Can Green Building Science Support Systems Thinking for Energy Education?

Abstract

Systems thinking (ST) is a foundational cognitive skillset to advance sustainability education but has not been well examined for learners prior to higher education. This case study research in rural middle schools in the Midwestern U.S. examines systems thinking outcomes of a place-based energy literacy unit focused on energy-efficient building design. The unit employs the science of energy-efficient, green buildings to illuminate the ways in which energy flows between natural and built environments. The unit emphasized electrical, light, and thermal energy systems and the ways these systems interact to create functional and energy-efficient buildings. This study focuses on three case study classrooms where students across schools (n = 89 students) created systems models as part of pre- and post-unit tests (n = 162 models). The unit tests consisted of student drawings, annotations, and writings, culminating into student-developed systems models. Growth from pre- to post-test was observed in both the identification of system elements and the linkages between elements. System elements included in the models were common classroom features, such as windows, lights, and temperature control, suggesting that rooting the unit in place-based teaching may support ST skills.

1. Introduction

Climate change education has made important strides in the U.S. [1]. Despite the need to reduce energy consumption, mitigate environmental degradation, and protect public health [2], energy education often remains abstract and disconnected from applied settings [3]. The built environment offers an applied context to teach energy ideas in science education. In the U.S., buildings contribute nearly 50% of total carbon emissions and use nearly 70% of all electricity produced [4,5] but are rarely a focus in science education [6]. Without making these connections for learners, it is challenging to advance environmental education in, about, and for green infrastructure [7].

This study is part of a broader initiative by an interdisciplinary team of built environment and educational researchers to make energy systems meaningful to youth. The team developed a place-based unit for middle school (MS), focusing on energy flow across natural and built environments [8]. While energy systems learning applies to various contexts, this unit focuses specifically on energy-efficient building design, which is a key part of the broader green building education movement [4]. Through green building design, MS students can explore energy flow between ecosystems and human systems (built environments). Thus, a systems thinking (ST) approach provides an essential framework for the unit, helping students to understand how energy systems, the built environment, and society interact and influence one another as part of a larger whole [9]. This paper investigates how students’ ST about how energy flows to, within, and from a building is enhanced through implementation of the unit. From the beginning of the unit to the end, pre- and post-tests reveal that students were increasingly naming systems elements and making links between them. Growth in demonstration of built environment elements was notable. System elements included in the models were common classroom features, such as windows, lights, and temperature control, suggesting that rooting the unit in place-based teaching may support ST skills.

1.1. Youth Energy Education and Rural Contexts

Youth and adults alike struggle to make sense of energy ideas, to understand how energy is used within the buildings they occupy daily, and to make informed decisions regarding energy use [10,11,12]. Understanding energy, however, is central to science education, as core ideas across all science domains require “a basic understanding of matter and energy” [9] (p. 104).

Energy is a dynamic concept because it constantly changes forms. It is typically observed indirectly through changes in objects or is conceptualized abstractly through mathematical formulations [13,14]. How students understand energy and the instructional supports needed for energy education have been extensively studied [15]. Learning progression research illuminates the ways in which student ideas about energy are arranged in a “complex network of ideas” [13] (p. 3), where interrelated energy ideas can co-develop [16,17]. This complex network conceptualization of energy learning allows for a fuller picture of energy understanding intertwined with social systems, which is a global learning objective of the United Nations [2].

The overlay of energy systems and society is important across K-12 education, but is a particular need in rural school districts, which are the target population of the current study. Rurality is not well defined within the U.S.—the U.S. Census Bureau defines rural as any area “not included within an urban area” [18] (p. 1). The state in which this study took place defines rural schools as schools with ≤600 students in the district [19]. However, critics of defining places by population numbers prefer to define rural as “small, close-knit places with intergenerational connections to land with a strong sense of pride, community history, and tradition” [20] (p. 126).

Rural districts are often underserved populations for STEM interventions [21]. Although STEM careers require K-12 students to engage in diverse and sustained learning experiences, rural students frequently lack access to the resources and opportunities necessary to prepare for STEM career pathways [22]. Simultaneously, rural communities disproportionately bear the environmental and health consequences of U.S. energy decisions—such as oil and gas extraction—which directly impact the people who live and work in these areas [23,24]. These same communities are also increasingly home to wind and solar farms. Being sites of both renewable and nonrenewable energy resources provides rural students unique opportunities to explore real-world examples within their communities. Being able to develop an awareness of this full energy spectrum is essential for rural students to learn about their future career possibilities, as well as becoming future advocates for the role of their communities in supplying and harvesting energy [25].

1.2. Youth Green Building Education

The built environment offers a compelling context for making energy flow tangible—from the point of resource extraction to the everyday use of electricity within buildings. Figure 1 illustrates the interdependent relationship between the buildings, the energy systems that support them, and the surrounding ecological context. The bidirectional arrow indicates a reciprocal relationship: while buildings and energy systems depend on ecological systems for resources and environmental conditions, they also affect those systems through emissions, extractions, and land use. To fully understand this flow of energy, learners must move beyond abstract energy concepts and ground their understanding in the physical features of both natural and built environments. This comprehension is further complicated by the multi-scalar nature of the built environment, which ranges from community-level energy grids down to individual components, such as lightbulbs and outlets. Causes and effects across these large-scale systems often span time and space, making visualization and conceptualization challenging for youth [26].

Frameworks for green building literacy (GBL) outline potential pathways to converge energy education and green building education [6]. The conceptualization of GBL identifies 14 categories of green building knowledge [6] that cross scales from buildings to infrastructure to ecology. Of these green building knowledge domains, the current work centers on “energy and atmosphere,” which aligns with green building professional standards. This domain comprises a key set of performance standards within the Leadership in Energy and Environmental Design (LEED) framework used by professionals that forefronts climate resilience and carbon assessments, decarbonization planning, and minimum energy efficiency requirements for buildings [27]. The current work thus examines the foundations for fostering green building knowledge centered on energy and atmosphere that are appropriate for middle school learners.

While sustainable design education has been a focus in practice [28,29,30] and higher education [31], fostering green building knowledge for youth is an emergent area of study in academic literature. Students in schools with certified green buildings have demonstrated higher green building knowledge than peers in conventional schools [32]. Other studies have shown that outdoor and eco-classroom environments can help students better explain solar energy systems [33]. However, for the dominant population of students without access to state-of-the-art green facilities, the processes for supporting the acquisition of green building knowledge are still emerging. Recent curriculum efforts have explored using digital tools to support green building education. The Concord Consortium has pioneered strategies for using simulations to support engineering design for energy efficiency, which yielded some early, but promising, results for knowledge outcomes from a small field study [34] and insights regarding self-regulated learning [35]. Similarly, Wu et al. [36] examined the potential for serious games to teach building energy consumption. Additionally, an upper elementary green roof unit has shown positive outcomes for students’ understanding of socio-hydrologic systems [37]. Other scholars have examined how environmental justice can be woven into green building education at the high school level, including co-design processes between students and community partners [38]. Despite growing interest in the field, the relationship between ST and MS students’ conceptual understanding about energy and green buildings remains underexplored.

