic, Google Cardboard was mailed to each student in the experimental group one week before the experiment. The teacher live-streamed the meetings for students to learn at their homes. All students knew about, had seen, or had been exposed to VR, but did not have experience of VR learning and had no prior knowledge of SV-IVR.
3.2. Measurement Scale Instruments (Tests and Questionnaires)
The measurement scale instruments adopted in the present study included the pre-test, post-test, learning attitudes questionnaire, self-regulation questionnaire, self-efficacy questionnaire, and cognitive load questionnaire. For the Likert 5-point scale in this survey, instead of using the “strongly agree to strongly disagree” description, we used a more precise numerical description, using a numerical scale from 1 to 5 to indicate the degree of approval (See
Appendix A).
A pre-test and a post-test were developed by three teachers who had taught landscape architecture education for many years. The pre-test aimed to evaluate students’ prior knowledge. It consisted of 10 multiple-choice items and 10 true/false items, with a perfect score of 100. The post-test was designed to evaluate students’ concepts and knowledge in the landscape architecture history course. The test also included 10 multiple-choice items and 10 true/false items, with a perfect score of 100. In addition, we consulted two experts in landscape architecture history to verify that the tests were adequate to assess students’ learning achievements for the unit.
The learning attitude questionnaire was adapted from the measure developed by Hwang, Yang, and Wang [
43]. It consisted of seven items with a 5-point Likert scale. Examples of questions include: “I would like to learn more and observe more in the course” and “I think learning landscape architecture is interesting and valuable”.
The self-regulation questionnaire with a total of 24 items was modified from Barnard et al. [
44]. It was further divided into six dimensions with a 5-point Likert scale: environment structuring (4 items), goal setting (5 items), task strategies (4 items), time management (3 items), help seeking (4 items), and self-evaluation (4 items). Examples of questions include: “I will summarize what I have learned in the course to see how well I understand the content” and “I will not reduce the quality of my learning because it is a landscape architecture course”.
The self-efficacy questionnaire was revised from Wang and Hwang [
45]. It consisted of eight items with a 5-point Likert scale. Examples of questions include: “I believe I will receive an excellent grade in this class” and “I’m confident I can do an excellent job on the assignments and tests in this course”.
The cognitive load questionnaire was adapted from the measure developed by Hwang et al. [
43]. It consisted of eight items with a 7-point Likert scale and was divided into two dimensions. Examples of questions include: “The learning content in this learning activity was difficult for me” and “During the learning activity, the way of instruction or learning content presentation caused me a lot of mental effort”.
3.3. The Landscape Architecture SV-IVR Learning System
In order to successfully employ SV-IVR in landscape education for students to learn the landscape architecture history curriculum, the current study developed a landscape architecture SV-IVR learning system, as shown in
Figure 1. This study made use of the EduVenture VR platform (
http://vr.ev-cuhk.net/, accessed on 10 May 2021) developed by a university in Hong Kong as the development tool (Jong et al. [
15]). Through this platform, we created a new stage to develop the SV-IVR system; the instructions of the buttons on the platform were easy to follow. The teachers created the learning materials on the EduVenture VR platform on the computer, and then the students could learn through the EduVenture VR app and Google Cardboard. The system consists of three modules: the learning material editing module, the database module, and the SV-IVR learning module. In the editing module, the teacher imported and edited the 360-degree panoramic photos and videos obtained through a panoramic camera (Insta 360) or other sources in advance (i.e., download from UtoVR:
https://www.utovr.com/, accessed on 10 May 2021). According to the teaching activities, the teacher was able to adequately arrange these materials and design the whole learning process through the editing stage in EduVenture VR. In addition, based on the requirements of a landscape architecture history course, the teacher could modify and design materials and set up quiz questions and interactive sessions. The interface of the SV-IVR learning material editing module for the teacher is demonstrated in
Figure 2. Through it, teachers could import teaching audio and questions on panoramic pictures or videos, set interactive sessions, and so on.
