Redesigning Building Thermal Science Education Through Inquiry-Based Experiential Learning

Abstract

Mastering building thermal science is essential for architectural professionals, as it supports the design of energy-efficient and thermally optimized buildings, which are critical for addressing the growing demands of sustainable architecture. However, traditional teaching methods often disconnect theoretical instruction from practical application, limiting students’ ability to apply core concepts in real-world scenarios. This study introduces a pedagogical reform that integrates design-oriented and inquiry-based experiments, hands-on physical activities, and field-based testing into the teaching of building thermal science. The revised curriculum focuses on applying theoretical principles in real architectural contexts, allowing students to directly design and experience thermal phenomena such as heat transfer and thermal resistance in building envelope structures. To evaluate the effectiveness of this reform, a control group using traditional confirmatory experiments (following predetermined instructions to complete experiments and validate the results) was compared to a reform group engaged in inquiry-based experimental learning. Over the course of three cohorts (2019, 2020, 2021), the reform group consistently outperformed the control group, with statistically significant improvements in average course grades. Specifically, the reform group had mean grade differences of 7.21 points higher in 2019, 4.55 points higher in 2020, and 5.83 points higher in 2021, as demonstrated by t-test results (p < 0.05). The reform group also exhibited more concentrated grade distributions, reflecting enhanced comprehension and retention of key thermal concepts. In addition to improved academic performance, students in the reform group demonstrated superior problem-solving abilities and a heightened awareness of energy conservation and sustainable design practices. This approach not only deepened their understanding of theoretical knowledge but also fostered a greater commitment to integrating sustainability into their architectural projects.

1. Introduction

Building thermal science is a required course for undergraduate students majoring in architecture. The purpose of this course is to help students understand the basic theories and technologies of thermal science and to apply the relevant knowledge flexibly in architectural design. This lays a solid foundation for improving the thermal environment quality of their design works in the future while also reducing building energy consumption [1,2]. The thermal environment of a building is an important component of its physical environment and is also a key criterion in China’s “Green Building Evaluation Standard” under the “Health and Comfort” category [3]. The quality of the building’s thermal environment not only affects the occupants’ health, thermal comfort, mood state, and efficiency in work and study but also has a significant impact on the building’s energy consumption [4,5,6,7,8,9]. With the development of the economy and the improvement of people’s living standards, increasing attention is paid to the comfort and health of the building’s physical environment. Thus, the thermal environment of buildings, along with energy conservation, has become one of the research focuses in architectural design and construction [3,6,10]. Therefore, as future architects, undergraduate students majoring in architecture need to master and flexibly apply the relevant knowledge of building thermal science in order to design buildings that are not only comfortable and healthy but also energy-efficient and environmentally friendly [11,12].

Experimental teaching is an important part of the building thermal science course. The main goal of the thermal science experiments is to allow students to personally operate equipment to test various aspects of the building’s thermal environment, thereby helping them to understand the relationship between the thermal environment and the building’s form, space, and structure. This deepens the integration of thermal science knowledge with architectural design concepts and techniques and encourages students to consciously apply it in their design projects [13]. However, in actual teaching, for various reasons, the experimental teaching of building thermal science has not played its role.

2. Current Status Analysis of Building Thermal Science Experimental Teaching

The undergraduate teaching of architecture in Chinese universities mainly centers around architectural design, with a focus on cultivating students’ visual thinking abilities. However, the teaching of technical courses related to architecture, such as building thermal science, is relatively weak [14,15]. The teaching of building thermal science experiments usually takes place after the completion of theoretical courses. The instructor first explains the experimental principles, procedures, and precautions, and then each group of students follows the steps to mechanically perform the validation experiments. Most of these activities were conducted with the primary goal of completing the task. As a result, students passively accept the experimental methods and results, making it difficult for them to independently think about and explore the underlying principles and methods of the experiments [16]. Throughout the experiment, students lack initiative and enthusiasm. Such experimental teaching fails to enhance students’ understanding and mastery of the relevant theories, nor does it improve their ability to apply theoretical knowledge to solve real-world problems. Moreover, it does not cultivate students’ awareness of energy conservation and environmental issues. Consequently, experimental teaching does not contribute to improving the overall effectiveness of course instruction as it should.

Furthermore, compared to architectural design courses, technical courses like building thermal science are often seen as relatively dull, with an abundance of theoretical derivations and calculations that increase the difficulty of learning. This creates a sense of fear toward the course among architecture students, leading to a lack of interest in technically focused courses taught using traditional methods. Even though students are aware that building thermal science is one of the core courses in green building technology—a trend that is inevitable in architectural development—and understand the importance of learning and applying theoretical knowledge to professional growth, they struggle to appreciate the practical value of the theory when they have not experienced it in application. As a result, they fail to grasp the deeper meanings of the related physical formulas, find it difficult to understand the role of construction measures in building insulation and thermal resistance, and, thus, show little interest in learning thermal science theories or summarizing and reflecting on experimental results. This ultimately affects the overall effectiveness of the building thermal science course.