Green building knowledge is operationalized in a variety of ways across scholars but can be understood as a multi-dimensional outcome. The Cole [6] framework for GBL collapses green building knowledge and skills into knowledge categories of factual, conceptual, and procedural knowledge. Where factual knowledge includes identification and recall of basic components of green building design (e.g., solar panel and recycled materials), a person with conceptual green building knowledge begins to demonstrate understanding of relationships between components—and between buildings and natural systems (e.g., how artificial lights create demand for nonrenewable energy sources). Procedural knowledge centers more on the ability to effectively design and operate green buildings. The current work emphasizes the knowledge domain of the GBL framework with a focus on factual and conceptual green building knowledge, where the latter will be understood through frameworks for ST.

1.3. Systems Thinking in K-12 Education

Understanding how energy is transferred and transformed from nature to built environments requires that students visualize and reason about systems. This process involves thinking about (and thinking with) systems. Systems thinking (ST) is widely recognized as a critical cognitive skill in sustainability education [39]. The topic of sustainability is often multi-scalar, crossing natural and human-made systems. When integrating ST within sustainability education, students (and adults) can consider how individual elements are a complex web of interconnections across systems [40]. Yet, ST is also one of the most difficult higher-order thinking skills to master. Individuals tend not to develop ST skills without intentional learning experiences that make the elements, relationships, and organization within systems explicit [41]. However, because the ST field is still evolving, there is no consensus on a unified framework for how learners develop systems thinking [39,42].

Numerous conceptual models have attempted to define systems thinking in educational contexts. For example, Hmelo-Silver and colleagues [43] proposed a framework called structures, behaviors, and functions (SBF) that focused on supporting student consideration of system structures and relationships, and the overall system behavior that resulted from system relationships that created system function. Assaraf and Orion [39] proposed a hierarchical model of systems thinking composed of four levels of growth, in which each new level builds upon the prior levels, yet there were few individuals within their studies that were able to achieve the higher levels. More recently, Cabrera and Cabrera [44] provided a definition of ST that builds from, and synthesizes across, prior ST literature. They consider the presence of natural, societal, and socio-ecological systems, and the mental models that individuals construct about these systems (i.e., how individuals think about the system). The Cabrera and Cabrera [44] DSPR framework distills competencies into four ST skills: making distinctions (D) between components and defining system boundaries, organizing those elements into constituent parts and processes within the system (S), locating and tracing relationships (R) between components, and representing a perspective (P) that includes people, objects, and ideas within the system.

1.4. Systems Modeling in K-12 Education

Modeling practice is considered essential to development of ST [44,45]. The U.S. Next Generation Science Standards (NGSS) identify modeling as a scientific practice in which “models are used to represent a system (or parts of a system) under study” [46] (p. 6). Both models and modeling have a rich history across the STEM disciplines, where models are used by STEM professionals to show current agreed upon understandings of the phenomenon by the discipline. Modeling is used as a learning and reasoning tool to develop that understanding. Just as with STEM professionals, within STEM education, models and modeling each serve different functions [47]. Within the K-12 classroom, models serve as depictions of scientific phenomenon, such as the molecular arrangement of water, physical models of cells, or the chemical equation of photosynthesis. Models are the agreed upon understanding of the phenomenon by disciplines. These are typically shown in textbooks for student learning. Modeling, however, is the process of students learning about a phenomenon. Through modeling, students externalize their mental models and use their expressed representation as a learning tool for the system.

While the practice of modeling can occur in different forms, such as computational and mathematical, of particular interest here are 2D diagrammatic systems models [48]. These are models that students draw with pencil and paper to externalize their mental systems model in response to a specific question or problem. We intentionally have students draw their models with paper and pencil, as the act of developing and using drawn models requires students to think about how they will represent the elements, show connections between elements, and organize elements. Drawing supports learning, especially when students are expressing their own ideas [49,50].

Within their drawings, students use words, numbers, and symbols to convey their current understanding. The power of developing 2D diagrammatic models is that students can decide how to express their knowledge of system objects, the relationship between those objects, and their current system perspective. As students review their expressed mental models, they are uncovering what they do (and do not yet) understand. Learning is evidenced by students’ expression of new mental models over time, in which new knowledge is integrated into their models [51]. As such, modeling theory is learning theory, and modeling artifacts are historical records of conceptual understanding at different moments in time [52].

1.5. Research Questions

In summary, the current work converges on two substantive content areas of energy systems and the built environment systems. Given the focus on conceptual understanding of energy flow across human-built and ecological systems, systems modeling served as an analytical tool to examine student learning outcomes about the interconnections between elements of natural and human systems. This study explored how and if rural students in grades 7–9 increased their demonstrations of ST from the beginning to the end of a place-based unit focused on energy-efficient building design. The research questions guiding this study were as follows:

• What changes in ST, if any, were evidenced in student systems models from the beginning to the end of a place-based unit on energy-efficient building design?
• Did, and how did, students demonstrate systems thinking (ST) in their systems models depicting energy flow between buildings and the natural environment?

2. Materials and Methods

This study employs a case study research design [53] in three rural middle schools in the Midwestern U.S. in which the “Energy and Your Environment” (EYE) unit was implemented and evaluated. The study was awarded ethics approval from the Institutional Review Board (University of Missouri Institutional Research Board (IRB) project #2090786) prior to beginning data collection activities. Consent forms for participation were distributed to all participants and signed. All schools, students, and teachers in this report have been anonymized or given pseudonyms.

2.1. Research Participants

The students and teachers involved in this study are shown in

Table 1. Study participants by time point, school, grade, and demographics.

	School 1 Faith			School 2 Charlotte	School 3 Sophia			All Schools
	A	B	Total	Total	A	B	Total	Total
Time Point
Pre-test	15	17	32	7	21	24	45	84 *
Post-test	15	16	31	7	20	20	40	78 *
Grade Level
7	15	-	15	5	-	-	-	20
8	-	17	17	3	-	-	-	20
9	-	-	-	-	24	25	49	49
Gender
Male	2	4	6	4	12	6	18	28
Female	10	10	20	4	6	7	13	37
No response	3	3	6	-	6	12	18	24
Ethnicity
White	10	8	18	4	16	13	29	51
Non-white	2	5	7	4	2	-	2	13
No response	3	4	7	-	6	12	18	25
Total students	15	17	32	8	24	25	49	89

Note: A and B refer to differing sections if the teacher taught the unit to more than one class period. * The pre-test and post-test rows include students who filled out at least one test at either time point, which is why these numbers differ. The final dataset included tests from n = 89 unique students.

. All three case study schools were located in rural Missouri and chosen based on rural designation and school agreement to video-recorded instruction. Two schools (School 1 and School 2) were classified as ‘rural and remote,’ while one school (School 3) was ‘town and remote’ [54]. School 1 is in an area with a total population of 3056 individuals [55]. The total number of students enrolled in K-12 is 468, with approximately 45% of the student body qualifying for Free and Reduced Lunch (FRL) [56]. School 2 is in an area with a population of 822 individuals [55]. The total number of students enrolled in K-12 is 118 students, with approximately 54% of the student body qualifying for FRL [56]. School 3 is in a town with a total of 11,520 individuals [55]. The total number of students enrolled in K-12 is 1867, with approximately 35% of the student body qualifying for FRL [56].

