Analysis of High School Students’ Argumentative Dialogues in Different Modelling Situations

Abstract

Few empirical studies in Science Education have investigated the contributions of integrating scientific practices such as argumentation and modelling. In this article, I examine the characteristics of high school students’ argumentative dialogues in different modelling situations. From this, I discuss the influences of modelling and the nature of each situation analysed on the characteristics of the students’ argumentative dialogues. One didactic unit consisting of sets of modelling activities in everyday, scientific and socio-scientific situations was applied in a regular class. The tool that describes argumentative dialogues in science teaching contexts across the varied and interrelated dimensions was applied to high school students’ argumentative dialogues that took place during modelling situations. Data collection (involving audio and video recording plus observations made by the researcher) revealed that students engaged in different argumentative dialogues, which were made up of different types of dialogic and meta-dialogic moves. Most of these moves were relevant and also contributed to the construction of knowledge in all modelling situations. The results also show that the nature of the situation can influence specific aspects of students’ argumentation, but such influence does not interfere with the quality of their argumentative dialogues; the argumentative dialogues are connected to persuasion, information sharing and sharing the same idea in all modelling stages; and the modelling influences the students to engage in quality argumentative dialogues that ultimately contributes to the construction of knowledge of different natures. Implications for future research and classroom practice are presented and discussed.

1 Introduction

The goal of Science Education is to contribute to the qualification of people who are scientifically literate. In order to achieve this goal, Science teaching should favour opportunities to learn science, which means to understand the relevant scientific concepts, so that the person may take up a critical stand within society if and when necessary; learn about science, which means understanding the issues regarding the philosophy, history and methodology of science and learn how to do science, which means to engage in activities similar to those in which scientists take part in the development of the scientific knowledge, leading the students to acquire both the scientific knowledge itself as well as learning about scientific practices (Hodson, 1992).

In this regard, recently, international documents, especially the American document A Framework for K-12 Science Education (NRC, 2012), and studies (e.g. Bybee, 2011; Ferraz & Sasseron, 2017; Krajcik & Merritt, 2012; Reiser et al., 2012) have stated that working in a way that is integrated with those scientific practices is considered essential for science teaching, namely preparing questions and defining problems; preparing and using models; planning and executing investigations; analysing and interpreting data; using mathematics and computational thoughts; constructing explanations and developing solutions; getting involved in argumentation supported by evidence and obtaining, appraising and communicating information, can all strengthen the students’ learning of and about Science and also help to develop their skills (for example argumentative skills, leading investigations and preparation of explanations).

Studies such as Clement, (2000) and Justi and Gilbert, (2002) have defended the use of modelling as a promising approach to favour the students’ engagement in different scientific practices in an integrated manner. This suggestion is supported by the results of work projects in which some of these practices were considered. For instance, the presentation and use of models (Clement & Rea-Ramirez, 2008; Maia & Justi, 2009a; Mendonça & Justi, 2013); the analysis and interpretation of data (Maia & Justi, 2009a; Mendonça & Justi, 2013); the use of mathematical reasoning and computational thought (Louca et al., 2011); engagement in the preparation of arguments backed up by evidence (Mendonça & Justi, 2013; Passmore & Svoboda, 2012); the construction of explanations and development of solutions (Louca et al., 2011; Mendonça & Justi, 2013; Oliveira et al., 2015) and the obtaining, appraisal and communication of information (Mendonça & Justi, 2013). Some of these studies also support conceptual learning and/or development of skills.

On the other hand, the integration of scientific practices of modelling and argumentation is still rarely empirically studied (Evagorou et al., 2020; Mendonça & Justi, 2013). In order to contribute to discussions on this topic, I examine the characteristics of high school students’ argumentative dialogues in different modelling situations. From this, I discuss the influences of modelling and the nature of each of the situations analysed on the characteristics of the students’ argumentative dialogues. This research could broaden the discussions in the Science Education teaching field about the integration of modelling and argumentation for construction of knowledge by students.

2 Theoretical Background

2.1 Argumentation and Modelling in Science Education

There has been growing interest in researching the issue of argumentation within Science Education, as can be seen in the growing number of publications addressing the matter (e.g. Albe, 2008; Martins & Macagno, 2022; Duschl & Osborne, 2002; Evagorou et al., 2020; Kelly, 2014; Khishfe, 2012; Osborne, 2016; Osborne et al., 2004; Osborne et al., 2013; Ryu & Sandoval, 2012; Sandoval & Millwood, 2008; Venville & Dawson, 2010). The same has been observed for the development of proposals of modelling-based teaching and/or the contributions of this approach to teaching (e.g. Barab et al., 2000; Baumfalk et al., 2019; Clement, 2008; Evagorou et al., 2020; Krajcik & Merritt, 2012; Louca et al., 2011; Maia & Justi, 2009a; Prins et al., 2016; Oliveira et al., 2015; Santos et al., 2020). This growing number of studies is due to the fact that both argumentation and modelling may contribute to students’ learning of science and about Science, as well as for skill development among the students—aspects pointed out in the studies above.

Most of these studies have investigated one single practice, and only a few of them focused on the analysis of the integration of modelling and argumentation (e.g. Evagorou, et al., 2020; Mendonça & Justi, 2013; Passmore & Svoboda, 2012; Puig et al., 2017). For instance, Passmore and Svoboda, (2012) developed a tool to analyse how students learn when participating in scientific modelling and argumentation practices, as well as to explore its usefulness in promoting argumentation. These authors understand modelling based on the elements of constructing, using, evaluating and reviewing models. To them, the objectives of argumentation in the context of modelling are to make sense of the phenomenon (using evidence to support claims and explanations about the phenomenon); to articulate understanding (constructing of arguments to communicate them to others) and to persuade (attempting to convince someone of the validity of ideas) (Berland & Reiser, 2009). In that study, data were collected from different levels of education and schools and from various scientific modelling situations, and their results indicate that argumentation occurred at all stages of modelling when students interpreted the phenomenon, sought evidence for the investigated question, related data to the model supported by justifications and, lastly, evaluated the model.

Mendonça and Justi, (2013) also identified the occurrence of argumentation in all stages of the scientific modelling situations when students built and evaluated their knowledge, although these steps were different from those focused on the work discussed earlier. For instance, the teaching activities about ionic bonding and intermolecular bonding were structured from the theoretical framework ‘model of modelling diagram’ proposed by Justi and Gilbert, (2002). The authors showed that, in the stages of production and expression of the model, argumentative situations were more related to the construction of explanations than to persuasion, while in the stages of testing and evaluation of the model, argumentative situations were rather more related to persuasion.

Puig et al., (2017) investigated the contributions of integration of scientific practices of modelling and argumentation in the learning of genetics curriculum content by biology students aged between 15 and 17 years. The teaching activities were structured from the theoretical framework ‘model of modelling diagram’ proposed by Justi and Gilbert, (2002). From this study, the authors found a greater number of modelling operations (104) than argumentative (83) used by students. Although argumentation plays a central role in the construction of representations, the main objective of the activity was to build a representation and not to promote argumentation among students. Puig et al., (2017) concluded that the scientific practices of modelling and argumentation partially contributed to students’ learning about the genetics curriculum content. Furthermore, they suggested the existence of a mutual contribution between these practices due to the overlapping of modelling and argumentation operations.

More recently, Evagorou et al., (2020) explored how elementary school students use their models while discussing a socio-scientific issue and investigated whether and how the argumentation process is linked to the modelling process. Their results showed that students mostly used evidence in discussions and, to a lesser extent, rebuttals, with the latter occurring mostly during the model evaluation stage. Moreover, the students elaborated better arguments in the model evaluation phase than in the other stages. The authors concluded that modelling can help students move beyond the provided information when they try to analyse, qualitatively relate, explain or evaluate the relationships with the produced model.

Most of these studies were carried out solely in scientific situations; they presented or considered the occurrence of argumentative situations in all stages of modelling and/or illustrated the contribution of integration of scientific practices of modelling and argumentation for learning scientific concepts. These contributions are highly relevant to our field; however, for the relationship between argumentation and modelling to advance, I suggest that it is necessary to conduct research that aims to also understand the argumentation raised by students when they participate in modelling situations that have different natures (for instance, scientific, socio-scientific and everyday) and levels of teaching. It is this gap that the current study attempts to fill.

2.2 Theoretical Framework of Argumentation

Argumentation is normally conceived as the product of discourse—namely the arguments (intended as claims supported by one or more reasons) provided in written or oral texts—or the process of dialogically presenting different opinions supported by reasons (Nielsen, 2013; O’Keefe, 1977; Rapanta & Macagno, 2016).

In this study, the process and product of argumentation are investigated. For this, the Martins and Macagno (2022)’s framework was used because it considers that arguments are elements of extended dialogues—and as dialogical units, they pursue different goals, originate from distinct presuppositions and shape different interlocutors’ roles and moves (i.e. turns or utterances defined by their dialogical purpose). Thus, it can contribute to characterize the high school students’ argumentative dialogues in modelling situations that have different natures. From this analyse, it is possible to understand the influences of modelling and the nature of each of the situations analysed on the characteristics of the students’ argumentative dialogues.

To Martins and Macagno (2022), the act of arguing is a social, verbal, non-verbal, and rational activity that may involve other dialogical and meta-dialogical moves, such as those to deliberate, persuade, share information, inquiry, clarify meanings, among others.

The framework relies on six aspects/steps of analysis, which are detailed below.

The aspect 1 of the tool: Selection of the discussions to be analysed favours the identification and delimitation of argumentative situations in dialogues, i.e. the identification and delimitation of dialogues in which argumentation occurs. The output of this step is a selection of the potentially (or actually) argumentative dialogical sequences, which allows determining their presence and the extent thereof within the activity, or in different dialogical contexts. It provides a first screening, narrowing the further analysis to specific dialogues and dialogical episodes (Martins & Macagno, 2022).

The aspect 2 of the tool: identification of the dialogic nature(s) of each activity helps identify whether students engage in a dialogue contributing to the joint objective, and thus, there has been learning or construction of knowledge. These categories narrow the types of dialogue to the ones that emerge from the educational literature as the types of pedagogical dialogues, aimed at building and evaluating curricular knowledge of social and/or scientific nature(s) (Martins & Macagno, 2022).

The types of dialogues and their objectives are information sharing dialogue (Dis), which has the objective of exchanging information; persuasion dialogue (Dpe), which aims to resolve a conflict of ideas; discovery dialogue (Dds), which has the objective of elaborating explanation(s) related to the topic under discussion; inquiry dialogue (Din), which has the objective of investigating/testing the validity of one or more hypotheses from evidence and deliberation dialogue (Ddl), which has the objective of proposing a solution/action to a problem under discussion, or solving such problem.

