Introduction

The Next Generation Science Standards (NGSS) have identified several cross-cutting topics related to all areas of science that integrate thinking skills from other disciplines. One of these concepts is Systems and Systems Thinking, which focuses on understanding a system well enough to predict how it will behave (National Research Council 2013). The challenges of the twenty-first century are complex and interdependent and require educational pedagogies to prepare students to respond to the interconnected economic, social, scientific, political, economic, and ethical aspects of a transition to sustainability (Everett 2008). Population growth, climate change, and resource consumption place increasing pressure on food, energy, and water systems, resulting in scarcities and inequities. Understanding the complex trade-offs and interlinkages between these environmental and human systems is at the heart of interdisciplinary environmental and sustainability (IES) programs and represents a quintessential application of systems thinking. These system linkages are typically obvious to instructors; however, teaching connections between socio-ecological systems and facilitating systems thinking abilities in students remains challenging (Vasconcelos et al. 2020). Environmental knowledge assessment may help us to determine what additional learning needs to be done in creating environmentally literate citizens, which is also an important public policy task (Sanford et al. 2017).

One difficulty for IES programs trying to cultivate systems thinking skills is the breadth of topics that are highly variable across courses, programs, and institutions (Vincent et al. 2013). The Food-Energy-Water (FEW) Nexus is a recent and inherently interdisciplinary paradigm that can provide scaffolding to facilitate systems thinking in students. Another difficulty to teaching systems thinking is the challenge of making implicit student mental models explicit for instructors. In other words, instructors may assume students are able to think in terms of systems without being able to concretely identify which, if any, interlinkages students understand. Drawing is a visual means of making internal, “hidden” mental models explicit (Jonassen et al. 2005). While drawings are often termed sketches, diagrams, visuals, models, concept maps, or illustrations, here, we adopt an inclusive definition and refer to drawings as “a learner-generated external visual representation depicting any type of content, whether structure, relationship, or process” (Quillin and Thomas 2015: 14). Some studies have used drawings to identify preconceptions and measure changes in learning, but it remains an underutilized tool in recognizing student’s systems thinking abilities. Studies argue that drawing should be recognized along with reading, writing, and speaking in science education to enhance engagement, foster learning, and increase ability to communicate science (Ippolito and Pazio 2019; Wilson and Bradbury 2016: 2). This is especially true considering different learning preferences that allow students to engage with materials in different formats (e.g., auditory, tactile) while also attending to the potential learning gains from multimodal instruction. Additionally, there is evidence suggesting that when students use multiple mediums and switch back and forth between them to communicate their ideas, students can further refine their thinking and ability to explain their mental models (Jewitt et al. 2001). Switching between different modalities can expand dimensions of knowledge but might also exert cognitive overload for some students.

Drawings have primarily been used in two contexts: to teach systems concepts and to assess student learning. Visual mapping tools, such as mind maps, concept maps, and visual metaphors, have been utilized as a teaching technique in academic disciplines, such as accounting, engineering, reading comprehension, biology, and medical science (Choudhari et al. 2021; Davies 2011). Mind mapping is defined as a graphic technique that allows students to imagine and explore associations between concepts in a non-linear way using words or images (FAO 2008). Concept mapping, on the other hand, is more structured and less pictorial. It provides a hierarchical tree structure among concepts, and cross-links using connective phrases, such as “leads to,” “results from,” or “part of,” are used to show connections among concepts (Davies 2011). In a case study examining student learning at university level in the UK, Hay and colleagues (2008) argued that the use of concept mapping as a visual tool puts the student as the central agent in learning and helps them make connections in their learning (as opposed to linear learning). Thus, it facilitates “meaningful learning” and adds to the quality of higher education teaching (Hay et al. 2008). Davies (2011:12) further elaborates on meaningful learning and argues that it “occurs when new perspectives are integrated into the knowledge structure and prior concepts of the student.” In a review article, Ainsworth and Scheiter (2021) describe how different forms of drawing can be integrated into teaching and learning and how visual representations (from drawing to current digital forms) can provide advantages for student learning (from reducing memory demands to externalizing information required for problem-solving).

In addition to teaching, visual tools are also used to assess student learning. A recent study used mind maps to assess student learning in medical education in Saudi Arabia to understand how the shift to online education due to COVID-19 and use of new pedagogies affected students’ learning experiences (Alsuraihi 2022). Alsuraihi’s (2022: 1) study found that “students believed that they gained skills like organizing and planning, decision making, and critical thinking” from the use of visual tools in assessment of their learning. There is a growing emphasis on alternative assessments of student learning in interdisciplinary programs. Alternative assessments, such as drawings, allow students of different learning preferences to demonstrate their understanding in different ways. Drawings also allow students “to create their own personalized representation of their knowledge and perceptions” that can incorporate different prior learning experiences, cultural understandings, and disciplinary teachings (Cronin-Jones 2005: 227). Drawing enables expression of ideas and feelings under conditions of limited time and vocabulary. Thus, it is also increasingly used in science education research (Chang et al. 2020).

