Revisiting secondary students’ ideas about air pollution. The challenge of particulate matter

Abstract

Many studies have researched students’ ideas about air pollution, basically focusing on nature and impact of gaseous pollutants on human health. However, recent research has highlighted the importance of the role of particulate matter air pollution for a good air quality in cities. This phenomenon is especially interesting for exploring the limits of the particulate model of matter at the mesoscale with secondary students. The purpose of this research is to investigate the ideas of 14–15 year-old students about polluted air in terms of its structure and its nature and how these ideas change after the implementation of a model-based teaching and learning Sequence. An interpretative qualitative approach is used to explore students’ ideas and how they change. Pre- and post-multimodal representations of 205 secondary students were analysed. Results showed that a sophistication of students’ ideas about the nature of polluted air after the teaching and learning sequence is not necessarily related to the sophistication of its structure. Also, students’ ideas at the mesoscale are varied and include a range of different semicontinuous or discontinuous conceptions.

---

Introduction

The particulate model of matter is undoubtedly one of the big ideas of science (Harlen, 2010) and plays an important role in students’ scientific literacy (Harrison and Treagust, 2002). This model is useful to describe, explain and predict natural phenomena (Oh and Oh, 2011) and, using it, students have to be able to interpret and interact with their environment. Usually, phenomena proposed in science classrooms are those which are paradigmatic and really useful to contribute to the development of a model (Izquierdo-Aymerich and Adúriz-Bravo, 2003), such as ideal gases or transitions from one state of matter to another (Liu and Lesniak, 2006). However, these tend to be far from daily phenomena.

One of those phenomena related to the particulate model of matter is air pollution. The World Health Organization defines it as “contamination of the indoor or outdoor environment by any chemical, physical or biological agent that modifies the natural characteristics of the atmosphere” (World Health Organization [WHO], 2021), which include gaseous pollutants such as carbon dioxide, related to global damages, or nitrogen oxides, related to local effects, but also solid pollutants that could be in suspension in the air such as particulate matter (PM). Thus, to understand air pollution, to act and to design solutions, students not only have to describe, explain and understand the behaviour of gases, but also how it is possible that solid pollutants are sustained in the air. Pollution caused by PM is especially interesting for working on the limits of the particulate model of matter, because understanding what they are and how they behave implies focusing on a phenomenon that occurs at the mesoscale, between 10⁻⁷ m (i.e. PM0.1, particulate matter from combustion) and 10⁻⁵ m (i.e. PM10, dust particles) (Solé et al., 2020).

This type of phenomenon at the mesoscale, such as air pollution caused by PMs but also others such as microplastics contamination in water, leads us to wonder whether the model of the structure of matter usually proposed in science classrooms has enough descriptive, predictive and explanatory power in this context. To address this issue, the purpose of this research is to investigate students’ ideas about air pollution, particularly regarding the nature and the structure of polluted air and, how these ideas change before and after a model-based teaching and learning sequence (TLS). The aim is not only to explore how the students’ ideas change in a TLS specifically designed with this goal, but also to provide insights for more adequate teaching of this model in the classroom.

The research in science education showed that explorations of students’ ideas about science topics are crucial for designing classroom activities. From the 1990s, a body of knowledge has been developed about student's ideas (see for example Driver et al., (1994)). However, the literature reported different approaches on how to understand the status and nature of those ideas and how they could be represented, including terms such as ‘conceptions’, ‘misconceptions’, ‘preconceptions’ and ‘mental models’ among others (Taber, 2013b). In the present research, we use the term ‘ideas’ to refer to those conceptions that we can interpret from the students’ drawings and explanations about a scientific topic from a broad perspective. As such we are not examining knowledges structures, as in the case of mental models, nor are we making judgements as in the case of misconceptions or alternative conceptions (Taber, 2013b).

Students’ ideas related to the nature of air pollution

Some authors have undertaken research on the conceptions of students of different educational levels about air pollution (Skamp et al., 2004; Myers et al., 2004; Boyes et al., 2007; Dimitriou and Christidou, 2007; Mandrikas et al., 2017; Tena and Couso, 2021). These studies are mainly focused on the students’ ideas about what airborne gaseous pollutants are (Thornber et al., 1999; Skamp et al., 2004; Boyes et al., 2007), how they behave (Mandrikas et al., 2017) or how these pollutants impact on the environment and living things (Skamp et al., 2004; Dimitriou and Christidou, 2007). However, there has been little research to study what students’ ideas are related to non-gaseous air pollutants such as particulate matter (PM) and its behaviour (Tena and Couso, 2021). Previous studies have focused on conceptions related to the global effects of air pollution, such as climate change, acid rain and the hole in the ozone layer caused by gaseous pollutants (Skamp et al., 2004; Boyes et al., 2007), but students' ideas related to the local effects of air pollution are unresearched, as is the case of PM.

Despite the fact that most previous studies are based on the ideas of primary school students (7 to 11 years) about air pollution, Skamp et al. (2004) conducted research with students from 11 to 16 years that indicates that there is a prevalence of ideas about some air pollution concepts across ages. Thus, in order to explore the challenges identified in the literature, we have considered studies including primary and secondary students.

Mostly, students associate their ideas about air pollution with its origin, highlighting transport and industry as the main sources (Thornber et al., 1999; Pruneau et al., 2005), but also viruses and pathogens (Dimitriou and Christidou, 2007; Tena and Couso, 2021). In terms of the gases present in polluted air that students named, the results of the closed questionnaire explored by Thornber et al. (1999) show that almost all students indicate that in the polluted air there are “extra” things that should not be there, and almost a fifth of their sample mention that carbon dioxide is only characteristic of polluted air. However, Myers et al. (2004) point out that as students grow most of them appreciate that carbon dioxide is also in unpolluted air.