From the first two aspects, it is possible to understand the overarching dialogical activity proposed by the teacher, the participants’ roles and the setting. This provides the overall ‘game’ the students are engaging in and is necessary for analysing how the individual moves relate to each other and to the overall activity (aspects 3 and 5) (Martins & Macagno, 2022).

The aspect 3 of the tool: identification of the nature of the move favours the identification of the dialogic and meta-dialogic moves expressed by the subjects in argumentative dialogues. Thus, the dialogical intention pursued by each move is identified. The objectives of the dialogic and meta-dialogic moves are: Dialogical eristic Move (DeriM), which aims to attack and defend the person; Dialogical deliberative Move (DdelM), which aims to request a decision or solution to a problem, and select or suggest a more coherent/appropriate solution/action to be considered; Dialogical Information Sharing Move (DishM), which aims to obtain or provide information in response to a request/question; Dialogical inquiry Move (DinqM), which aims to investigate the validity of hypotheses from evidence or request their investigation; Dialogical persuasive Move (DperM), which aims to convince the other of the validity of an idea or request that this be done, or resolve a conflict of ideas in order to analyse each idea or request that the conflict be resolved; Dialogical persuasive dispute Move (DpdiM), which aims to resolve a conflict of ideas in order to select which idea(s) is/are the most appropriate or coherent, or request that this conflict be resolved; Meta-Dialogical Move clarification Move (MDclaM), which aims to clarify the meaning of the purpose of other moves, or seek clarification of this meaning; Meta-Dialogical Move to clarify the meaning of an idea (MDcmiM), which aims to clarify the meaning of an idea present in other moves, or seek clarification of that meaning; Meta-Dialogical context-setting Move (MDctsM), which aims to clarify whether the subject has any prior knowledge considered relevant to the discussion, or seek clarification on this and Meta-Dialogical consensus Move (MDconM), which aims to express the sharing of the same idea or ensure that this happens.

The aspect 4 of the tool: characterization of the structure of argumentation in persuasive and dispute persuasive dialogic moves, namely the ones that by definition involve the exchange of arguments of the tool. It is important to understand the structure of the arguments and their function, as this can contribute to a better understanding of the argumentation of individuals, that is, how they question, refute, defend and sustain their ideas/positions (Walton, 1990, 2006; Martins & Macagno, 2022).

This step allows characterizing aspects related to the argumentative structure of such moves in argumentative dialogues, being them: attack may be directed at the conclusion (or part of it), the reasons (all or one of them) or the relationship between reasons and conclusion and direct and indirect support relationships between arguments and between reasons and conclusions. In the direct support relationships, the reasons and arguments presented specify or detail the reasons or arguments previously presented. On the other hand, in indirect support relationship, the reasons and arguments presented are related to the topic of the dialogue. This relationship is weaker than the direct one, as it only provides other information related to the topic of the dialogue, instead of specifying or detailing the information presented in the previous movement.

The aspect 5 of the tool: relevance of dialogic move¹ captures the relationship between an individual move (aspect 3) and the overall dialogue—both considered as a collective goal pursued by the participants (aspect 2) and as a succession of interrelated moves on a specific topic (aspect 4). This step allows identifying whether and to what extent a move contributes (being relevant, partially relevant and irrelevant) to the common joint goal (which are related to each activity) and the dialogical proposals advanced by the previous move (Martins & Macagno, 2022).

Three criteria are used were used to analyse the relevance. Topic and pragmatic coherence analyse whether the move contributes to the development of a joint dialogical activity. Finally, inferential distance assesses whether the interlocutor expresses premises for drawing the conclusion or instead takes for granted premises that require several inferences to be retrieved (Martins & Macagno, 2022).

The aspect 6 of the tool: contribution to the construction of knowledge in dialogic and meta-dialogic moves enables assessing how a relevant move can increase the common ground between speakers and consequently for the construction of knowledge. Thus, relevant moves tend to contribute to developing a ground by confirming, completing, deconstructing hypothesis or resolving a doubt (Dialogic Moves) or by explaining or correcting a concept (meta-dialogic moves). An irrelevant move either repeats what is already granted or cannot be related to what the interlocutors hold, accept and negotiate at the given point of the dialogue. A partially relevant move can contribute to developing a ground shared among the interlocutors that emerges from the discussion if it expresses new or specific assumptions about the topic under discussion, either dialogically or meta-dialogical (Martins & Macagno, 2022).

The tool does not make it possible to identify the concepts or knowledge constructed by the students. For this, the analyst must identify, in the moves that generated contributions to the construction of knowledge, the concepts or knowledge constructed and expressed by the students, and assess whether they are coherent with those present in the curriculum.

2.3 Theoretical Framework of Modelling

In this study, Gilbert and Justi, (2016a)’s proposal was used because it brings theoretical and empirical contributions both for the learning of and about science as also for the development of skills. According to Gilbert and Justi (2016b), modelling is a creative, complex, cyclical, non-linear and dynamic process of creating, expressing, testing and evaluating models.²

The stage of creating the proto-model³ is characterized by the elaboration of the model’s proposal, according to the following steps. The subject should, in a creative manner and using critical thinking, deal with the following elements: definition of the goal(s) for which the model shall be established; obtaining information regarding the institution to be modelled either through external and/or internal sources and the identification of the bases with which it appears to be possible to establish analogies or mathematical resources that are appropriate for the problem considered.

The stage of expressing the proto-model is characterized by the communication of the proto-model to another subject through some method of representation, which can be verbal, gestural, concrete, visual, symbolic, virtual or based on combinations of these—a recurring point within an educational context (Gilbert et al., 2000).

The testing stage is characterized by the analysis of the adequacy of the model to the objectives that developed it. Such an analysis can be made through mental⁴ and/or empirical testing. If the model does not positively respond to the tests, then this shall be modified or rejected. In this latter case, the subject shall start the process over again.

Finally, the evaluation stage is characterized by the analysis of the scope and limitations of a model. For this reason, the subject should use the same model in another circumstance, different from the one for which it was designed. In this view, the argumentation, which is the focus of this paper, is present in the entire modelling process, as the appropriate justification of the statements is essential so that meanings may be assigned to the models and also to persuade other people about the use and validity thereof (Gilbert & Justi, 2016b).

This modelling view was selected for this study, because (i) it is coherent with the current goals for the teaching of science; (ii) it is wider in scope than other views of modelling as used in school contexts (such as, for example those of Clement (1989), and Wells, Hestenes and Swackhamer (1995), discussed in Gilbert and Justi, (2016a)); (iii) it has generated contributions as a structural reference (model of modelling (v2)⁵) to give support to learning of and about science (for example Gilbert & Justi, 2016b; Maia & Justi, 2009a; Santos et al., 2020) and to the development of skills and analogue reasoning (for example Andrade & Mozzer, 2017; Maia & Justi, 2009b; Mendonça & Justi, 2013) and (iv) it considers argumentation to be a process that can occur at all stages of modelling (for a review, see Mendonça & Justi, 2013; Gilbert & Justi, 2016d).

Up to the start of this study that led to the present work, there had already been empirical investigation into how students argued in modelling activities (Gilbert & Justi, 2016c) in scientific situations from Gilbert and Justi, (2016b)’ framework. Such research generated relevant contributions for our field (Mendonça & Justi, 2013). However, there is a need for research studies that use this framework in situations of different natures, such as those with everyday and socio-scientific natures. This is because a citizen is a person capable to have an opinion with regard to the problems faced by our society that can be of either a social and/or scientific ilk (NRC, 2012).

In this case, it does not suffice for science teaching to offer opportunities so that the individual may undergo personal development and know how to apply scientific knowledge. It is also necessary that they have the opportunity to develop and know how to apply scientific knowledge, as well as articulating them to scientific knowledge—which would contribute so that they could take a position and/or make decisions when faced with conditions that would require such attitudes. In the same way, it would not be enough for the researchers to understand just the argumentation of students in scientific situations. It is also necessary to widen the scope of investigations, to include other types of situations. This is because it could contribute to bring more elements about students’ argumentation and also may generate contributions for the structuring of modelling activities that involve situations of different natures.

3 Research Questions

Considering that it is important to understand students’ argumentative dialogue in different modelling situation, in this study, I address the following research questions:

• 1. What are the characteristics of high school students’ argumentative dialogues in everyday, scientific and socio-scientific modelling situations?

• 2. Are there differences in the characteristics of high school students’ argumentative dialogues when they participate in everyday, scientific and socio-scientific modelling situations? If there are differences, what are they?

• 3. What are the influences of the nature of each situation of modelling on the characteristics of high school students’ argumentative dialogues?

• 4. What are the influences of modelling on the characteristics of high school students’ argumentative dialogues?

4 Methods

4.1 Study Context and Data Collection

The data were collected in a high school class consisting of 40 students (between 17 and 18 years old) from a public state school in Brazil from video and audio recordings, field notes of the researcher and copies of all written activities produced by the students over the 18 lessons of 50 min each for the didactic unit. The students in this class were not familiar with modelling activities, so, participating in this type of activity was something completely new to them. For this, some aspects of the modelling, associated with doing science, were explained to the students during the activities.

This class was chosen because of the teacher’s availability and the consent given by the teacher, students and parents to conduct the study. Moreover, the selection of the students’ grade level was due to the fact that the thematic present in the didactic unit—plastic—was foreseen in scope of the school’s curriculum.

A didactic unit was built with activities related to modelling in three situations, which have distinct natures described below. These activities were built using the Model of Modelling (v2) proposed by Gilbert and Justi, (2016b):

• i) Everyday situation, the students have the opportunity to the construct and express a proto-model to explain the operation of a machine to sell soft drink cans (activity 1, stages of creating and expressing the proto-model of Model of Modelling (v2)). In activity 2, the students should to test the model in order to analyse its adequacy to the objectives that developed it in light of the new evidence (test stage of Model of Modelling (v2)). Finally, in activity 3, the students have to evaluate the model in another situation, namely the functioning of an automated teller machine (ATM) (evaluation stage of Model of Modelling (v2)).

• ii) Scientific situation, whose objective is the construction of models to explain the behaviour of certain plastic objects (common supermarket plastic bag, TV casing and tyre). Initially, the students collect information/data about the flexibility of common supermarket plastic bag and TV casing (activity 4, stage of creating the proto-model of Model of Modelling (v2)). Then, they develop and explain proto-models that account for the flexibility of these plastic objects (at the submicroscopic level) (activity 5, stage of expressing the proto-model of Model of Modelling (v2)). Thereon, they test whether their models can explain the observed phenomena in face of the data collected on the response of common supermarket plastic bag and TV casing when heated indirectly and directly by a flame (activity 6, test stage of Model of Modelling (v2). Finally, the students evaluate whether their models can explain the behaviours of tyre (its flexibility and its reaction to direct and indirect heating) (activity 7, evaluation stage of Model of Modelling (v2)).

