The Use of the Constructivist Teaching Sequence (CTS) to Facilitate Changes in the Visual Representations of Fifth-Grade Elementary School Students: A Case Study on Teaching Heat Convection Concepts

Most primary school students, although they grasp the scientific concepts of heat convection at the macroscopic level, commonly fail to visualize those concepts. Therefore, our research aims to enact a constructivist teaching sequence (CTS) to restructure students' visualization changes, ultimately enabling them to synergize macroscopic and sub-microscopic levels of understanding and visual representation. This study has employed a case study, combining qualitative and quantitative data to obtain an in-depth explanation. The quantitative data represent the percentage of the students' visual representation category and their understanding of pattern changes before and after the intervention. Meanwhile, the ways students presented their thoughts about a concept based on their visual representation are presented via qualitative data. All data come from the participants, comprising 69 fifth-grade elementary school students at one public school in Indonesia. Our research findings show that students' understanding of heat convection at both macroscopic and sub-microscopic levels improved to scientific conception, after undertaking the learning process using CTS. In addition, the use of CTS fostered a level of visual representation change regarding "construction" that dominated compared with other approaches: students shifted their visual representations from the varying styles of undefined drawing (UD), non-microscopic drawing (NMD), or no drawing (ND), to partial drawing (PD) and scientific drawing (SD).

Keywords: Constructivist teaching sequence (CTS); Heat convection; Macroscopic level of understanding; Sub-microscopic level of understanding; Visual representation

Copyright comment Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Heat convection is a concept that is more observable than other heat transference phenomena. For example, the flow patterns involved in water boiling can be studied via unaided eyes. In these contexts, we can recognize three levels of possible representation: macroscopic, sub-microscopic, and symbolic representation. The use of our unaided eyes here is categorized to a macroscopic level that refers to observable, tactile, and sensorial phenomena. In multiple contexts, the existence of observable concepts, theories, and principles that are suitable for the macroscopic level leads to the sub-microscopic level; such concepts as molecules, electrons, and atomic movements reside at the sub-microscopic level. Meanwhile, the symbolic level represents macroscopic phenomena using mathematic equations, graphs, mechanisms, analogies, and formulae (Johnstone, [32]). Because it is observable and involves all three levels, the concept of heat convection is useful for presenting multiple types of scientific representation.

In addition, because the scientific concept of heat convection comprehensively involves all three types of representation, we can consider how words and pictures can be used as analytical tools to reveal students' understanding of scientific concepts (Opfermann et al., [42]). Involving primary school students in specific concepts such as heat convection opens the opportunity for them to visualize abstract ideas. However, positioning the subject matter alone is insufficient; teaching deep structure that leads to sub-microscopic representations often requires a learning process that utilizes an appropriate instructional model (Treagust & Chittleborough, [56]) that contributes to students' ability to create sustainable construction from the macroscopic to the sub-microscopic levels. For this purpose, this study's research has used the constructivist teaching sequence (CTS) for the construction of visual representation among students at a primary school.

The use of CTS is strongly connected to constructivism-based science as the intervention (Widodo, [61]). The exploration of students' initial knowledge becomes a pivotal foundation for their comprehension of upcoming specific content, in this case the concept of heat convection. In this research project, CTS aims to restructure students' visualization changes, and ultimately they are able to synergize to sub-microscopic levels of conceptions (Duit et al., [20]). This study has been designed as a response to prior education outcomes: Most learners have tended to fail in efforts to visualize scientific concepts even when they grasp the concepts at the macroscopic level. Moreover, in the primary school context, students are generally understood to have limited abilities to think about abstract concepts.

Based on the above background, we have developed three research questions on how CTS can facilitate changes in visual representations of heat convection:

• How well do primary school students understand heat convection at the macroscopic and sub-microscopic levels?
• How well can primary school students create visual representations in the context of heat convection?
• In what ways do primary school students create visual patterns of change in the context of heat convection?