These authors have reported many difficulties in understanding the phenomenon of gaseous air pollution, but also many difficulties in understanding “normal” air (Myers et al., 2004). For young students, in terms of nature, it is challenging to understand that we are surrounded by a colourless, odourless, tasteless substance that has particular properties such as density (Driver et al., 1994), and as they grow older the challenge is to understand that this substance is a mixture of gases that can be differentiated (Skamp et al., 2004; Tena and Couso, 2021).

Taking into account that normal air has already been shown to be a challenging substance to describe, it is expected that new difficulties will arise for students when having to fit into their alternative model of air the new information regarding air pollution that they receive both at school and from out-of-school activities (Thornber et al., 1999).

Students’ ideas about the structure of matter applied to the PM air pollution phenomenon

These previous studies have explored students’ ideas related to the nature of gaseous air pollution. But, despite the fact that students may mention the name of particular gases, such as CFCs (chlorofluorocarbons), carbon monoxide, or carbon dioxide (Thornber et al., 1999), it is probable that they do not recognize them as particulate entities at the submicro scale (Skamp et al., 2004; Tena and Couso, 2021). Accordingly, regarding the specific topic of air, the research done by Vilardo et al., (2017) shows that, prior to a classroom intervention, one third of their students represent air at the submicro level but none of them drew it correctly. This signals the importance of research and practice focusing not only on the nature of air pollution, but also on its structure. To reflect on the structure of air pollution implies delving into the particulate model of matter and scales.

When the literature explores the particulate model of matter, it usually focuses on the degree of discontinuity of matter and on whether or not students represent an underlying structure between particles (Talanquer, 2009; Hadenfeldt et al., 2014), among other things such as the movement of these particles or their interactions in the upper levels (Moltó et al., 2021). Thus, in his review Talanquer (2009) defines three dimensions, from novice to advanced learners, of the implicit assumptions that students have about the structure of matter: continuity, granularity (embedding) and corpuscularity (vacuum). Hadenfeldt et al. (2014), including the work of Talanquer (2009) among others, define 5 levels of sophistication related to the structure and composition of matter, also in terms of the particulate concept. These ideas are circumscribed at the submicro scale, but what happened at the other levels of description such as the mesoscale, and how these ideas could be connected with other levels, is a challenge for various reasons.

One of the difficulties associated with the corpuscularity of matter is the polysemic use of the term ‘particle’. In teaching proposals, the term ‘particle’ could be assimilated both to molecules or atoms (i.e. a particle of CO₂) and to small portions of a body (i.e. a dust particle), or even at higher educational levels to subatomic particles (i.e. in expressions such as ‘subatomic particles called nucleons’) (Blanco and Prieto, 1996; Taber, 2013a). In our discussion, this issue is very important because in the air pollution context we are adding a new meaning for the term ‘particle’ related to pollutant particles in suspension.

On the other hand, traditionally one of the objectives of secondary education in chemistry has been to develop micro-macro connections thinking (van Berkel et al., 2009; Gkitzia et al., 2020), connecting the properties of matter at the macroscale to its internal structure at the submicroscale and “scale literacy” (Gerlach et al., 2014). In their research, Gerlach et al., (2014) showed that students, no matter how experienced, were anchored in the human realm, and in consequence, all non-visible small objects were perceived as being similar in size (i.e. cells and viruses or bacteria).

Despite the fact that this is a key conceptual area, there are some phenomena that we cannot explain in terms of micro-macro thinking (Meijer et al., 2009), such as the transmission of a virus, the behaviour of surface friction, topics related to nanotechnology or particulate matter air pollution. For this type of phenomenon an intermediate level is necessary, the mesoscopic level (Besson and Viennot, 2004; Meijer et al., 2013). Thus, we advocate explicitly including this intermediate level, describing three different levels: macro, meso and submicro. The macro level is related to visible or measurable phenomena closely connected to the human scale (Meijer et al., 2013). The meso level is related to structures between 10⁻¹ and 10⁻⁹ m. The submicro level is related to the development of models such as atom or electron distribution (Gilbert and Treagust, 2009). Likewise, the submicro level encompasses different sublevels that add complexity to the models: the atomic-molecular level refers to atoms and molecules and the subatomic level refers to the internal structure of the atom (Caamaño, 2020). In the review of the literature, we usually find the mesoscopic level included in the micro level, including examples such as multimolecular structures (Caamaño, 2020) or colloids (Gilbert and Treagust, 2009).

Despite the consensus that both the structure and scale are important, very few studies have been conducted on students’ ideas about air pollution that include both dimensions. In previous research, Tena and Couso (2021) explored primary students’ (10 to 12 year-olds) ideas about air pollution in terms of its nature and its structure identifying two main challenges: understanding that air is a mixture and moving towards the idea of air (and polluted air) as a non-continuous substance. In the light of these results, and all of the above, we have researched secondary students’ ideas about air pollution.

Models, target model and modelling

To develop a sophisticated particulate model of matter implies a progressive development of ideas in which some of them could act as stepping stones for more advanced ideas to be developed in the future (Campbell et al., 2016). Thus, a convenient strategy is to define target models suitable for each educational level (Rea-Ramirez, 2008). In order to establish the key ideas of the target model, Duit (2007) proposed the Model of Educational Reconstruction, in which didactical content is constructed by taking into account both scientific ideas and previous research on teaching and learning of a particular topic. Following the Model of Educational Reconstruction, the definition of the target model comprised the following:

– Analysis of the subject matter based on public understanding of science publications (e.g.Basagaña (2018)) and interviews with experts in different fields: air pollution, epidemiology, public health and air pollution sensors developers.