As for the database module, it is a web-based cloud database (
http://vrhost.ev-cuhk.net/locale/stage, accessed on 1 April 2021; it requires a password) which contains the data of the four aspects shown in
Figure 1. Students’ personal information, grades, and recorded data were stored in this database. Teachers could check students’ learning performance by accessing the database. To carry out the teaching smoothly, the teacher could check, modify, and understand students’ latest learning status and situation through the system at any time.
In the SV-IVR learning module, students conducted SV-IVR learning with mobile phones and Google Cardboard. They entered the virtual environment according to the course arrangement where they immersed themselves in the landscape design works or architecture and followed the system instructions to learn, answer questions, and complete designated learning tasks (see
Figure 3).
In the landscape architecture SV-IVR learning system, students experienced a systematic cyclical learning process with level design. According to the learning content designed by the teacher, students entered the scenes of each unit and followed the audio instructions (AI text-to-speech conversion by IFLYTEK) from the background to experience and learn in the system. For instance, in the scenes of Roman landscape architecture, students visited Roman architecture and landscapes as if they were standing in front of these buildings and enjoyed the enchantment of the landscape designs with the audio introduction of Roman landscape architecture. After students visited the Roman landscape architecture, within a limited period of time, the system automatically popped up the quiz questions about the landscape design. Students were required to answer the questions according to what they had learned within a limited time. Only when they answered all the questions correctly, they could proceed to the next unit. Otherwise, they had to study the unit again until they obtained the correct answers.
Figure 4 illustrates the interface when students answer the quiz question.
Figure 5 shows the interface for the selection of units. This cyclical learning process (see
Figure 6) can ensure that students learn each unit seriously. Only by carefully listening and visiting, students can receive the corresponding knowledge to answer the questions and enter the next unit to complete the learning.
The system also provided a number of interactive sessions. The teacher promptly proposed thought-provoking questions, and then guided students to answer them during the learning process. Students were then required to record their answers in the virtual environment and upload them to the back-end database, as shown in
Figure 7. Besides, the teacher set up a scavenger hunt during the learning process. When the system introduced the classic landscape designs, it described the appearance of the work, and then guided the students to find it in a hidden location of the virtual scene (see
Figure 8). Through this interaction, it was expected to stimulate students’ proactivity and activeness, enabling them to obtain multi-level feedback and reflections, and guiding them to carry out high-level thinking.
This immersive, interactive, multi-level, and cyclical learning process was expected to enhance students’ deep motive, reduce their surface strategies, and decrease the ineffective learning in the landscape education course. This learning process also aimed to improve upon the one-way knowledge transfer from teachers in the past. In the context of COVID-19, the combination of the problems faced by landscape architecture courses, and the fact that there is little research on SV-IVR in landscape architecture studies, means that this approach can be considered as an innovation. We hope it will be helpful for landscape design education.
3.4. Experimental Procedure
Figure 9 shows the experimental process. A total of 140 students were randomly divided into two groups. At the beginning of the experiment, the students in the two groups were required to take the pre-test and complete the pre-questionnaires of self-regulation, learning attitudes, and self-efficacy, which took 50 min. It is worth noting that students did not know which group they were assigned to before the experiment and, in order to maintain the accuracy of the experiment, we did not tell students that there were two learning styles before the experiment. Furthermore, the two groups were separated so that the students did not know how the class would be conducted until it started.
Moreover, the teacher live-streamed and briefly introduced the course, the SV-IVR learning system, the operational guidance, the precautions, and so on. During the experiment, the experimental group adopted the SV-IVR approach and learned with a Google Cardboard in the landscape architecture SV-IVR learning system. After entering the learning system, the SV-IVR contexts were presented through a smartphone. On the other hand, the control group learned with the conventional technology-supported learning approach which refers to the traditional multimedia approach to class, such as PPT presentations, videos, pictures, etc., presented through a computer screen. A 50-min online instruction session was given by the teacher. After the learning activity, all students took the post-test and completed the questionnaires. The time required for each segment is also shown in
Figure 9.