This study focuses on heat convection as the education concept for three reasons. First, heat convection is one of the key concepts in physics that can represent a dynamic process (Chiou, [14]). Because the particle movement of water can be observed clearly through macroscopic representation, it provides an underpinning for students to develop their own appropriate mental models for the heat convection concept, which enables them to make meaning from what they observe. Second, science education researchers have conducted ground-breaking work to identify useful ways to articulate the concepts of heat and temperature (Erickson, [23]; Kesidou et al., [33]; Lewis & Linn, [37]), whereas few studies exploring concepts of heat convection are understood by primary school students. The third reason involves the ways students in primary school attempt to investigate what they can visualize when studying heat convection concepts. Because there is scant empirical evidence that reveals how changes in students' visual representation abilities occur, this study investigates that research gap, in the context of heat convection.

As a foundation for the current study, the research team has undertaken a review of the prior literature. Chiou ([14]) has investigated the mental models of physics students who were working to understand heat convection. His findings describe seven mental models which physics students use to present heat convection. The third model, for example, discusses the way in which convection currents emerge immediately after heating commences, and describes the currents' ability to evenly distribute absorbed heat energy within a liquid. In a different context, Park et al. ([43]) have investigated the impacts of integrating engineering into students' conceptual understanding of heat transfer, including convection. Their research reveals that some students in grades 4–9 were aware of the way in which reducing the volume of a space could reduce the rate of convection on a given object inside that space. This means that the students understood temperature difference to be among the factors that affect the rate of convection. In addition, Schnittka and Bell ([48]) have investigated students' conceptual changes regarding heat transfer, in an eighth-grade classroom utilizing engineering design as part of the learning process. One of their findings is that students conceptualized that cold transfers from cold areas to warmer ones. This in turn led to their conception that particle movement occurs in heat convection. Ultimately, prior research on students' conceptualizations of heat convection emphasizes that students comprehend the convection process at the macroscopic level.

Visual representation is a key in teaching, learning, and communicating science and is among the primary methods scientists use to communicate their thoughts (Ainsworth et al., [4]; Fiorella & Zhang, [24]). In addition, visual communications are used to simplify complex science concepts, presented in the forms of illustrations, pictures, diagrams, animations, and physical models (Rau, [46]). Across multiple contexts, visual representation is pivotal because it supports a comprehensive skill that drives the brain to automatically respond to several forms of communication that involve the generation of imagery (Guida & Lavielle-Guida, [26]). In addition to providing training in several forms of communication, visual representation performs three functions in educational activities: choosing existing knowledge, organizing that knowledge around a concept, and integrating the resulting internalized elements to provide external forms (Quillin & Thomas, [45]).

Because it provides these multiple benefits, visual representation is widely used in the learning process. Empirical evidence demonstrates that instructional design which incorporates visual representation provides more benefits within the classroom learning process than the design which solely relies on verbal representation (Fiorella & Zhang, [24]). In these instructional settings, students must respond to external representations, such as words, pictures, or both, that describe a concept, and explore the visual representations. They also experience the ways in which scientists use multiple literacies in constructing and documenting knowledge: reading, writing, and speaking are integrated with the visual representations scientists create (Krajcik & Sutherland, [35]). Moreover, students can use visualization to deepen their understanding of a concept and understand additional presentations in detail (Kozma & Russell, [34]). Thus, teachers can simultaneously guide students to acquire scientific visual literacy, see its relevance, and respect teachers' explanations (Enyedy, [22]).

While visual representation can be infused into instructional models in the classroom, it can also function as an effective strategy to achieve a specific goal (Van Meter & Garner, [59]). For example, when students are asked to read a text and describe what they understand, they tend to discuss abstract concepts in terms of observable ones (Ainsworth, [3]).

For the current study, the researchers have selected an instructional approach with the goals of providing students with an appropriate understanding of convection concepts, and providing ways in which they can effectively visualize the related abstract ideas. With these goals in mind, this study sets out to provide a detailed description, explore students' understanding at the sub-microscopic level through their visual representations, and examine how their patterns of understanding change after they are taught using a method designed to improve their visual representations.