The three participating teachers at case study schools taught the EYE unit in 1–2 class periods (if 2 periods, they are labeled A and B in Table 1) in the 2023–2024 academic year. While Charlotte (School 2) and Sophia (School 3) ran the curricular unit within a single timeframe, Faith (School 1) implemented the unit at two different time points in the academic year for her Sections A and B. Charlotte implemented the unit with one class period while Sophia and Faith each ran the unit in two different class periods. Charlotte (School 2) and Faith (School 1) both had over ten years of teaching experience, and both indicated that green building concepts were a new content area for them and their students. Sophia (School 3) had seven years of high school teaching experience. Sophia had previous experience with the curriculum content as a doctoral student in the lab that developed the EYE unit. Table 1 shows demographic information for each instructional section involved in the study. Supplementary File S1 offers a more in-depth description of each case study teacher, their school, and how they enacted the unit in their unique context.

2.2. The Energy and Your Environment (EYE) Curriculum

The Energy and Your Environment (EYE) curriculum is a six-week unit designed to support energy education for formal middle school science classrooms by employing green building science [8] (see links to open-access curricular materials in Supplementary File S2). This standards-aligned unit fosters place-based education by using the school building to enhance students’ systems thinking about energy consumption and energy flow between buildings and Earth systems. It consists of four modules (Figure 2). Module 1 introduces energy systems and energy transfer in the school building and introduces students to the engineering design challenge they will complete throughout the unit. Module 2 introduces learners to the ways in which natural and artificial lighting in the classroom relates to energy consumption. Module 3 extends this learning into heating, ventilation, and cooling (HVAC) systems. The unit then culminates in Module 4, in which students participate in a team-based engineering design project where they design an energy-efficient one-room schoolhouse.

ST is supported throughout these modules, as energy is framed as a system wherein energy transfers and transforms. The specific forms of energy of focus in EYE are electrical, light, and thermal energies that readily pertain to green building concepts. To support classroom teachers in implementing systems thinking, the curriculum design team created a system model (Figure 3) that was used to guide the curriculum content and served as a resource for teachers. Figure 4 shows images of unit enactment, including building tours, light metering, and the engineering design outcomes.

A key pedagogical strategy in the unit was the use of systems models, where students were asked to make annotated drawings and write about energy systems in buildings (see Section 2.3 below).

2.3. Data Collection

Student systems models were developed through a pre- and post-unit student test administered via paper (Supplementary File S3). The test was informed by Mislevy et al. [57] and based on our previous work on model-based learning [33]. It was designed to elucidate student ST through drawing, annotations, and writing, where students were encouraged to include the elements of the system, the relationships between elements, and their overall reasoning. A panel of experts in assessment, energy literacy, and modeling were consulted about the initial design of the test. Once the test was written, it was submitted for review by the same panel of experts prior to use in the classroom. The test was administered at the beginning and end of the EYE unit by participating case study teachers.

In total, 84 pre-tests and 78 post-tests were collected, comprising a total of 162 student models from a total of 89 unique participants (n = 73 who took both pre- and post-tests). The paper tests were scanned and digitized for analysis. Students (n = 25) across the case study classrooms participated in end-of-unit interviews, which were used as secondary data to supplement interpretations of results. Additional data sources included video recordings of classroom sessions, teacher interviews, and student artifacts generated throughout the unit.

2.4. Data Analysis

The multi-phase qualitative content analysis was conducted using a grounded, inductive approach [58]. While no predetermined theoretical framework guided coding, the system model developed for the curriculum (Figure 3) served as an analytic anchor for identifying relevant system elements and relationships. The mixed-methods analyses began with iterative and collaborative qualitative content analysis, where final sub-codes were quantified for data representations and examined statistically.

First, two researchers holistically and independently examined the case study data, creating extensive memos for each case study classroom and generating detailed analytic memos. These memos captured broad observations and emerging patterns within and across case study sites. The researchers then conducted a side-by-side comparison of the memos to illuminate both shared themes and differing interpretations of the data.

The team used a consensus-based process [59] to determine the final coding scheme based on the codebooks of two independent coders. Then, a third member of the research team was trained on the codebook and coding process and applied the final codebook to the full dataset. Both Excel spreadsheets and MAXQDA 24 (Release 24.10.0) analytical software were used for coding management and analysis, with separate tabs and segments created to organize tests by teacher and pre- to post-test time points. This structure enabled tracking of coding distribution within and across classrooms and allowed for flexible reorganization and visualization of data.

After all student models (n = 162) were coded, code frequencies were then tabulated. The unit of analysis was the unique student model at a given time point. Each student was only counted once per code, even if their test contained more than one example of the code. Thus, code counts represent unique students whose tests featured the element or relationship. These code frequencies were used to create a comprehensive diagram that is both grounded in the data and organized by the unit model (Figure 3). Frequencies of code applications were then converted to percentages to examine pre- and post-test data side-by-side, enabling observation of trends over time. Code categories that included 5% or less of students were removed from data representations. The McNemar test was then used for paired data (pre- versus post-test for each student; n = 73) using dichotomous variables (1 = code present; 0 = code not present) to examine statistical differences in the codes pre- versus post-test. Data from each case study were first analyzed separately within the case to understand any key differences between case study teachers (see Supplementary File S4). The results highlighted here are from the final stage cross-case analysis.

3. Results

Results are presented by first mapping codes over the systems model developed for the unit (Figure 3) and then examining the code frequences for pre- and post-tests side-by-side. The finalized codebooks were divided into two broad themes of individual elements (Appendix A) and energy flow through the system (Appendix B) and are shown as a systems model in the following section. Statistical analyses supporting the figures can be found in Appendix C.

3.1. Codes Mapped over Systems Model

The EYE unit was developed with a systems model that depicted the interactions between ecology, infrastructure, and buildings (Figure 1). The unit tests asked students to create one drawing that focused on how electricity arrives at and interacts with their classroom and a second drawing that focused on electrical, light, and thermal energy working together in a classroom (see Supplementary File S3). The final pre- and post-test codes (Appendix A and Appendix B) could be mapped over the unit systems model, where Figure 5 shows the relationships that emerged as major themes in the qualitative coding. Figure 6 and Figure 7 show detailed findings using a diagrammatic visualization of codes, depicting system elements (represented by boxes) and the linkages between elements (represented by arrows). Percentages for individual elements and their linkages were assigned shades in 20% increments to show which of the elements and links were most prominent at the pre-test (Figure 6) versus the post-test (Figure 7). For example, the element “Power Plant” was found in 21–40% of the tests pre-unit and in 41–60% of the tests post-unit.

The next layer added to these diagrams is the student identification of individual elements as either energy-efficient (EE; represented by “+”) or non-EE elements (represented by “−”). For example, artificial lights might be annotated by students as being EE because of the energy-efficient light bulbs, or they could have been annotated as non-EE compared to daylight. Percentages for EE and non-EE elements are shown in blue shades in 6% increments ranging from lighter (lower percentage) to darker (higher percentage) to show which parts of the model were most prevalent across student tests from pre- to post-test. This visualizes, for example, the growth in the percentage of students identifying artificial lights as non-EE.

Taken together, the Figure 6 and Figure 7 diagrams show the parts of the unit’s system model that are ‘lighting up’ for learners collectively at pre- versus post-test conditions. More detailed descriptions of each individual element, and the energy flows between them, are given in the codebooks in Appendix A and Appendix B. The sections below go in more depth on the key shifts evidenced in student tests from the beginning to end of the unit. In the sections to follow, the diagrams above are presented under two main themes:

(1)

The Energy Grid

a.