In this paper, we took an exploratory approach using drawings to uncover undergraduate student systems thinking capabilities of the FEW Nexus and to assess student learning in introductory environmental studies and sciences courses in US higher education institutions. The goal of our research was to study how students understand the FEW Nexus using drawings to elicit connections between resource systems. We had three research questions and several accompanying hypotheses:

  1. (1)

    Which connections are students making between food, energy, and water resource systems?

  2. (2)

    Does drawing improve student ability to make connections between food, energy, and water systems?

  • H1: Students are able to think of more FEW Nexus connections when prompted to draw than when relying on verbal explanation alone.

  1. (3)

    Does the level of abstraction influence student ability to make interlinkages between food, energy, and water systems?

  • H2: Students who draw more specific or concrete drawings are able to identify more connections than students who draw less specific or abstract representations of the FEW Nexus.

This study contributes to a growing body of research using drawings to understand student learning and systems thinking. Few studies have focused on undergraduate students or students in interdisciplinary programs, preferring to concentrate on K-12 learners in disciplinary settings (e.g., biology, geology, physics, environmental education) (Cronin-Jones 2005; Wilson and Bradbury 2016). As far as the authors know, we are the first to use drawings as a tool to elicit student understanding of the FEW Nexus in a higher education research setting.

The Food-Energy-Water Nexus as a paradigm for systems thinking

The FEW Nexus emerged after the 2008 global financial crisis and gained popularity as a research and development paradigm that accounts for synergies, interlinkages, and trade-offs between food, energy, and water resources (Leck et al. 2015; Platts et al. 2022). Whereas previous management approaches focused on a single sector, the FEW Nexus seeks to alleviate unintended side-effects and negative trade-offs by holistically considering the connections between resource systems across social and ecological systems (Al-Saidi and Elagib 2017). A FEW Nexus approach seeks to provide food, energy, and water security in the face of increasing population and consumption levels and climate change impacts (Katz et al. 2020). Integral to the FEW Nexus is the connections between ecological and social systems; however, a recent review of FEW Nexus projects found a central focus on water systems, often at the expense of food and energy systems, with limited integration of social issues related to health and sustainable development (Platts et al. 2022). Despite these shortcomings, the majority of IES courses incorporate food, energy, and water as critical topics (Horne et al. 2023). Due to its inherent reliance on systems integration, the FEW Nexus can serve as a paradigm to anchor systems thinking across socio-ecological contexts. We therefore used the FEW Nexus as a commonly covered IES phenomenon to assess student systems thinking skills.

Mental models as visual depictions of systems thinking

Science education often teaches individual system components that students must then connect to gain a holistic representation of the entire system, including interlinkages between individual components. Instructors, therefore, require students to develop systems thinking capabilities, defined here as the ability to see the entire systems, its individual parts, and how these parts relate to the whole (Cloud 2005; Grohs et al. 2018). Mastering systems thinking allows students to understand interrelationships between system parts and to comprehend the dynamic nature of the system as it undergoes change (Brandstädter et al. 2012). Systems thinking can support student ability to analyze, hypothesize, and predict system responses to various interventions, enabling more holistic decision-making in natural resource management contexts. While systems thinking is critical within interdisciplinary fields, teaching and assessing systems thinking capabilities among students remains challenging. This is, in part, due to the hidden, internal nature of student thinking and a lack of methods for assessing systems thinking ability (Brandstädter et al. 2012).

Constructivist learning theories recognize that students actively participate in their learning and continuously build mental frameworks, referred to as schemas, as they explore and understand the world (Bruning et al. 1995; Piaget 1970) and as they activate their imagination and engage with cultural experiences and beliefs (Forcino 2013). They use schemas and the relationships between schemas to construct mental models that allow them to holistically understand a phenomenon or system (Dauer et al. 2013). To create a mental model, students must select information from prior learning (e.g., personal experience, coursework, cultural practices), process and organize this information, and then integrate these elements into an explanatory model (Quillin and Thomas 2015). With these multiple steps, there is certainly room for error and misinterpretation, especially when acknowledging that all models are inexact interpretations of the world that are constantly being refined as people learn (Jonassen et al. 2005). While learners use mental models to explain how the world works, these models can be inaccurate, incomplete, not scientific, and unstable (e.g., we forget details, our models change) (Greca and Moreira 2000). In spite of this, mental models are extremely useful as learners attempt to understand connections between system components and how systems behave. Moreover, what is learned by memory alone is easily forgotten, and even when learning is possible, it may not occur, or students do not learn what was intended (Illeris 2012). Identifying and refining mental models is one way to encourage students to improve their systems thinking skills. In contrast to memorization, mental models and visual tools can facilitate meaningful learning (Hay et al. 2008).