To improve the teaching effectiveness of building thermal science courses, foster students’ interest in learning architectural theoretical courses, and enhance their ability to apply knowledge to solve problems, many scholars have made various efforts and attempts at teaching reform, achieving some success. For instance, Zhang from the South China University of Technology divided the basic principles of building thermal science into four parts: conduction, convection, radiation, and (non) steady-state heat transfer. He employed a teaching method that integrates “basic theory + engineering technology + design methods + application examples” for each section [1]. This approach effectively combines theory and practice, helping to improve students’ problem-solving abilities. However, it still falls short in encouraging independent thinking and exploratory learning among students, and the specific teaching outcomes were not further discussed. Professor Yuchuan Chen from Guizhou University adopted an experiential teaching model for building physics experiments [17], integrating experiments into the actual architectural environment. This allowed students to directly experience the real building environment, thereby igniting their enthusiasm for learning architectural theoretical courses and enhancing their analytical skills. However, the outcome of such teaching reform was only qualitatively introduced, stating that it significantly improved students’ scientific innovation abilities and social practice skills, leading to favorable experimental teaching outcomes. Wang from Guangxi Arts University set up building physics experiments focused on architectural design [18], combining qualitative and quantitative analyses during the experiment process to cultivate students’ perceptual and rational thinking. This method helps in applying building physics knowledge to architectural design, though the specific quantitative analysis of teaching outcomes was not provided. To deepen students’ understanding of building thermal science concepts and theories, enhance their interest and sense of responsibility in learning building physics, and foster practical skills, collaborative work abilities, and innovation, Guo, Zhu, and others from the Dalian University of Technology have reformed building thermal science experiments by combining thermal science experiments with architectural design [19]. This teaching reform has been implemented for three years, with students generally responding positively.

Based on the above analysis, the main reasons for the unsatisfactory teaching outcomes of “building thermal science” and other architectural theoretical courses are that, compared to architectural design, the content appears dull and challenging. Additionally, traditional experimental teaching has not effectively contributed to the overall teaching of these courses. The unsatisfactory outcomes of experimental teaching are largely due to the failure of traditional experimental methods to incorporate the principles of linking theory with practice and encouraging inquiry-based learning. The principle of linking theory with practice is a crucial principle in human cognition and learning, and it should be an essential teaching principle as well. The ancient Greek Sophists believed that theory without practice and practice without theory are both meaningless. Pestalozzi also emphasized the importance of “knowledge and its application”, pointing out that “knowing and doing are so closely linked that if one ceases, the other ceases as well”. Ushinsky also noted that “hollow, baseless theories are of no use at all. Theory cannot be separated from practice, and facts cannot be separated from thought”. The principle of inquiry-based learning emphasizes that students are the main agents of learning, and mastering knowledge ultimately depends on their observation, thinking, and operation. Instructors should not and cannot take over this process. If an instructor hastily provides the answers to students or completes tasks on their behalf, it can lead to students’ dependency, affecting their ability to think independently and complete tasks. In more severe cases, it may cause them to lose confidence in learning, making them fearful of difficulties, lazy, or less competent. Inquiry-based learning requires that the teaching process stimulates students’ initiative, encourages hands-on participation, and develops their ability to solve problems independently [20]. Motivation arises from needs and is the direct cause and internal driving force of behavior. All human behavior and activities are driven by motivation, including students’ learning behavior, which is also governed by motivation. The aspiration and level of learning motivation directly affect students’ learning behavior and outcomes [21]. American educational psychologist D.P. Ausubel pointed out that learning motivation can improve learning outcomes, and the knowledge acquired by students can further enhance their learning motivation [22]. The underlying philosophy of the reform is that the experimental process should be a dynamic experience where students apply knowledge to solve real-world problems, fostering a stronger connection between theory and practice. Traditional confirmatory experiments often fall short in this regard, as key steps—such as explaining experimental principles, methods, and even results—are typically led by the instructor. As a result, students follow predetermined steps passively, limiting their ability to independently apply theoretical concepts or engage in critical problem-solving. This passive approach not only fails to stimulate students’ interest in theoretical knowledge but also diminishes their initiative and enthusiasm for learning.

To address these limitations, the reformed experimental course in building thermal science was introduced, aiming to immerse students in hands-on learning experiences that directly link fundamental thermal science principles with practical architectural design. Researchers from Penn State Architectural Engineering have been actively involved in the design and development of this teaching reform, contributing to both the planning and implementation stages. The goal was to create an inquiry-based experimental learning process that comprehensively covers key knowledge areas while encouraging students to design experiments and analyze and solve real-world problems using the theoretical frameworks they have learned. This approach was designed to increase student motivation, deepen their understanding of thermal science, and enhance the overall effectiveness of theoretical instruction. The teaching reform covered all major topics in the thermal science course, while we present the central topics related to building envelopes and indoor thermal environment in this paper. Quantitative and qualitative analyses were conducted to assess the outcomes of the reform. By comparing the final grades of students who participated in the reformed experimental teaching with those who followed the traditional methods, the results clearly demonstrated the benefits of this new pedagogical approach. The reformed course not only improved academic performance but also fostered greater problem-solving skills and a deeper engagement with the material, proving to be a more effective method for integrating theory with practice.

3. Methods for Experimental Teaching Reform

Given the issues present in traditional experimental teaching of building thermal science, many teaching teams have undertaken various reforms. However, no teaching team has yet conducted a quantitative analysis of the effectiveness of their reforms. Therefore, it is still necessary to keep reforming the experimental teaching of building thermal science. The aim is to change students’ passive attitudes towards building thermal science and related experimental courses, enhance their understanding, mastery, and application of relevant theoretical knowledge, stimulate their learning motivation, and increase their interest and enthusiasm in studying building thermal science and other theoretical courses. Furthermore, a quantitative analysis of the effectiveness of the teaching reform will be conducted to identify the strengths and weaknesses of the experimental teaching reform, allowing for further optimization of the teaching approach.