Natural resources and produces electricity

b.

Transfers electricity

c.

Environmental impacts of built environments

(2)

The Energy-Efficient Classroom

a.

How the classroom uses electricity

b.

Light energy and building design

c.

Thermal energy and building design

d.

Energy flow through the classroom

e.

Describing energy-efficient buildings

3.2. The Energy Grid

Module 1 of the EYE unit covered “energy systems” (see Figure 2, unit overview), which included an overview of renewable and nonrenewable energy sources that power the built environment. Students’ pre- to post-tests showed growth in describing the energy grid—from raw sources to methods of transmission. The sections below explore the changes in student tests in the categories of “natural resources,” “produces electricity,” “transfers electricity,” and “environmental impact,” as seen in Figure 6 and Figure 7.

3.2.1. Natural Resources and Produces Electricity

Pre-tests demonstrated a generally low grasp of natural resources being used for electrical energy. While more than 70% of pre-tests showed the sun as a source of light energy, the pre-tests had low percentages of codes for fossil fuels (20%), wind (17%), and other (such as water; 5%) being used as sources of electrical energy. In turn, the pre-tests also had low percentages of power plants (24%), wind turbines (23%), solar panels (18%), and other (such as dams; 4%) being used as sources of electrical production (Figure 8). While the percentage of post-tests featuring power plants increased by 18 percentage points (p < 0.05; Figure 8), the incorporation of solar energy was the largest shift from pre- to post-test with a 23 percentage point increase in mentions of solar panels (p < 0.001) and an 18 percentage point increase in the code for sun as a source of light energy (p < 0.05). This shift to include solar panels is exemplified in the pre- and post-tests by Austin (Figure 9) and Quincy (Figure 10) in Faith’s class (School 1). Another student in Faith’s class, Skylar, showed growth as well. Her pre-test mentioned an “outside force supplying energy,” but her post-test described how the “solar panel redirects the sunlight to the bulb for energy, curtains blackout extra sunlight, [the] electrical box directs energy to [the] outlet, [and] insulation holds temperature.” Paired-sample statistical tests examining differences in “natural resources” and “produces electricity” codes from pre- to post-test are reported in Appendix C,

Table A3. ‘Natural resources’ and ‘produces energy’ codes paired-samples proportions tests, pre- vs. post-test.

Code Label	Sub-Code (If Applicable)	Difference in Proportions (Pre–Post-Test)	Asymptotic Standard Error	Z	Sig. (Two-Sided)
Natural Resources	Fossil Fuels	0.068	0.049	1.387	0.166
	Wind	0.123	0.043	2.714	0.007 *
	Sun (L)	−0.123	0.058	−2.065	0.039 *
	Other	−0.959	0.023	−8.367	0.000 **
Electrical Production	Power Plant	−0.192	0.066	−2.746	0.006 *
	Wind Turbine	0.164	0.043	3.464	0.001 **
	Solar Panel	−0.260	0.064	−3.657	0.000 **
Natural Resources → Electrical Production	Fossil Fuels → Power Plant	−0.027	0.033	−0.816	0.414
	Wind → Wind Turbine	0.123	0.043	2.714	0.007 *
	Sun (L) → Solar Panel	−0.192	0.054	−3.300	0.001 *

(L) = Light energy; * p < 0.05; ** p < 0.001.

.

In interviews, two of the case study teachers discussed how they enhanced the EYE unit with additional foundations for energy production. Sophia described an in-depth conversation her class had regarding power plants and how they generate power. She also utilized a storybook to explain the process that she said students “loved.” This focus was showcased in her students’ models and drawings, as they demonstrated a more detailed depiction of energy generation via energy flow on their post-tests (see Supplementary File S4 that shows results by teacher). Charlotte (School 2) said that the greatest struggle when introducing the EYE unit was the lack of background knowledge her students had regarding energy components and vocabulary. She engaged students with stations, vocabulary cards, and collaborative activities to help increase their content knowledge before starting the more hands-on parts of the unit.

3.2.2. Transfers Electricity

At the beginning of the unit, students were already demonstrating an understanding of transmission of electricity into buildings. The tests asked students to draw and explain how energy gets into their classrooms, and many pre-tests clearly showed power lines (65%) and/or underground wires (69%; Figure 11). It was unclear from most tests whether students truly understood the infrastructure of buried electrical lines or if they were making a diagrammatic guess that electricity arrives from underground. Many such drawings showed the wires going directly to a wall outlet. In post-tests, students shifted toward power lines (81%; p < 0.05) over underground wires (37%; p < 0.001; Figure 11), likely because the EYE unit visuals emphasized aboveground power lines. However, it is possible that student concepts of energy transmission sharpened from vague lines in the ground to power lines as identifiable features of the energy grid. The idea of a transformer was represented in roughly equal proportions pre- to post-test, with just under a third of students including the idea of a transformer, breaker, or electrical box playing a role in transferring electrical energy (Figure 11). From pre- to post-test, students increasingly showed links between power plants and power lines (increasing from 8% to 31% of tests demonstrating this link; p < 0.001; Figure 11). This showed that students were beginning to connect the dots between energy production and transmission. The pre- and post-test drawings from Ren (Faith’s class) exemplify a student test that showed this kind of growth (Figure 12). Paired-sample statistical tests examining differences in “transfers electricity” codes from pre- to post-test are reported in Appendix C,

Table A4. ‘Transfers electricity’ codes paired-samples proportions tests, pre- vs. post-test.

Code Label	Sub-Code (If Applicable)	Difference in Proportions (Pre–Post-Test)	Asymptotic Standard Error	Z	Sig. (Two-Sided)
Electrical Transference	Power Lines	−0.137	0.056	−2.357	0.018 *
	Underground Wires	0.288	0.068	3.772	0.000 **
Transformer	-----	0.041	0.060	0.688	0.491
Electrical Production → Electrical Transference	Power Plant → Power Lines	−0.219	0.056	−3.578	0.000 **
	Wind Turbine → Underground Wires	0.041	0.030	1.342	0.180
	Power Lines → Underground Wires	0.137	0.062	2.132	0.033 *
	Power Lines → Transformer	−0.027	0.043	−0.632	0.527
Electrical Transference → Electrical Transference	Underground Wires → Transformer	0.082	0.054	1.500	0.134
	Transformer → Underground Wires	0.014	0.053	0.258	0.796
	Transfers Electricity → Uses Electricity	−0.027	0.019	−1.414	0.157

* p < 0.05; ** p < 0.001.

.

3.2.3. Environmental Impacts of Built Environments

The test prompted students to include environmental impacts in their drawings but left it open for students to determine what to show and how to show it (see Supplementary File S3). There was significant growth in the inclusion of environmental impacts as part of the energy system from pre- (38%) to post-test (71%; p < 0.001; Figure 13). The explicit link between power plants and environmental impact also increased significantly by 21 percentage points from pre- to post-test (p < 0.05), with students increasingly describing how energy production impacts the natural environment. Paired-sample statistical tests examining differences in “environmental impact” codes from pre- to post-test are reported in Appendix C,

Table A5. ‘Environmental impact’ codes paired-samples proportions tests, pre- vs. post-test.