The importance of constructivism as a pivotal approach in science teaching and learning can be considered from the viewpoint of its role in supporting higher-order thinking skills (Boddy et al., [10]; Skamp, [50]). This viewpoint derives from constructivist theory, which recognizes how students learn new knowledge by constructing their own understanding in synergy with their existing views. Indeed, the constructivist approach succeeds when students can change their conceptions from inaccurate to scientific ones. To attain accurate conceptions, students generally need to be engaged to observe, think, and experience the phenomena of study (Bächtold, [6]). However, the fundamental activity in the constructivist approach is how the teacher explores and establishes students' preconceived ideas before carrying out an effort to teach new concepts, because new knowledge might not override students' prior learning or life experiences (Baviskar et al., [8]; Boddy et al., [10]).

A limited amount of previous research has specifically focused on the learning cycle in terms of CTS. Research examining constructivist teaching and learning environments indicates that such approaches help students organize their thinking (Iofciu et al., [29]), manifest conceptual change (Tekos & Solomonidou, [54]), and develop more sophisticated epistemological beliefs about scientific knowledge (Chang, [12]). Additional prior research examining the learning process in three different countries—Finland, Germany, and Switzerland—demonstrates that two key aspects of constructivist teaching, structured knowledge acquisition and fostering autonomy, improved students' motivation in a physics classroom (Beerenwinkel & von Arx, [9]). Thus, previous research indicates that the constructivist learning environment is beneficial to improving not only students' thinking skills and conceptual changes but also their motivation.

Therefore, the current study utilizes the CTS learning model, whose process is rooted in constructivism-based science as the intervention. To facilitate students' ability to perform mental model-building, and specifically changes in how they create visual representations, CTS adopts learning cycles that explore pre-conceptions, restructure those conceptions, and apply and evaluate new conceptions. These activities aim to obtain ongoing refinement of knowledge, ultimately achieving disequilibrium (Martin, [39]). In the current study, as Fig. 1 above (Widodo, [61]) illustrates, the stages of the learning cycle consist of (1) the introduction stage, to prepare and ignite students' motivation to learn; (2) the exploration stage, to investigate students' initial knowledge about the upcoming topic; (3) the restructuring stage, to facilitate students' changes in conception; (4) the implementation stage, to apply the new concepts students learn; and (5) the evaluation stage, to review and evaluate new conceptions to encourage students to compare their newest knowledge with their initial one (Duit et al., [20]). Table 1 shows the steps of the current study's teaching and learning process. In addition, Appendix 1 Table 7 provides the course materials and pedagogical components used in this study's student–teacher interactions.

Graph: Fig. 1CTS learning cycle

Table 1 Students' activities in the CTS-based learning process

The CTS learning model is useful to drive conceptual changes (Posner et al., [44]; Tekos & Solomonidou, [54]). These changes can follow several patterns, including dissatisfaction, intelligible, plausible, and fruitful processes. These patterns can be explained as follows:

• Dissatisfaction is a phase in the learning process that begins students' knowledge exploration. In the current study's learning activity, the students first brainstorm the concepts based on their prior experiences. Then, with their teacher's guidance, the students discuss common phenomena. At this phase, teaching the concept of heat convection has only focused on macroscopic representation; however, providing students with guiding questions to activate their sub-microscopic conceptual level enables them to begin feeling cognizant of the inadequacies in their preconceptions. This dissatisfaction stems from the restructuration of their conceptions.
• The intelligible phase during CTS occurs as a result of the restructuration and application of newly constructed ideas. In this phase, the students begin understanding a new concept via reinforcement and revision regarding the preconceptions they previously held.
• The plausible phase occurs when students accept a new concept; this phase continuously evolves to improve the way students process new information and embrace a scientifically accepted viewpoint.
• The fruitful phase occurs in the last part of the exercise, when the students gain new ideas and understanding from their learning activities. At the end of the CTS structured learning, the students are guided to identify and compare their conceptual changes from the beginning to the end of the educational activity.

The current study adopted a case study design to investigate the use of CTS in elementary-school science teaching, rather than using a controlled treatment. The case study design is suitable for the educational setting, which does not permit tight controls or experimental manipulation (Anderson & Arsenault, [5]; Kuo et al., [36]). This study employs both quantitative and qualitative data to obtain in-depth explanations (Creswell & Plano Clark, [17]). The quantitative data represent two different types of information: the percentage of the students' visual representation category and the ways in which their understanding patterns change before and after the intervention. Meanwhile, the qualitative data describe how students present their thoughts about a concept, based on their visual representation revealing what is on their minds regarding the lesson (Quillin & Thomas, [45]).