Drawing can be a tool to make internal mental models explicit for external educators and/or researchers (Gobert and Clement 1999; Jonassen et al. 2005). Drawing is an especially important tool for novice learners to develop metacognition as they create and refine their mental models (Dauer et al. 2013), but drawings are rarely used to assess learning and systems thinking in college level interdisciplinary programs.

Drawings as a tool to assess learning and systems thinking

Drawing can be a powerful tool to facilitate learning, build systems thinking capacity, and assess mental models. Several theories support this inference, including the cognitive model of drawing construction (Van Meter and Firetto 2013) and the cognitive theory of multimedia learning (Mayer 2014). The cognitive model of drawing construction states that learners must build connections between their drawings and verbal explanations, thereby making their thinking/mental models explicit (Van Meter and Firetto 2013). Additionally, this theory argues that students who create visualizations are often more effective than verbal only explanations in supporting knowledge acquisition and assimilation (Van Meter and Firetto 2013). For example, in an experiment with elementary school physics classes, students who drew a pictorial representation of energy conservation had a significantly more detailed conceptual understanding of the law of energy conservation than students who wrote about their understanding (Edens and Potter 2003). Similarly, undergraduate biology students recalled 50% more content when they studied by generating drawings (i.e., visually modelling concepts) compared to using only visual studying techniques (i.e., reading through notes) (Heideman et al. 2017). Importantly, visual assessments have helped lower performing students reduce the achievement gap, thereby raising overall student achievement (Aylward 2010; Dauer et al. 2013).

The cognitive theory of multimedia learning posits that students learn more effectively when instructors use both words and visuals compared to verbal only instruction (Mayer 2014). According to Mayer (2014), the brain processes visual and auditory information using separate channels, and if both channels are presented with information, the brain is able to process more information than if only one channel was engaged. In other words, when paired with reading or verbal explanations of systems, students often exhibit greater knowledge through drawings. For example, fifth grade geology students who generated drawings while reading an explanatory text demonstrated greater understanding of spatial and causal system aspects than students who only read the text (Gobert and Clement 1999). Similarly, the process of generating a concept map in a sophomore environmental methods course resulted in a greater understanding of connections for high achieving learners (Proctor and Bernstein 2013). Some studies indicate that drawing reduces cognitive load; however, it is possible that pairing visual with verbal modes of explanation could result in cognitive overload whereby students demonstrate less learning and systems thinking capabilities than if one modality was selected (Heideman et al. 2017). The Proctor and Bernstein (2013) study of sophomore students found that concept mapping did not universally improve student ability to make connections, and students who were more engaged (i.e., following instructions, reflecting on the mapping process) demonstrated the most learning compared to less engaged students (i.e., those that did not deeply engage with and reflect upon the mapping activities). This might also speak to different learning preferences and familiarity with engaging with different mediums (e.g., visual, auditory, tactile) to acquire information. Given this information, we hypothesized that students would make or express more connections between food, energy, and water systems when drawing compared to using only verbal explanations of connections.

Interestingly, the approach to drawing appears to be important in student learning. Abstract drawings use symbols and representations of objects that are concise but removed from physical references. More grounded drawings (here, we use the term “concrete”) are more similar to the objects we experience in daily life (Koedinger et al. 2008). Previous studies have found that more concrete representations of phenomena are useful when solving simple problems. In contrast, more abstract representations can be more useful when solving more complex problems (Koedinger et al. 2008). Importantly, the approach to representation (abstract versus concrete) does not necessarily correspond to concrete or abstract thinking. Previous studies have examined the approach to drawing along a spectrum of concrete to abstract representation. Weinrich and Sevian (2017) found that undergraduate chemistry students displayed increasing levels of abstraction in their drawings as learning occurred throughout the semester; however, it is also possible for students to create excessively abstract representations of phenomena by “parroting” (or simply repeating) key system components heard in courses without comprehending deeper connections, indicating surface-level comprehension (Weinrich and Sevian 2017). A study of fourth and eighth grade biology students in Germany revealed that the level of directedness impacted student systems thinking; a highly directed concept mapping approach (i.e., those that give students concepts and linking words) resulted in higher systems thinking scores than non-directed approaches (i.e., where concepts and linking words are not provided) (Brandstädter et al. 2012). We can therefore infer that the level of abstraction influences student systems thinking. Our sample was composed of students enrolled in introductory IES courses, or novice learners in IES topics. Unlike concept mapping methodology where students are trained specifically to draw and incorporate concepts provided in a list, we did not train students or provide a list of concepts. We therefore hypothesized that students who chose to draw more concrete examples of the FEW Nexus would be able to make more connections than students who adopted an abstract approach, believing that introductory students would be likely to “parrot” key systems components in their drawings without understanding more nuanced connections.