4. Results
4.1. Test and Questionnaire Results
The reliability of the test and questionnaires were verified using Cronbach’s alpha. The scale for the interpretation of Cronbach’s alpha values, according to George and Mallery [
45] is: >0.9 excellent; >0.8 good; 0.7 acceptable; 0.6 questionable; and >0.5 poor. The reliability of the results obtained in the present research is good according to this scale [
35], as shown in
Table 1.
The Cronbach’s alpha values of the pre-test and post-test were 0.88 and 0.90, respectively, showing acceptable internal consistency (Cortina, 1993) [
42].
The Cronbach’s alpha value of the learning attitude questionnaire was 0.83.
The self-regulation questionnaire’s overall Cronbach’s alpha value was 0.92, and the Cronbach’s alpha values of the six dimensions were 0.95, 0.92, 0.93, 0.87, 0.96, and 0.94, respectively.
The Cronbach’s alpha value of the self-efficacy questionnaire was 0.92, and that of the cognitive load questionnaire was 0.96.
This means that the reliability of all of the questionnaires was good, and we could use them to conduct the survey.
4.2. Analysis of Learning Achievements
To explore the effectiveness of the proposed learning approach on students’ learning achievement, ANCOVA was used to exclude the difference between the prior knowledge of the two groups.
The Shapiro–Wilk test was applied to calculate the normality of the data obtained in the study. The result of this test was 0.97 (p = 0.23), implying that the data exhibited a normal distribution. Moreover, Levene’s test for determining homogeneity of variance was met (F = 3.11, p > 0.05), showing that the assumption is tenable and that there was no significant difference in the variance of the two groups. In addition, the assumption of the homogeneity of regression slopes was verified, indicating that one-way ANCOVA could be performed (F = 0.26, p > 0.05).
Table 2 shows the ANCOVA results of the learning achievements. The adjusted means and standard error were 74.71 and 3.45 for the experimental group, and 65.9 and 3.59 for the control group. According to the result, there was a significant difference between the post-test scores of the two groups (
F = 10.84,
p < 0.05). The experimental group attained significantly higher scores on the post-test than the control group, suggesting that students adopting the SV-IVR learning system had significantly better learning achievements than those adopting the conventional technology-supported learning approach. Moreover, the effect size (
η2) of the learning achievements was 0.62, implying a small to medium effect size (Cohen, 2013) [
46].
4.3. Analysis of Learning Attitudes
In order to examine the effectiveness of the proposed learning approach on students’ learning attitudes, ANCOVA was performed to exclude the difference between the pre-questionnaire of the two groups. Levene’s test of determining homogeneity of variance was met (F = 9.16, p > 0.05), specifying that the assumption is tenable and that there was no significant difference in the variance of the two groups. In addition, the assumption of the homogeneity of regression slopes was verified, showing that one-way ANCOVA could be performed (F = 0.025, p = 0.87 > 0.05).
Table 3 shows the ANCOVA results of the learning attitudes. The adjusted means and standard error were 4.23 and 0.10 for the experimental group, and 3.83 and 0.11 for the control group. According to the result, a significant difference in the post-questionnaire scores was found between the two groups (
F = 6.79,
p < 0.05), indicating that the SV-IVR learning system can significantly enhance students’ learning attitudes compared with the conventional technology-supported learning approach.
4.4. Analysis of Self-Regulation
To investigate the effectiveness of the proposed learning approach on students’ self-regulation, ANCOVA was conducted to exclude the difference between the pre-questionnaire of the two groups.
The Shapiro–Wilk test was performed to calculate the normality of the data obtained in the study. The result of this test was 0.96 (p = 0.32), showing that the data were normally distributed. Besides, Levene’s test of determining homogeneity of variance was met (F = 0.71, p > 0.05), implying that the assumption is tenable and that no significant difference was found in the variance of the two groups. The assumption of the homogeneity of regression slopes was also verified, showing that one-way ANCOVA could be applied (F = 0.56, p > 0.05).