3.1. Design of the Experimental Teaching Process

To change the current state of building thermal science experimental teaching and to implement the principles of linking theory with practice and inquiry-based learning, the school (Liaoning University of Science and Technology) established an experimental teaching reform group within the School of Architecture and Art Design. This group consists of four members, including the main lecturers of the building thermal science course and dedicated experimental teachers. The university is located in a cold climatic zone defined by the Code for Design of Civil Buildings (GB50009) in China with a long heating season lasting up to five months. It falls within the Dwa classification of the Köppen–Geiger climate classification system. The building thermal science course is offered in the autumn semester, coinciding with the heating season when there is a significant temperature difference between indoor and outdoor environments. Furthermore, our thermal science laboratory (consisting of a few full-scale mockups) has configurable features on wall and window components and intentionally left some exterior walls uninsulated to facilitate experiments related to the thermal environment. During the heating season, there is a noticeable difference in the interior surface temperature of walls with and without insulation, providing excellent conditions for on-site detection experiments of the heat transfer coefficient of building envelope structures. Based on the analysis of the current state of building thermal science experimental teaching and years of teaching experience, combined with the experimental conditions of our laboratory, the reform group decided to change the traditional validation experiment of on-site detection of the heat transfer coefficient of building envelope structures into a design experiment that can cover and integrate the main knowledge points of building thermal science.

The detailed course outline and associated experimental teaching details (including both reform and traditional experiments) are shown in

Table 1. Experimental workflow before and after teaching reform.

Week #	Topics	Experiments
		Reform	Traditional
1	Outdoor Weather and Thermal Environments	Outdoor weather data collection and analysis
		Pose a question such as, “Are there any relationships or patterns among typical weather parameters?” Students may leverage their understanding of psychrometric charts and climate zone features to collect data under specific sky conditions and times of day. They will analyze the collected data to draw conclusions about weather patterns and relationships among parameters.	Provide students with detailed instructions on using weather stations to measure solar radiation, temperature, wind speed, and humidity. After collecting the data, students will compare their results with provided theoretical weather patterns.
2	Thermal Science Basics in Architecture	Measure the thermal properties of materials in the mockups
		Present questions like, “What kind of construction materials normally have low thermal conductivity?” and “Which materials are better at storing heat?” Students will set up their hypotheses, measure the samples, and then justify their choices and assumptions based on the measured data.	Provide instructions for measuring the thermal properties of various exterior wall and window components. Students will follow the instructions to measure temperature changes and thermal performance, comparing their results to theoretical values to validate the properties of the materials.
3	Heat Transfer Mechanisms	Demonstrate heat transfer methods in the mockups
		Pose a question such as, “How can we change the conductive/convective/radiative heat transfer within fixed temperature differences?” Students will design experiments to demonstrate conduction, convection, and radiation using different configurations. They will measure temperature changes and heat flow, testing the effectiveness of their strategies.	Set up stations within the lab to demonstrate conduction, convection, and radiation. Provide detailed instructions for measuring temperature changes and heat flow in each scenario. Students will follow these instructions and compare their measurements with theoretical expectations to validate their understanding of heat transfer mechanisms.
4–6	Building Envelopes and Indoor Thermal Environment	Envelope thermal performance on indoor thermal sensation
		Pose a question such as, “How can we alter envelope components to improve thermal comfort in a given position?” Students will design their own experiments to measure envelope thermal performance, and then configure envelope components to improve thermal comfort levels in various scenarios within the mockup room. They will analyze the rationale via measured envelope properties and collected data to examine their design interventions.	Provide detailed instructions and instruments for measuring envelope properties and assessing thermal comfort in different positions within a mockup room. Students will report their measurements and thermal sensations and compare their findings with theoretical calculations or models.
7	Passive Design Strategies	Solar gain and shading in the mockups
		Present a question like, “How can we utilize solar gains near window zones in winter?” Students will design experiments using various materials, reflectors, and instruments within the mockups. They will measure internal temperatures and assess the effectiveness of different passive design strategies, proposing the best solutions based on their findings.	Provide instructions for measuring internal temperatures and the effects of shading devices within mockups. Students will follow these instructions to assess the impact of passive design strategies on thermal performance, comparing their results with theoretical predictions.
8	Review and Final Exam	N/A

.

The typical routine of each experiment is as follows: All experiments were typically conducted under conditions of a large temperature difference between indoor and outdoor environments during the heating season. Instructors proposed questions related to different topics and also provided sufficient instruments, sensors, samples, and necessary hardware and tools for students. Before each experiment, students participating in the teaching reform were required to visit the laboratory (a few full-scale mockups) to conduct hands-on activities in terms of measurements, assessments, and documents, and analyze the reasons behind the phenomena. With guidance from teachers in this process, they designed and conducted the experiments to examine their hypotheses or strategies, forming a teaching model that integrates “theory guiding experiment, experiment reinforcing theory”. On the other side, the students who did not participate in these inquiry-based experiments still work on traditional confirmatory experiments. These students follow detailed instructions to perform experiments and validate predetermined results or theories, ensuring they understand fundamental concepts but without the same level of active problem-solving or creative design involved in the inquiry-based approach.

These experimental teaching reforms aimed to cultivate students’ interest in learning building thermal science and other theoretical courses, strengthen their understanding and mastery of thermal science knowledge, and enhance their ability to apply knowledge to solve practical problems. The design of the experimental teaching process is illustrated in Figure 1.

3.2. Theoretical Basis of the Experiment (In Building Envelop Session of the Course)

Based on the heat transfer process of building envelope structures, different on-site detection methods for the heat transfer coefficient of building envelope structures can be obtained from different perspectives.