Materials and methods

Participant recruitment and sample

During fall 2020–spring 2021, we interviewed 114 students at ten US institutions with interdisciplinary environmental programs. We recruited higher education institutions using professional connections and the Association of Environmental Studies and Sciences (AESS) listserv. From this voluntary sample, we selected institutions that represented a diversity of environmental programs across small liberal arts (SLAC), teaching and research (M, R2), and research-intensive (R1) universities (see Supplemental Table 1). To capture the range of responses, we interviewed students before, during, or after completion of entry level environmental courses. We recruited students on a voluntary basis by offering gift card compensation for their time and effort. For ease of analysis, the first five students who were interviewed and consented to drawing were included in this sample (6 of the 114 students declined to draw in response to the interview questions). The demographics of our subsample of 50 students reflected the demographics of our larger sample. Of the 50 students whose drawings were analyzed, 75% were female, 66% were white, and first (28%), sophomore (26%), junior (24%), and senior (22%) year students were almost equally represented. The majority of students grew up in the US (94%) with the Northeast, Southwest, and Midwest being the most common locations students had lived, respectively. Forty-six percent of students were majoring in natural resources, followed by social sciences and humanities (22%), science and math (18%), arts (8%), business (2%), and other unique majors (4%). Almost all students had previously completed at least one science college-level course (96%), which could include earth sciences, physics, chemistry, biology, and social sciences.

Interview process

To capture how students understood the FEW Nexus, we asked them to describe their experiences with food, energy, and water systems, as well as the connections between these resource systems and commonly taught environmental topics (i.e., agriculture, climate change, waste and pollution, soil health) (Table 1). During these virtual interviews, we asked students to verbally explain the connections(s) between food, energy, and water. We then asked students to draw the connections between these three resource systems before describing their drawing to the interviewer. Students were given no further instructions on their drawings unless clarification of the task was needed. Before beginning their drawing explanation, students held their images up to the camera such that the interviewer could follow their explanation. Interviews were transcribed verbatim and stored and analyzed in NVivo 1.6 © alongside corresponding images of student drawings (which students shared via email immediately following their interview).

Table 1 Overview of coding approach for verbal and visual descriptions of FEW Nexus connections
Full size table

Analysis and validation

Two coders analyzed interview transcripts for FEW Nexus components and connections and level of detail (Table 1). FEW Nexus codes included individual resource components (i.e., food only, energy only, water only) and connections between resource components (e.g., water to food, water to energy, food to energy). During a pilot phase of a larger research project, two coders developed and pre-tested a coding framework to capture FEW Nexus components and connections. We used an anthropogenic interpretation of the FEW Nexus (given that IES courses focus on coupled human and natural systems) whereby food, energy, and water were systems that provide resources to humans (e.g., as opposed to provisioning resources to wildlife). For example, students occasionally described the sun as the basis of energy for ecosystems, such as this statement: “[Y]our ultimate source of energy, as life, is the Sun.” Even though the term “energy” was used, this would not be coded under “energy” as it was not directly connected to human consumption in some way. When coding, an explicit reference to food, energy, or water was needed to be included under the FEW Nexus. Simply restating the words food, energy, or water was not enough. While some students may be able to make implicit connections between FEW Nexus systems and other IES knowledge areas, here, we measure direct and explicit connections so we do not overestimate student experience and comprehension. For further detail on the FEW Nexus coding schema, please find more details in (Horne et al. 2023).

As previously described, students were first asked to verbally describe the connections between food, energy, and water before then being asked to visually represent these connections and explain their drawings. Verbal-only responses were coded for FEW Nexus components and connections using a similar coding structure as Horne et al. (2023), as well as the number of times each reference was made. For example, if a student described multiple ways that water can be used to create energy (e.g., dams using water to produce hydropower, water needed to cool nuclear energy reactors, water that produces steam during solar energy production), that would count as multiple “water to energy” connections.

For analyzing the drawings, corresponding pieces of the transcript were also coded for FEW Nexus components and connections and the number of each reference. Interpretation of these explanations was aided by looking at the drawings, especially as students referenced their drawings while describing the FEW Nexus using words such as “this,” “that,” and “it.” Transcripts were also coded for the number of arrows students drew (as an additional way to quantify the number of connections). While students were not required to draw arrows, the majority of students chose to include one-way or two-way arrows to connect FEW Nexus resource systems. Finally, we coded the drawings for level of detail on a spectrum from abstract to concrete based on the definitions provided by Weinrich and Sevian (2017) (Figs. 1, 2, 3). Although these terms may have other applications in other contexts, especially with regard to different disciplines represented in FEW-Nexus scholarship and education, differentiating between simpler versus more complex drawings in this way allowed for explanation of qualitative analysis of responses. Abstract drawings were those that included only words and arrows, while concrete examples specified scenarios that included elements like farms, dams, houses, transportation infrastructure, and natural features (e.g., rivers, mountains, sun). Drawings that included a few words and pictograms were coded as a medium level of abstraction.Footnote 1

