The augmented reality learning model based onsituational reflection mechanism is beneficial tostudentsScientific inquiry learning performance, Behavior pattern and Higher order thinking Influence
（Effects of a contextualised reflective mechanism-based augmented reality learning model on students’ scientific inquiry learning performances, behavioural patterns, and higher order thinking）
2022
1. Concept analysis
1. AR learning based on contextualized reflection mechanismModel:This study proposes a contextualized reflection mechanism to promote students’ learning in AR science Explore reflective thinking in learning situations, as well as students’ learning performance and higher-order thinking tendencies. In this model, the learner's inquiry learning process is cycled through a contextualized reflection mechanism, which is divided into four stages: conceptual understanding, reflective cognition, in-depth exploration and knowledge construction (Figure 1). The model is designed to promote learners' learning performance, interaction, and propensity to engage in higher-order thinking. Researchers use AR to provide students with an immersive experience and promote their deep thinking, thereby promoting understanding of scientific concepts during the learning process (Lin et al, 2020). Elaish et al. (2019) showed that using AR technology can create virtual AR learning scenes and provide a more convenient contextualized teaching environment. The contextualized AR mechanism can provide a knowledge construction process related to the learning context and help to deeply understand the nature of scientific knowledge (Lin et al, 2022). Therefore, contextualized AR mechanisms can be embedded in scientific inquiry to solve the difficulties of scientific inquiry due to lack of motivation and separation of the inquiry environment from the real world. By combining digital and interactive learning content with real-world environments, AR environments can be a powerful way to help students understand complex phenomena in scientific inquiry. In this study, we propose reflective scaffolding as sufficient scaffolding to enable students to actively engage in reflective thinking during scientific inquiry.
2. Scientific inquiry:Scientific inquiry refers to an active process that supports students’ understanding of the nature of science (i.e., understanding the value of science) and develops them as scientists ofscientific process skills(Ekici & Erdem, 2020;Wu & Wu, 2020). Scientific process skills refer to the skills of using scientific methods to design experiments, investigate, collect, and analyze science-related data (Sandoval, 2005). The understanding of the nature of science is to understand the nature of science from the perspective of scientific characteristics, such as its provisionality, creativity and experience, as well as the role of the scientific community. Scientific inquiry activities emphasize student-centered scientific learning activities (Lin et al., 2020). After experiencing these related scientific inquiry activities, learners constructed knowledge concepts related to science, and then established an understanding of the nature of scientific knowledge (Peten, 2021). These studies demonstrate that students benefit greatly from scientific inquiry, which is fundamental to the process of scientific knowledge generation and discovery. Such a process also reflects the common nature of scientific inquiry, which is focused on helping students develop evidence-driven reasons to support their scientific claims.
3. Behavior pattern: Behavior pattern is the structure, content and regular behavior series of people's daily activities with motivation, goals and characteristics. It is the stereotype of behavioral content and methods, the "externalization" of life values, and expresses people's action characteristics and behavioral logic. Observed from the perspective of time, a certain behavioral pattern is the program structure of activity time allocation; viewed from the perspective of space, it is the distribution of location and scope of activities. The specific type of a person's behavior pattern is restricted by external environmental conditions, the role played by the person himself, and life values.
4. Higher-order thinking:Higher-order thinking refers to mental activities or cognitive abilities that occur at higher cognitive levels. In the classification of teaching objectives, they are analysis, synthesis, evaluation and creation. Higher-order thinking is the core of higher-order abilities, mainly referring to innovation ability, problem-solving ability, decision-making ability and critical thinking ability. High-order thinking ability embodies the new requirements for talent quality in the knowledge era and is a key ability to adapt to the development of the knowledge era.
5. Reflective thinking promotes scaffolding: can be defined as a reasoning process that has the potential to help learners gain new understanding in the process of solving problems and gain active learning experience (Chen et al, 2019). Reflective scaffolding can enhance students' in-depth understanding of concepts and help them overcome learning difficulties by actively examining their exploration process, thereby helping them make progress (Chen et al., 2020; Hsia & Hwang, 2020)
2. Research content
1. Research background:Although AR environments provide opportunities to promote inquiry-based learning among students, for most students, completing science without appropriate support Exploration tasks remain a challenge. Research evidence demonstrates the potential of using reflective scaffolding to engage students in developing deep conceptual understanding and construction when applying scientific inquiry. Therefore, we designed an AR learning model based on a contextualized reflection mechanism to help students complete scientific inquiry tasks.