Code Label	Difference in Proportions (Pre–Post-Test)	Asymptotic Standard Error	Z	Sig. (Two-Sided)
Environmental Impact	−0.342	0.062	−4.642	0.000 **
Power Plant → Environmental Impact	−0.205	0.075	−2.611	0.009 *

* p < 0.05; ** p < 0.001.

.

The identification of environmental impacts is seen through multiple representations, including clouds labeled as “pollution” or “greenhouse gases,” deforestation and cutting of trees, and animals dying. Post-test drawings from Karson (Faith’s class; Figure 14) and Aaron (Sophia’s class; Figure 15) exemplify this code, as they added pollution near the energy sources in their drawings. Oakley (Charlotte’s class) took another approach, drawing how power lines affect trees (Figure 16). In addition to environmental concerns, students discussed cost concerns for energy sources. Riley (Faith’s class) noted the importance of energy-efficient buildings “to keep the electric bill low.”

In teacher interviews, Charlotte (School 2) expanded on the full-class discussions that her students had about climate change. She described her previous lessons on climate change as an “afterthought,” explaining how there is not a great place to fit them into her usual coursework. She expressed excitement to expand upon information that was relatively new for her students, so problem-solving discussions regarding carbon emissions, pollution, and environmental concerns were an important part of her approach. Student outcomes after taking part in the EYE unit highlighted the emphasis certain teachers made on the environmental impacts of built environments. Sophia (School 3) discussed how her 9th graders had a strong background in understanding fossil fuels and their contribution to greenhouse gas emissions from prior classes. However, based on the results of their pre-tests, she had the lowest percentage of students mentioning environmental impact (31%), whereas 47% of Faith’s students (School 1) and 43% of Charlotte’s students noted environmental impact during the pre-tests. That said, Sophia’s class also had the highest gains in this code, as their percentage points for environmental impact increased by 42, with 73% of students demonstrating this code in the post-tests (see Supplementary File S4).

3.3. The Energy-Efficient Classroom

The diagrammatic results in Figure 6 and Figure 7 show a grey box entitled “classroom” that contains the codes for the electrical demands and building elements within classrooms. This box represents the building as a system unto itself, full of rich interconnections between design elements, energy flow, and natural resources like the sun. The sections below describe how students demonstrated growth in their conceptualizations of energy flow through the classroom.

3.3.1. How the Classroom Uses Electricity

Communicating components within the classroom that demand electricity was straightforward for students. Nearly every test (pre and post) included artificial light (99%), and outlets were also a common feature of drawings (90% pre-test and 83% post-test; Figure 17). Further, student tests showed that most students were highly aware that buildings have heating, ventilation, and/or air conditioning (HVAC; 88% pre-test and 82% post-test; Figure 17). Interestingly, these code percentages decreased for each classroom in the post-test assessment, but not significantly (p > 0.05; Figure 17). This is possibly a reflection of student attention being focused on the new material they had learned for the post-tests rather than the material that they already knew. However, some students, like Taylor (Sophia’s class; Figure 18) and Casey (Charlotte’s class; Figure 19), elaborated on the classroom electrical demands from pre- to post-test. Additionally, each classroom showed gains in understanding that artificial lights produce not just light energy, but thermal energy as well (p < 0.05; Figure 17), with Charlotte’s students gaining 15 percentage points, Faith’s students gaining 13 percentage points, and Sophia’s class gaining 3 percentage points in this code (see Supplementary File S4). Paired-sample statistical tests examining differences in “uses electricity” codes from pre- to post-test are reported in Appendix C,

Table A6. Electricity use in the classroom (‘uses electricity’) codes paired-samples proportions tests, pre- vs. post-test.

Code Label	Sub-Code (If Applicable)	Difference in Proportions (Pre–Post-Test)	Asymptotic Standard Error	Z	Sig. (Two-Sided)
Electrical Use	Artificial Light (L)	0.000	0.019	0.000	1.000
	Artificial Light (T)	−0.110	0.053	−2.00	0.046 *
	Outlets	0.068	0.045	1.508	0.132
	HVAC	0.027	0.047	0.577	0.564

(L) = Light energy; (T) = thermal energy; * p < 0.05.

. The following sections present further growth in understanding the building features that affect energy efficiency.

3.3.2. Light Energy and Building Design

Module 2 of the EYE unit covers light energy and includes a building tour for students to examine windows in their own building, followed by hands-on lessons with artificial light calculations and the measuring of daylight through the windows with light meters. After experiencing the EYE unit, students showed strong growth in understanding the role of windows in a building’s energy efficiency. Pre-tests showed that over half of the students (67%) drew and understood the importance of windows for daylight, while 79% demonstrated this importance in the post-test. Over time, students demonstrated a more detailed and nuanced understanding of how window design and window placement affect light energy in the classroom. The “window design (L)” code included elaborations on how window design features can permit or obstruct light energy, with mentions of curtains or window tinting. While roughly a third of students showed window design features impacting light energy in the pre-test, 59% of students showed this understanding in the post-test (p < 0.001; Figure 20). Post-test drawings from Ren (Faith’s class; Figure 21) and Hallie (Sophia’s class; Figure 22) showed examples of the elaborations on window design in the post-test compared to the pre-test, where items like curtains, sunlight, and artificial light energy were newly annotated in post-drawings. Additionally, while no student identified sunlight as being related to energy efficiency in the pre-test, 24% of students annotated their post-tests with an understanding of this connection (see Figure 6 and Figure 7). Paired-sample statistical tests examining differences in “light energy” and “window design” codes from pre- to post-test are reported in Appendix C,

Table A7. ‘Light energy’ and ‘window design’ codes paired-samples proportions tests, pre- vs. post-test.

Code Label	Sub-Code (If Applicable)	Difference in Proportions (Pre–Post-Test)	Asymptotic Standard Error	Z	Sig. (Two-Sided)
Natural Resources	Sun (L)	−0.123	0.058	−2.065	0.039 *
Electrical Use	Artificial Light (L)	0.000	0.019	0.000	1.000
Building Elements	Windows (L)	−0.110	0.066	−1.633	0.102
	Window Design (L)	−0.274	0.071	−3.536	0.000 **
Natural Resources → Building Elements	Sun (L) to Windows (L)	−0.137	0.065	−2.041	0.041 *
	Sun (L) to Window Design (L)	−0.082	0.038	−2.121	0.034 *
Electrical Use → Building Elements	Artificial Light (L) to Windows (L)	−0.055	0.051	−1.069	0.285

(L) = Light energy; * p < 0.05; ** p < 0.001.

.

Several case study teachers discussed the way sunlight and windows were incorporated into their teaching of the EYE unit, which supports interpretation of students’ improvement in this area. Both Faith (School 1) and Sophia (School 3) discussed cardinal directions and window placement during their school building tours. Sophia showed students a large room with no windows to demonstrate the impact that windows can have on a room’s lighting. She also did stations in her classroom to teach students more about window placement and how the sun impacts windows. Faith and Sophia’s dedication to this part of the unit was shown in their individual classroom results, as Faith’s students gained 13 percentage points in sunlight and 18 percentage points in sunlight transferring through windows, while Sophia’s students gained 26 percentage points in sunlight and 16 percentage points in sunlight transferring through windows. Additionally, Sophia’s students showed the greatest gains in understanding about window design features impeding sunlight, with a 10-percentage-point increase, along with the greatest gains in understanding that light energy put off by artificial light sources can leave the classroom through windows, having a 14-percentage-point increase. Supplementary File S4 shows these student results by teacher.