Fig. 1
figure 1

Example of student drawing coded as abstract as the student uses only words and arrows to illustrate connections between food, water, and energy. This drawing was coded for “energy to food,” “water to food,” “food to water,” and “food.” Notice that while the student has arrows connecting food to energy, they are referring to caloric energy rather than fossil fuels or renewable energy sources. As we are using the FEW Nexus to mean energy security, we did not code caloric energy under “energy”

Full size image
Fig. 2
figure 2

Example of student drawing coded as a medium level of abstraction as the student uses a combination of pictograms and words to illustrate connections between food, water, and energy. This drawing was coded for “food,” “energy,” “water,” and “water to energy”

Full size image
Fig. 3
figure 3

Example of student drawing coded as concrete (or a specific example) as the student uses only pictograms to depict a specific scenario to illustrate connections between food, water, and energy. This drawing was coded for “food,” “energy,” and “water”

Full size image

Coders worked collaboratively to discuss and refine the coding schema until a 90% agreement was reached before coding interviews independently (Schreier 2014). To ensure coder agreement over time, we double coded 10% of interviews at the end of analysis (Elo et al. 2014). Calculations were completed using NVivo 1.6 © coding comparisons and Excel (number of agreements divided by the total number of codes) to ensure consistent interpretation of the coding schema.

Results

We answer our research questions sequentially, beginning with the FEW Nexus components and connections students are familiar with and ending with results from comparing student drawings to their verbal responses.

What connections are students making between FEW Nexus resource systems?

Students are most likely to describe needing water for food production, closely followed by energy requirements for agriculture (Figs. 4 and 5). Most commonly, students discussed needing water to grow crops; however, students also described the large water footprint needed to support animal agriculture, especially converting cows to beef. Industrial agriculture was the most discussed food production system, and students described the energy inputs needed to support machinery for irrigation, planting, harvesting, processing, and transportation to consumers (Table 2). Closely related, students described how energy resources are essential in purifying water for human consumption (e.g., wastewater treatment plants) and transporting water to consumers. Students seem to firmly understand that the food and water they consume come “from away,” and therefore require significant energy inputs to transport. Overwhelmingly, students described water as a source of energy through hydro-electric dams. Food systems rarely influenced water or energy systems in student explanations of the FEW Nexus. When students described the impacts of food systems on water and energy systems, they talked about runoff from agricultural production as a source of water pollution and growing corn to be converted to biofuel energy.

Table 2 Summary of FEW Nexus connections, frequencies, and exemplar quotes from students. Percentages have been rounded for ease of interpretation
Full size table
Fig. 4
figure 4

Frequencies of FEW Nexus coded references, including individual components (i.e., food, water, energy) and connections (e.g., food to water, energy to food, water to energy) (N = 191)

Full size image
Fig. 5
figure 5

Summary of FEW Nexus components and connections. Arrows and circles represent the frequency of each element proportionally

Full size image

Some missing connections stand out. Almost no students in our sample described the influence of food or energy production on the quantity of available water resources. While students described agriculture as a large consumer of water, they did not go on to describe the influence that has on water quantity and availability for other purposes. Instead, they were more concerned with water quality issues associated with agricultural use of water resources (i.e., pollution from industrial agriculture). A handful of students mentioned that water is important for energy production (e.g., nuclear, solar), but overwhelmingly, hydropower was the primary use of water in energy production. Again, almost no students commented on how energy production could influence water availability.

Of the students describing individual FEW Nexus components (i.e., food systems, water systems, energy systems) without making connections to other resource systems, 42% of student references were to food systems, followed by 30% of references to energy systems and 29% of references to water systems (N = 191). In other words, food systems were twice as likely to be described as an individual component of the FEW Nexus compared to energy or water systems. It appears that food systems is an anchor of the FEW Nexus for students, and this may explain why food systems were most often impacted by water and energy systems in student interviews, rather than food systems impacting water and energy resources.

We compared coding frequencies across demographic variables (collected in a survey completed at the end of each interview). Demographic information collected included the following: gender identity, ethnicity, age, year in college, regional location of college, major, and experience growing up in rural to urban settings. We observed two patterns in our data in relation to year in college and major, but we remind readers of our small sample size and suggest future studies to investigate these trends. Firstly, third and fourth year students were more likely to make more connections after drawing than first and second year students (Table 3). First and second year students were more likely to make fewer or the same number of connections when asked to draw. Third and fourth year students were also much more likely to draw a higher number of connections compared to their verbal responses. This might indicate that students learn how to use drawing as a tool to explain connections as they progress through the curriculum. Secondly, students majoring in agriculture and natural resources or science and math were more likely to describe a greater number of connections after drawing compared to social science majors (Table 4). This might indicate disciplinary differences in the use of drawings in learning. Our social science major sample was only 11 students (compared to 23 and 9 majoring in agriculture and natural resources or science and math, respectively), and further study would need to explore drawing by degree major to fully understand this trend.