2. Research purpose:They did not consider reflective scaffolding among the factors perceived to be important in the relationship between technology-assisted teacher support and technology-embedded (e.g., AR and mobile) science inquiry As a key mechanism, its impact on AR-based scientific inquiry has also not been tested. Contrary to their findings, this study focuses on studying the effect of using reflection mechanisms to cope with AR-based scientific inquiry challenges and provides some suggestions for improving students' scientific inquiry effects. Furthermore, unlike Bidarra and Rusman (2017), who only considered AR and reflection as a possible technology and pedagogy among various options without practical application, this study incorporated reflective activities into AR-based activities, To further explore the impact of augmented reality learning models based on contextualized reflection mechanisms on promoting more complex levels of inquiry.
In order to test students' scientific inquiry learning outcomes, it is important to measure students' learning performance, behavioral patterns, and perceptions of higher-order thinking tendencies.
3. Researchmethod:Using quasi-experimental and lagged sequence analysis methods, 81 people The scientific inquiry learning performance, higher-order thinking and behavioral patterns of sixth grade students were studied.
4. Research questions:(1) Whether the scientific inquiry learning method based on contextualized reflection mechanism can improve students'learning performance a>?
(2) Is the cognitive load of students learned through the scientific inquiry learning method based on the contextualized reflection mechanism lower than that of traditional AR science? Students who learn through inquiry learning methods?
(3) Does the scientific inquiry learning method based on the contextualized reflection mechanism promote students’higher-order thinking tendencies?
(4) What are the differences in behavior patterns between the scientific inquiry learning method based on the situational reflection mechanism and the traditional AR scientific inquiry learning method? ?
5. Research results:The experimental results show that this method improves students' inquirylearning performance and higher-order thinking tendencies (Problem-solving tendencies and metacognitive awareness).
In addition, evidence from this study also shows that students who learned using this approach displayed more observation, comparison, exploration, and reflective behavioral patterns during fieldwork than students without contextualized reflection mechanisms. This means that this method can provide a good reference for the teaching design of future AR-based scientific inquiry activities.
3. Experimental design
1.Experimental subjects:This study recruited 81 students from two classes in a primary school in the south Sixth grade students.
None of the participants had any previous experience with AR technology. One class was randomly selected as the experimental group, using the proposed method, and the other class was selected as the control group, using the traditional AR-based query method.
There were 41 students in the experimental group, with an average age of 11.25 years and a standard deviation of 1.60.
There were 40 students in the control group, with an average age of 11.70 years and a standard deviation of 0.98.
Students from both groups were randomly divided into several groups, each group consisting of 4 to 5 students. To ensure that all participants had the same AR learning experience, they were required to complete an AR-based science course in order to qualify for a series of more advanced courses. Therefore, all participants had previous exposure to AR learning projects that used mobile devices for teaching and learning (i.e., projects took up more than one-third of class time).
2. Measurement tools:This study uses knowledge test performance and questionnaire surveys to examine students' learning performance. The pretest and posttest contained questions validated by three scientific experts.
The pre-test and post-testinclude 10 multiple-choice questions (5 single-choice questions and 5 double-choice questions).
The pretest is designed to measure students’ prior knowledge of conceptual understanding and common mistakes in life. Basically, these students had no prior scientific concepts or related knowledge. The post-test is designed to evaluate students' final knowledge learning level and includes three subscales: knowledge conversion, error correction, and life application. The difficulty analysis results show that the average difficulty average of the two groups is 0.86. The sample questions that appear in each test are as follows: (1) Pre-test sample question: What is a flower with both stamens and pistils called? (2) Post-test question: What is the reason why the light green sepals of newly opened lilies disappear?
Higher-Order Thinking Tendency QuestionnaireThis questionnaire is modeled on the original Student Higher-Order Thinking Tendency Scale (Lai & Hwang, 2014), including Complex Problem Solving Tendency (6 items), Three subscales: metacognitive awareness (10 items) and creative tendencies (6 items). The questionnaire included 22 items, rated on a 5-point Likert scale.
Cognitive Load QuestionnaireIn order to further explore whether the introduction of AR learning resources in classroom teaching activities will affect students’ cognitive load, Wang et al. developed the Cognitive Load Questionnaire
(2018) and Chang et al. (2018) were modeled in this study. The questionnaire includes 10 items scored on a 5-point Likert scale. This study aims to measure students' cognitive load after participating in AR scientific inquiry learning activities from two dimensions: psychological load (5 items) and mental effort (5 items). The pretest of cognitive load refers to the load experienced by learners in the first week when learning basic science knowledge in AR-based instruction.