• iii) Socio-scientific situation, the students have the opportunity to construct (activity 8) and express a proto-model (activity 9) to solve the problem of the accumulation of plastics in a fictitious community, but with characteristics analogous to several real communities in Brazil (stages of creating and expressing the proto-model of Model of Modelling (v2)). Then, they test the model using the knowledge that they have developed during the modelling activities in a scientific situation, by assessing whether it can explain social, economic, environmental and ethical aspects, among others (test stage of Model of Modelling (v2)). Finally, the students evaluate whether the model can solve the problem of the accumulation of different types of plastics (activity 11, evaluation stage of Model of Modelling (v2)).

All activities were made in groups at the school science laboratory. The students were divided into six freely formed groups, which were organized in the six benches available in the science laboratory.

The activities were carried out in groups, as this can encourage them to express their ideas to colleagues, discuss and modify their views and build meanings (Clement, 2008; Lemke, 1997). Furthermore, modelling activities generally involves complex problems to be analysed individually.

The teacher and the researcher who participated in the data collection were already Masters in Science Education and graduated in chemistry when the data collection took place. They presented a large and diverse experience as teachers, having already conducted teaching activities based on modelling and argumentation.

During the data collection, an audio recorder was placed on each group. In addition, cameras were placed next to groups 3 and 6 (G3 and G6) because, during activities, the seven students from each of these groups tried to express and evaluate their own ideas and the ones of their colleagues. The researcher combined direct observations with the information contained in her field notes, audio and videos recordings.

The researcher conducted an active observation (according to Spradley, (1980)), as she adjusted the video cameras and audio recorders, produced field notes and made interventions to help the teacher carry out the teaching process and meet possible group requests.

All classes were described by the researcher based on videos, audios and their field notes, focusing on the pre-selected groups. The descriptions show the different types of classroom discussions, the written records of the activities carried out by the groups, the drawings and photographs of the models produced by the groups and the field notes made by the researcher throughout the data collection. The data of each selected group was transcribed. In order to carry out the transcriptions, all video and audio recordings were watched and listened to again. In the following stage, the analytical tool was then used to analyse the data from G3 and G6.

4.2 Data Analysis

Based on the description of the whole process and the transcripts, the argumentative dialogues analysis tool was used to analyse the data from G3 and G6. The six aspects/steps of tool are summarized in Table 1.

Table 1 Aspects and descriptions of analysis

Initially, the data on modelling in the everyday context were analysed independently by two researchers. First, the potentially argumentative situations in which students engaged in a dialogue were identified and selected (aspect 1 of the tool). Secondly, the type(s) of dialogue the activity favoured the engagement of students was/were identified, from the objectives of the activity (aspect 2 of the tool). Third, considering the first two steps and limiting the analysis to the potentially argumentative situations detected through aspect 1, the dialogic and meta-dialogic moves expressed by the students were identified (aspect 3 of the tool). Fourth, the aspects related to the argumentative structure of persuasive and dispute persuasive dialogic moves expressed by the students in argumentative situations (aspect 4 of the tool) were characterized. Then, the step of relevance of the dialogic moves in which students had engaged within argumentative situations was detected (aspect 5 of the tool). Finally, the students’ moves contribute to the construction of knowledge were assessed (aspect 6 of the tool).

After these steps, the researchers met to triangulate the data for each stage of this analysis. The initial reliability achieved among the researchers was 81% (percentage agreement between the two judges) (Cohen et al., 2011). The researchers observed that the following categories of the tool needed to be better defined: (i) Dialogical persuasive Move (DperM) and Dialogical persuasive dispute Move (DperM) that constitute aspect 3 of the tool: identification of the nature of the move; (ii) relevant and partially relevant that make up aspect 5 of tool: relevance of dialogic move. For example, in the new definitions about DperM and DperM, dialogical persuasive refers to moves assessing the strengths and weaknesses of an idea, while the persuasive dispute moves are advanced for showing that a given idea is better or has greater explanatory power, than that of another party in a given context. While the dialogical persuasive intends to grasp the development and expression of arguments for co-constructing an interpretation, explanation, or perspective, the dispute one is simply aimed at destroying the interlocutors’ argument by providing underminers or specific attacks to the reasons on which it is based (Martins & Macagno, 2022).

After the refinement of these analysis categories, the reliability reached among the researchers was 90%.

The data analysis process for modelling in scientific and socio-scientific situations took place in the same way as that of modelling in the everyday situation, that is the data was analysed independently by two researchers and then triangulated.

The initial reliability reached among the researchers was 88% for modelling in scientific situation (percentage agreement between the two judges) (Cohen et al., 2011). The researchers sometimes failed to differentiate between DdisM and DperM by using the definitions of these moves during the analysis. For this, new definitions were created for the DdisM. In this move, the initial situation is needed to elaborate an explanation for a fact or phenomenon, the objective of dialogic move is to request or prepare explanation(s) and the objectives of the participants are to request or propose hypotheses related to the elaboration of an explanation, individually or collectively, from the sharing of previous ideas to explain a fact/phenomenon.

After the refinement of these analysis categories, the reliability reached among the researchers was 93% for modelling in the everyday and scientific situations.

Finally, the initial reliability reached among the researchers was 93% for modelling in socio-scientific situation (percentage agreement between the two judges) (Cohen et al., 2011). At this stage, researchers had difficulty, for example for distinguishing MDclaM from MDcmiM. Thus, new definitions were created for each of these: (i) MDclaM, which the initial situation is a misunderstanding or some problem of lack of clarity related to some previous move. The objective of this move is to clarify the meaning of the purpose of other moves or seek clarification of this meaning, and the objective of the participants is to clarify or seek clarification on the point in question; (ii) MDcimM, which the initial situation is a misunderstanding or some problem of lack of clarity regarding the meaning of an idea. The objective of this move is to clarify the meaning of an idea present in other moves or seek clarification of that meaning, and the objective of the participants is to clarify or seek clarification on an idea under discussion.

At this stage, after the refinement of these analysis categories, all cases of disagreement between researchers were reviewed until a consensus was reached for each one of the modelling situations.

From the data analysis, graphs were constructed for G3 and G6 that show the absolute frequencies of types of dialogues; types of dialogic and meta-dialogic moves; structural aspects of argumentation in persuasive and dispute persuasive dialogic moves; relevant dialogic and meta-dialogic moves; interventions of the researcher and teacher and contribution of moves to knowledge construction. Therefore, a simple count was made of the number of times a value was observed for each category that constitutes the above elements.

Through this analytical process, it was possible to characterize the high school students’ argumentative dialogues in modelling situations that have different natures. From this, they can enable the understanding of influences of modelling and the nature of each of the situations analysed on the characteristics of the students’ argumentative dialogues.

In the ‘Results and Discussions’ section, I presented the graphs together with the discussions about them, which are linked to the research questions.

4.2.1 Example of Application of the Tool

To help readers understand the application of the tool during the analysis, an example is presented.⁶ This example is expressed by Aline⁷ (member of G3) during the modelling in an everyday situation when the teacher asked students to perform activity 1: building knowledge in a different way (aspect 1 of the tool: selection of the discussions to be analysed). The objective of this activity was to develop a model that would explain how a soft drink selling machine works. Therefore, it contributed to the creation and expression of the model(s) (stages 1 and 2 of the Model of Modelling Diagram, v2), as well as favouring the occurrence of discovery dialogue (Dds), whose main objective was to elaborate an explanation related to the topic under discussion (aspect 2: identification of the dialogical nature(s) of each activity). The following excerpt drawn from this situation was analysed through the other aspects, considering the information provided by the use of aspect 1 and aspect 2:

Aline: (1)⁸ Do you think it would be cool if we had elves in the machine? (2) Imagine a bunch of little creatures in the machine and they get your soft drinks and juice!

Aspect 3: Identification of the nature of the move. (1) Dialogical persuasive Move (DperM) and (2) Dialogical discovery Move (DdisM). The student asked for an evaluation of the hypothesis, as she had doubts about its validity, and then expressed the hypothesis that explains the softdrink machine works.

Aspect 4: Characterization of the structure of argumentation in DperM: the question requested an external assessment (attack attempt).

Aspect 5: Relevance of the Dialogical Move: DdisM is relevant, as there was topic coherence: the way a soft drink machine works is a topic of the dialogue Operation of a Machine that sells Soft Drinks; and pragmatic coherence with the main objective of the dialogue: to propose a model to explain the operation of the Machine that sells Soft Drinks. DperM⁹ is relevant because there was topic coherence since it was the same topic as DdisM; and pragmatic coherence with the main objective of the dialogue because the movement expressed a specific objective that contributed to the objective of DdisM.

Aspect 6: Contribution to the construction of knowledge in dialogical moves: DperM expressed a new hypothesis to explain how the soft drinks machine works. DperM expressed a request to evaluate the hypothesis, which contributed to the construction of knowledge.

5 Discussion of the Findings

I have divided this section into five topics, which are associated with the aspects 2, 3, 4, 5 and 6 of tool. This may help the reader to understand more deeply the results and their discussion.

5.1 Type of Dialogue in Modelling Situations

In Figs. 1 and 2, I show the frequency of type of dialogue in everyday (bars in solid black), scientific (bars in solid dark grey) and socio-scientific (bars in solid light grey) modelling situations for groups 3 and 6. Such patterns were used to build all the figures related to students.

Fig. 1

Frequency of types of dialogues in everyday, scientific and socio-scientific modelling situations experienced by group 3

Fig. 2

Frequency of types of dialogues in everyday, scientific and socio-scientific modelling situations experienced by group 6

In Figs. 1 and 2, it is noted the occurrence and the same frequencies of inquiry Dialogue (Din) in all situations and groups. I assess that this is connected to the stages of testing and evaluating the model of Model of Modelling (v2), as they favour the investigation of the validity of the model from evidence in the situation where it was created, and in a new circumstance, respectively (Gilbert & Justi, 2016b).

Another aspect that can be observed from the Figs. 1 and 2 is the occurrence or not of Deliberation (Ddl), Discovery (Dds) and Information Sharing (Dis) Dialogues in the groups. As I see it, this difference could be related to each teaching situation.

The modelling in the socio-scientific situation favoured Ddl (Deliberation Dialogue) and not Dds (Discovery Dialogue), as the groups had to prepare a model to try and solve the problem of accumulation of plastics, and not an explanation for this problem, as requested by the activity 9: Prepare a model to try to solve the problem of accumulation of plastics, based on recycling. Your model should contain all the aspects you consider relevant for solving the problem (stage of creating and expressing the proto-model of Model of Modelling (v2)).