Table 3 The number of students (N = 50) who made fewer, more, or the same number of connections with their drawing compared to their verbal only explanation by year in college
Full size table
Table 4 The number of students (N = 43) who made fewer, more, or the same number of connections with their drawing compared to their verbal only explanation by major category. We had one business major, four arts majors, and two “other” majors in our sample, which are not represented in this table due to low numbers
Full size table

Does drawing facilitate connections between FEW Nexus resource systems?

There is some evidence to support our hypothesis that drawing helped students make more FEW Nexus connections than relying on verbal explanation alone (see Table 5 for examples of students who made more and fewer connections after drawing). Fifty percent of students made more connections when asked to draw the relationship between food, energy, and water systems compared to relying solely on verbal explanation (25 students). Thirty-two percent of students made more connections using only verbal explanations (16 students), and 16% (eight students) made the same number of connections using verbal and then a drawing to explain linkages. For students who increased the number of connections while drawing, an average of 4.6 additional connections were made with a range of 1–13. For students who made more connections verbally, 3.1 fewer connections were described when using drawing as a medium with a range of 1–10. In conclusion, it appears that drawing facilitated systems thinking in half of our sample, while decreasing the number of connections in a third of students.

Table 5 The first row demonstrates a student whose first verbal response identifies several connections between food and water related to industrial crop and animal agriculture. While there are possibly vague references to energy in the verbal only explanation, the response was not clear enough to identify energy in the coding. In comparison, the verbal response combined with the drawing show a greater number of connections that more fully incorporate energy into food production. In the second example, the student makes more connections in her verbal only explanation. Italics indicate where text was coded
Full size table

Does the type of drawing influence the number of FEW Nexus connections students make?

We found no discernible pattern linking the number of FEW Nexus connections to the type of drawing (i.e., abstract, concrete, in the middle), and we therefore refute our hypothesis that students who draw concrete FEW Nexus examples are able to identify more connections than students who draw abstract representations of the FEW Nexus. In fact, students who created drawings that depicted more specific scenarios tended to make fewer connections when drawing compared to their verbal explanation. This result should be interpreted with extreme caution as only eight drawings (16%) were classified as concrete, opposed to 36% abstract drawings and 46% hybrid drawings. Interestingly, when students were given almost no guidelines to construct their drawings, they default to using arrows to signify connections between resource systems (82%). This suggests that, while drawing and modelling approaches vary across disciplines, students absorb some disciplinary norms for mapping and modelling even if not explicitly part of their course instruction.

Discussion and conclusion

This study contributes to the growing body of literature examining drawing as a tool to support student learning in the classroom. Despite its utility, drawing has rarely been used to assess student systems thinking capabilities in IES programs, nor have drawings been widely used to elucidate student understanding of FEW Nexus systems connections. We found that drawing can be a helpful tool to understand how students connect food, energy, and water resources. Additionally, drawing increased the number of connections made compared to a verbal only explanation in half of our student sample. Here, we unpack our findings and suggest practical implications for instructors to increase student understanding of FEW Nexus connections and to implement drawings as classroom tools.

Food as the anchor to the FEW Nexus

Students have a firm grasp on water and energy systems impacts to food production; however, the effects of food systems on water and energy systems are less frequently understood. Connections between water and energy systems were commonly described, and food systems appeared to act as an anchoring concept for students to comprehend the FEW Nexus. Several possible interpretations of these findings exist. Because we utilized “Food-Energy-Water Nexus,” students may have started with food systems simply due to its placement as the first word in this term, thereby making it the anchor system for students. Similarly, students may have drawn things that were easier to draw, rather than what they first associated with the FEW Nexus (Neumann and Hopf 2017). It could also be that food systems are more visible to our student sample than water and energy systems. For example, though we use water and energy every day, much of the production process is “invisible” in daily lives. Hmelo-Silver et al. (2014) found that novice learners, such as students, more easily identify structures in a system than behaviors within or functions of a system because they are more salient. In the context of our study, many students in the United States turn on a switch to acquire electricity and water; however, food must be shopped for and prepared in individual households, making food systems potentially more visible in daily life. Perhaps food-related events were also recently in the media; high media coverage events, such as a nuclear disaster, can influence perceptions and issues salience (Neumann and Hopf 2017). Additionally, students conceptualized energy differently in their drawings, such as including caloric energy (see Fig. 1 for an example) instead of energy resources. This is likely due to inconsistent presentation of energy concepts and limited linking across disciplines. For example, students often struggle to transfer energy concepts (e.g., caloric, trophic, kinetic, mechanical) between chemistry and biology even though it is the “same” energy (Kohn et al. 2018).