If we start from the concept of heat flux in the envelope structure, under certain temperature differences on both sides of the envelope, the heat flux within the structure is inversely proportional to its thermal resistance, i.e.,

$$q^{″}={\displaystyle \frac{{\mathsf{\theta }}_i-{\mathsf{\theta }}_e}{R_s}}$$

(1)

where $q^{″}$ is the heat flux (W/m²); ${\mathsf{\theta }}_i$ and ${\mathsf{\theta }}_e$ (K) are the interior and exterior surface temperatures, respectively; and $R_s$ (m²·K/W) is the total thermal resistance of the envelope structure.

Adding the convective heat transfer resistance on the inner and outer surfaces of the envelope structure $R_i$ and $R_e$ (m²·K/W), the total thermal resistance $R_0$ (m²·K/W) of the structure can be calculated, and, further, thermal transmittance, the $U$ value (W/(m²·K)), can be determined:

$$R_0=R_i+R_s+R_e$$

(2)

$$U={\displaystyle \frac{1}{R_0}}={\displaystyle \frac{1}{R_i+R_s+R_e}}$$

(3)

Since $R_i$ and $R_e$ are considered constants in thermal calculations and can be found in relevant references, it is only necessary to measure the heat flux $q^{″}$ within the envelope structure and the surface temperatures ${\mathsf{\theta }}_i$ and ${\mathsf{\theta }}_e$ on the inner and outer sides of the envelope to calculate the $U$ value. The on-site detection instrument for the heat transfer coefficient of building envelope structures is designed based on this principle to detect the $U$ value on-site.

If we made such an assumption that heat transfer in the envelope structure is one-dimensional and steady-state (although the heat transfer in the envelope may not always be purely one-dimensional and steady-state, calculations based on one-dimensional steady-state are sufficient for engineering needs), using the principle of equal heat flux through each layer of materials (including surface boundary layers) in the envelope structure [23,24]:

$$q^{″}={\displaystyle \frac{t_i-t_e}{R_0}}={\displaystyle \frac{t_i-{\mathsf{\theta }}_i}{R_i}}$$

(4)

Such that,

$$R_0={\displaystyle \frac{t_i-t_e}{t_i-{\mathsf{\theta }}_i}}{R}_i$$

(5)

$$U={\displaystyle \frac{1}{R_0}}={\displaystyle \frac{t_i-{\mathsf{\theta }}_i}{t_i-t_e}}·{\displaystyle \frac{1}{R_i}}$$

(6)

Thus, as long as the indoor air temperature $t_i$ (K), the outdoor air temperature $t_e$ (K), and the inner surface temperature of the envelope structure ${\mathsf{\theta }}_i$ are measured, the total thermal resistance $R_0$ and total thermal transmittance $U$ of the structure can be calculated.

3.3. Experimental Preparation and Integration into Teaching Reform

The on-site detection experiment for the heat transfer coefficient of building envelope structures is conducted after the completion of theoretical instruction. Before the experiment, students have already studied the basic principles and processes of heat transfer in building envelope structures. Therefore, they were aware that under the same outdoor climate conditions and indoor heating conditions, different building envelope structures will result in different indoor thermal environments. Before the experiment, students were asked to experience the temperature differences in various locations within the thermal laboratory’s mockups, such as near exterior walls, exterior windows, and interior walls. The instructor introduces the construction of the mockup’s exterior walls and asks students to review the theory of heat transfer in building envelope structures, analyze the reasons for the thermal sensation differences in various locations, and design experiments to verify their analysis.

3.3.1. Analysis of Thermal Sensation Differences

During the experimental class, each group of students discusses their analysis of the reasons for the thermal sensation differences in various locations within each mockup in the laboratory. Human thermal sensation is influenced by ambient air conditions and the mean radiant temperatures of all indoor surfaces. Under identical indoor air settings and distributions, variations in thermal sensation are primarily driven by differences in mean radiant temperature that are determined by the surface temperatures. Under the guidance of the instructor, and considering the specific spatial environment of the mockup and the construction of the corresponding envelope structures at different locations, students gradually reach a consensus. They conclude that the main reason for the thermal sensation differences in various locations within the mockup was the different surface temperatures of the exterior walls, exterior windows, and interior walls. Further analysis revealed that the difference in surface temperatures between the exterior walls and exterior windows is due to the difference in their total heat transfer coefficients, while the difference in surface temperatures between the exterior walls and interior walls is due to differences in their construction and the exterior air temperatures.

3.3.2. Experiment Design

Based on the above analysis, students identify that the direct cause of the thermal sensation differences in various locations within each full-scale mockup is the different surface temperatures of the corresponding envelope structures, which can be verified with a suitable temperature measuring instrument. The fundamental cause of the thermal sensation differences is the difference in the total heat transfer coefficients of the corresponding envelope structures. How can this analysis be verified experimentally? After group discussions and evaluations by the instructor, students were guided step-by-step to apply theoretical knowledge to solve practical problems. This process led them to identify two methods for detecting the heat transfer coefficient of building envelope structures, as described in the Section 3.2: (1) Direct measurement using the on-site heat transfer coefficient detection instrument for building envelope structures. (2) Indirect measurement based on the principle of one-dimensional steady-state heat transfer, by measuring the indoor and outdoor air temperatures and the surface temperature at a specific point on the envelope structure and then calculating the corresponding heat transfer coefficient of the envelope structure.