– Review of previous research studies about students’ ideas of air pollution.

– Preliminary research study about 14 to 15 year-old students’ ideas about air pollution conducted by the authors (Solé et al., 2020).

From this process, we have defined the core ideas of the target model about air pollution for secondary students (14–15 years old) described in Fig. 1.

Fig. 1 Core ideas of the target model about air pollution addressed in the TLS designed.

Modelling is a privileged instructional approach to help students master the target model (Clement, 2000). It consists of successive and progressive approximations in which students express their initial ideas about a phenomenon, test their ideas, review their ideas based on their new observations and information and finally express a consensus model (Schwarz et al., 2009; Couso and Garrido-Espeja, 2017). The focus on modelling is due not only to its epistemic importance as a crucial scientific practice (Osborne, 2014) but also to the fact that it is an instructional approach that has been shown to be helpful for the design of model and modelling-based teaching and learning sequences (Windschitl et al., 2008; Hernández et al., 2015).

Objectives of the research

In this research we explore students’ ideas about air pollution before and after participating in a model-based TLS designed to promote the development of core ideas about air pollution defined in Fig. 1. Based on the foregoing, we defined the following research questions:

• What are 14 to 15 year-old students’ ideas about polluted air in terms of its structure and its nature?

• How do these students’ ideas change after the implementation of a model-based teaching and learning sequence?

Context and methods

Research context

The present study has been developed in the context of a citizen science project called ‘Projecte ATENC!Ó’ (https://www.projecteatencio.cat) in Barcelona. The project was approved by the Parc de Salut Mar Clinical Research Ethics Committee (approval number: 2018/7968/I).

In this context, a TLS was designed following a Design-Based-Research (DBR Collective, 2003) methodology, through an iterative process that consists of iterative cycles of design, implementation in real classrooms and evidence-based modifications during the 2018–2019 and 2019–2020 academic years. The TLS consisted of 12 classroom hours divided in two modules (6 classroom hours each). The first module was based on the modelling cycle (Couso and Garrido-Espeja, 2017) and aimed to develop a sophisticated idea about the nature and the structure of polluted air, focusing on the behaviour of particulate matter at the mesoscale (activities A10 and A11 in the Appendix I). The core ideas of the target model that students were expected to grasp are included in Fig. 1 and a brief description of the full activities can be found in Appendix I, relating each core idea with classroom activities. The second module was inquiry-based (Windschitl et al., 2008) and aims that students design and conduct their own research about air pollution in their high school. The TLS is published in the open repository of the Universitat Autònoma de Barcelona: https://ddd.uab.cat/record/201543?ln=ca.

Data collection

Throughout the 2019–2020 academic year, a total of 12 high schools from Barcelona and its metropolitan area interested in the air pollution issue participated in the project with students aged 14 to 15. Of these, only 4 participating high schools were able to complete all the TLS in the first term of the course due to the lockdown caused by the COVID-19 pandemic. The 205 students from these 4 public high schools make up the convenience sample selected for data gathering and analysis purposes.

We collected data from the individual productions of the students answering the first and the last activity in the first module of the TLS. We asked students to draw and describe in written form how they conceived of the air in a hypothetical sample of polluted air. Drawings, and not only written explanations, are used because sketching is a powerful assessment tool that can show the complexity of students’ ideas that other assessment instruments do not often capture (Cooper et al., 2017). Furthermore, to promote the use of their ideas at different scales, we asked them to draw and describe the polluted air both as seen with the naked eye (Question A) and how they imagine it inside (Question B).

All participants and their legal guardians were informed about the main objectives of the research. Furthermore, all the teachers participating in the implementation of the TLS were informed about the objectives of the research and signed informed consent. On the other hand, preliminary results were shared with students and teachers in the final event of the project.

Data analysis

In order to answer our research questions, we decided to explore students’ ideas using an interpretative qualitative approach. To classify students’ productions about polluted air, a system of coding was created using the constant comparative method (Kolb, 2012), considering both categories emerging from the data, and categories identified in previous research, including the categories coding system proposed by Tena and Couso (2021) in the case of primary students in the same context of polluted air.

The coding system is shown in Table 1 and is formed by two dimensions: students’ ideas regarding the nature of polluted air and students’ ideas regarding the structure of polluted air.

Table 1 Categories used to codify students’ representations of air with pollution in terms of its structure and its nature

Subdimension		Category
Dimension: Students’ ideas regarding the nature of polluted air
Type of components	Components present in the normal atmosphere	Representations that include the major components present in the normal atmosphere, such as nitrogen, oxygen or argon.
		Representations that include secondary components present in the normal atmosphere, such as carbon dioxide.
		Representations that include minority components present in the normal atmosphere, such as noble gases.
	Components not present in the normal atmosphere	Representations that include components related to pathogens, viruses or bacteria.
		Representations that include inert components related to human activity, such as components derived from industrial activity or traffic.
		Representations that include inert components not directly related to human activity, such as pollen or dust.
Number of components		Representations that only include one component of air with or without pollution.
		Representations that include more than one component of air with or without pollution.