3.3.3. Thermal Energy and Building Design

EYE Module 3 addresses thermal energy in buildings, with an emphasis on the ways in which thermal energy travels through windows and leaks in the building envelope (along with how design strategies like insulation and double-pane windows can support energy efficiency). The tests asked students to identify and explain the components that keep a classroom at a constant temperature. As mentioned previously, students began the unit with a high level of awareness about HVAC systems. As they learned more about thermal design, identification of HVAC systems dropped insignificantly by 6 percentage points, while significant increases were evidenced in insulation (46% to 78%; p < 0.001) and thermal energy traveling through windows was trending upward (46% to 58%; p > 0.05; Figure 23).

Post-tests had a stronger emphasis on the role of windows and doors in the transfer of thermal energy in buildings. Mention of windows or doors being included in the system of thermal energy flow to and from the classroom increased, but not significantly, by 12 percentage points (p > 0.05; Figure 23), and recognition of window design features (double panes and curtains) impeding thermal energy flow significantly increased by 17 percentage points (p < 0.001; Figure 23). By the end of the unit, students were increasingly depicting the role of the sun moving through windows and impacting the thermal conditions inside the building as well, increasing from 23% to 27%, though this increase was not significant (Figure 23). Paired-sample statistical tests examining differences in “thermal energy” and “building design” codes from pre- to post-test are reported in Appendix C,

Table A8. ‘Thermal energy’ and ‘building design’ codes paired-samples proportions tests, pre- vs. post-test.

Code Label	Sub-Code (If Applicable)	Difference in Proportions (Pre–Post-Test)	Asymptotic Standard Error	Z	Sig. (Two-Sided)
Natural Resources	Sun (T)	−0.068	0.065	−1.043	0.297
Electrical Use	HVAC	0.027	0.047	0.577	0.564
Building Elements	Windows (T)	−0.137	0.085	−1.581	0.114
	Window Design (T)	−0.164	0.043	−3.464	0.001 *
	Insulation	−0.288	0.060	−4.200	0.000 **
	People (T)	0.000	0.034	0.000	1.000
Natural Resources → Building Elements	Sun (T) → Windows (T)	−0.123	0.077	−1.567	0.117
	Sun (T) → Window Design (T)	−0.055	0.027	−2.000	0.046 *
	Sun (T) → Insulation	−0.137	0.049	−2.673	0.008 *
Electrical Use → Building Elements	HVAC → Windows (T)	−0.027	0.067	−0.408	0.683
	HVAC → Insulation	−0.082	0.074	−1.095	0.273

(T) = thermal energy; * p < 0.05; ** p < 0.001.

. Charlotte’s students showed the most gains in these areas, with a 14-percentage-point increase in recognition of the sun putting off thermal energy, a 43-percentage-point increase in recognition that windows allow for thermal energy flow, and a 43-percentage-point increase in recognition that window features can impede thermal energy flow (see Supplementary File S4).

Even though “insulation” was included in the assessment word bank in both pre- and post-tests (Supplementary File S3), building insulation was featured much more prominently in student post-tests versus pre-tests. Drawings and writings about insulation increased by 32 percentage points (p < 0.001), and students began to recognize how insulation not only blocks thermal energy from the sun (6% to 21%; p < 0.05), but also how it keeps thermal energy from the HVAC system within the building (21% to 31%; p > 0.05; Figure 23). For example, Ren (Faith’s class) drew an arrow pointing to a section view of the wall in her post-test drawing (Figure 21), identifying “insulation keeping thermal energy in and keeping outside air out.” However, Sophia’s students showed the most gains in understanding how thermal energy from the HVAC system can both travel through windows (20% to 33%) and be kept inside by insultation (11% to 30%; see Supplementary File S4). Furthermore, in the pre-tests, only 2% of students identified insulation as energy-efficient, whereas 15% of students identified this connection in the post-tests (see Figure 6 and Figure 7).

Table 2. Sample student writings for “Identify and explain how the features of your classroom keep the temperature constant throughout the year,” pre- versus post-test.

Student (Teacher)	Pre-Test Response	Post-Test Response
Harrison (Faith)	“Turning the [thermostat] down or open a window.”	“The A/C or windows and [insulation] The fan can help as well.”
Elizabeth (Faith)	“The electricity from powerlines and the electrical box contains energy and is given energy to the [thermostat] so the teacher can keep a constant [temperature] year round.”	“The [insulation] traps heat and the teacher uses the air conditioner or heater to keep the classroom at a constant value or [temperature]. Some lights also cause a little bit of heat.”
Christian (Sophia)	“Electricity gives off heat and tries to keep the same temperature the same by using different amounts of energy.”	“The insulation do well to keep heat in and keep heat out. HVAC provides heating or cooling to also get it around the right temperature.”
Leah (Sophia)	“Inside of a classroom there is AC and heating machines.”	“It depends on where you put the windows because certain spots in the room bring in different amount of thermal energy.”
Oakley (Charlotte)	“By allowing air to flow during summer months, it allows for heat to more easily escape and maintain a more mild temperature. Fans also help with too much heat. During winter months however, to heat the room we use heaters and vents.”	“Insulation traps in heat, the classroom has insulation. A heater may be used or an air conditioning unit, one to heat the room and the other to cool it. By using insulation less heat escapes to the outside world and keeps the room at a more constant temperature, but by using artificial heat and artificial coolers we are able to hold the classroom at a more consistent temperature.”
Alexis (Charlotte)	“Insulation.”	“It can keep it warm or cool from the insulation in the walls, the windows, heaters, AC, humidifiers, etc. or the light bulbs that get hot and warm a room.”

elucidates these descriptive statistics with examples of how students across case study schools enhanced their explanations over time to the write-in test question: “Identify and explain how the features of your classroom keep the temperature constant throughout the year” (see the prompt in Supplementary File S3).

3.3.4. Energy Flow Through the Classroom

The tests also asked students to map energy flow (electrical, light, and thermal) through the classroom before and after the EYE unit and to then describe their drawings in words.

Table 3. Sample student responses when asked to explain energy flow in the classroom, pre- versus post-test.