3.3.3. Experiment Implementation

To verify the analysis of the fundamental causes of thermal sensation differences in various locations within each mockup, students in the reform group were divided into two large groups voluntarily using different measurement tools and methods. One group uses the direct measurement method, and the other group uses the indirect measurement method for on-site detection of the heat transfer coefficient of building envelope structures. This basic procedure followed the in situ measurement and analysis method defined in the previous study [25]. The first group used a heat transfer coefficient detection instrument to measure the heat transfer coefficients of exterior walls with insulation, exterior walls without insulation, and exterior windows. The detection instrument used was the JXJ-1 model from Beijing Hongou Chengyun Technology Co., Ltd., Beijing, China, as shown in Figure 2.

When using this instrument to measure the heat transfer coefficient of building envelope structures, both the heat flux at the measurement point and the surface temperatures on the inner and outer sides of the envelope structure near the measurement point must be measured simultaneously. Both sensors feature a specialized adhesive that ensures they remain securely attached to the surface throughout the experiment. This design enhances the accuracy of temperature measurements by maintaining consistent thermal contact. Figure 3 shows the use of this instrument to measure the heat transfer coefficient of building envelope structures. After the measurements were completed, the heat flux and surface temperatures at each measurement point were extracted from the instrument, and the relevant parameters were input into the software on a computer. The software calculates and directly outputs the detection results.

Table 2. Partial detection report of exterior insulated walls using the JXJ-1 on-site detection instrument.

Channels and Constants
Inner channel: 02 04; Outer channel: 12 13; Heat flow channel: 07; ${\mathit{R}}_{\mathit{i}}=0.11 {\mathrm{m}}^2\mathrm{K}/\mathrm{W}$$; {\mathit{R}}_{\mathit{e}}=0.04 {\mathrm{m}}^2\mathrm{K}/\mathrm{W}$
Calculation Results
Heat transfer coefficient: $\mathit{K}=0.782$

shows one example of the experimental results for the heat transfer coefficient $K$ of exterior walls with insulation.

The second group used the AZ8703 thermometer and hygrometer (manufactured by Hengxin Technology Ltd., Taiwan, China, as shown in Figure 4) to measure the indoor and outdoor air temperatures: $t_i$ and $t_e$. They used the Raytek ST60XXAP portable infrared thermometer (shown in Figure 5) to measure the surface temperatures, ${\theta }_i,$ of exterior walls with insulation, exterior walls without insulation, and exterior windows. By calculating the results using the formulas or inputting the measured parameters into a calculation program, they obtained the heat transfer coefficient corresponding to each measurement point on the envelope structure.

Table 3. Air temperatures inside and outside the exterior insulated wall, interior surface temperature, and corresponding K value were measured at different times (one example was measured with the JXJ-1 on-site detection instrument).

Value	10:00	10:10	10:20	10:30	10:40	10:50	11:00	Average
$t_e (℃)$	−8.4	−8.5	−8.4	−8.5	−8.2	−8.5	−8.0	−8.4
$t_i (℃)$	13.7	13.9	13.9	13.8	13.9	13.9	13.9	13.9
${\theta }_i (℃)$	11.7	11.9	11.9	11.8	11.9	11.9	11.9	11.9
$K (\mathrm{W}/({\mathrm{m}}^2\mathrm{K}\left)\right)$	0.82	0.81	0.82	0.82	0.82	0.81	0.83	0.82

shows the measurement and calculation results for the heat transfer coefficient $K$ of exterior walls with insulation.

Although the results obtained by the two measurement methods were not identical, a comparison of the heat transfer coefficients measured for different envelope structures in different mockups (as shown in Figure 6) indicates that both methods could measure the K-values accurately, with an acceptable error percentage (<5%) relative to the given K-values. By conducting experiments based on their theoretical analysis and using their chosen measurement methods, the instructors taught the associated knowledge and concepts about building thermal science and also highlighted the logic process behind the experimental design.

3.4. Control Experiment Grouping and Data Analysis Methods

To evaluate the effectiveness of the experimental teaching reform, students in the building thermal science course were divided into two groups each year: a reform group and a control group. The reform group receives design-based experimental teaching for the on-site detection of the heat transfer coefficient of building envelope structures, while the control group continues with traditional confirmative experimental teaching for the same experiment. Students’ final grades (0–100) were used as the main metrics to evaluate and compare the learning outcomes. The grading policy for this 8-week building thermal science course evaluated students based on attendance and participation (10%), experiments (10%), homework completion (15%), and exam performance (60%). The experiments were team-based, with evaluations based on the completion and quality of the reports, and its rubric was the same for both control and reform groups. Homework assignments were also the same for all students, which mainly focused on calculations, simulations, and theoretical knowledge and understanding. The exam consisted of paper-based assessments that included closed-book and closed-note multiple-choice questions, calculations, and open-ended questions, requiring students to apply theoretical concepts to practical scenarios in architecture and building thermal science. This grading policy, along with the format of homework and exam questions, has been established and adopted in our department over the past decade, with minor variations made each year to ensure relevance and rigor.

3.4.1. Control Experiment Grouping

The experimental teaching reform has been conducted for three years since 2021, with traditional experimental teaching for the control group running concurrently. Three cohorts of undergraduate architecture students have participated in this teaching activity. The reform group and control group were randomly selected from the classes of each cohort. Students in each class were further divided into experimental groups of four. The experimental activities were conducted in groups.

The number of classes in each cohort was determined based on student numbers and teaching resources. The principles for forming classes in each cohort were as follows: First, the students were divided into two main groups based on gender. Each of these groups was then further categorized into three levels—high, medium, and low—based on their GPAs (ranging from 0 to 100) in fundamental architectural theory courses. These courses, which cover a series of core architectural knowledge, were used as a baseline metric for all participating students. Afterward, to meet the required number of classes (n), students within each gender group were evenly distributed into n smaller groups according to their GPAs. A class was then formed by selecting one small group from each level within both the male and female groups. This process ensured that each cohort was divided into n classes, with each class maintaining a balanced gender ratio, as well as comparable entrance exam scores and learning abilities.