---

Subdimension	Category
Dimension: Students’ ideas regarding the structure of polluted air
Continuity of matter	Continuous visions: representations in which air with or without pollution is represented as a continuous substance with no underlying structure.
	Semicontinuous visions: representations in which air with or without pollution is represented as a continuous substance with small particles embedded, understanding the term ‘particle’ in a broad form (particle with the same characteristics as the substance, molecules, atoms, etc.).
	Discontinuous representations: representations in which air with or without pollution is represented as small particles, understanding the term ‘particle’ in a broad form (particle with the same characteristics as the substance, molecules, atoms, etc.), without any supporting material between them.
Scale of matter	Macroscale representations: representations of air with and without pollution that refer to the structure of matter that can be observed with the naked eye, without the need to use a magnifier or a microscope, and closely connected to the human scale.
	Mesoscale representations: representations of air with and without pollution that refer to particles, with or without the same characteristics as the substance, which could be seen with a magnifier or a microscope.
	Atomic-molecular scale representations: representations of air with and without pollution that refer explicitly to atoms or molecules.
	Subatomic scale representations: representations of air with and without pollution that refer explicitly to subatomic particles, such as electrons, protons or neutrons.

---

In addition to the coding system proposed by Tena and Couso (2021) for primary students, we have detailed some other categories to include more sophisticated ideas of high school students. For the dimension related to the ideas regarding nature, we have specified the categories of the components present in the normal atmosphere, because high school students are expected to mention them specifically, especially regarding well-known components such as nitrogen, oxygen or carbon dioxide (Skamp et al., 2004).

To analyse the nature of polluted air, we used students’ answers to both questions A and B as data. However, for the sake of capturing the ideas on the structure of polluted air, we have only used the answers to question B, in which students have to report their views on the continuity and scale of polluted air.

For each dimension, students’ ideas were categorized according to the different categories of Table 1. Following this analysis, each student was coded with one category related to the subdimensions continuity of matter, scale of matter and number of present components. However, each student could be categorized with more than one category related to the components present in the air, because they are not mutually exclusive.

When it was necessary quantitative methods were used and, in particular, co-occurrence analysis. For each dimension, the co-occurrence analysis cross the two subdimensions defined in Table 1 in order to identify relations among them.

To increase the reliability of the analysis and to reduce researcher bias, an investigator triangulation of category assignment by the three authors was used (Carter et al., 2014; Cohen, Lawrence, Morrison, 2018). The first author made an initial categorization and selected a representative sample of students’ representations to share and discuss with the other authors. When the categorization of this sample was agreed, the first author used the criteria established for the whole group.

Results and discussion

Students’ ideas regarding the nature of polluted air

In terms of nature of polluted air, the target model expected in final students’ productions includes representations with more than one component and a diversity of components as large as possible (e.g. polluted air as a mixture of gases, including gases present in normal air, gases associated with human activity and PM).

Regarding the number of components of polluted air, initial productions include representations of polluted air as a single substance (41%) and representations of polluted air as a mixture of substances (50%), with 9% of students not addressing the nature of polluted air in their representations (see Fig. 2). After the TLS, it is remarkable that all students’ productions specify components of polluted air. Although representations that present polluted air as a mixture of substances increased to 64%, productions showing polluted air as a single substance still amounted to 36%. These results agree with previous findings which signal the challenge of understanding air, and consequently polluted air, as a mixture (Skamp et al., 2004).

Fig. 2 Example of students' representations and their explanations regarding the “Number of components present” subdimension.

Regarding the components present in polluted air, in initial representations only 16% include at least one component present in normal atmosphere. This idea agrees with Thornber et al. (1999), who point out that students understand air pollution as extra things in the air which are not usually there, forgetting the components of normal air, such as nitrogen and oxygen. After the TLS, this number increases and 28% of representations include components present in normal atmosphere.

A deeper analysis about these components shows that, initially, 30% of representations include secondary components, which are basically CO₂, and only 16% of representations include major components such as nitrogen or oxygen (see Fig. 3). Such representations are coherent with the common idea that air pollution is basically CO₂, due to its relation with global warming. After the TLS, the number of representations which include major components almost doubles (29%). Also, the number of representations which include secondary or minor components increases (37% secondary components and 15% minority components). It can be observed that there is an increase in students’ productions that include gases present in normal atmosphere when representing polluted air, which shows that air pollution is not being understood as something that replaces normal air.

Fig. 3 Examples of students’ representations and their explanations regarding the “Components present in normal atmosphere” subdimension. Keywords marked in bold written in these representations were used to categorize.

On the other hand, in polluted air we have to consider those components which are not present in normal atmosphere (such as those produced by human activities, pathogens or other pollutants not related to human activities). Initially, 20% of representations include components related to human activity, such as industry or traffic smoke, which according to previous research are the main air pollutants declared by students (Pruneau et al., 2005). Also, there are some representations that include pathogens (11%) and other components not related to human activity (8%) (see Fig. 4). After the TLS, the number of representations which include air pollutants related to human activity increases to 41%, with these types of pollutant being the main component included in students’ representations of polluted air. Final representations which include pathogens or components not related to human activity are anecdotal. Despite the fact that usually pathogens, viruses and bacteria are not considered as air pollutants in traditional research studies, the context of the Sars-Cov-2 pandemic could have affected students’ perception.

Fig. 4 Examples of students’ representations and their explanations regarding the “Components not present in normal atmosphere” subdimension.

Additionally, we have to mention that almost the same number of representations at the beginning (32%) and after the TLS (39%) do not specify the exact origin of polluted air, for example writing that ‘polluted air is made up of polluted/dirty particles’ or ‘made up of dirty air’. The fact that “particle” is a well-recognized polysemic concept (Bucat and Mocerino, 2009), particularly in the topic of air pollution (PMs are particulate matter but the air is also made up of a different scale of particles) (Solé et al., 2020), could be one of the reasons why the number of representations, even after the TLS, is large.

The co-occurrence analysis crossing students’ views of the number and type of components present in polluted air is included in Fig. 5. Percentages of initial and final representations for each category are shown in light and dark grey, respectively.