The modelling in socio-scientific situation also favoured Dis (Information Sharing Dialogue). It requested that the students provide information about the advantages and disadvantages of the recycling process (questions of activity 10: What are the advantages of the recycling process? Why are they advantages? What are the drawbacks of the recycling process? Why are they drawbacks?) to test whether the models based on recycling, as proposed, can solve the problem of accumulation of plastics (stage of testing the model of Model of Modelling (v2)).

On the other hand, the modelling in the scientific situation favoured the occurrence of Dds (Discovery Dialogue) rather than Ddl (Deliberation Dialogue). This is because the groups had to prepare models that could explain the different behaviour patterns of plastics when they are subjected to attempts to fold them (for example the question of activity 5 in this situation is: Prepare models that explain, at a submicroscopic level, what happened to the common supermarket plastic bag and the television casing after the attempt to fold each object (stage of expressing the proto-model of Model of Modelling (v2)). For this, first they had to get engaged in Dis (Information Sharing Dialogue) to predict what would happen to certain plastics on attempting to fold and heat them.

The modelling in the everyday situation favoured the occurrence of Dds (Discovery Dialogue) and did not favour Dis (Information Sharing Dialogue), as the groups had to create and express a model to explain the operation of a machine to sell cans of soft drinks (command for the activity 1: Prepare a model to explain how vending machine selling canned soft drinks works. Explain all the characteristics of your model, in writing) (stages of creating and expressing the model of Model of Modelling (v2)). As this was a well-known situation, the students of both groups did not have to predict what would happen to function of a vending machine for canned soft drinks in certain situations, in the same way that they did not need to obtain any additional information before creating, testing or assessing the models—an aspect that would contribute towards the engagement in a Dis (Information Sharing Dialogue), as was the case in the other teaching situations.

Finally, it can be seen in Figs. 1 and 2 the non-occurrence of Persuasion Dialogue (Dper) in all modelling situations and groups. This could have occurred because this activity is investigative in nature (Gilbert & Justi, 2016b). Therefore, its main goal is not that of solving a conflict of ideas. This does not mean that modelling has not contributed towards the groups getting involved in Dialogic Moves that would benefit the solution of conflict of ideas.

5.2 Types of Dialogic and Meta-dialogical Moves in Modelling Situations

In Fig. 3, I present the frequency of types of dialogic and meta-dialogical moves in everyday, scientific and socio-scientific modelling situations for.

Fig. 3

Frequency of types of dialogic and meta-dialogical moves in everyday, scientific and socio-scientific modelling situations experienced by group 3

In Fig. 3, it can be seen that, during the modelling in the everyday situation, the students of G3 frequently pursued:

• i. Propose explanation(s) (expression of the DdisM (Dialogical discovery Move) 21 times). For example, Aline proposed the following hypothesis to explain how a vending machine selling canned soft drinks works (activity 1, stages of creating and expressing the proto-model of Model of Modelling (v2)) in the first argumentative situation that the group got involved in: Do you think it would be cool if we had elves in the machine? Imagine a bunch of little creatures in the machine and they get your soft drinks and juice!

• ii. Convince a colleague that an idea is valid and solve a possible conflict of ideas, seeking to analyse each idea (expression of the DperM (Dialogical persuasive Move) 29 times). For example the hypothesis raised by Aline to explain the operation of a vending machine for selling canned soft drinks was assessed by Louise during activity 1 (stages of creating and expressing the proto-model of Model of Modelling (v2)): I don’t think this way of explaining is right for a chemistry class. I think we have to explain it in a realistic way;

• iii. Investigate the validity of hypotheses from evidence (expression of the DinqM (Dialogical inquiry Move) 18 times). For example, during the test stage of the model of Model of Modelling (v2), the students have investigated the validity of the model as proposed to explain the operation of the vending machine selling canned soft drinks, in terms of being able to explain two key observations: (1) the soft drink is not served after keeping the machine idle and then inserting the coin; (2) the machine does not serve the drink immediately on switching if off for 2 h and then switching it on again and inserting the coin. However, after some minutes, the machine starts to work and, after a certain elapsed time, by inserting a coin, the soft drink is served (activity 2). This can be seen from an excerpt of one of the argumentative situations in which the students have got involved:

Isabelle: Our model can explain the observations (DinqM). It is as if the system of the machine got locked when switched off, and for this reason it does not work. (DperM)

Aline: And also because the machine is a computer, and shall not be processing to take the client’s money. The most that can happen is for the coin get trapped in the machine and the client loses their money, as the machine does not have a ‘brain’ to give the product to the client, and neither to return the client’s money. If the client wants the money back, they would have to wait for the machine to be connected to the outlet, and then for the system to start running again, to see if the machine would recognize the coin and maybe return the money to the client so that they could start a new operation. (DperM)

In this excerpt of the argumentative situation, Isabelle analysed of the validity of the model (DinqM) and provided reason to support the analysis. Next, Aline provided others reason to try support the idea/position that the group’s model was able to explain the observations.

• iv. Request information (Dialogical Information Sharing Move (DishM)) and investigation the validity of hypotheses from evidence (Dialogical inquiry Move (DinqM)). For example, the question of activity 3: Is the model of your group able to explain the operation of a cashpoint? requests that the students look into the validity of the model to explain the operation of a vending machine selling canned soft drinks, in another circumstance: the operation of a cashpoint (stage of evaluating the model of Model of Modelling (v2)). The evaluation of the group 3 model is presented below:

  Our model is able to partially explain the functioning of an ATM, as it has a computerized financial system containing some sort of calculator, which is similar to an ATM; a system for recognizing bills and coins, which is also similar to the recognition of cards by ATMs. On the other hand, our model is not capable of explaining the functioning of the card and the balance and statement operations, since it does not work via card, nor does it have these operations.

In relation to the students of G6, they pursue during the modelling in an everyday situation (Fig. 4):

• i. Look into the validity of the hypotheses from evidence and request their investigation (expression of the DinqM (Dialogical inquiry Move) 9 times); and

• ii. Convince the colleague of the validity of an idea and request that this be done, as well as solve a conflict of ideas, seeking to conduct an analysis of each idea and requesting that this conflict be solved (expression of the DperM (Dialogical persuasive Move) 16 times).

Fig. 4

Frequency of types of dialogic and meta-dialogical moves in everyday, scientific and socio-scientific modelling situations experienced by group 6

For example, during the testing stage of the model of Model of Modelling (v2), the students were asked to verify the validity of the model in order to explain the operation of a vending machine selling canned soft drinks, in terms of it being able to explain observation 2 (the machine does not serve the drink immediately, on switching it off for two hours). They were also asked to defend the model they proposed. This can be seen below, based on an excerpt of one of the argumentative situations in which the students got involved:

Maggie: Can our model explain it? Why?¹⁰{Activity Question (student)}. (DinqM)

Camille: Our model can explain it (DinqM), as, when the machine is switched on again, it shall need some time to reconnect all the functions. Therefore, it shall only take the cash when it recognizes that the whole system is working perfectly. (DperM)

{Students agree with Maggie}. (MDconM (Meta-Dialogical consensus Move))

In this piece of dialogue, it can be noted that Camille investigated the validity of the model and tried to support to it through reasons.

On the order hand, in the scientific situation, the members of G3 frequently pursued to (Fig. 3):

• i. Obtain and provide information in response to a request/question (expression of the DishM (Dialogical Information Sharing Move) 45 times);

• ii. Convince a colleague about the validity of an idea and request that this is done, as well as to solve a conflict of ideas seeking to look into every idea and request that this conflict be solved (expression of the DperM (Dialogical persuasive Move) 30 times) and

• iii. Express the sharing of one same idea and check that this has indeed happened (expression of the MDconM (Meta-Dialogical consensus Move) 17 times).

Such moves can be made evident based on a small excerpt of the argumentative situation in which the students have got involved during activity 4 (stage of creating the proto-model of Model of Modelling (v2)):

¹¹R: What do you think will happen with the common supermarket plastic bag and the TV casing, when they are subjected to an attempt to fold them? {Activity Question (R)}. (DishM).

Beatrix: I think that the common supermarket plastic bag will fold, and the TV casing will break, if we exert a very strong force. (DishM)

R: But will the TV casing really break? Why? (DperM)

Beatrix: As the casing is more compact, then it shall get harder, meaning that if we apply a very strong pressure, it will break. In the case of the common supermarket plastic bag, as it is thinner and more malleable, it shall fold. (DperM)

{Students agree with Beatrix}.(MDconM)

In this section of dialogue, it is noted that Beatrix provided the information when the group was asked to predict what would happen with the common supermarket and the TV casing when they are subjected to attempts to fold them. She also gave reasons in support of her idea that the common supermarket plastic bag would fold, and the TV casing would break if there was the application of a very strong force when their idea had been questioned by R. Then, Beatrix’s colleagues said that they shared the same idea as Beatrix.

In activity 5 (stage of expressing the model of Model of Modelling (v2)), the G3 expressed the following model to explain the flexibility of the common supermarket plastic bag and the TV casing at the end of the argumentative situation:

The common supermarket plastic bag fold, as the distance between its molecules is greater, which helps them to move more easily. On the other hand, TV casing are difficult to fold as the distance between their molecules is smaller, which makes them more difficult to move.

The most common moves carried out by the students of G6 were similar the students of G3 in the scientific situation (Fig. 4):

• i. Obtain and provide information in response to a request/question (expression of the DishM (Dialogical Information Sharing Move) 26 times);

• ii. Convince a colleague about the validity of an idea and request that this be done, as also solve a conflict of ideas seeking an analysis of each idea, and request that this conflict be solved (expression of the DperM (Dialogical persuasive Move) 27 times);

These moves can be seen as follows, based on one of the argumentative situations in which the students have got involved, with the aim of predicting what would happen with the common supermarket bag and the TV casing, when subjected to attempts to fold them (activity 4, stage of creating the proto-model of Model of Modelling (v2)):

Maggie: What do you think will happen to the TV casing and common supermarket plastic bag? (DishM)

Rachel: The TV casing will break when it is folded. (DishM)

Julie: The common supermarket plastic bag will fold (DishM). I do not agree that the TV casing will break when it is folded, as it is quite rigid. (DperM)

Danielle: I do not agree, as if we exercise force upon it, it will break. (DperM)

{Students agree with Rachel and Danielle}. (MDconM (Meta-Dialogical consensus Move))

In this section of the dialogue, it is noted that Rachel and Julie provided information in response to a request/question from Maggie. It is also noted that there was a conflict of ideas between students, which was solved based on an analysis of the ideas under discussion.