Student (Teacher)	Pre-Test Response	Post-Test Response
Norville (Faith)	“The heat flows out the [window] and lets in cool breeze.”	“Thermal energy is flowing throughout the classroom because the body heat of every person. Light energy flows into the classroom by the sun shining. Electrical energy is going to the outlet from the telephone pole.”
Skylar (Faith)	“There is an outside force supplying energy.”	“Solar panel redirects the sunlight to the bulb for energy, curtains blackout extra sunlight, electrical box directs energy to outlet, insulation holds [temperature]. “
Calloway (Sophia)	“Lights and smartboard is connected to wires which let them work. Air purifier connected to outlet and outlets provides energy for air purifier.”	“Wires are connected to the lights and outlet. The light produce light and heat for the classroom. The outlet takes in a plug and produces electricity for the smartboard and computers which also produces light and heat.”
Connie (Sophia)	“Light is flowing through windows, into the plants, and once the plant dies, it goes back outside and energy is recycled.”	“My drawing is showing energy flowing through a classroom. The light and thermal energy flows in through the windows mostly, flows through the classrooms, and bounces into the hallways and leave through multiple exits. Electrical energy enters through the power plant and flows into the building before it is converted to light or thermal.”
Devan (Charlotte)	“An electric pole flowing energy through the lights and out.”	“It is a building with lights and energy can flow out [cause] of the windows and doors and in energy boxes.”
Casey (Charlotte)	“Ceiling lights with wires. Air conditioning flowing in. Sunlight from the window. Technology on whiteboard.”	“It is showing electrical, thermal, and light energy. It is flowing thermal energy into from the windows, it is flowing through with light energy, and electrical energy is flowing out because we’re using it.”

shows sample student writings for the prompt, comparing pre- and post-tests. Some students, like Connie (Sophia’s class) and Devan (Charlotte’s class), described how energy can both enter and leave the classroom (Table 3). Connie specifically stated that “Electrical energy enters through the power plant and flows into the building before it is converted to light or thermal.” Here, Connie described light, thermal, and electrical energy flow in her elaborate post-test response (Table 3). Devan, on the other hand, began with power lines in his pre-test response and increasingly incorporated building elements into the post-test (Table 3).

3.3.5. Describing Energy-Efficient Buildings

The final part of the tests contained a write-in portion that asked students to describe an energy-efficient building and the purpose of energy efficiency. In the pre-tests, students demonstrated limited ability to describe an energy-efficient building, as they often described it as a building that created energy or were entirely unsure of how to describe it.

The responses in

Table 4. Student responses when asked to explain energy-efficient buildings, pre- versus post-test.

Student (Teacher)	Pre-Test Response	Post-Test Response
Riley (Faith)	“It is a building that for its [size] the electric bill is low.”	“An energy-efficient building is a building that can stay at a [constant] temperature by using minimal amounts of energy. It is important so the electricity bill is low.”
Jessica (Faith)	“No idea what that is. It sounds environmentally friendly.”	“A building that uses technology with little energy waste. It matters because it causes a lot of pollution.”
Cayden (Charlotte)	“I’ll look into it. I think it is energy and how it builds up.”	“It is a building you use less energy in.”
Alexis(Charlotte)	“Houses with solar panels.”	“A building with more curtains, LED bulbs, more energy-efficient things, etc.”
Connie (Sophia)	“An energy-efficient building is a building that uses energy efficiently without wasting it. This matters because it is good for the environment to use efficiently.”	“Energy-efficient buildings are buildings that do not use a lot of energy and they matter because energy is finite and we need to protect our Earth by conserving it.”
Calloway (Sophia)	“School because it uses a lot of electricity and wires in order for kids to learn.”	“An energy-efficient building is a building that runs [effectively] on energy while not exerting a lot of energy making it have a cheaper electrical bill.”

highlight sample student written explanations of why energy-efficient buildings are important and support the increase in understanding of what it means to be an energy-efficient building across the pre- and post-tests. Jessica (Faith’s class) had “no idea what that is,” but wrote “it sounds environmentally friendly” (Table 4). In the post-tests, students increased their vocabulary and descriptions, often indicating a reduction of energy use to be a major component in energy efficiency. Students, like Jessica and Riley (Faith’s class; Table 4), referred to cost concerns and environmental concerns to explain why energy-efficient buildings are important. Overall, the students connected a reduction of energy use with the reduction of natural resource use and fossil fuel production, and the overall increased health of the environment. Jessica’s understanding of an energy-efficient building shifted in the post-test, where she described an energy-efficient building as “a building that uses technology with little energy waste. It matters because it causes a lot of pollution” (Table 4). Post-test responses included better integration of building features as well.

While Faith and Sophia’s classes emphasized the energy benefits of solar panels in post-tests, Charlotte’s class recognized the inefficiency of the HVAC system. Additionally, Charlotte’s focus on energy efficiency design vocabulary and definitions may have enhanced her students’ ability to respond to this question in the post-tests.

4. Discussion

This study explored the progression of student ideas from the beginning to the end of a place-based curricular unit focusing on energy flow between built and natural environments. The qualitative content analysis of 162 pre- and post-test systems models for 89 rural students (grades 7–9) revealed the mental models of learners at the beginning and end of the unit. Teacher interviews supported the interpretation of student data. Below, the findings of the current study are linked with precedent scholarship.

4.1. Thinking with Systems

Nordine et al. [60], in their review of teaching energy as a system, promote the idea of curriculum in which students learn to trace energy forms within and across systems, which they call a “systems transfer approach” (p. 178). Tracing energy across systems supports students in making sense of dynamic changes within systems connected to energy. The results of the current study illuminate these kinds of traces across built and natural systems.

The first research question for this study was: What changes in ST, if any, were evidenced in student systems models from the beginning to the end of a place-based unit on energy-efficient building design? The work here contributes to the growing evidence that model-based learning can support ST for youth in grades 7–9 [45,61]. Our findings can be interpreted considering the Cabrera and Cabrera [44] DSRP model. Our pre- to post-test results suggest that students increased in the distinctions (D) represented in their models, as evidenced in the “Individual Elements” Codebook (Appendix A). The ability to identify and make distinctions of visible and hidden system components is foundational to organizing and linking components [39,43]. Results also suggest that students organized their represented components into systems (S) and included relationships (R) in which they linked and traced the flow of energy from the external to the internal environment (“Energy Flow” Codebook, Appendix B). In addition, students began to consider the presence of feedback loops, as evidenced by consideration of environmental impacts resulting from energy system flow. Incorporating feedback loops into a systems model requires recognizing hidden system dimensions, such as system levels and time scales [39]. Yet, 6–12th graders and undergraduate students rarely consider feedback within a system [62]. Considering that relationships lead to system feedback and the impact that feedback has on system function is a critical steppingstone for systems thinking development [39]. Our results suggest that, in the post-test, students were beginning to consider feedback from energy systems.

4.2. Energy Flow Between Built and Natural Environments

Tsurusaki and Anderson [63] found that middle-school-aged students rarely considered the relationship between human engineered systems and natural systems or the impact that human activities have on natural systems. When specifically asked about energy, students only understood that some appliances (such as a dishwasher) use “less energy”—they did not consider that there was an impact on natural systems from the energy that was consumed by the appliance. There was no difference in understanding between age levels (elementary, middle, or high school) or demographics (rural, urban, or suburban). Energy systems, which span societal and natural systems, are a “black box” [63] (p. 427) for students in which they have no knowledge of how they contribute to the energy system or how societal involvement in the system impacts nature. The work in the current study sought to open this “black box” for students so they can begin to build their mental models of the complexity of energy systems and their location within that system.

The second research question asked: Did, and how did, students demonstrate systems thinking (ST) in their systems models depicting energy flow between buildings and the natural environment? Reviewing the changes in “energy flow” codes across the results evidenced the shifts that occurred for learners from pre- to post-test. Particularly, results showed that students not only improved their connections on the energy grid (especially regarding solar energy and the flow from power plants to transmission lines) but also their systems models of how energy flows within the classroom. By the post-test, more students were weaving together natural elements (e.g., the sun) and building elements (especially windows and insulation) to explain how classrooms have functional light and maintain thermal comfort. Importantly, learners also connected this system back to impacts on the natural environment. Interestingly, students collectively reduced their depictions of fossil fuels, wind energy, and “other” energy sources (e.g., dams) in the post-tests. The strong emphases in the EYE unit of sun orientation and solar energy may have redirected student focus ahead of the post-test.