Fig. 5 Co-occurrence analysis between the components present and the number of components of the nature of initial and final students’ representations. Percentages of initial and final representations for each category are shown in light and dark grey, respectively.

The co-occurrence analysis shows that the number of representations that mention the same polluted air or polluted particles as the only component of polluted air is maintained before and after the TLS. However, we observe a decrease of representations in which polluted air is only made up of CO₂, overcoming one of the common misconceptions related to air pollution.

In representations which include more than one component, the final productions are richer than the initial ones for two reasons. Regarding the components present in normal atmosphere, in final representations there is an increase in major, secondary and minority components presented, breaking with the idea that air pollution is something that replaces normal air. Furthermore, there is an increase in representations which mention human activities as a source of air pollution. Otherwise, after the implementation, pathogens and viruses and components not related to human activity mostly disappear in agreement with the fact that these two possible forms of air pollution are not addressed in the TLS.

Students’ ideas regarding the structure of polluted air

In order to establish the structure of polluted air we have analysed the answers to question B. That question was expected to appeal to students’ ideas about continuity and scale compatible with a particulate model of matter. In terms of structure, the target model expected in final representations considers polluted air as a semicontinuous entity at the mesoscale (e.g. air in the background with PM in it) or, otherwise, representations that show polluted air as a discontinuous entity at the atomic-molecular scale (e.g. molecules of some gases and PM with no underlying structure).

Regarding continuity, before and after the TLS the majority of students represent polluted air in semicontinuous (42% in initial and 46% in final representations) or discontinuous forms (41% in initial and 44% in final representations) (see Fig. 6), and there is no essential change between pre and post representations. As such, the air pollution phenomenon per se is shown to be a powerful context in which students, even before the TLS, represent matter in non-continuous forms, avoiding the view of continuity that is a common misconception in the field (Talanquer, 2009). Furthermore, after the TLS there is a decrease in students who represent polluted air as a continuous substance, from 17% to 7%, showing a sophistication of the ideas according to the target model described.

Fig. 6 Examples of students’ representations and their explanations regarding the “Continuity of matter” subdimension.

Regarding the scale, the majority of students represent polluted air at the mesoscale before and after the TLS, showing an increase in this type of representations in their final productions (60% and 68%, respectively). Similarly, the number of representations at the atomic-molecular scale was maintained between initial (19%) and final (20%) productions. On the other hand, in initial productions 21% of students represent polluted air at the macroscale. However, in final productions this number had decreased to 8% of students (see Fig. 7). Despite the fact that in initial representations there is a large number of representations at the mesoscale, the TLS that is focused on PM air pollution seems to slightly increase the representations in which polluted air is represented at the mesoscale, sophisticating the representations from macroscale to mesoscale.

Fig. 7 Examples of students’ representations and their explanations regarding the “Scale of matter” subdimension.

The co-occurrence analysis crossing students’ views on both continuity and scale is included in Fig. 8 and allows a deeper understanding of the studied phenomenon. Initial representations are shown in light grey and final representations are shown in dark grey.

Fig. 8 Co-occurrence analysis between continuity and scale of the structure of the initial and final students’ representations. Initial representations are shown in light grey and final representations are shown in dark grey.

Semicontinuous representations at the mesoscale are predominant in both initial (36%) and final (41%) productions, being one of the options of the target model in terms of structure discussed above. The large number of representations in this configuration (semicontinuous at the mesoscale) suggests the need to go beyond macro–micro thinking (Meijer et al., 2013), including meso thinking explicitly.

However, the number of students’ productions at the mesoscale using discontinuous representations is also remarkable before and after the TLS. This includes, for example, representations showing PMs with no underlying substance such as air. Despite the fact that these representations are apparently more sophisticated, because they are discontinuous and overcome one of the learning challenges of the particulate model of matter (Talanquer, 2009), the scale of representation (at the mesoscale instead of at the atomic-molecular scale) shows that there is an added challenge that has to be addressed. As we discussed in the literature review, these intermediate ideas could act as stepping stones to continue progressing towards a more sophisticated understanding of the particulate model of matter (Campbell et al., 2016).

On the other hand, the major change from pre to post students’ productions is the decrease in continuous representations at the macroscale. This suggests that the implementation of the TLS helped students to represent the inside of matter as non-continuous.

Conclusions

Even though the particulate model of matter is one of the key scientific models and, as such, there is a great deal of research on students’ ideas of matter, when these ideas are applied to a challenging new phenomenon at the mesoscale, such as air pollution caused by PMs, we still need further research about them. The independent analysis of the students’ representations in terms of the nature and structure of polluted air allows us to understand better the main opportunities and obstacles of teaching and learning about this important phenomenon.

Although there is a difference between the nature of polluted air in students’ representations before and after the TLS, the same level of improvement is not observed when analysing them in terms of structure. This implies that interventions able to help students to sophisticate their ideas regarding the nature of polluted air do not necessarily help them to sophisticate their ideas regarding its structure. For example, after the TLS a student could mention that polluted air is composed of nitrogen, oxygen, carbon dioxide and PMs, or even write these gases using molecular formulas (N₂, O₂, CO₂) but, at the same time, they could represent all these gases in a continuous manner. These results provide insights about the relationship between nature and structure of polluted air. Ideas related to the nature of polluted air are shown to be easier to learn in sophisticated ways than ideas related to its structure. Teaching and learning interventions that focused only on components of air pollution could not be expected to produce understandings regarding its structure. This is problematic because ideas related to the mesoscale and the particulate model of matter are crucial to understand the phenomenon and to act and design solutions to the problem, e.g. understanding why many councils recommend watering construction sites during air pollution episodes. Hence, to reflect with students about the submicro-, meso- and macro-scale and the continuity/discontinuity of matter at each scale when compared to the other should be an explicit aim in teaching interventions on this and related topics.