All data collected during this study has come from the participants, 69 fifth-grade elementary school students from one public school in Indonesia. The researchers' decision to take one school is based on the ease of accessibility to the specific elementary school, because it is categorized as a pilot school project. The participants included 28 male and 41 female students, aged from 10 to 11 years old.

The research team for this study developed one problem to reveal students' understanding of heat convection. The problem consists of five tiers of questions about heat convection as study instruments (Anam et al., [1]). The first and third questions comprise the content tier and consist of multiple choices: four answer options, in which the researchers provide three specific choices, and students may propose an alternative as the fourth choice. The first question asks students to predict the phenomenon that will occur during an experiment; this aims to investigate their macroscopic level of representation. The third question asks students to explain the reasons underlying the phenomenon that emerges in the experiment; this aims to examine their sub-microscopic level of representation. Meanwhile, the second and fourth questions comprise the confidence tier, which measures students' certainty levels in making choices during the content tier. The second and fourth questions provide two options for students to choose between: "sure" and "not sure." Finally, the fifth question comprises the drawing tier and asks students to provide a visual representation based on the scientific concepts they have studied. Appendix 1 Table 7 presents the problem and all five question tiers.

To validate the study instruments, the researchers consulted three experts in science education who hold doctoral degrees: two lecturers and an elementary school teacher. The research team asked these experts to evaluate whether the study instruments fulfill key indicators, such as content appropriateness. In addition, we conducted validity testing; Kendall's Tau score is 0.82, which indicates the instruments have strong internal validity. To further assess the validity of the instruments, we tested them empirically using the Pearson product-moment correlation coefficient and obtained a score of 0.73 with p < 0.05. Meanwhile, we tested the reliability of the instruments using Cronbach's Alpha; the alpha score is 0.652, indicating that the instruments are reliable.

Using the study instruments, the researchers obtained students' responses to the problems. The responses show the students' understanding of heat convection at the macroscopic and sub-microscopic levels, via the confidence level questions and the visual representation exercise. We analyzed students' responses via three approaches. First, we analyzed students' understanding at the macroscopic and sub-macroscopic levels. Here, we analyzed students' initial (pre-test) and final (post-test) responses via frequency analysis. Each student response that represented understanding at these two levels was counted. To assess the significance, we used the chi-square test, carried out by using SPSS 23 (p < 0.05). Second, we coded the results of the visual representation tier to categorize the students' drawings into six levels: scientific drawing (SD), partial drawing (PD), misconception drawing (MD), undefined drawing (UD), non-microscopic drawing (NMD), and no drawing (ND). Appendix 2 provides a detailed explanation of these categories; we used a rubric to code students' visual representation, derived from the explanation of each category of drawing. This rubric was evaluated by two experts in physics (Appendix 4 Table 9). Third, we grouped and coded all changes in students' visual representations to five phases: construction, revision, complementation, static, and disorientation. The grouping aims to classify whether each student's mental model for visual representation has shifted. The determination of visual representation changes was conducted by analyzing shifts in each participating student's initial drawing compared with their final drawing, as Table 2 illustrates.

Table 2 The pattern and level of students' visualization changes

The explanation of students' understanding begins with presenting the data at the macroscopic level about heat convection before and after the intervention. In the initial part of the exercise, students observed how wood powder moved in circular motions, following the flow of heated liquid. Table 3 provides these data. The results indicate that the majority (66%) of participating students initially anticipated that the closest water particles to the heat source would rise, and the furthest particles would replace their positions. The students attained this conception before they formally learned it within the education process; thus, these students demonstrated desirable prediction capacities prior to observing the phenomenon during the laboratory activity. After involving students to learn heat convection via the CTS model, findings indicate their conception of particle movement improved appropriately to the observable phenomenon, with 87% identifying the phenomenon correctly. However, this difference is less significant compared with the results for other choices. Our investigation here clearly shows the enhancement of students' conception as a result of their participation in the activity, which facilitated their observation of the visible phenomenon.