Table 4 shows the ANCOVA results of students’ self-regulation. The adjusted means and standard error were 3.68 and 0.13 for the experimental group, and 3.10 and 0.12 for the control group. According to the result, there was a significant difference in the post-questionnaire scores between the two groups (
F = 9.84,
p < 0.05). The experimental group attained significantly higher scores on the post-questionnaire than the control group, suggesting that the SV-IVR learning system can significantly increase students’ self-regulation compared with the conventional technology-supported learning approach.
The current study further conducted analysis on the six dimensions of the post-questionnaire (i.e., goal setting, environment structuring, task strategies, time management, help seeking, and self-evaluation). The results indicated that the experimental group achieved a significantly higher score than the control group on goal setting (AM = 3.60, SE = 0.08), environment structuring (AM = 3.23, SE = 0.11), task strategies (AM = 3.53, SE = 0.15), time management (AM = 4.20, SE = 0.13), help seeking (AM = 4.06, SE = 0.10), and self-evaluation (AM = 3.89, SE = 0.11).
4.5. Analysis of Self-Efficacy
The Shapiro–Wilk test was adopted to examine the normality of the data obtained in the study. The result of this test was 0.98 (p > 0.05), suggesting that the data exhibited a normal distribution. In addition, Levene’s test of determining homogeneity of variance was met (F = 1.65, p > 0.05), showing that the assumption is tenable and that no significant difference was found in the variance of the two groups. The assumption of the homogeneity of regression slopes was also confirmed, indicating that one-way ANCOVA could be performed (F = 0.26, p > 0.05).
Table 5 shows the ANCOVA results of students’ self-efficacy. The adjusted means and standard error were 3.28 and 0.12 for the experimental group, and 3.27 and 0.11 for the control group. According to the result, no significant difference in the post-questionnaire scores was found between the two groups (
F = 0.01,
p > 0.05), implying that there was no significant difference in the self-efficacy of students adopting the SV-IVR learning system and those adopting the conventional technology-supported learning approach.
4.6. Analysis of Cognitive Load
A
t-test was performed to analyze the cognitive load of students adopting the SV-IVR learning system and those adopting the conventional technology-supported learning approach.
Table 6 shows the
t-test results of students’ cognitive load. The means and standard deviations of the post-questionnaire scores were 3.34 and 1.55 for the experimental group, and 3.28 and 1.25 for the control group. No significant difference was found between the two groups, indicating that there was no difference in students’ cognitive load between the SV-IVR learning system and the conventional technology-supported learning approach.
5. Discussion
The findings indicated that, compared to the conventional technology-supported learning approach, the landscape architecture SV-IVR learning system can result in better learning achievements, learning attitudes, and self-regulation during the pandemic. However, there were no significant differences in the self-efficacy and cognitive load of the two groups.
The present study aimed to integrate the SV-IVR technology into landscape architecture education in the hope of bringing some new changes for landscape education and to keep using SV-IVR, a convenient VR technology, in the field of education. The SV-IVR tool used in this study overcame the problem of unrealistic VR in the past, as well as the high cost and difficulty of designing teaching materials. Furthermore, employing SV-IVR in landscape education is a comparatively new attempt. The results of the current study also indicated that it has a positive influence on the teaching effects in landscape architecture education. In terms of the SV-IVR technology, its effectiveness has been verified again in a new curriculum.
For learning achievements, the study verified the effectiveness of the proposed landscape architecture SV-IVR learning system. Compared to students adopting the conventional technology-supported learning approach, the experimental group achieved higher scores and better learning performance, indicating that the system is effective in landscape education. Based on SV-IVR, the current study developed this immersive, interactive, multi-level and cyclical game-based learning system. With this system, students can ensure the integrity and completion of their learning. This kind of mechanism with its multi-level design can effectively ensure that students can achieve the learning goals pre-determined by the teacher after completing each small unit. If students do not achieve the goal of a unit, they cannot progress to the next unit. Through immersive learning, students can imperceptibly acquire the learning materials and obtain a higher level of experiential learning. Therefore, it enhances students’ comprehension and experience in the course content and inspires their learning motivation, which is in line with Jong et al. [
15].