Our results show that some students' chemical representations and explanations present a lack of coherence among the macro, micro and symbolic levels, as defined by Johnstone (1993). For example, a student who writes the molecular formula of oxygen “O₂” (symbolic level) as a component of polluted air could be representing this substance at any scale (macro, meso or submicro). This means that students' expression of chemical formulas as symbolic representations is not necessarily related to these students' interpretation of matter at a submicroscopic level in terms of molecules or in terms of discontinuity (Taber, 2013a). Thus, in order to help students to sophisticate their ideas, teachers should go beyond the use of symbolic representations, and promote classroom discussions in terms of scale and continuity of polluted air.

Additionally, students’ understandings of the mesoscale have been shown to be varied, including a range of different conceptions in which students could represent polluted air in semicontinuous or discontinuous forms. This is related to the fact that matter at the mesoscale has properties close to those at the macroscale (e.g. some PMs could be observed with a magnifier) but also properties close to the submicroscale (e.g. the movement of these PMs). The TLS focused on the behaviour of such PMs at the mesoscale included an activity where students have to order different images of different sizes, according to activities that research has shown to be successful to improve scale literacy (Gerlach et al., 2014). The TLS was shown to be useful to decrease students' representations at the continuous and macroscale and to increase those representations at the mesoscale. However, the link between meso- and submicro-scale should be addressed in order to promote the ability to relate different scales. In accordance with the results, from a Design-Based-Research perspective (DBR Collective, 2003) a new version of the TLS was designed and published in an open-access repository (https://ddd.uab.cat/record/201543?ln=ca). As such, we consider that introducing phenomena at the mesoscale in secondary school is useful for students to test their ideas regarding the particulate model of matter, and this fact should be explicitly addressed with students.

Finally, the necessity to address these challenges in classrooms is not only because we have observed students’ limitations in their ideas of sophisticated versions of the particulate model of matter. Apart from air pollution, there is a growing variety of phenomena that involve reasoning at the mesoscale, such as microplastics contamination (Raab and Bogner, 2021), nano devices (Tretter et al., 2006) or the transmission of nanoparticles or viruses. Understanding all these relevant phenomena is part of students’ scientific literacy and citizenship and requires a sophisticated understanding of the particulate model of matter that clearly focuses on both nature and structure at different scales.

Conflicts of interest

There are no conflicts to declare.

Appendix 1. Sequence of activities of the TLS in relation to the phase of the modelling cycle and the core ideas addressed

Phase of the modelling cycle	Core ideas	Sequence of activities
1. Anchoring phenomenon	Idea 1, 2, 3	A1. The air pollution phenomenon is presented through a real news item entitled “Children exposed to air pollution in schools could have more risk of suffering overweight and obesity”. The context of the project is presented and students are asked what they think about air pollution in general terms.
2. Asking for the explicit expression or use of their initial model		A2. Students are asked to draw and describe in written form the air from two hypothetical samples: one polluted and the other without pollution. They have to describe the samples as seen with the naked eye and how they imagine them inside.
3. Empirically testing the model	Idea 1	A3.1. In small groups, they are asked to compare the different explanations of clean air in order to test their ideas.
4. Generating new points of view		A3.2. The explanations of all small groups are shared and a final consensus model of clean air is structured.
5. Facilitating the structure in a final consensus model
1. Anchoring phenomenon	Idea 2	A4. To connect their ideas about clean air with polluted air, it is asked if they think that the air of their city is exactly as they explained in the previous activity.
2. Asking for the explicit expression or use of their initial model	Idea 2.1	A5.1. An experiment to reproduce the combustion of a vehicle is proposed. In small groups, students have to burn a piece of paper and observe what happens with the watch glass above the smoke. Prior to reproducing the experiment, they are asked to think about their hypothesis and their explanations.
3. Empirically testing the model		A5.2. The experiment is conducted and students have to observe what happened with the watch glass using a microscope. From the results observed, students are asked to reconstruct their ideas about what smoke is.
4. Generating new points of view		A6. A video of how scientists study air pollution in cities is shown and an analogy with the experiment conducted is made.
4. Generating new points of view	Idea 2.2	A7. From the explanations made in A2 and an extract from a scientific paper, air pollution caused by gases is discussed.
4. Generating new points of view	Idea 2.3	A8. Is CO2 an air pollutant in cities? From the air quality indicators shown on TV news (O3, NO2 and PM10 are shown as air quality indicators), it is discussed whether or not CO2 is an air pollutant.
5. Facilitating the structure in a final consensus model	Idea 2	A9. To explain a final consensus model about air pollution.
4. Generating new points of view	Idea 3.1	A10. From an extract of a public understanding of science paper which explains the impact of PM on human health, we focus on what PM air pollution is like. In order to reflect on the scale and size of PM, students are asked to order some images according to the size of what is represented. These images include a grain of salt, a virus, the principal molecules which compose air, PM1 and PM10.
5. Facilitating the structure in a final consensus model		A11. After reflecting on the size of PMs, they are asked how it is possible that PMs are suspended in the air. A video of movement of solid particles that include increasing and decreasing magnifications is seen. Finally, with the guidance of the teacher, ideas related to PM are summarised.
2. Asking for the explicit expression or use of their initial model	Idea 3.2	A12. Related to activity A10, they are asked what they think the effects of air pollution on human health are.
3. Empirically testing the model		A13. Students are asked to classify the effects of air pollution expressed in the different human systems and the mechanism that they think causes the illness.
4. Generating new points of view		A14. After watching a video and reading part of a paper about the effects of air pollution on human health, students have to rewrite the previous activity with the new ideas.
5. Facilitating the structure in a final consensus model
5. Facilitating the structure in a final consensus model	Idea 1, 2, 3	A15. Students are asked to draw and describe in written form the air from two hypothetical samples: one polluted and the other without pollution. They have to describe the samples as seen with the naked eye and how they imagine them inside. Finally, students are asked to reflect on how their explanations have changed from activity A2.