Table 3 Students' understanding at the macroscopic level

Note: p ≤ 0.05, LoS level of significance, *Significant, NS not significant, N number of students

Next, to investigate the participating students' understanding at the sub-microscopic level related to the observable phenomenon, the researchers asked about the factors underlying the phenomenon. Our investigation reveals that 43% of students proposed during the pre-test a correct reason related to particle movement. This percentage of participants understood that hotter water will have more tenuous particles or become lighter than cooler water, the result being that hot water will rise and cooler water will sink. The students also shared that they had not previously discussed or been taught about the changes in particle position when water boils. Prior to this exercise, they had only formally learned about the concepts of water boiling and sea winds and had not yet studied what happens during the boiling or cooling processes of liquid and gas.

Due to the instructional approach that had only focused on providing the boiling water and sea wind examples of heat convection phenomena, many of the students provided the answer that did not match with scientific concepts—or misconception (Talanquer, [52]). In turn, this suggested that even though it is directly observable, the concept of how heat convection occurs needs to also be introduced in the form of visual representation, rather than simply telling students to memorize the common examples of heat convection as illustrated in Table 4. In different situations, students also understood that when water is boiled, its particles will spread out. This phenomenon can be observed through the size increase when an object is heated, whereas no difference is observable when the object is unheated. Generally, this analogy works as a supporting tool to provide students' answer, facilitate changes in their conception, and resolve misconceptions. It also helps students build a relationship between a new concept and their initial understanding and guides them to organize their knowledge so that they can understand new concepts better (Çoruhlu, [16]; Rule & Furletti, [47]).

Table 4 Students' understanding at the sub-microscopic level

Note: p ≤ 0.05, LoS level of significance, *significant, NS not significant, N number of students

We next asked students to provide visual representations to demonstrate their understanding of heat convection in response to a given question. Visual representation is a suitable approach to measure students' true understanding (Dikmenli, [18]). Table 5 presents the data on how accurate students' visual representations were, in the form of frequency analysis before and after the intervention. In addition, Fig. 2 shows samples of the types of visual representations.

Table 5 Analysis of students' drawing based on visual representation categories

Note: p ≤ 0.05, LoS level of significance, *significant, NS not significant, N number of students

Graph: Fig. 2Samples of the types of students' visual representation

The result of the pre-test shows that UD dominated students' visual representation and that initially, no SD was present. A small number of students provided visual representations that fit the categories of PD and MD. Our analysis reveals that CTS helped students improve their visual representation, enabling most participating elementary school students to learn the related concepts of particles and their characteristics. However, the improvement in students' MD post-test is an interesting finding that requires further explanation. We here consider that students can acquire holistic knowledge and prevent any restraints that may hinder their learning progress through visual representation (Dikmenli, [18]). To produce visual representations that are appropriate to scientific concepts, students need to improve their understanding at both the macroscopic and sub-microscopic levels, because these specific levels reflect the students' depth of understanding (Quillin & Thomas, [45]).

Our findings also indicate that visual representation helps reveal students' thinking processes regarding a scientific concept; a description of the macroscopic and sub-microscopic levels does not suffice to help them understand. If elementary school students are not supported with sub-microscopic representation models, conceptual misunderstandings arise because they construct their own knowledge, which is often at odds with the scientifically accepted view. Thus, CTS here helps students visualize abstract concepts and may lead to their gaining an understanding that agrees with that accepted by the scientific community (Taber, [51]).

This section addresses how changes occurred in students' visual representations before and after the intervention. Our findings indicate a shift in the students' mental model of visual representation dealing with the sub-microscopic level. Table 6 describes the changes in students' visual representations, from the initial to the final condition; as the data illustrate, student visual representations beginning at three levels (construction, revision, and static) underwent significant differences, whereas others did not.

Table 6 Number of visual representation changes and significance of the differences

Note: p ≤ 0.05, LoS level of significance, *Significant, NS not significant, N number of students

We found that the construction level accounted for the most common changes among the students' visual representations. This group of students initially produced visual representations fitting the categories of ND, NMD, and UD. However, after the intervention, their visual representations improved in PD and SD. This indicates appropriate construction for what students understood at the macroscopic and sub-microscopic levels, and their mental models for visual representation.