• iii. Propose and request explanation(s) (expression of the DdisM (Dialogical Information Sharing Move) 19 times). For example, during the stage of expressing of the model of Model of Modelling (v2) (activity 5), the students were asked to come up with a hypothesis or model to explain the behaviour of the common supermarket bag and the TV casing when subjected to attempts to fold them. Maggie made the following hypothesis: The size of the carbon chain is different in these plastics. The molecules of the TV casing are larger than those of the supermarket bag. This explains why the casing is less flexible than the common supermarket plastic bag.

In relation to the socio-scientific situation, the students of G3 frequently pursued (Fig. 3):

• i. Proposal of solution(s) and selection of the most appropriate solution, as well as request that these actions are indeed carried out (expression of the DdelM (Dialogical deliberative Move)16 times);

• ii. To convince the colleague of the validity of an idea and request that this be done, as well as solve a conflict of ideas by seeking to analyse each idea and request that this conflict be solved (expression of the DperM (Dialogical persuasive Move) 25 times);

For example, during the activity 8 (stage of creating the proto-model of Model of Modelling (v2)), it is possible to observe such moves as shown by the students from an excerpt of one of the argumentative situations in which they got involved:

Aline: Guys, look here. We need to choose a plastic object to be recycled, but which one shall we choose? {Activity question (student)}. (DdelM)

Some students: A tire. (DdelM)

Louise: I disagree, as a tire does not melt but it carbonizes when heated directly on a flame. (DperM)

{Students agree with Louise}. (MDconM (Meta-Dialogical consensus Move))

In this excerpt, it is noted that some students have made the decision to choose the tyre to be recycled, when the group was asked to decide which plastic object (a common supermarket bag, a TV casing or a tyre) could be recycled. Soon afterwards, Louise came up with a reason to attack the choice made by her colleagues. Next, the colleagues agreed with Louise.

• iii. Obtain and provide information in response to a request/question (expression of the DishM (Dialogical Information Sharing Move) 13 times). For example, during the stage of evaluating the model based on recycling as a solution to the problem of accumulation of plastics (activity 11), the students were asked to provide information about the advantages of the recycling process. Students Gisele and Beatrix said:

  I think that it is the issue of the environment, the reduction of thermoplastics in the environment.

The students of G6 also frequently pursued to:

• i. Propose solutions and select the most appropriate solution, as also request that these things be done (expression of the DdelM (Dialogical deliberative Move) 22 times) and

• ii. Convince a colleague about the validity of an idea and request that this be done, as also solve a conflict of ideas, seeking to analyse each idea and request that this conflict be solved (expression of the DperM (Dialogical persuasive Move) 38 times);

For example, during the stage of expressing of the model based on recycling to solve the problem of accumulation of plastics (activity 9), the students proposed possible objects that could be produced through the recycling of common supermarket bags (DdelM) and also assessed whether this would reduce the amount of plastic in society (DperM). This can be seen in the extract of the argumentative situation below:

Maggie: We could transform the common supermarket plastic bags, turning them into plastic bottles, or containers for putting food in (DdelM). However, by doing this, we shall not be reducing the quantity of plastic in society, as we would be producing even more plastic. (DperM)

Julie: I don’t agree, as the plastic would thus be removed from the environment. (DperM)

Rachel: By using the same material to produce new items, there would be a reduction of the presence of plastic in society. (DperM)

{Students agree with Rachel and Julie}. (MDconM (Meta-Dialogical consensus Move))

• iii. Obtain and provide information in response to a request/question (expression of the DishM (Dialogical Information Sharing Move) 20 times). For example, during the stage of evaluating the model based on recycling as a solution for the problem of accumulation of plastics (activity 11), the students were asked to provide information about the disadvantages of the recycling process. To them, the disadvantages would be: Expenses incurred through use of water that shall not be reused or treated, and very high costs to carry out a fully sustainable type of recycling.

From Figs. 3 and 4, it is noted that both groups expressed the Dialogical information Sharing (DishM), inquiry (DinqM), Persuasive (DperM) Moves, and the Meta-Dialogical consensus Move (MDconM) and Meta-Dialogical Move to clarify the meaning of an idea (MDcmiM) in all situations. This means that, during the all modelling situations the students in both groups shared information and analysed the validity of the hypothesis based on evidence, and exploited every idea or solution in depth, appraising their strengths and weaknesses based on arguments, stating if they agreed with the ideas of other colleagues, seeking to make it clear that both shared the same idea, and thus clarified the meaning of an idea under discussion.

My assessment was that the occurrence of DishM (Dialogical Information Sharing Move) could be linked to the fact that, in order to create, test and assess models, the groups would need to obtain and provide information. For example, during the stage of creating model of Model of Modelling (v2), the students in both groups obtained and provided information about the institution to be modelled through external sources (empirical experiments, and information provided by the teacher and the researcher) and also internal sources such as previous ideas of those participating in the process. However, the sharing of information was not limited to this stage, as modelling is a cyclical and dynamic process (Gilbert & Justi, 2016b). Thus, in the stages of expressing, testing and evaluating the model of Model of Modelling (v2), the students in both groups shared information in a move to select the modes of representation for stating the model, checking the appropriateness of the model for the purposes for which it was devised, and assessment of the scope and limitations of the model thus created.

I already expected that groups expressed DinqM (Dialogical inquiry Move), because, as mentioned before, modelling is an investigative activity. Therefore, it would be totally unexpected if they did not seek to investigate the validity of any hypothesis raised, based on evidence, or request that this be done in the stages of testing and evaluating the model of Model of Modelling (v2).

On the other hand, the occurrence of DperM (Dialogical persuasive Move) is linked to the fact that argumentation is one of the cognitive processes inherent to the reference for modelling as adopted in this study (Gilbert & Justi, 2016d). In this regard, for instance, in the stage of creating the model of Model of Modelling (v2), the DperM was expressed by G3 and G6, for example during the analysis of information to be selected to create and express the models, to explain the operation of the vending machine for canned soft drinks, and the behaviour of different types of plastics when subjected to folding attempts and to direct and indirect heating by a flame, as also to solve the problem of accumulation of plastics.

In the stage of testing the model of Model of Modelling (v2), the DperM was expressed by G3 and G6 during the analysis of the models when faced with observations either made or provided, related to the operation of the vending machine for canned soft drinks (modelling in an everyday situation), and the behaviour patterns of different types of plastics when these were subjected to attempts to fold them, and also to direct and indirect heating by a flame (modelling in the scientific situation), as in the recycling process (modelling in the socio-scientific situation). In the latter, the students analysed of the models when faced of the social, economic, environmental and ethical aspects, among others.

Finally, in the stage of evaluating the model of Model of Modelling (v2), the DperM was expressed by G3 and G6 during assessment of the models in a different circumstance: the operation of a cashpoint (modelling in an everyday situation); reaction of the tyre when subjected to attempts to fold it, and also direct and indirect heating by flame (modelling in the scientific situation) and a solution for the problem of accumulation of TV casing units and tyres in society (modelling in the socio-scientific situation).

Thus, considering these aspects and data on the frequencies of DperM as shown in Figs. 3 and 4, it can be said that both groups had sought to convince their colleagues about the validity of an idea and also request that this be done. They also had sought to solve a conflict of ideas, seeking to consider each idea, assessing its strong and weak points, based on arguments, reasoning and questions, and also requested that this conflict be solved in all modelling situations.

I evaluate that the occurrence of MDconM (Meta-Dialogical consensus Move) and MDcmiM (Meta-Dialogical Move to clarify the meaning of an idea) is due to the fact that modelling has well-defined goals and stages (Gilbert & Justi, 2016b), and this came from the fact that the activities that the groups carried out also have such aspects. This contributed to the groups stating if they agreed with the ideas raised by colleagues and shared the same ideas; clarifying the meanings of ideas considered right or doubtful/vague and requesting that these things might be done. Without performing these actions, groups may have difficulty to achieve the activity objectives and proceeding with the modelling steps.

In Figs. 3 and 4, it is also noted that the groups did not express the Dialogical eristic Move (DeriM) in all situations. This means that the students in G3 and G6 did not try to attack or defend colleagues, but rather the arguments that they produced throughout all the modelling situations. This result may indicate that modelling can contribute to promoting discussions in which students construct and assess ideas based on arguments.

Furthermore, a larger frequency, and not identical, of certain dialogic moves in both groups can be seen in these figures, that could be related to the each situation of the modelling. For example, DishM (Dialogical Information Sharing Move) was expressed by students in both groups, more often in the scientific situation than in the others. This could have happened as a result of the complexity of this type of situation, which is greater than the other, because of the nature of the knowledge thus constructed. Consequently, students had to share information frequently in order to that knowledge could be built.

In the everyday situation, the groups could use social knowledge, which is part of their culture and their daily life to explain the operation of a vending machine for canned soft drinks. In the case of the socio-scientific situation, the groups have not had to construct scientific knowledge, but just to apply this knowledge, together with knowledge of social nature, to solve the problem of accumulation of plastics in society. On the other hand, in the scientific situation, the members of G3 and G6 had to construct and use exclusively scientific knowledge to be able to solve the problem under discussion. Furthermore, the complexity of the scientific situation was made even bigger because of the fact that the science studied was chemistry, which is a subject composed by abstract concepts that may not be familiar to the students.

I also posit that the greater frequency of DishM in the scientific situation compared to other situations could also be connected to the fact that, in this type of situation, the researcher and the teacher were quite clear about what types of knowledge the students should construct and apply, to explain the ways in which different types of plastics behave. Conversely, in the everyday and socio-scientific situations, they only had an understanding of the possible knowledge that the students could construct and use to explain the operation of a machine to sell canned soft drinks and solve the problem of accumulation of plastics, especially in the everyday situation, as in this situation the students could only construct and use social knowledge.

MDconM (Meta-Dialogical consensus Move) was also expressed more frequently by the students of both the groups, in the scientific situation than in the other situations. This was due to the complexity of this type of situation being higher than for the others. As I see it, the complexity of the nature of the situation contributed to the members of these groups expressing their agreement with the idea of their colleagues in order to make it clear that they shared the same idea, so they could progress in this teaching situation.

On the other hand, groups showed more often the DdelM (Dialogical deliberative Move) in modelling during the socio-scientific situation. This occurred because, in this type of situation, the groups had to get engaged in Ddl (Deliberation Dialogue) to solve the problem being discussed. For the goal of the dialogue to be reached, it was necessary that they proposed solutions/actions and then selected the most appropriate one, as also requested the proposal of solutions/actions or chose one out of many possible to solve the problem of accumulation of plastics. This can also explain why the groups did not show expression of DdisM (Dialogical discovery Move). In this move, the participants request or prepare explanation(s).