Based on AR-type systems and AR features, a coding scheme for AR-based scientific inquiry learning behavior patterns was developed. Hwang et al. (2018) modified the research results of Lin et al. (2013), and this study adopted the research results of Lin et al. (2013) to obtain a reasonable coding scheme for behavioral pattern analysis. Lag sequence analysis (LSA), a method used to understand learners' behavioral patterns in AR environments (Cheng & Tsai, 2013; Lin et al., 2013; Wang et al., 2014) conducted a Study to elucidate changes in student behavior after incorporating contextualized reflection mechanisms into AR-based scientific inquiry. In order to evaluate the effectiveness of the AR learning model and LSA based on the contextualized reflection mechanism, we classified all students' learning behaviors recorded in the archives according to the coding scheme. Based on this coding scheme, the final behavioral pattern coding scheme is obtained. The experimental group and the control group adopted the same AR-based scientific inquiry learning behavior coding scheme, learning behavior code and reflection process.
3. ExperimentProcess: Figure 1 illustrates the experimental process. In the first week, all students received AR-based teaching training before pre-prediction to use AR resources to familiarize themselves with AR equipment and AR learning resources needed in subsequent classes. During the second week, students had to complete a pre-test and pre-questionnaire. All students are taught by the same teacher, who is mentored by a science expert teacher with 16 years of teaching experience and receive skills training in using AR for science learning (over 80 hours in 20 lessons).
Figure 1. Experimental process
4. Experimental tools: Development of AR scientific inquiry learning environment based on contextualized reflection mechanism
This study proposes a contextualized reflection mechanism to promote students' reflective thinking in AR scientific inquiry learning situations, as well as students' learning performance and higher-order thinking tendencies. In this model, the learner's inquiry learning process is cycled through a contextualized reflection mechanism, which is divided into four stages: conceptual understanding, reflective cognition, in-depth exploration and knowledge construction (Figure 2). The model is designed to promote learners' learning performance, interaction, and propensity to engage in higher-order thinking.
Figure 2. Learning cycle in research-based learning based on contextualized reflection mechanism
Figure 3. Structure of an “AR-like” system
4. Results
(1) Learning performance
Levene's equality of variances test was conducted on the academic performance, and no violation was found (F = 2.54, p > 0.05). In addition, the homogeneity of regression slopes between groups was confirmed, indicating that it is appropriate to use analysis of covariance F = 2.36 (p > 0.05). Therefore, a one-way analysis of covariance (ANCOVA) was performed on the posttest using the pretest scores as the covariance.
Learning performance raw data:
Analysis of covariance for student posttest:
Table 1.The impact of ANCOVA on students’ academic performance
| Group |
N |
Mean |
SD |
After adjustment M |
standard error |
F |
h2 |
| test group |
41 |
90.682 |
2.382 |
90.683 |
0.425 |
245.758 |
0.757 |
| control group |
40 |
81.200 |
3.031 |
81.200 |
0.430 |
*p < 0.05.
Results: The ANCOVA results are shown in Table 1. The adjusted mean of the post-test scores of the experimental group was 90.683 and that of the control group was 81.200. In other words, the students in the experimental group performed significantly better than the students in the control group (F = 245.758, p < 0.05, η2 = 0.757). This means that students who use the AR scientific inquiry learning method based on the contextualized reflection mechanism show significantly better academic performance than students who use the traditional AR scientific inquiry learning method.
(2) Cognitive load
Before ANCOVA, the homogeneity of variance assumptions and the homogeneity of regression coefficients were tested. Levene's test for equality of variances is not significant (F = 0.36, p > 0.05).
Therefore, the assumption of homogeneity of variances is not violated. In the results of the intra-group homogeneity test, the homogeneity of the regression coefficients is not significant, indicating that excluding the influence of the pre-test data, ANCOVA analysis can be continued to compare the post-test data of the two groups.
Cognitive load posttest raw data:
Analysis of covariance for cognitive load posttest:
Table2. ANCOVA results of cognitive load
| Group |
N |
Mean |
SD |
After adjustment M |
standard error |
F |
h2 |
| test group |
41 |
3.650 |
1.33 |
3.650 |
0.202 |
0.235 |
0.003 |
| control group |
40 |
3.512 |
1.23 |
3.512 |
0.202 |
Results: It can be seen from Table 2 that the corrected mean and standard error of the experimental group are 3.650 and 0.202 respectively, and the corrected mean and standard error of the control group are 3.512 and 0.202 respectively. . There was no statistically significant difference between the two groups (F = 0.235, p > 0.05, η2= 0.003).