The reform and control groups selected from the three cohorts of students who have undergone the experimental teaching are as follows: The reform group consisted of 22 students from the 2019 cohort, class 1 (out of 3 classes); 33 students from the 2020 cohort, classes 1 and 2 (out of 4 classes); and 45 students from the 2021 cohort, classes 3 and 4 (out of 4 classes). The control group consisted of 44 students from the 2019 cohort, classes 2 and 3; 33 students from the 2020 cohort, classes 3 and 4; and 46 students from the 2021 cohort, classes 1 and 2.

3.4.2. Data Analysis Methods

To assess the effectiveness of the teaching reform, we first computed general descriptive statistics for the reform group and the control group’s test scores for their “building thermal science” course. This included calculating the mean, standard deviation (SD), and sample size (n) for the grades. These descriptive statistics provided an initial overview of the performance of students in both the control and reform groups across the three cohorts. Secondly, to ensure the validity of subsequent statistical tests, we assessed the equality of variances between the control and reform groups within each cohort using Levene’s test (null hypothesis: $H_0:{\sigma }_1^2={\sigma }_2^2; H_a:{\sigma }_1^2\ne {\sigma }_2^2$, with a significance level of 0.05). Levene’s test is robust to deviations from normality and tests the null hypothesis that the variances of the groups are equal. A non-significant result (p > 0.05) indicates that the variances are approximately equal, while a significant result (p ≤ 0.05) suggests that the variances are unequal. Subsequently, based on the results of Levene’s test, we conducted an independent two-sample t-test to compare the grades between the control and reform groups within each cohort (null hypothesis: $H_0:{\mu }_1={\mu }_1; H_a:{\mu }_1\ne {\mu }_2$, with a significance level of 0.05). If Levene’s test indicated equal variances, we used the standard t-test. If it indicated unequal variances, we used Welch’s t-test, which adjusts for the unequal variances. The t-test allowed us to determine whether the differences in student grades between the control and reform groups were statistically significant. By following this data analysis approach, we aimed to discern whether the course reform design led to significantly different grade outcomes, thus providing insight into the effectiveness of the reform. Key metrics from our analysis included the mean scores, standard deviations, p-values from Levene’s test, and p-values from the t-tests. These metrics enabled us to comprehensively assess and compare the performance of students in the control and reform groups across the three cohorts.

4. Results

4.1. Baseline GPA Comparison and Grouping Validation

Based on the data provided in

Table 4. Statistics of Baseline GPAs for three cohorts of students in the reform class and control class.

	Pre- and Post-Teaching Reform Comparison	Number of Individuals	Average	Standard Deviation
Class of 2019	Control group	44	69.61	11.396
	Reform group	22	71.50	11.198
Class of 2020	Control group	33	78.94	13.601
	Reform group	33	79.30	12.970
Class of 2021	Control group	46	73.52	9.809
	Reform group	45	72.49	10.820

, the baseline GPAs for the three cohorts in both the reform and control groups appear to be relatively similar. For the 2019 cohort, the control group had an average GPA of 69.61 compared to 71.50 in the reform group, while in the 2020 cohort, the averages were 78.94 for the control group and 79.30 for the reform group. For the 2021 cohort, the control group had an average GPA of 73.52, slightly higher than the 72.49 average for the reform group. The mean values of standard deviation in each cohort also indicate similar variations in GPA distribution, further supporting the argument that the groups are comparable in terms of their general academic performance prior to any teaching interventions.

Focusing on

Table 5. The t-test of students’ baseline GPAs for three cohorts of students in the reform group and control group.

		Levene’s Test for Equality of Variances		t-Test for Equality of Means
		F	Significance	t	Sig. (Two-Tailed)	Mean Difference	Standard Error Difference	95% Confidence Interval
								Lower Limit	Upper Limit
Class of 2019	Assume Equal Variances	0.001	0.974	−0.638	0.526	−1.886	2.959	−7.797	4.025
	Assume Non-Equal Variances			−0.41	0.525	−1.886	2.941	−7.819	4.046
Class of 2020	Assume Equal Variances	0.408	0.525	−1.111	0.912	−0.364	3.272	−6.899	6.172
	Assume Non-Equal Variances			−1.111	0.912	−0.364	3.272	−6.900	6.172
Class of 2021	Assume Equal Variances	0.036	0.850	0.477	0.634	1.033	2.164	−3.267	5.333
	Assume Non-Equal Variances			0.477	0.635	1.033	2.166	−3.273	5.338

, which presents the results of the t-test for the baseline GPAs of all participating students, we can see that the significance values (i.e., p-values) for each cohort are all well above the 0.05 threshold, indicating that there are no statistically significant differences (i.e., failed to reject the null hypothesis, $H_0:{\mu }_1={\mu }_1$) in the baseline GPAs between the reform and control groups. For example, the significance value for the 2019 cohort is 0.526, for 2020 it is 0.912, and for 2021 it is 0.634. These high p-values suggest that the mean differences observed between the groups’ baseline GPAs are not significant, and we can reasonably conclude that the student’s overall learning abilities, as reflected in their GPA performance, were similar across both groups.

In summary, these results demonstrate that the grouping of students into reform and control groups was effective in creating comparable baseline conditions. The absence of significant differences in the students’ prior GPAs ensures that any future differences observed in the thermal science course grades are likely attributable to the teaching reform interventions, rather than pre-existing differences in student abilities. This strengthens the validity of the subsequent analysis comparing the outcomes of the teaching reforms.