Different frequencies and types of dialogic and meta-dialogical moves are observed between the groups investigated when comparing Figs. 3 and 4. For example, the students of G3 have shown the Meta-Dialogical context-setting Move (MDctsM) in all situations, but the students from G6 did not show the same pattern. On the other hand, the students from this group have increased the frequency of DperM throughout the modelling situations and have expressed this move 16, 27 and 38 times in modelling in everyday, scientific and socio-scientific situations, respectively. The same standard was not observed in G3, as the students in this group expressed the DperM 29, 30 and 25 times during everyday, scientific and socio-scientific modelling situations, respectively. These results may indicate that each group conducted the argumentative dialogues, and the issues addressed in these, in different ways.

5.3 Structure of Argumentation in Persuasive and Dispute Persuasive Dialogic Moves in Modelling Situations

I now present Figs. 5 and 6¹² seeking a deeper discussion of the dialogic moves that have argumentative structure the moves in argumentative dialogues: Dialogical persuasive Move (DperM) and Dialogical persuasive dispute Move (DpdiM). In these figures, I present the frequency of structural aspects of the groups’ argumentative moves in each modelling.

Fig. 5

Frequency of structural aspects of argumentation in persuasive and dispute persuasive Dialogic Moves in everyday, scientific and socio-scientific modelling situations experienced by group 3

Fig. 6

Frequency of structural aspects of argumentation in persuasive and dispute persuasive Dialogic Moves in everyday, scientific and socio-scientific modelling situations experienced by group 6

In Figs. 5 and 6, it is noted that, in all the modelling situations, both groups sought: to attack the conclusion (or part thereof), the reasons (one or all) and the relationship between reasons and conclusion, through questions, arguments or reasons and frequently sustain their ideas, hypotheses and solutions based on a relationship of direct support.

In one of the argumentative situations, which the G3 students got involved during modelling in the everyday situation (activity 1, stage of creating the model of Model of Modelling (v2)), the students presented hypotheses to explain the payment system in the vending machine for canned soft drinks. These hypotheses have been queried/attacked through reasons, as also defended and sustained through a relationship of direct support:

Aline: Ah! I have had a great idea. We could explain the card system, and thus we would eliminate the idea of paying with cash. Like a bus payment card.

Beatrix: The person enters the shopping centre, and receives a card.

Nina: The card is more interesting, as the client could then top it up with as much credit as the client wishes.

Aline: This is more interesting, because, as you top it up with money, there is no need to carry bills and coins with you, you just put the card against the machine.

Louise: I think it would be easier if we proposed a mechanism so that the machine could receive cash; otherwise, we would have to explain the mechanism of the card and also how to top up the card. This is very complex.

Beatrix: But the card could be topped up at a snack bar. This would mean that we would not have to prepare yet another mechanism for our machine.

In this extract of argumentative situation, students Nina and Aline expressed their reasons which would give direct support to the hypothesis of explaining the payment system for the vending machine selling canned soft drinks, using a card similar to that used on the bus. On the other hand, Louise presented an argument to attack the use of this hypothesis, as also to directly back up the use of the cash hypothesis to explain the payment system. Next, Beatrix came up with an argument to lend direct support to the use of a card similar to that used on bus transport, to explain the payment system for the vending machine selling canned soft drinks, and thus to attack Laura’s proposed argument.

In stage of testing the model of Model of Modelling (v2), I observed aspects similar to those presented above. In the extract of argumentative situation below, which the G6 students got involved during modelling in the socio-scientific situation (activity 10), the students investigated the validity of the model(s). Initially, Danielle requested the analysis of the validity of the model and Raquel did this by considering the aspects that constitute it. Maggie disagreed with Raquel’s position and provided reason to lend direct support the idea/position that the group’s model was able to explain the possibility of job creation for the community in the plastic bag recycling process. In addition, this reason attacked the position/idea of Rachel. Next, Julie disagreed with Maggie’s argument. She expressed an argument to lend direct support the idea/position that the group’s model was unable to explain the possibility of job creation for the community from the plastic bag recycling process. This argument convinced the colleagues.

Danielle: Does your model consider the possibility of generating jobs for the community in the recycling process of this plastic object? {Reading aloud the question for the activity}

Rachel: No.

Maggie: YES,¹³our model considers employment generation, as there are people to separate the garbage, wash and dry the bags .

Julie: In our model the people will do these jobs voluntarily. So, we don’t consider generating jobs for the community in the process of recycling the bag.

{Students agree with Julia}.

Regarding the stage of evaluating the model of Model of Modelling (v2), G6 students sought to attack conclusion and reasons through arguments and reasons and to lend direct support their ideas during modelling in the scientific situation (activity 7). These aspects can be seen in the excerpt below:

Camille: Our TV casing model is able to explain the behaviour of the tire, as it burned when subjected to direct heating in the flame.

Julie: I disagree. The TV casing model consists of intramolecular bonds. This makes it able to explain the behaviour of objects that are inflexible. The model to explain the behaviour of flexible objects must have intermolecular bonds. The tire is flexible. Thus, the model to explain the tire flexibility must have only intermolecular bonds. Therefore, our TV casing model is not able to explain the tire flexibility.

Rachel: I don’t agree with you, Julie. The tire model must have intramolecular bonds because the tire burned when subjected to direct heating in the flame. However, the amount of intramolecular bonding in the tire structure must be less than that of a TV casing. These aspects explain why the tire burned when subjected to direct flame heating and its fallibility.

{Students agree with Rachel}.

In this excerpt, Camille evaluated the models tested in activity 6. Julie disagreed with Camilla’s conclusion and provided argument to lend direct support her position. Furthermore, this argument attacked the Camille’s position. Next, Rachel expressed a new argument to lend direct support the idea/position that the tyre model has intramolecular bonds. This argument attacked the Julie’ argument and convinced the colleagues.

These results and those presented in the Figs. 5 and 6 show that the students attempted to express moves, to convince the other party about the validity of an idea and also to solve conflict of ideas in a consistent way during the stages of Model of Modelling (v2) in all investigated situations.

Both groups sought to defend and sustain their ideas through relationships of direct support more often than to attack them in all modelling situations (Figs. 5 and 6), which may have contributed to the construction of ideas or knowledge. This result may indicate that modelling situations could favour the construction of ideas or knowledge. Activities, in which there are well-defined positions such as a mock trial, the frequencies of attacks and support relationships are high, as was observed in the work by Martins and Justi (2019). However, the high occurrence of the attacks does not contribute much for the construction of an idea in a co-operative manner, with regard to a certain subject, as shown in the academic works by Felton et al., (2015) and Asterhan and Babichenko, (2015). In such studies, the authors noticed that learning and the skill in argumentation as shown by the students were lower when the teacher or researcher asked them to take a position before they got involved in a dialogue and kept their positions during the dialogue, then when they asked the students to solve a problem by co-operation.

I did not identify any pattern on analysing the nature of each modelling situation and the characteristics of the argumentative structure the moves (DperM (Dialogical persuasive Move) and DpdiM (Dialogical persuasive dispute Move)) in argumentative dialogues, comparing Figs. 5 and 6. On the other hand, I observed that, in the case of G6, there was a gradual rise in the frequency of attacks, and also of relationships of support, and DperM during the all modelling situations (Figs. 4 and 6). These results could suggest that the students of G6 perceived the need to explore and requested that each idea or solution, as proposed during the modelling situations should be exploited.

5.4 Relevance of Dialogic and Meta-dialogic Moves in Modelling Situations

In order to better understand the characteristics of groups’ argumentative dialogues, the influences that could be established between their engagements in argumentative dialogues and the modelling in the different situations investigated, I now present Figs. 7 and 8. These figures show the frequency of the relevance of dialogic and meta-dialogic moves expressed by the groups in everyday, scientific and socio-scientific modelling situations.

Fig. 7

Frequency of the relevance of dialogic and meta-dialogic moves in everyday, scientific and socio-scientific modelling situations experienced by group 3

Fig. 8

Frequency of the relevance of dialogic and meta-dialogic moves in everyday, scientific and socio-scientific modelling situations experienced by group 6

Figures 7 and 8 show that both groups frequently expressed relevant dialogical and meta-dialogic moves in all situations. This could have happened as a result of the stages of Model of Modelling (v2) having such well-defined goals (Gilbert & Justi, 2016b), which helps the students seek to achieve the main and specific goals (that are related to those of each activity) of the dialogues, which they are involved in.

For example, in stage of creating the model of Model of Modelling (v2), Aline (student of G3) expressed: Do you think it would be cool if we had elves in the machine? Imagine a bunch of little creatures in the machine and they get your soft drinks and juice! This sentence was classified as relevant, as there was topic coherence: the way a soft drink machine works is a topic of the dialogue Operation of a Machine that sells Soft Drinks and pragmatic coherence with the main objective of the dialogue: to propose a model to explain the operation of the Machine that sells Soft Drinks (activity 1, modelling in the everyday situation).

In stage of testing the model of Model of Modelling (v2), Julie (student of G6) expressed: In our model the people will do these jobs voluntarily. So, we don’t consider generating jobs for the community in the process of recycling the bag. This sentence was classified as relevant, as there was topic coherence of the dialogue: extension of the model (to try to solve the problem of plastic accumulation from recycling) in relation to the possibility of generating jobs for the community in the plastic bag recycling process; and pragmatic coherence with the main objective of the dialogue: to investigate/test the validity of the model from evidence (activity 10, modelling in the socio-scientific situation).

Regarding the stage of evaluating the model of Model of Modelling (v2), Rachel (student of G6) expressed: I don’t agree with you, Julie. The tire model must have intramolecular bonds because the tire burned when subjected to direct heating in the flame. However, the amount of intramolecular bonding in the tire structure must be less than that of a TV casing. These aspects explain why the tire burned when subjected to direct flame heating and its fallibility. This sentence was classified as relevant, as there was topic coherence of the dialogue: extension of TV casing and supermarket bag models in another circumstance; and coherence with the main objective of the dialogue: to investigate/test the validity of models in another circumstance from evidence (activity 7, modelling in the scientific situation).