(3) Higher-order thinking tendencies
Higher-order thinking raw data:
Higher-order thinking covariance analysis:
Table3. ANCOVA results of higher-order thinking
| Group |
N |
Mean |
SD |
After adjustment M |
standard error |
F |
h2 |
| test group |
41 |
4.275 |
1.301 |
4.275 |
0.197 |
4.564 |
0.055 |
| control group |
40 |
3.682 |
1.192 |
3.683 |
0.195 |
Results:Table 3 lists the ANCOVA results of higher-order thinking tendencies. There was no significant difference in Levene's test for equality of variances and regression homogeneity test (i.e. F = 4.09, p > 0.05; F = 3.59, p > 0.05). Assuming homogeneity of the regressions suggests that ANCOVA analysis can be performed. In the pre-test, the differences in variables between the experimental group and the control group were not significant. The corrected mean and standard error of the experimental group were 4.275 and 0.197 respectively, and the corrected mean and standard error of the control group were 3.682 and 0.195 respectively. The results showed that the post-test scores of the experimental group were higher than those of the control group (p < 0.05). The effect size (η2) of higher-order thinking tendency is 0.055, which is a medium effect size.
(4) Learning behavior patterns
Using the coding scheme summarized in Table 1, we coded the teaching process according to the principle of recording a behavioral pattern change. To study the behavioral patterns of the two groups, a series of lagged sequence analyzes (LSAs) were performed. As shown in Figure 4, LSAs revealed several important behavioral patterns of students E2, E3, E4 and C3 in the experimental group, which explained the experiment group of students’ learning process. Under the AR scientific inquiry learning method based on the contextualized reflection mechanism, students tend to observe key phenomena or characteristics in the field, which will lead to two main behavior patterns: comparing characteristics or explaining phenomena (E2→E3 and E2→ E4). Likewise, students can compare characteristics and then explore problem inquiry tasks (E3→E4), explain phenomena (E3→E5), think and reflect (E3→C3), or ask questions for learning help, which can trigger discussions with peers. discussion.
Compared with the behavioral pattern transition diagram of the experimental group, the behavioral pattern transition diagram of the control group shows that the main behavioral patterns include answering tests in the system and reading test feedback(Fig. 5). Students in the control group can learn through a linear process, including reading the learning content, subsequently taking notes in the system, and observing on the spot, which may lead to explanations (R1↔R2→E2→E5). They will then focus on the test module and ask questions if needed (Q1↔Q2→C1→C2↔C3). The Thinking and Reflection Behavior Pattern (C3) is located at the end of the Behavior Pattern Transition Chart, indicating that students in the control group tended to view thinking and reflection as a task rather than a method that facilitates their inquiry and discussion.
Figure 4. Behavioral pattern transition diagram of the experimental group.
Figure 5. Behavioral pattern transition diagram of the control group.
５、结论
The main contribution of this study is that it combines the scientific inquiry learning method based on the contextualized reflection mechanism with AR technology and allows students to interact with it. Experimental results show that compared with conventional AR technology
Compared with the inquiry learning method, this method improves students' inquiry learning performance and higher-order thinking tendencies (i.e., complex problem-solving tendencies, metacognitive awareness and creativity tendencies), and reduces cognitive load. Compared with traditional classroom science education, the method proposed in this article provides the following two key reflection stages for developing well-designed AR-based scientific inquiry activities. First, combining augmented reality with guidance on non-invasive contextualized reflection can help students face difficulties understanding abstract scientific concepts by connecting 3D supplementary materials to real-world tasks, rather than just imagining them with 2D materials. . To effectively apply this approach, teachers need to provide non-intrusive guidance and prompts, as well as AR-based 3D science models, to enable students to identify inquiry gaps and explore answers on their own.
Second, the AR scientific inquiry learning method based on the contextualized reflection mechanism can help students deeply participate in scientific inquiry activities with a positive attitude by guiding students to participate in reflective cognition. Therefore, teachers should not only encourage students to interact and develop their learning knowledge, but also let students learn how to identify and analyze problems, how to manage their own learning plans, and how to reflect during the scientific inquiry process. A deeper understanding of the positive cycle formed by AR scientific inquiry learning based on the contextualized reflection mechanism can help students think deeply, discover their own knowledge gaps, and think proactively to find appropriate solutions and overcome challenges in scientific inquiry practice.
This study may have some limitations. First, a self-report questionnaire was used to investigate students' metacognitive awareness and problem-solving tendencies, which refers to students' tendencies and concepts of these two factors rather than their abilities. Second, it is necessary to follow the proposed method over a longer period of time and conduct relevant experiments to elucidate the impact of the method on higher-order skills such as creativity and critical thinking, which are more difficult to acquire in scientific inquiry.
https://www.tandfonline.com/doi/abs/10.1080/10494820.2022.2057546