4.2. General Findings upon Observations

After three years of implementing the experimental teaching reform, several positive outcomes have been achieved in the following aspects:

(1)

Increased environmental awareness in the reform group: Students in the reform group demonstrated higher environmental awareness compared to those in the control group. During the design process, students in the reform group considered the building environment, particularly the thermal environment, more thoroughly, compared to the control group. They were more inclined to choose building materials with low production energy consumption, ease of recycling, and minimal environmental impact at the end of their lifespan. Additionally, they favored the use of renewable energy sources when selecting energy systems.

(2)

Improved application of theoretical knowledge: Students in the reform group showed a greater ability to apply theoretical knowledge than those in the control group. They were more enthusiastic about participating in student innovation and entrepreneurship training programs, as well as national green building competitions. For example, a team from the 2020 reform group used the design experiment to compete in the National College Physics Experiment Competition, and students from the 2021 reform group participated in the National College Students Energy Conservation and Emission Reduction Social Practice, and Science and Technology Competition by designing insulation for building envelope structures.

(3)

Higher evaluation of the experiment in the reform group: The reform group students rated the experiment more favorably than those in the control group. Anonymous surveys conducted among students revealed that those in the reform group rated the experiment higher across several aspects: fostering creativity, thinking ability, practical skills, understanding the application value of theoretical knowledge, increasing interest in learning architectural theory courses, improving collaboration and teamwork, enhancing problem-solving abilities, and increasing environmental awareness. The standard deviation of the ratings in the reform group was also lower, indicating a higher and more consistent level of satisfaction with the experiment.

4.3. Data Analysis on Learning Outcomes upon Final Grades

4.3.1. Descriptive Statistics Results

The descriptive statistics for the three cohorts (2019, 2020, and 2021) in the course reform experiments are based on the comparison between the reform and control groups, as presented in

Table 6. Statistics of course grades for three cohorts of students in the reform group and control group.

	Pre- and Post-Teaching Reform Comparison	Number of Individuals	Average	Standard Deviation
Class of 2019	Control group	44	77.20	9.328
	Reform group	22	84.41	7.620
Class of 2020	Control group	33	74.70	6.396
	Reform group	33	79.24	5.874
Class of 2021	Control group	46	73.30	10.172
	Reform group	45	79.13	10.047

and Figure 7.

In the 2019 cohort, the reform group achieved a higher average score (84.41) compared to the control group (77.2). The reform group also had a larger proportion of students with “Excellent” grades (22.7%) compared to the control group (15.9%). Both groups had similar percentages in the “Good” and “Average” categories, but the reform group had a lower failure rate and a better overall grade distribution, suggesting the reform measures positively impacted student performance.

For the 2020 cohort, the reform group once again exhibited a slightly higher average score of 79.24 compared to the control group’s 74.70. A higher percentage of students in the reform group achieved “Excellent” grades (12.1%), compared to only 6% in the control group. The reform group also had a higher proportion of students in the “Good” category (51.5%) than the control group (48.5%), indicating a modest shift towards higher performance in response to the reforms. Meanwhile, the control group had a greater distribution of students in the “Average” category (33.3%), reflecting a more evenly spread performance without as many top-achieving students.

In the 2021 cohort, both the control and reform groups had comparable average scores, with the control group at 73.30 and the reform group at 72.49. The control group had a slightly higher percentage of students in the “Good” and “Average” categories, but the reform group had fewer failing students (2.2%) compared to the control group (8.7%). This lower failure rate suggests that, while the reform group’s average score was slightly lower, the intervention helped more students pass the course, minimizing failure rates while maintaining a solid grade distribution overall.

In brief, the reform group generally exhibited better performance across all cohorts. There was a consistent pattern of higher “Excellent” grades and lower failure rates in the reform group, indicating that the teaching reforms had a positive effect on student outcomes, especially in reducing the failure rate in later cohorts. The control group, on the other hand, tended to have a more balanced distribution between “Good” and “Average” grades, without the same improvements seen in the reform group.

4.3.2. t-Test Results

The t-test results (as shown in

Table 7. The t-test of “building thermal science” course grades for three cohorts of students in the reform group and control group.

		Levene’s Test for Equality of Variances		t-Test for Equality of Means
		F	Significance	t	Degree of Freedom	Sig. (Two-Tailed)	Mean Difference	Standard Error Difference	95% Confidence Interval
									Lower Limit	Upper Limit
Class of 2019	Assume Equal Variances	2.003	0.162	−3.134	64	0.003	−7.205	2.299	−11.797	−2.612
	Assume Non-Equal Variances			−3.353	50.430	0.002	−7.205	2.149	−11.519	−2.890
Class of 2020	Assume Equal Variances	0.088	0.768	−3.007	64	0.004	−4.545	1.512	−7.565	−1.526
	Assume Non-Equal Variances			−3.007	63.542	0.004	−4.545	1.512	−7.566	−1.525
Class of 2021	Assume Equal Variances	0.874	0.352	−2.619	89	0.010	−5.829	2.225	−10.251	−1.407
	Assume Non-Equal Variances			−2.617	88.039	0.010	−5.829	2.227	−10.255	−1.403

) provide strong evidence that the reform group consistently outperformed the control group across all three cohorts in the “building thermal science” course. The negative t-values for each cohort indicate that the reform group achieved higher average scores, with statistically significant differences (p < 0.05) in all cases. Specifically, the reform group scored an average of 7.21 points higher than the control group in 2019, 4.55 points higher in 2020, and 5.83 points higher in 2021. These consistent improvements in the reform group indicate that the teaching reform was effective in raising student performance.