Students who began in the ND category were unable to represent either macroscopic or sub-microscopic visual representation, as Fig. 2C illustrates. Moreover, NMD drawings represented inappropriate understandings of the sub-microscopic level. Students whose initial representations fit NMD mostly drew the direction of particle motion for hot water and cold water and complemented these representations with texts, for example, stating that heated water particles are lighter while cold water particles are heavier (Fig. 2B). Although the understanding these drawings indicate is appropriate to the scientific concepts, these texts do not meet standards for appropriate visual representation. In the UD category, students' representations showed unclear visuals for both particle and convection flow. In this category, an illustration of the liquid particles was mostly lined up in three horizontal groups as a representation of point A, which can be interpreted as dense particles, while in representing point B the student depicted the particles as not evenly spread out, which might be translated into "expanding" (Fig. 2A). Although students provided their own models, their visual representations were less comprehensible and were often at variance with the scientifically accepted view.

However, the quality of visual representations improved after the learning process, when students were able to provide drawings that met the categories of PD and SD. In the PD category, most students drew similar visual representations for both hot and cold water, representing water particles as equidistant from each other, although they showed an understanding of the particle movement by presenting appropriate directional indications for particles in hot (point A) and cold (point B) water (Fig. 2E). In the SD category, students drew the convection flow as moving upward and downward. Students appropriately illustrated that the hot water contains expanding particles and that the number of particles in hot water is lower than that in cold water. Meanwhile, the particles of the cold water appear dense, and they have a higher quantity. This indicates that the students were able to understand the concept of water density even though it had not yet been formally taught in the classroom. The students demonstrated an understanding that hot water is less dense than cold water at the same volume; this type of visual representation fits the SD category.

The participating students improved their levels of visual representation, although this only progressed in one or two categories at the revision level. Several changes in visual representation occurred, from ND to UD, NMD to UD or MD, UD to MD, and MD to SD. At this level, some students were able to revise their mental models of visual representation toward other models that might not align with scientific conceptions. For example, a student's mental model shifting from NMD to UD indicates a revision from macroscopic visualization (Fig. 2B) to microscopic visualization (Fig. 2A) in terms of patterns the student had not yet understood.

In an additional type of revision of visual representation, from MD to SD, some students demonstrated an understanding that hot water particles are less dense than cold water particles before the intervention, presenting this understanding at the sub-microscopic level in their visual representation. Their drawings indicate they possessed a mental model about particle density, although their illustrations are inconsistent with the accepted scientific viewpoint. In these visual representations, the particles are shown as dense and numerous in boiling water and the space between particles reduces, while the distance between a sparser number of particles expands in cold water (Fig. 2F). If the process had followed the student's way of thinking, the hotter the water the heavier the particles would have been, and hot particles would have sunk, while cold water would have been lighter and cold particles would have risen to the surface. However, after the intervention, some students were able to revise their visualizations and understanding of the scientifically accepted model. These revisions occurred as a result of altering the students' conception and ability to visualize convection flow based on the description of densities in cold and hot water (Fig. 2D).

At the static level, no improvement of visual representation occurred. Our findings indicate that some students did not improve their visual representations, as in the ND, UD, and MD categories. In the ND category, one student did not provide a visual representation. This result suggests the student lacked an understanding of heat convection. In the UD category, some students successfully showed they grasped the macroscopic and sub-microscopic conceptions, but they were not able to successfully adjust their visual representations. In addition, the researchers observed a consistent pattern of MD, in which some students whose scientific understanding was incomplete, produced flawed visual representations (Fig. 2F). Thus, several stages in the CTS process left students' conceptions static, showing they retained the same conceptions as during the initial conditions.

At the disorientation level, our findings show a dramatic change; students initially provided an illustration depicting hot water particles with bigger circles and cold ones with smaller dots. After the intervention, their visual representations shifted from the type in Fig. 2F to that in Fig. 2B. In Fig. 2F, students' understanding and visual representation were both inconsistent with scientific conception. In practice, the learning process failed to support students in restructuring existing concepts, applying new concepts, and evaluating new concepts. As a result, the construction of their visual representations showed incomprehensible patterns.