Regardless of the cause, instructors can take away several practical implications regarding student understanding of FEW Nexus connections. Firstly, instructors may need to more explicitly explain the influence of food systems on water and energy resources, especially related to quantity and quality. This could include further details on energy and water production processes so that students have a more in-depth understanding of where their water and energy comes from. With a deeper understanding of these systems, the impacts of food resources on energy and water may become more obvious to students and create a more accurate mental model from which to understand systems connections. Secondly, instructors can help students develop interdisciplinary systems thinking capabilities by making connections between energy concepts presented across disciplines, thereby illustrating that energy concepts are all connected. Co-developing and co-teaching courses across programs could facilitate this learning. Thirdly, instructors using drawings in their teaching should strive to ensure that students know what types of energy representation are being solicited. For example, many students immediately thought of caloric energy through food webs; however, for the purposes of our research, we were interested in things like chemical energy, nuclear energy, and thermal energy. Directions for drawing exercises should consider how scientific terms are used colloquially as well to try to ensure clarity (“food as fuel” is a somewhat common idea in health and wellbeing, but “fuel” is likely to have a different application in many science courses).

Drawings as a helpful tool for students and instructors

Drawings helped half of the students increase the number of connections between food, energy, and water systems; however, some students made fewer connections when asked to draw compared to their previous verbal only explanation. The level of abstraction in student drawings had no effect on systems thinking capabilities in our study. Previous studies identified drawing as a useful tool to increase student performance related to learning, systems thinking, and depth of knowledge. For example, a study of undergraduate students learning about the circulatory system revealed that students who generated drawings were more able to transfer their knowledge to new contexts (though they performed more poorly on recall) (Zhang and Fiorella 2021). It appears that drawing supports complex skill building (e.g., systems thinking, transference of knowledge) more effectively than more simple cognitive tasks (i.e., recall). This is partially supported by our study whereby half of students improved their systems thinking through drawing despite being unfamiliar with the term “FEW Nexus.” Additionally, previous research suggests that drawing can result in increased student motivation to learn, can support deeper learning by making mental models of systems explicit (and thereby provide instructors with information on important gaps in knowledge, misconceptions, etc.), and can be a means to overcome shortcomings in presented materials (Ainsworth and Scheiter 2021). There is also some evidence to suggest that drawing may help reduce the achievement gap for students from historically marginalized backgrounds (Aylward 2010), students with lower skill levels (Bobek and Tversky 2016), or non-traditional students (Van Der Veen 2017). This may be because drawings represent internal mental models, or ways of knowing, that may reflect different types of knowledge (e.g., cultural knowledge, local knowledge, Indigenous knowledge) that are often not acknowledged in mainstream classroom activities, texts, etc. Furthermore, drawing is an active learning activity (i.e., one that requires students to actively play a role in their learning rather than passively absorbing information); active learning approaches have reduced the achievement gap in previous studies (Theobald et al. 2020). Finally, some studies share feedback from students that drawing assignments are fun (Van Der Veen 2017); this element of enjoyment can help motivate students to persist in classes and majors.

While drawing may increase student performance in many contexts, it is important to acknowledge that drawing can produce cognitive overload where the brain is overwhelmed by multiple tasks resulting in negative implications for learning outcomes (Heideman et al. 2017). This could explain why some students in our study made fewer connections when drawing. Additionally, even when the same components and connections are described in both verbal and visual modes, a reshaping of knowledge across modalities has occurred that may not be captured in our approach to analysis (Kress et al. 2001). We therefore suggest that drawing be used as a complementary tool by instructors and students. In other words, instructors might consider using drawing as a means of assessing student learning alongside more traditional assessment methods, such as verbal and written explanations. One of the strengths of drawing is being able to represent spatial elements of systems (including connections between systems), as well as processes (e.g., cycles) that might be more challenging for students to describe verbally or through written communication. We therefore suggest that a logical place to start incorporating drawings into teaching could be helping students understand cycles, such as nutrient cycling; scales, such as local to regional impacts; or connections between systems, such as the FEW Nexus.