Levene’s test for equality of variances also showed that we could assume equal variances across the reform and control groups for all three cohorts, as the p-values from Levene’s test were all above 0.05 (0.162 for 2019, 0.768 for 2020, and 0.352 for 2021). This ensures the validity of using the standard t-test, which confirms the significant differences in performance between the groups.

In addition to higher average scores, the reform group’s grade distributions were more concentrated, as indicated by comparable or slightly smaller standard deviations across all cohorts. For example, in the 2020 cohort, the standard deviation for the reform group was 5.87 compared to 6.40 for the control group, indicating that the reform group’s scores were more consistent and less dispersed. This pattern held across other cohorts as well, suggesting that the reform not only improved overall performance but also led to more uniform results.

The reform was particularly effective in elevating the grades of students who were initially in the “Good” and “Average” categories. The differences in the percentage of students achieving “Excellent” grades between the reform and control groups were modest, but a more significant impact was observed among students with weaker initial performance. The reform’s hands-on, practical approach seems to have supported these students in improving their understanding and application of the material, which translated into better grades. Such improvement was limited, however, among the students who could originally achieve “Excellent” no matter which groups they were in, as they were already self-motivated and had a good understanding of the knowledge.

To summarize, the t-test results across the three cohorts demonstrate that the experimental teaching reform significantly improved student performance. The reform group consistently achieved higher average scores and exhibited more concentrated grade distributions compared to the control group. Levene’s test confirmed the appropriateness of assuming equal variances, and the statistically significant mean differences across all cohorts highlight the effectiveness of the teaching reforms.

These results suggest that the reform’s focus on problem-solving, experimental design, and hands-on learning successfully enhanced students’ ability to apply theoretical knowledge to practical challenges. This was especially beneficial for students who initially struggled with the course material, as the reform seemed to better engage and support their learning, ultimately narrowing performance gaps. The insights gained from this teaching reform can provide valuable guidance for improving other theoretical courses, particularly by integrating practical applications to foster deeper understanding and improve student outcomes.

5. Conclusions

The results of our study demonstrate that the grouping of students into reform and control groups was fair and effective, as confirmed by the t-test analysis of baseline GPAs. Across the three cohorts, the t-test results showed no statistically significant differences in students’ overall academic abilities prior to the course (p > 0.05), ensuring that any observed differences in course outcomes were a result of the teaching reform, not inherent differences in student performance. This foundational validation of grouping provides a solid basis for assessing the impact of the experimental teaching reform on the “building thermal science” course. The subsequent t-test analysis of the course grades revealed that the reform group consistently outperformed the control group in all three cohorts, with statistically significant improvements (p < 0.05) in average scores ranging from 4.55 to 7.21 points higher. These findings clearly demonstrate that the reform, which transformed traditional confirmatory thermal science experiments into inquiry-based, hands-on exploratory learning experiences, was effective in enhancing students’ academic performance. The more focused distribution of grades within the reform group, as indicated by comparable or smaller standard deviations, further suggests that this teaching method not only improved overall performance but also helped students of varying abilities to achieve more consistent results.

The teaching reform allowed students to engage deeply with key concepts, in the learning topics about envelopes and indoor thermal environment, like thermal resistance, heat transfer processes, and material properties, reinforcing theoretical knowledge through practical application. This approach significantly improved students’ ability to analyze and solve complex problems, as evidenced by their improved performance in course assessments. By experiencing real-world temperature variations and designing their own validation experiments, students were encouraged to think critically, fostering a deeper understanding and appreciation of building thermal science. This also stimulated their curiosity and enthusiasm, which heightened their awareness of environmental issues and the practical application of theoretical knowledge.

From the perspective of research limitations, we made significant efforts to control potential influencing factors and maintain consistency between both groups. This included ensuring uniformity in lecture content, teaching methods, and assessment criteria. Additionally, our sampling procedure took into account student backgrounds and prior academic performance. However, we recognize that other factors, such as varying student workloads and differing contact hours with instructors, which were not documented in our evaluation process, may have influenced the results. Furthermore, students’ levels of motivation and interest in the subject matter, which can significantly affect learning outcomes, were not measured. We will carefully document these potential influencing factors in future evaluations.

Furthermore, certain logistical challenges were also encountered, such as the time required for experimental preparation. However, by coordinating office hours and providing additional support from specialized experimental tutors, we were able to overcome these obstacles without disrupting the course’s progress. This collaborative effort not only ensured the smooth implementation of the teaching reform but also created opportunities for more personalized student support. The three-year implementation of the experimental teaching reform has proven its effectiveness in enhancing the learning experience and improving student outcomes. Given the statistically significant improvements in student performance, the reform group has decided to expand this design-based experimental teaching model across the entire cohort, and also to the other curricula that incorporate experimental teaching (e.g., building mechanics). Additionally, an instrument-free session that could serve as a supplement to the current teaching reform is may be added in the future. Students will be asked to design an experiment that does not require any devices, to measure some key parameters (for instance, measuring indoor air temperature by recording the time-elapse of hot water getting cold) before they conduct hands-on experiments using the instrument. This instrument-free session would facilitate students’ interests and involvement in the experiments, and also help them comprehend the theoretical knowledge better.

In conclusion, this reform not only enhances academic performance but also equips students with practical skills in green building design and environmental awareness, making it a valuable pedagogical model for future curriculum development. By continuing to integrate practical applications with theoretical knowledge, we aim to prepare students to address real-world challenges in architecture and building sciences with creativity, critical thinking, and environmental responsibility.