Finally, our analysis of changes in visual representation explored representations that altered from PD to SD; at this complementation level, we observed a change of visual representation from Fig. 2E to D. In this category, all students demonstrated valid conceptions at the initial condition of both the macroscopic and sub-microscopic levels. They understood water particle movements at low and high temperatures and also understood that colder water has a higher particle density than hot water. Nevertheless, they drew identical visual representations of both hot and cold water, in which water particles were depicted at the same distance from each other. After the CTS learning process, however, these students complemented their visual representations such that the drawings became appropriately aligned with scientific concepts. Their adjusted visualizations show water particles at low temperatures as more compact than those at high temperatures, indicating their mental model of convection became aligned with the accepted scientific viewpoint.

This section focuses on our findings in the context of this study's CTS-based learning intervention, regarding students' understanding of heat convection at the macroscopic and sub-microscopic levels, their visual representations, and the patterns of change in their visual representations.

Our findings indicate that after completing the learning process, the proportion of students' understanding at the macroscopic and sub-microscopic levels changed. Most of the participants demonstrated scientific conception at both levels of understanding after the CTS-based learning process. Our analysis indicates the students did not possess strong preconceptions at the sub-microscopic level before the intervention; instead, their understanding level was at the macroscopic level. In other words, they simply understood the concepts based on their experiences. The participating students tended to use these perceptions at the macroscopic level when attempting to understand the concept of phase transitions. This plays a generative role, particularly in disorganized and incomplete knowledge discourse. Macroscopic-level thinking does not support the development of critical thinking, because it generates systemic correspondence between two separable domains: analytic thinking and creativity (Clement, [15]). Consequently, we consulted other studies indicating that understanding would improve when students learned science at the sub-microscopic level (Boz & Boz, [11]; Harrison & Treagust, [28]). The current study's CTS-based learning process involved students in experimentation, demonstration, and context introduction. This learning process strengthened participating students' understanding of particle concepts at the sub-microscopic level. Via this approach to explanation, the students became able to construct appropriate verbal, visual, and symbolic representations (Tang et al., [53]).

The second finding of the current study is that students were able to provide a visual representation relating to scientific concepts after completing the CTS-based learning process. The research team integrated CTS methods with the development of visual representation, to enable students to think explicitly and factually (Einarsdottir et al., [21]). Consequently, although not all students were able to present scientific visual representations, the largest proportion of the final visual representations consisted of scientific drawing (SD). The explanation for this result is that integration of visual representation via CTS strengthens scientific concepts (Merino & Sanmartí, [40]; Selley, [49]), and this supports students in collecting various experiences to show how their ways of thinking have developed (Einarsdottir et al., [21]; Haney et al., [27]). Students can use visual representation to hypothesize, interpret data, and undertake other related activities (Ainsworth et al., [4]; Quillin & Thomas, [45]). Further, teachers can use visual representation to construct and promote minds-on and hands-on learning (Glynn & Muth, [25]). As a consequence, learning activities become meaningful as science is demonstrated to be a subject that can be portrayed via various visual representations (Banda et al., [7]).

The research team next focused on visual representation changes, and the ways in which students revised their visual representation from their preconceptions to their ultimate understandings. We are interested in explaining why the construction level was dominant compared with other types of visual representation changes. The reason for this involves not only whether students' revisions reached the final two categories of PD and SD, but also that scientifically, participating students altered their conceptual changes according to several patterns, including the dissatisfaction, intelligible, plausible, and fruitful processes (Posner et al., [44]). Specifically, students whose final representation reached the SD category achieved multiple patterns of conceptual change. This suggests that the CTS learning model is particularly relevant to "theory restructuration," which enables students to change their intuitive, synthetic mental models to a scientific model (Vosniadou, [60]). Interestingly, our findings indicate that the revision level was dominated by a change from UD to MD. Ultimately, studying not only enriched or completed students' knowledge, but also alleviated misconceptions through ontological change (Chi, [13]; Lin et al., [38]). Prior research indicates that most elementary school students commonly do not possess a consistent structure of knowledge and that they tend to collect their knowledge and integrate what they have gathered to construct new knowledge (diSessa & Sherin, [19]). Although the CTS approach facilitates the restructuration of students' conceptions, confusion regarding the nature of the sub-microscopic level in this process may lead to students' inability to visualize some entities (Harrison & Treagust, [28]; Tuckey & Selvaratnam, [58]). In the context of patterns of conceptual change, this situation leads to the intelligible process (Posner et al., [44]) in which reinforcement and revision form misconceptions.