In addition, for learning attitudes, the results implied that the experimental group had better learning attitudes than the control group. With the innovative SV-IVR in landscape education, it is novel and interesting for most landscape architecture students. Thus, it can stimulate students’ learning proactivity and motivation well, which improves their learning attitudes. In particular, compared to the boring one-way method of knowledge transfer, this novel VR approach can undoubtedly bring new experience and perception to students, and provide greater potential for teaching and learning in landscape education. The findings are also in line with those of Chen, Chai, Jong, and Jiang [
47] and Chen, Hung, and Yeh [
48].
Furthermore, for self-regulation, students adopting the landscape architecture SV-IVR learning system showed better self-regulation than those adopting the conventional technology-supported learning approach. During the pandemic, due to the different learning environments, some of the students have decreased their concentration levels and cannot pay attention to their learning. This problem has been improved through the SV-IVR learning system. As this system needs students to immerse themselves in the virtual world to carry out the learning process, with the related setting, it can help them carry out learning tasks well and control their self-regulation, which is in accordance with Chen and Hsu [
49]. According to the findings, with this learning system, there were significant differences in the six dimensions of the self-regulation questionnaire between the two groups, namely goal setting, environment structuring, task strategies, time management, help seeking, and self-evaluation. This also shows that students’ abilities in these six dimensions have been improved by learning in the landscape architecture SV-IVR learning system. This system helps students carry out their learning process more consciously, improve their learning behavior, enhance their self-observation, realize their learning objectives, and conduct higher-level thinking and deeper learning. Students can then take more control of their own learning, which is in line with Chen and Hsu [
49] and Zimmerman, Schunk, and DiBenedetto [
50].
According to the interviews with the students in the experimental group after completion of the experiment, some of them said that they had a clear goal of passing the quiz and moving on to the next unit when they were using the SV-IVR learning system. However, some students said that such goal setting was done with the system rather than as a spontaneous behavior. In terms of the environment, some students said that learning through SV-IVR would make them more determined to find a quiet learning environment to study in, but some students said that there was no difference. These results deserve further investigation and analysis to improve the accuracy of the data. However, in general, there are some differences between the two groups of students.
Lastly, no significant differences in the self-efficacy and cognitive load of the two groups could be found in this study. In terms of self-efficacy, one possible explanation is that a short period of studying may not increase students’ self-efficacy. Huang et al. [
6] and Yang et al. [
29] also implied that students need more time and learning experience to develop their capabilities. All students in the experimental group were exposed to this learning system for the first time, so a 50 min learning experience may not have been sufficient to improve their self-efficacy, and self-efficacy needs to be cultivated. Especially for emerging technology, there is still a great deal of uncertainty; hence, students may spend more time developing their self-efficacy. Such results were also found in Huang’s study [
8] which showed that the change in students’ science self-efficacy was not significant after the VR learning activity. A prior study [
51] also found insignificant self-efficacy results, which they believe may be due to environmental and technological reasons. This gives future researchers new ideas that may take longer to explore, for example, by performing an extended experiment to develop students’ self-efficacy. Meanwhile, due to the limitations of the experimental method, measurement method, and experiment time, the measurement of self-efficacy may not be as accurate. The relevance of the two learning approaches may be limited, and due to cost constraints, this aspect will be further explored in the future.