Acknowledgements

This research was conducted in the PhD Education program at the Universitat Autònoma de Barcelona and funded by Spanish Government under the ESPIGA Project (PGC2018-096581-B-C21), the ACELEC group (2017SGR1399) and “la Caixa” Foundation (ID100010434) under the RecerCaixa program (2017ACUP00274). C. S. was supported by a predoctoral contract under the ESPIGA Project (PRE2019-087419).

Notes and references

1. Basagaña X., (2018), Els efectes de la contaminació en les persones: què en sabem I de què busquem evidències? Ciències, 35, 28–34.
2. Besson U. and Viennot L., (2004), Using models at the mesoscopic scale in teaching physics: Two experimental interventions in solid friction and fluid statics, Int. J. Sci. Educ., 26(9), 1083–1110.
3. Blanco Á. and Prieto T., (1996), Algunas cuestiones sobre la comprensión de la Química desde la perspectiva de las “ideas de los alumnos”, Investigación En La Escuela, 28, 69–78.
4. Boyes E., Myers G., Skamp K., Stanisstreet M. and Yeung S., (2007), Air quality: A comparison of students’ conceptions and attitudes across the continents, Compare, 37(4), 425–445.
5. Bucat B. and Mocerino M., (2009), Learning at he Sub-micro Level: Structural Representations, in Gilbert J. and Treagust D. (ed.), Multiple Representations in Chemical Education, Dordrecht: Springer, pp. 11–29.
6. Caamaño A., (2020), Enseñar química. De las sustancias a la reacción química, Barcelona: Graó.
7. Campbell T., Schwarz C. and Windschitl M., (2016), What We Call Misconceptions May Be Necessary Stepping-Stones Toward Making Sense of the World, Sci. Teach., 83, 69–74.
8. Carter N., Bryant-Lukosius D., Dicenso A., Blythe J. and Neville A. J., (2014), The use of triangulation in qualitative research, Oncol. Nurs. Forum, 41(5), 545–547.
9. Clement J., (2000), Model based learning as a key research area for science education, Int. J. Sci. Educ., 22(9), 1041–1053.
10. Cohen L., Lawrence M. and Morrison K., (2018), Research Methods in Education, New York: Routledge.
11. Cooper M., Stieff M. and DeSutter D., (2017), Sketching the Invisible to Predict the Visible: From Drawing to Modeling in Chemistry, Top. Cogn. Sci., 9(4), 902–920.
12. Couso D. and Garrido-Espeja A., (2017), Models and Modelling in Pre-service Teacher Education: Why We Need Both, in Hahl K., Juuti K., Lampiselkä J., Uitto A. and Lavonen J. (ed.), Cognitive and Affective Aspects in Science Education Research. Selected Papers from the ESERA 2015 Conference, Dordrecht: Springer, pp. 245–261.
13. Dimitriou A. and Christidou V., (2007), Pupils’ understanding of air pollution, J. Biol. Educ., 42(1), 24–29.
14. Driver R., Squires A., Rushworth P. and Wood-Robinson V., (1994), Part III: Materials and their properties, Making sense of secondary science: Research into children's ideas, London: Routledge, pp. 137–229.
15. Duit R., (2007), Science Education Research Internationally: Conceptions, Research Methods, Domains of Research, Eurasia J. Math. Sci. Technol. Educ., 3(1), 3–15.
16. Gerlach K., Trate J. M., Blecking A., Geissinger P. and Murphy K., (2014), Investigation of Absolute and Relative Scaling Conceptions of Students in Introductory College Chemistry Courses, J. Chem. Educ., 91, 1526–1537.
17. Gilbert J. and Treagust D., (2009), Introduction: Macro, Submicro and Symbolic Representations and the Relationship Between Them: Key Models in Chemical Education, in Gilbert J. and Treagust D. (ed.), Multiple Representations in Chemical Education, Dordrecht: Springer, pp. 1–8.
18. Gkitzia V., Salta K. and Tzougraki C., (2020), Students’ competence in translating between different types of chemical representations, Chem. Educ. Res. Pract., 21(1), 307–330.
19. Hadenfeldt J. C., Liu X. and Neumann K., (2014), Framing students’ progression in understanding matter: a review of previous research, Stud. Sci. Educ., 50(2), 181–208.
20. Harlen W., (2010), Principles and Big Ideas of Science Education, Hatfield: Association for Science Education.
21. Harrison A. G. and Treagust D., (2002), The Particulate Nature of Matter: Challenges in Understanding the Submicroscopic World, in Gilbert J., De Jong O., Justi R., Treagust D. and Van Driel J. (ed.), Chemical Education: Towards Research-based Practice, Dordrecht: Kluwer Academic Publishers, pp. 189–212.
22. Hernández M. I., Couso D. and Pintó R., (2015), Analyzing Students’ Learning Progressions Throughout a Teaching Sequence on Acoustic Properties of Materials with a Model-Based Inquiry Approach, J. Sci. Educ. Technol., 24, 356–377.
23. Izquierdo-Aymerich M. and Adúriz-Bravo A., (2003), Epistemological Foundations of School Science, Sci. Educ., 12, 27–43.
24. Johnstone A. H., (1993), The development of chemistry teaching: A changing response to changing demand, J. Chem. Educ., 70(9), 701–705.