It can be seen Figs. 7 and 8 that the students of G3 expressed only one irrelevant move in modelling scientific situation, while the students of G6 did not express any in the same situation. On the other hand, in modelling in everyday and socio-scientific situations, the members of both the groups more frequently showed partially relevant and/or irrelevant moves when compared with the modelling in the scientific situation. Initially, I consider that the nature of situation could have had an influence upon the relevance of the dialogic moves expressed by the groups. However, the scientific situation showed a degree of complexity higher than the others, as I mentioned before. This could have contributed to students more frequently expressing partially relevant and/or irrelevant moves than in other situations. Based on this dilemma, I have drawn up Fig. 9, which explores the frequency of teacher and researcher intervention in the modelling of everyday (represented by the use of a black-and-white shade), scientific (represented by the use of a white-and-dark-grey shade) and socio-scientific situations (represented by the use of a white-and-light grey shade).

Fig. 9

Frequency of teacher and researcher intervention in everyday, scientific and socio-scientific modelling situations experienced by groups 3 and 6

From Fig. 9, it can be seen that the researcher and teacher intervened more frequently in modelling scientific situation than in the others. This could have happened because the complexity of this type of situation is higher than for the others, and also the fact that the teacher and the researcher have greater clarity about what types of knowledge the students should construct and apply in this type of situation compared with the others, as discussed before. The fact that such interventions have occurred and, as a result, given greater guidance to the high school students’ reasoning could have contributed to the students showing less expression of partially relevant and/or irrelevant moves (Figs. 7 and 8).

With regard to the frequencies of interventions of the researcher and the teacher in the everyday and socio-scientific situations, it can be seen in Fig. 9 that they are much lower than those that occurred in the scientific situation. I assess that the fact that such interventions may have occurred with a lower frequency in the everyday and socio-scientific situations may have contributed to the students expressing more partially relevant and/or irrelevant moves than in the scientific situation (Figs. 7 and 8).

Finally, looking at Figs. 7 and 8, it is noted that the students of G6 expressed less partially relevant and irrelevant moves than the students of G3. This could suggest that the students of this group have shown better development of their argumentative dialogues, in terms of the quality of argumentative moves, than those in G3—an aspect which, in turn, could have had an influence on the construction of knowledge in the situations investigated. To evaluate this statement, I present Figs. 10 and 11 in the next topic.

5.5 Contribution to the Construction of Knowledge in Dialogic and Meta-dialogic Moves in Modelling Situations

Figures 10 and 11 show the frequency of contribution to the construction of knowledge in dialogic and meta-dialogic moves expressed by groups during the modelling situations.

Fig. 10

Frequency of contribution to the construction of knowledge in dialogic and meta-dialogic moves in everyday, scientific and socio-scientific modelling situations experienced by group 3

Fig. 11

Frequency of contribution to the construction of knowledge in dialogic and meta-dialogic moves in everyday, scientific and socio-scientific modelling situations experienced by group 6

In Figs. 10 and 11, it is noted that the dialogical and meta-dialogical moves expressed by both groups frequently contributed to the construction of knowledge in all modelling situations. The calculation of the relative frequency in percentage¹⁴ indicated that 93.0% and 98.5% of the moves expressed by the students in G3 and G6, considering all the teaching situations as analysed, generated contribution to the construction of knowledge.

It is also observed that the students expressed few moves that did not contribute to the construction of knowledge in the scientific situation (one in the case of G3 and two for G6). I consider that this may have happened due to the intervention of the researcher and the teacher, which was more frequent in this situation than in the others (Fig. 9). The fact that such interventions have occurred frequently in the scientific situation may have contributed for greater guidance of the high school students’ reasoning and therefore helped the members of both groups to show few moves that do not help the construction of knowledge. As discussed in the works of Passmore and Svoboda, (2012), Mendonça and Justi, (2013) and Puig et al., (2017), production of modelling activities with clear goals, explicitly stated in the commands of the questions, and knowing how to lead them may have a particularly important influence in the construction of knowledge by the students.

In everyday modelling situation, I identified, in the moves that generated contribution to the construction of knowledge, that G3 and G6 constructed and applied knowledge related to the selection, delivery, refrigeration, payment, change and the activation operations, of a machine to sell soft drink cans.

On the other hand, in scientific modelling situation, I identified, in the moves that generated contribution to the construction of knowledge, that G3 and G6 constructed and applied knowledge related to plastic structures; the physical and chemical properties of different types of plastics and of organic compounds.

Finally, in socio-scientific modelling situation, I identified, in the moves that generated contribution to the construction of knowledge, that G3 and G6 constructed and applied knowledge related to plastic structures; the physical and chemical properties of different types of plastics; the recycling process of a type of plastic and other solutions that could help to solve the problem of accumulation of plastics in society.

On conducting an analysis of Figs. 10 and 11, it is observed that students of G6 showed fewer moves that did not contribute to the construction of knowledge than the students of G3 in all situations. This result supports the statement I presented previously that the students of G6 developed their argumentative dialogues better, in terms of the quality of argumentative moves, than those in G3, thereby generating more contributions for the construction of knowledge in the situations investigated, compared to the students of G3.

6 Conclusions

The first research question addressed in this paper is: What are the characteristics of high school students’ argumentative dialogues in everyday, scientific and socio-scientific modelling situations? The results show that students engaged in different argumentative dialogues seeking to achieve different goals, including those of exchanging information (Information Sharing Dialogue (Dis)), elaborating explanation(s) or hypotheses of a phenomenon under discussion (Discovery Dialogue (Dds)), investigating/testing the validity of one or more hypotheses/models from evidence (Inquiry Dialogue (Din)) and proposing a solution/action to a problem under discussion or solving such problem (Deliberation Dialogue (Ddl)) in everyday, scientific and socio-scientific modelling situations.

In these argumentative dialogues, students expressed different types of dialogic and meta-dialogic moves in different modelling situations, namely Dialogical deliberative (DdelM), discovery (DdisM), information Sharing (DishM), inquiry (DinqM), persuasive (DperM), persuasive dispute (DpdiM) Moves, Meta-Dialogical clarification Move (MDclaM), Meta-Dialogical Move to clarify the meaning of an idea (MDcmiM), Meta-Dialogical context-setting Move (MDctsM) and/or the Meta-Dialogical consensus Move (MDconM). Thus, they have expressed different argumentative moves mainly to (i) request a decision or solution to a problem, and select a more coherent/appropriate solution/action to be considered; (ii) request and prepare explanation(s); (iii) obtain and provide information in response to a request/question; (iv) investigate the validity of model from evidence and request their investigation; (v) convince others of the validity of an idea/model and request that this be done, as well as resolve a conflict of ideas to analyse each idea and request that the conflict be resolved; (vi) resolve a conflict of ideas in order to select which idea/s is/are the most appropriate or coherent and request that this conflict be resolved; (vii) clarify the meaning of the purpose of other moves and seek clarification of this meaning; (viii) clarify the meaning of an idea present in other moves and seek clarification of that meaning; (ix) clarify whether the individual has any prior knowledge considered relevant to the discussion and seek clarification on this and/or (x) express the sharing of the same idea and ensure that this happens.

Furthermore, the students expressed moves that sought to defend and sustain their ideas through relationships of direct support more often than to attack them in in everyday, scientific and socio-scientific modelling situations. Most of these moves were indeed relevant and also contributed to the construction of knowledge in different modelling situations.

These results show that both groups expressed quality argumentative moves, and consequently, they got engaged in quality argumentative dialogues in the different modelling situations evaluated. Therefore, engaging students in modelling situations tends to make a contribution to students conducting argumentative dialogues with quality. This is an important conclusion to the field of Science Education, as previous studies have investigated the occurrence of argumentative situations in the modelling stages (Mendonça & Justi, 2013; Passmore & Svoboda, 2012) or the quality of argumentation in only one kind of situation (Evagorou et al., 2020).

The second research question addressed in this paper is: Are there differences in the characteristics of high school students’ argumentative dialogues when they participate in everyday, scientific and socio-scientific modelling situations? If there are differences, what are they? The results show that there were differences in the characteristics of argumentative dialogues that the groups got engaged in everyday, scientific and socio-scientific modelling situations.

The first difference observed was that the frequency and expression of dialogic moves were also different between the groups investigated. The students in G3 showed the Meta-Dialogical context-setting Move in all the situations, which was not the case with the members of G6. On the other hand, the students in this group increased the frequency of Dialogical persuasive Moves during the modelling situations. The same standard was not observed in G3. These results show that each group of students conducted the argumentative dialogues in particular way, producing different questions and answers. This could have had an influence on the frequencies and expressions of different dialogic and meta-dialogic moves between the groups.

The second difference noticed was that, in the case of G6, there was a gradual increase in the frequency of attacks, and the relation of direct support during the modelling situations, as also a rise in frequency of the Dialogical persuasive Move during these situations. These results show that the participation in the modelling situations contributed to the students in G6 perceiving the need to exploit and suggest that each proposed idea or solution be evaluated in depth during the process experienced. This pattern, however, was not observed in G3.

The third difference observed was that the members of G6 showed less partially relevant and irrelevant moves than those of G3. This could suggest that the students in G6 had better development of their argumentative dialogues, in terms of the quality of argumentative moves, than those of G3.

The fourth difference noticed was that the students of G6 showed less dialogic and meta-dialogic moves that did not contribute to the construction of knowledge, compared to G3 in all situations. This result backs up the statement previously mentioned that the members of G6 have developed their argumentative dialogues better than those of G3, which may have contributed more intensely to the construction of knowledge, in the situations as investigated.

Based on the points here mentioned, it is concluded that the differences found between the different groups are linked to the way in which each group developed argumentative dialogues throughout the process, producing different questions and answers to solve the problems posed to them. It is also concluded that the engagement of students in modelling that involve situations with different natures may not favour a gradual improvement of the quality of argumentative moves and, consequently, in the quality argumentative dialogues that they got engaged.

The third research question addressed in this paper is: What are the influences of the nature of each situation of modelling on the characteristics of high school students’ argumentative dialogues? The results show that nature of the modelling in the everyday situation influenced the students to get engaged in Discovery Dialogue (Dds) and did not get engaged Information Sharing Dialogue (Dis) and Deliberation Dialogue (Ddl) as the groups had to create and express a model to explain the operation of a machine to sell cans of soft drinks. As this was a well-known situation, the students of both groups did not have to predict what would happen to function of a the vending machine selling canned soft drinks in certain situations, in the same way that they did not need to obtain any additional information before creating, testing or assessing the models—an aspect that would contribute towards the engagement in an Information Sharing Dialogue (Dis), as was the case in the other teaching situations. Furthermore, the students of both groups did not have to propose a solution/action to a problem under discussion or solve such problem—an aspect that would contribute towards the engagement in a Deliberation Dialogue (Ddl).

The nature of this situation influenced the students to express the Dialogical discovery (DdisM) and deliberative (DdelM) Moves, as they sought to request or prepare explanation(s) to explain the operation of a machine to sell cans of soft drink, and request a decision between possible explanations and select a more coherent/appropriate explanation to be considered.