On the other hand, there was no significant difference in students’ cognitive load, which means that students using the SV-IVR learning system and students using the conventional technology-supported learning approach had the same perception of cognitive load. In other words, students who learned with SV-IVR did not reduce or increase their cognitive load. The results of this study are consistent with the research of many scholars [
38,
52] and suggest that using the SV-IVR learning system in design education is not troublesome for students and that students can control this system well. Moreover, the experimental group will not have an extra cognitive load when using this system to learn the landscape architecture course. Consequently, this learning system can be adopted by landscape architecture students, which is consistent with Huang and Liaw [
52]. Furthermore, according to the students’ interview records, most of the students from the two groups expressed similar perspectives and perceptions of the system regarding their cognitive load. This was also in line with a prior study [
38] that integrated the two-tier test strategy into VR.
However, we also note that many scholars’ VR learning studies have shown changes in cognitive load. For example, a study [
53] argued that students with certain learning style preferences must bear a greater cognitive load in VR learning in order to achieve the same learning outcomes as other students. Moreover, in another study [
26], students who used VR for library guide learning activities developed higher cognitive load and therefore perceived students would work harder. Due to the limitation of the experimental method, measurement method, and experimental time, the cognitive load data in this study may have defects, which will be further investigated in future studies.
6. Conclusions
In summary, the major contribution of this study is that, different from previous VR learning, this study selected a more accessible, and more convenient SV-IVR, and then developed a landscape architecture SV-IVR learning system to help the teaching and learning in landscape architecture education. In the field of landscape architecture education, the combination with SV-IVR is still a research gap, so this study also calls for more researchers to examine the relationship between SV-IVR and landscape architecture education in the future. In addition, with the sudden COVID-19 pandemic, campuses are closed, which also inspires the future development of education. Educational technology is a development trend. During the pandemic, there are several problems in landscape education such as students’ decreased attention, poor learning effects, low motivation, difficulty in grading, and lack of contexts required for the courses. The learning system developed in this study can solve the abovementioned problems well, and provide effective assistance to landscape architecture education during the pandemic.
In addition, the interactivity of the learning system also helps students learn actively. In the past, VR learning only focused on presenting students with views, but interactive sessions were added to this SV-IVR learning system. Students can candidly control the angle and content they want to see in the system. At the same time, according to the thought-provoking questions raised by the teacher, students can record their answers and upload them to the database. The system also guides students to learn in the learning process with audio instructions from the background and provides feedback and guidance according to the different choices of the students. The setting of some interactive sessions similar to a scavenger hunt also adds a great deal of fun to the learning system; students can also leave messages and feedback in the system, and they can answer questions and take quizzes and an examination, which greatly reduces the possibility of cheating. Teachers can easily grasp students’ learning situation from the back-end database. It solves the online education problem of the difficulty of assessing students’ learning situation during the COVID-19 pandemic. As a result, this kind of interaction can effectively improve students’ proactivity and enable them to learn actively, making them actively think, experience, and imagine, gain deeper insights and inspirations, carry out high-level thinking, and ultimately enhance their learning performance. For landscape architecture students, the proposed SV-IVR learning system in this study is feasible and effective. The fact that SV-IVR technology increases students’ learning performance has also been proved by many researchers [
27,
29].
However, the limitations of the present study need to be noted. First of all, future studies can consider conducting a long-term experiment. For example, researchers can conduct the experiment for a semester and verify the sustainability of the experimental results. The stability and sustainability of the research findings are the issues that we need to further explore. Second, it is worth trying to apply SV-IVR to more courses in landscape education. Future studies can conduct teaching experiments to confirm its effects on different courses. Besides, researchers can consider expanding the sample size of the experiment to involve more students and further improve the accuracy of the experimental results. Third, the selection of equipment may influence the experimental results. In this study, some students mentioned that cardboard goggles will be easily damaged. As a result, the effects of different equipment can be considered in future studies. Moreover, novelty is also an issue worthy of our attention. Since the novelty brought by new technologies may affect students’ real perceptions, future research should consider how to overcome this problem. Last, factors such as different learning styles, different characteristics, academic performance, and gender can also be taken into account to further expand the scope and depth of research.