We provided little guidance in our instructions to students regarding their drawing. While some students may have viewed this as freedom to express themselves, some students may have easily been overwhelmed by the lack of direction. In a study of fourth and eighth grade biology students, students performed better at drawing systems when provided with a list of concepts and linking words to direct their attention toward the underlying system; this was especially true for the fourth grade students (Vasconcelos et al. 2020). While college students are older than Vasconcelos’ study subjects, it may be useful for instructors to provide some structure for students in introductory courses as they are exposed to new concepts. Examples of structure could include word banks with key concepts and discussing disciplinary norms around symbol usage, such as arrows, to ensure consistent interpretation across students and the instructor (Vasconcelos et al. 2020). Directed drawings may also allow for easier grading, a common concern among instructors using drawing as a medium for assessment in the classroom (Reusser et al. 2012). Some students will be uncomfortable with drawing as a skill, especially as some disciplines use drawings and diagrams with varying frequencies and levels of abstractness. To overcome this initial discomfort, drawings for low stakes assessments (at least to begin with) may build student familiarity (and help them find their “visualization voices”) and confidence with the process, eventually enabling higher stakes assessments with drawings (Dauer et al. 2013; Van Der Veen 2017).

Though not examined in this study, there is evidence to suggest that having students switch between modes of explanation (e.g., verbal to drawing to verbal) may result in more robust mental models as students revisit and refine their ideas (Kress et al. 2001). In a study of four Year 7 science students, Jewitt and colleagues (2001) observed that having students use multimodal means of expressing their ideas (i.e., writing, drawing) addressed different aspects of their understanding related to cell biology. It may be that in some instances, students in our sample gravitated toward visually describing certain aspects of their understanding while using verbal explanation for different aspects of their knowledge. Instructors might consider mixing up the modes students use to represent their ideas and use small group work to have students explain written and/or drawn ideas to switch modalities within a lesson. As described by Jewitt et al. (2001), students engaged in active learning select, integrate, and adapt information, a process that also involves transforming information across modalities. Different modes of expression extend different dimensions of knowledge, and switching between modalities requires sophistication and integration (Kress et al. 2001). It could be that switching between modalities was a higher cognitive load for some students, especially if they are unused to visually communicating their ideas. Instructors should consider what modality is most appropriate for a specific learning outcome as different forms of representing knowledge and learning place different demands on student cognition (Jewitt et al. 2001).

Limitations and future research

While our sample represented the diversity of IES programs, our sample was more female, white, and students were primarily from the US. We did not collect any survey data related to past drawing experience inside or outside of the classroom, which could have helped explore the link between experience, approach to drawing (i.e., abstract or concrete), disciplinary background, and instructional practices experienced. We therefore have no knowledge of if and how students have used drawing in higher education settings before the interviews. Some disciplines may use diagrams and drawing exercises more frequently (e.g., natural sciences, math) than others (e.g., social sciences), though there is great diversity across curricula. We detected a trend indicating this pattern might exist in that social science majors described fewer connections after drawing, but further work is needed to tease out the influence of disciplinary learning on student drawings.

Importantly, it could be that students are better able to express their knowledge of connections through drawing or it could be that students are making further connections while drawing as they continue to think about FEW connections due to the question order. In other words, because we asked students to first verbally explain connections and then draw, it may be that the verbal explanations served as a primer that the drawings built upon to increase the number of connections. We are not able to parse out these two possibilities in our study, but further research could use an experimental design where verbal and drawing explanations are solicited in different orders.

Many questions were raised from our exploratory study that point to future research needs. We suggest that future work experiment with framing the FEW Nexus using a different order of concepts (i.e., WEF, EFW, WFE, EWF) to test whether food systems are a universal anchor for students or merely a result of word arrangement. There is some evidence that the first concept drawn or written (such as in a free list exercise) is the most salient and/or important concept (Quinlan 2019). Capturing the timing and order of concepts students draw, such as through using a LiveScribe pen or camera, may provide interesting insight into concept salience and illuminate the systems thinking process in real time. Additionally, a LiveScribe pen or camera could show if students begin with a more concrete representation and then switch to a more abstract representation as they try to represent connections across contexts. A concrete approach to drawing that shows a specific context (e.g., farming, hydropower) can only reflect a limited number of phenomena at the same spatial scale. For example, students may draw agricultural fields with wind turbines and an irrigation system to represent food, energy, and water. It becomes challenging to use that context to represent a larger number of connections between systems across different geographies and spatial scales. Students may consciously or unconsciously realize this as they approach their drawing response and shift to a more abstract form of representation (i.e., with words, arrows), resulting in a hybrid approach to drawing. We observed no pattern when examining the number of connections in comparison to the drawing approach. This could be an important distinction: the level of abstraction in students’ drawings does not necessarily reflect an abstract way of thinking about the FEW Nexus. Our findings also connect to the insights from Ainsworth and Scheiter (2021) about the changing modes of drawing and how that affects student learning and assessment. Future studies may also focus on the missing links from drawing and visual representation and compare it to the IES curriculum to provide recommendations on how to teach these missing connections (see Shepardson et al. 2009). Future work could compare the impact of directed and nondirected approaches to drawings on student learning. Such a comparison could consider the level of abstraction in student systems representation to better understand if different approaches to drawing impact systems understanding.