The results of the current study align with findings with upper elementary students who experienced a water literacy unit focused on green roofs [37]. Students drew systems models across the unit and evidenced enhanced understanding of socio-hydrologic systems—with increases in both individual elements and the ability to combine elements [37]. The current study shows that older students are doing more than combining elements—they are accomplishing multi-scalar systems models that shift from the building out to the energy grid. This finding aligns with learning progression research illuminating how student ideas about energy flow are integrated into “complex networks of ideas” [13] (p. 3).

4.3. Green Building Knowledge

This work contributes to the emerging research on green building literacy [6], where there is much to be learned about effective green building education for youth and the general public. While some early work has shown the benefits for building occupants who experience green buildings [32,64], curricular interventions have the potential to expand green building education far beyond this small group. Most sustainable building curriculum evaluation has occurred in higher education with undergraduate students already interested in sustainable design professions [29]. Other work has focused on learner engagement in virtual green building games [65] or focused on engineering design processes [34]. Green building knowledge categories have been conceptualized in theory [6], but operationalizing this outcome proves challenging. In place of pursuing multiple-choice green building knowledge tests, the current study used an assessment that integrates systems modeling, ultimately combining verbal and visual data for each learner over time. This assessment technique lends to a more nuanced qualitative analysis that illuminates ST with factual and conceptual green building concepts. In this way, the current study contributes methodologically to the study of green building knowledge outcomes.

One interesting learning outcome in the current study was increased demonstrations of the sun’s role in building design, which was notable across both the light and thermal energy findings. This was especially evident in the codes related to window design (orientation, shading, double-pane glass, etc.), including the connections students made to the sun, artificial lights, and HVAC performance. The interaction of the sun with building design and materials is an advanced concept more commonly covered in college classrooms (i.e., architecture and engineering students learning to model the impacts of daylight on illumination and thermal conditions inside the building). The current study shows how the foundations of this understanding can be provided to middle and early high school learners. Our previous work on a solar energy unit suggests that these foundations can begin in upper elementary with targeted curriculum and may be enhanced by physical examples of solar panels on the school campus [33].

4.4. Learning Rooted in Place

Prior work suggests that youth struggle to apply energy learning from the science classroom into daily life [15]. A unique feature of the EYE unit was the use of the students’ own school building in learning activities. Our results showed that 7–9th graders had an overall better command of the term “energy efficiency” by the end of the EYE unit. This is an expected outcome of a unit that focuses intensely on the concept. However, increased understanding of the term shows that scientific classroom learning can translate to application beyond the classroom.

One highly related study involved middle school renewable energy education with place-connected experiential activities that resulted in more positive perceptions of renewable energy [66]. The current work builds on the Buldur et al. [66] study by examining knowledge outcomes across a larger energy system. Interestingly, in the current study, most areas of growth were elements of the system that were tangible in students’ daily lived experiences, like windows, lights, and temperature control, indicating that ST may begin with the most familiar elements in one’s mental map. This finding suggests that place-based units employing the nearby environment have potential to support ST. These results fall into the larger discourse of place-conscious environmental education [67].

The current study was situated in the context of rural schools in the Midwestern U.S. The current dataset did not allow comparison to non-rural schools to understand if rural students demonstrated unique patterns. While guidance is scarce for climate change pedagogy in rural communities, one study in rural areas of New York and Kentucky illuminates the potential for advancing climate education by addressing local values [68]. The EYE unit in the current study emphasizes the science of energy flow and allows students and teachers to arrive with a diversity of values (e.g., human health, economics, environment, etc.). The grounded analysis in the current study, with the resultant codebooks, did not surface themes specifically related to rural environments. However, future work will analyze rural teachers’ experiences in the EYE unit, exploring how teacher connection to place may impact how teachers incorporate energy systems into the unit. Future work can also explore the impact of place-based energy units on urban and suburban learners who may have better access to green infrastructure but also may be more distanced from energy sources and production facilities.

4.5. Limitations and Future Directions

This study is limited to three case study rural schools in the Midwestern U.S. As such, care should be taken in transferring findings to other contexts. Further, there was variation in case study teachers in terms of years of experience and pre-familiarity with the curriculum, where one teacher (Sophia) had worked in the lab where the EYE unit was developed. Teacher and school contexts are elucidated in Supplementary File S1, and comparisons of student data by teacher are available in Supplementary File S4. Nevertheless, these teachers represent some of the natural variation that will occur in naturalistic settings when a curricular unit is enacted. Our data are also limited to the timeframe in which the unit was taught, where future longitudinal studies would be needed to understand the stickiness of learning content over time.

Additionally, the current work emphasizes the paper tests with supporting data from teacher interviews but does not feature student interview data. Analyzing the full set of student tests (n = 162 models; n = 89 students) allowed for a larger sample size than what can be accomplished with a study design that employs interviews or focus groups. The current analysis builds on an initial, smaller analysis that included pre- and post-tests paired with student interviews (n = 33) [69] that was beyond the scope of reporting in the current paper. The current work can inform future, increasingly quantitative mixed-methods work with higher sample sizes and control groups.

Future studies can additionally integrate both computational and systems thinking into sustainability education. While the current EYE unit contains several multimedia elements (e.g., videos and Phet simulations), it remains dominantly an analog unit relying on hands-on activities and physical modeling. The next evolution of the unit is being developed with increasingly sophisticated digital tools. These tools will include interactive simulations that allow students to manipulate design options and observe how those decisions impact system performance [70]. Recent work shows that such digital environments support ST education by helping learners iteratively test ideas and visualize dynamic feedback [71]. Integrating computational thinking, including algorithmic reasoning and ability to break complex problems into smaller parts, can further strengthen students’ understanding of how systems function at different levels. It also prepares them to engage with open-ended environmental challenges that involve data use and uncertainty [72]. Providing digital tools for green building literacy at the middle school level expands prior work on green building games, such as Reinhart et al. [73], to better understand systems thinking as an outcome developed alongside green building science.

5. Conclusions

To be conversant in 21st century global environmental issues, learners need to understand how individual choices impact societal and natural systems [74]. The built environment provides a compelling portal for investigating human impacts on Earth systems and provides a way to ground energy education in applied, everyday settings, such as the student’s own classroom. Energy flow in and through built natural environments is a particularly rich subject matter for rural learners who are likely to see energy infrastructure such as mines, fracking sites, wind turbines, and power plants in their own communities. Rural learners may also be less likely to experience certified green buildings that tend to cluster in urban centers, which means that curricular interventions like the EYE unit have potential to expand exposure of green building concepts to learners outside of major cities. The current study shows that such curriculum has the potential to seed foundational energy efficiency concepts in connection with the surrounding energy grid and natural environment. Such learning not only introduces students to career paths in green design and energy infrastructure, but it may also have practical implications for everyday use of buildings. A green building literate citizenry may be empowered to make better energy choices within buildings and advocate for energy-conscious design in their own communities.