In relation to the modelling in the scientific situation, it influenced the students to get engaged in Discovery Dialogue (Dds) rather than Deliberation Dialogue (Ddl). This is because the groups had to prepare models that could explain the different behaviour patterns of plastics when they are subjected to attempts to fold them or to indirect and direct heating in the flame. This justifies the students having expressed the Dialogical discovery Move (DdisM). For the groups to prepare the models, they first had to get engaged in Information Sharing Dialogue (Dis) to predict what would happen to certain plastics on attempting to fold and heat directly and indirectly in the flame.

Furthermore, the results show that nature of this situation influenced the students of both the groups to express more frequently the Dialogical Information Sharing Move (DishM) than in the other situations. This could have happened as a result of the complexity of this type of situation, which is greater than that of the other situations, because of the nature of the knowledge thus constructed. Thus, students had to share information frequently so that knowledge could be built and express their agreement with the idea of their colleagues in order to make it clear that they shared the same idea, so they could progress in this teaching situation. In the everyday situation, the groups could use social knowledge, which is part of their culture and their daily life to explain the operation of a vending the vending machine selling canned soft drinks. In the case of the socio-scientific situation, the groups have not had to construct scientific knowledge, but just to apply this knowledge, together with knowledge of social nature, to solve the problem of accumulation of plastics in society. On the other hand, in the scientific situation, the members of G3 and G6 had to construct and use exclusively scientific knowledge to be able to solve the problem under discussion. Furthermore, the complexity of the scientific situation was made even bigger because of the fact that the science studied was chemistry, which is a subject composed by abstract concepts that may not be familiar to the students.

Finally, the results show that the nature of the socio-scientific situation influenced the students to frequently show the Dialogical deliberative Move, as they had to get involved in the Deliberation Dialogue to solve the problem of accumulation of plastics in society. So in order for the goal of the dialogue to be reached, it was necessary that they proposed solutions and selected the most appropriate, and also requested the proposal of a solution and/or action, or the choice of one among the possible solutions or actions.

I did not identify any patterns by analysing the nature of each modelling situation and the characteristics of the argumentative structure the moves (DperM (Dialogical persuasive Move) and DpdiM (Dialogical persuasive dispute Move)) in argumentative dialogues nor did I observe any patterns on analysing the nature of each modelling situation and the relevance of the students’ moves and their contributions to the construction of knowledge in argumentative dialogues. This is because most of these moves were indeed relevant, and also contributed to the construction of knowledge in everyday, scientific and socio-scientific modelling situations.

Based on the points I raised, I conclude that the nature of the modelling situation could indeed have an influence on specific aspects of students’ argumentation, without this influence interfering with the quality of argumentative dialogue thereof. This is because all the situations investigated contributed to the students expressing quality argumentative moves, i.e. which were relevant to the dialogues and contributed to the construction of knowledge. Therefore, it seems that there is no situation better than the other, in terms of what they have contributed to the quality of argumentation. To support this claim, further studies are needed to investigate influences of the nature of situations of modelling on the characteristics of students’ argumentative dialogues.

I consider that the use of situations with different natures within teaching could contribute to students having opportunities to engage in different dialogues and dialogic and meta-dialogic movements in argumentative situation, in a process of construction of knowledge. Such opportunities may mean that the students can be capable of actively participating in argumentative dialogues of different natures and with different goals, which could contribute to their qualification as citizens.

The last research question addressed in this paper is: What are the influences of modelling on the characteristics of high school students’ argumentative dialogues? The results show that the modelling has influenced the engagement of students in the Inquiry Dialogue (stages of testing and evaluating the model of Model of Modelling (v2))—something that was expected, due to the fact that modelling is an investigative activity, as stated by Justi and Gilbert, (2002). On the other hand, this type of activities has not helped them get involved in Persuasion Dialogue, as the goal is not to solve a conflict of ideas, but rather testing and evaluating the validity of a hypothesis based on evidence.

In addition, the results show that modelling has influenced the showing of Dialogical information Sharing (DishM), inquiry (DinqM), Persuasive (DperM) Moves, and the Meta-Dialogical consensus Move (MDconM) and Meta-Dialogical Move to clarify the meaning of an idea (MDcmiM). In other words, the modelling has contributed to the students: (i) sharing information; (ii) investigating the validity of a hypothesis/model from evidence, and requesting their investigation (iii) convincing the other of the validity of an idea/model and requesting that this be done, as also resolving a conflict of ideas to analyse each idea and requesting that the conflict be resolved; (iv) stating whether they agree with the idea of another colleague, and seeking to make it clear that both share the same idea and (v) clarifying the meaning of the purpose of other moves and seeking clarification of this meaning. The students of the groups investigated showed these moves in all situations. Therefore, this study expands the conclusion of Passmore and Svoboda, (2012) and Mendonça and Justi, (2013) since the students’ argumentative dialogues have characteristics of sharing information, convincing the other party and sharing the same ideas in all modelling stages, as well as investigating of the validity of hypotheses from evidence in the stages of testing and evaluating the model of Model of Modelling (v2).

The occurrence of DishM could be linked to the fact that, in order to create, test and assess models, the groups would need to obtain and provide information. Thus, in the stages of creating, expressing, testing and evaluating the model of Model of Modelling (v2), the students in both groups shared information in a move to obtain and provide information about the institution to be modelled through external sources and also internal sources such as previous ideas, select the modes of representation for stating the model, checking the appropriateness of the model for the purposes for which it was devised and assessment of the scope and limitations of the model thus created.

The expressed DinqM could be linked to the fact that modelling is an investigative activity. Therefore, it would be totally unexpected if both groups did not seek to investigate the validity of any hypothesis raised, based on evidence, or request that this be done in the stages of testing and evaluating the model of Model of Modelling (v2)—which is in agreement with the definitions of these stages of modelling proposed by Gilbert and Justi, (2016b).

The occurrence of DperM is linked to the fact that argumentation is one of the cognitive processes inherent to the reference for modelling as adopted in this study (Gilbert & Justi, 2016d). Thus, both groups sought to convince the other of the validity of an idea/model and request that this be done, as also resolve a conflict of ideas to analyse each idea and request that the conflict be resolved in the stages of creating, expressing, testing and evaluating the model of Model of Modelling (v2).

The occurrence of MDconM (Meta-Dialogical consensus Move) is due to the fact that modelling has well-defined goals and stages (Gilbert & Justi, 2016b), and this came from the fact that the activities that the groups carried out also have such aspects in all modelling stages. These aspects influenced the groups to express if they agreed with the ideas raised by colleagues and to request that these things be done. Without performing these actions, the groups could have difficulty to achieve the goals of activities and proceed with the modelling stages.

The modelling influenced students not to express the Dialogical eristic Move. Thus, this process does not favour students to attack or defend colleagues, but rather the arguments that they produced. This result may indicate that modelling can contribute to promoting dialogues in which students construct and assess ideas based on arguments.

The modelling has also influenced the students frequently pursued to (i) attack the conclusion (or part thereof), the reasons (all, or at least one of them) and the relationship between reasons and conclusion, through questions, arguments and reasons and (ii) defend and sustain their ideas/hypotheses/solutions, based on a relationship of direct support throughout all the teaching situations considered. Thus, the modelling influenced students’ argumentative dialogues to have characteristics of convincing the other party about the validity of an idea and also of solving conflict of ideas in a consistent way.

Furthermore, both groups sought to defend and sustain their ideas through relationships of direct support more often than to attack them during the stages of Model of Modelling (v2) in all modelling situations. Therefore, this study expands the conclusion of Evagorou et al., (2020). In the study by these authors, refutation occurred mostly in evaluation stage of modelling.

This type of activity also influenced students from both groups to frequently show relevant moves during the stages of Model of Modelling (v2) in all modelling situations. This result shows that modelling may play a part in making students show relevant moves in argumentative dialogues.

However, I must also stress that this high frequency is not linked only to modelling, but also to the interventions made by researchers and by the teacher, which were essential to help the students during the problem-solving processes. As I discussed in the Methodological Aspects section, both the teacher and the researcher had prior knowledge with regard to modelling and argumentation. Thus, the results show that the role played by the teachers within modelling has an influence upon the quality of the high school students’ argumentative dialogue and, as a result, on the construction of their knowledge. It is therefore essential that teachers also have such knowledge that is capable of leading this kind of activity, as discussed by Mendonça and Justi, (2013). For this reason, I assess that it is relevant to include discussions on argumentation and modelling in initial and ongoing teacher training courses.

Finally, the results show that modelling influenced the students got engaged in dialogic and meta-dialogic moves that contributed to the construction of knowledge. Indeed, 93.0% and 98.5% of the moves shown by the students in G3 and G6 during the stages of Model of Modelling (v2) in all the teaching situations as analysed have generated contribution for the construction of knowledge. Therefore, I conclude that the participation of students in modelling tends to favour students to engage in quality argumentative dialogues that generate a contribution to the construction of knowledge of different natures: social, scientific and socio-scientific. This conclusion corroborate the idea that working on scientific practices in an integrated way, within the science classroom, could strengthen the construction of knowledge and quality argumentative dialogues, as theoretically mentioned by the US document A Framework for K-12 Science Education (NRC, 2012).

7 Implications

I evaluate that the implications of this research for teaching are related to the recommendation that the teacher construct and apply modelling activities that involve distinct situations in different sciences subjects. This is because participation in such activities may contribute to science as students: (i) have opportunities to get engaged in different argumentative dialogue, within a process of construction of knowledge and (ii) get engaged in quality argumentative dialogues that generate a contribution to the construction of knowledge, by offering them opportunities to express different dialogic and meta-dialogical moves; attack the conclusion (or part thereof), the reasons (all, or just one of them) and the relation between reasons and conclusion through questions, arguments and reasons; defend and sustain their ideas, hypotheses and solutions, mainly setting out from a relation of direct support.

These opportunities could offer ideal conditions so the students may act critically and thoughtfully in the argumentative dialogues that involve different natures or theme, which could contribute to their qualification as citizens.

Finally, I assess that the implications of this study for the research involve the need for additional studies, that aim, for instance, to carry out analysis of argumentative dialogues (or the quality thereof): (i) in modelling that involve different socio-scientific topics, as also different everyday and scientific topics; (ii) in everyday, scientific and socio-scientific situations based on one single didactic approach, different from what was used in this study and (iii) in distinct levels of education (primary and above) and science subjects (biology and physics). Such research studies could indeed contribute to understanding of if and how different investigative approaches and contextual factors, such as prior knowledge, familiarity and personal relevance of the problem, have influenced the students’ argumentative dialogues and the relations between argumentation and modelling.