This study's other findings regarding visual representation changes indicate that some students' drawings were categorized at the static, disorientation, and complementation levels. Most students at the static level presented UD in both their initial and final responses. The problem of student work remaining static may derive from a lack of understanding of macroscopic and sub-microscopic representation during several CTS phases, perhaps leaving the students unclear about what they were expected to visualize (Nelson, [41]). Surprisingly, the emergence of the disorientation level in the current study suggests this situation is related to confusion regarding the role of the macroscopic representation presented in the CTS-based learning process. In other words, the learning process's macroscopic representations (e.g. experiments and experiences) did not help these students arrive at the insightful observations that lead to the formal understanding necessary to construct scientific visual representations (Treagust et al., [57]). In contrast to the static and disorientation levels, at the complementation level, students easily achieved successful scientific visual representations because of a strong foundation of macroscopic and sub-microscopic understanding. Based on these findings, the scientific teaching process should integrate all three levels—macroscopic, sub-microscopic, and symbolic representation—to maximize students' learning outcomes (Lin et al., [38]). Incorporating all of these various representation types accommodates these diverse levels of understanding (macroscopic, sub-microscopic, and symbolic) and avoids the flaws of relying on a single teaching material (such as a textbook alone) to inform students' conceptual construction of heat convection (Chiou, [14]).

The current study has demonstrated that the process of creating visual representation in the context of CTS enables students to further explore individual concepts during the transition from one mode to another (Ainsworth, [2]; Tippett, [55]). In this process, students assimilate and accommodate knowledge. Assimilation occurs when students use their preconceptions to grasp a new concept, while accommodation occurs when they start to change their conceptions (Duit et al., [20]). Furthermore, research Jansoon et al. ([31]) have conducted reveals that students understand science material better when they use their visual representation skills at their respective levels. Via multilevel representations, students' representations can also be connected to each other (Treagust et al., [57]).

In addition, aiming for sub-microscopic representation in learning does not mean that students need to become science experts; rather, their learning must be facilitated coherently in school to improve their science mastery (Merino & Sanmartí, [40]). Learning should enrich students' experiences. In this process, teachers should motivate students to visualize, communicate, test, and reorganize the knowledge they are learning; this is the true goal of scientific activity in schools (Izquierdo-Aymerich & Adúriz-Bravo, [30]). With this in mind, teachers need to pay sufficient attention to their students' ability to represent their understandings not only at the macroscopic level, but also at the verbal, visual, and symbolic levels.

For supporting data can consider the appendixes and can refer to previous article made by author.

The authors declare no competing interests.

Please see Table 7.

Table 7 Course materials and pedagogical components in terms of convection concept in the context of the CTS of learning cycles

Question: The main question about the conception.

Fiyya is conducting an experiment by heating water in a clear container which there is contains wood powder. The aim of this experiment is to know how the movement of water is represented by the movement of the wood powder. For more details, look at the picture below!

What will happen to that experiment?

• The closest water with the heat source will rise and the far ones will be above it.
• The closest water with the heat source will rise and the far ones will replace their positions.
• The near and far water from the heat source will stay in its position or have no movement at all.
• If you have your own answer, please write it here.

Why can it happen in that experiment?

• The hotter water will have the same arrangement of particles with the cooler one and there are no changes in position in both water conditions.
• The hotter water will have more dense particles or become heavier than cooler one; therefore, the particles of hot water will go down and cooler water will go up.
• The hotter water will have more tenuous particles or become lighter than the cooler one, the result hot water will go up and cooler water will go down.
• If you have your own answer, please write it here.

Based on your explanation, how do you draw the flow and particle of water at points A and B (in the circle provided) in that experiment?

Please see Table 8.

Table 8 Students' drawing categories showing the level of visual representation

Please see Table 9.

Table 9 Rubric of visual representations