25. Kolb S., (2012), Grounded Theory and the Constant Comparative Method: Valid Research Strategies for Educators. J. Emerg. Trends Educ. Res. Policy Stud., 3(1), 83–86.
26. Liu X. and Lesniak K., (2006), Progression in children's understanding of the matter concept from elementary to high school, J. Res. Sci. Teach., 43(3), 320–347.
27. Mandrikas A., Stavrou D. and Skordoulis C. (2017), Teaching Air Pollution in an Authentic Context, J. Sci. Educ. Technol., 26(2), 238–251.
28. Meijer M., Bulte A. and Pilot A., (2009), Structure–Property Relations Between Macro and Micro Representations: Relevant Meso-levels in Authentic Tasks, in Gilbert J. and Treagust D. (ed.), Multiple Representations in Chemical Education, Dordrecht: Springer, pp. 195–213.
29. Meijer M., Bulte A. and Pilot A., (2013), Macro–Micro Thinking with Structure–Property Relations: Integrating ‘Meso-levels’ in Secondary Education, in Tsaparlis G. and Sevian H. (ed.), Concepts of Matter in Science Education. Innovations in Science Education and Technology, Dordrecht: Springer, pp. 419–436.
30. Myers G., Boyes E. and Stanisstreet M., (2004), School students’ ideas about air pollution: Knowledge and attitudes, Res. Sci. Technol. Educ., 22(2), 133–152.
31. Oh S. P. and Oh S. J., (2011), What Teachers of Science Need to Know about Models: An overview, Int. J. Sci. Educ., 33(8), 1109–1130.
32. Osborne J., (2014), Teaching scientific practices, J. Sci. Teach. Educ., 25, 177–196.
33. Pruneau D., Richard J. F., Langis J., Albert G. and Cormier M., (2005), The evolution of children's ideas on pollution in the framework of experiential and socioconstructivist activities, Int. J. Environ. Sustainable Dev., 4(1), 17–34.
34. Raab P. and Bogner F., (2021), Conceptions of university students on microplastics in Germany, PLoS One, 16(9), 1–25.
35. Rea-Ramirez M. A., (2008), Determining Target Models and Effective Learning Pathways for Developing Understanding of Biological Topics, in Clement J. and Rea-Ramirez, M. A. (ed.), Model Based Learning and Instruction in Science, Dordrecht: Springer, pp. 45–58.
36. Schwarz C. V., Reiser B. J., Davis E. A., Kenyon L., Achér A., Fortus D., Shwartz Y., Hug B. and Krajcik J., (2009), Developing a learning progression for scientific modeling: Making scientific modeling accessible and meaningful for learners, J. Res. Sci. Teach., 46(6), 632–654.
37. Skamp K., Boyes E. and Stainstreet M., (2004), Students’ Ideas and Attitudes about Air Quality, Res. Sci. Educ., 34, 313–342.
38. Solé C., Tena È. and Couso D., (2020), ¿Qué modelo de materia explica la contaminación? Explorando el modelo corpuscular en la mesoescala. Alambique: Didáctica de Las Ciencias Experimentales, 101, 30–36.
39. Taber K. S., (2013a), Revisiting the chemistry triplet: Drawing upon the nature of chemical knowledge and the psychology of learning to inform chemistry education, Chem. Educ. Res. Pract., 14(2), 156–168.
40. Taber K. S., (2013b), Modelling Learners and Learning in Science Education, Dordrecht: Springer.
41. Talanquer V., (2009), On cognitive constraints and learning progressions: The case of “structure of matter”, Int. J. Sci. Educ., 31(15), 2123–2136.
42. Tena È. and Couso D., (2021), What Is City Air Made of? An Analysis of Pupils’ Conceptions of Clean and Polluted Air, in Levrini O. (ed.), Engaging with Contemporary Challenges through Science Education Research, Dordrecht: Springer, pp. 133–148.
43. Tretter T., Jones M., Andre T., Negishi A. and Minogue J., (2006), Conceptual boundaries and distances: Students’ and experts’ concepts of the scale of scientific phenomena. J. Res. Sci. Teach., 43(3), 282–319.
44. The Design-Based Research Collective, (2003), Design-Based Research: An Emerging Paradigm for Educational Inquiry, Educ. Res., 32(1), 5–8.
45. Thornber J., Stanisstreet M. and Boyes E., (1999), School Students’ Ideas about Air Pollution: Hindrance or Help for Learning? J. Sci. Educ. Technol., 8(1), 67–73.
46. van Berkel B., Pilot A. and Bulte A., (2009), Micro-Macro Thinking in Chemical Education: Why and How to Escape, in Gilbert J. and Treagust D. (ed.), Multiple Representations in Chemical Education, Dordrecht: Springer, pp. 31–54.
47. Vilardo D., MacKenzie A. and Yezierski E., (2017), Using Students’ Conceptions of Air To Evaluate a Guided-Inquiry Activity Classifying Matter Using Particulate Models, J. Chem. Educ., 94(2), 206–210.
48. Windschitl M., Thompson J. and Braaten M., (2008), Beyond the scientific method: Model-based inquiry as a new paradigm of preference for school science investigations, Sci. Educ., 92(5), 941–967.
49. World Health Organization, (2021), Air pollution, [online], available at: https://www.who.int/health-topics/air-pollution [accessed 20 December 2021].

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d2rp00117a

---