Introduction

Research has long shown the relevance of science classroom discourse in education (Lemke 1990; Mortimer and Scott 2003), as informed by the ideas of Vygotsky (1978), which state that social interactions are fundamental to human learning. Discourse in the classroom, understood as language-in-use (Cazden 2001), has become an essential tool for science teaching (Kelly and Green 2019). Among the different research fronts that have emerged over the last few decades, we highlight the development of methodologies that are able to establish a connection between discourse within the science classroom and the sociocultural backgrounds of its members (Howe and Abedin 2013).

This is by no means a new concern in the area of science education (Lemke 2001). Over the last decade, the school-age population has become more and more diverse. This includes a series of actions and challenges presented to teachers, curricula and government authorities, in the promotion of equity in learning opportunities (Shea and Sandoval 2020). For the research community, there are also significant challenges looming ahead, especially from methodological issues, with which this paper is concerned.

A major challenge aimed at equity issues in science education is the development of discourse analysis capable of showing how macro-contextual elements of society are in fact defined ways of construction of scientific knowledge in the classroom (Franco and Munford 2020; Gómez Fernández 2019; Howe and Abedin 2013). Neglecting this aspect means understanding science education as a transmission of an aseptic and decontextualised form of knowledge (Lemke 2001). The naive idea that scientific concepts and their epistemic practices are neutral still lingers (Shea and Sandoval 2020).

In spite of this, translocal meanings from different sociocultural contexts are intertwined with the scientific enterprise. This aspect of the Nature of Science is well documented in relation to the work of scientists (e.g. Schiebinger 1993), but we need to have a better understanding of how this happens with the “outsiders” who learn science or have access to scientific information. How to develop methodologies in order to analyse the relationships between different contexts of human life and students’ engagement in scientific discourse practices in the classroom? This is an important question for the twenty-first century regarding equity in science education.

In a similar direction, another methodological challenge concerns new scenarios of increasing diversity in science classrooms. In recent years, analyses such as these have become more complex, as the sociocultural backgrounds within the classrooms become less and less predictable (Vertovec 2010): they are dynamic and constantly under negotiation (Blommaert 2015). Phenomena such as significant migratory movement, the advent of the Internet and the increased use of smartphones (Vertovec 2010) have brought on significant changes to the nature of our interactional life in different parts of the world. Sociocultural borders have been thinner, affecting the way we act and the way we communicate. The idea that a classroom is a “pure” community, with well-defined characteristics or a stable culture, must now be cast aside (Blommaert and Backus 2013).

If the classroom’s diversity is becoming less and less predictable, research needs to develop tools to understand science classroom discourse in a more fluid way, when interactions take place in complex and diverse contexts. However, many of the tools for science classroom discourse analysis do bring limited responses to these demands. Recent tools have given emphasis to factors such as the frequency of certain codes that have been created to establish categories for types of discourse in the classroom and/or characterise what the teachers and students are saying (Vrikki et al. 2018; Bansal 2018; Lee and Irving 2018; Wagner and González-Howard 2018). The analyses have shown common features of discourse in the classroom (as is typical) or uncommon features (something that rarely appears), or even variations between different classes and different contexts. Despite such progress and such significant results, the design of research into these tools is limited, to take into account the intercontextual relations in the classroom and their growing diversity. Assuming interactional features of the “other” based on the frequency of occurrence has become an unfeasible task (Blommaert 2015).

In this article, we present the hourglass approach as an analytical metaphor that can contribute to responding to such demands. We use the hourglass as a metaphor as it has proved to be useful for organising and describing a research design, based on ethnography in education, to analyse discourse practices in the science classroom, considering intercontextual relationships. For this, the hourglass approach aims to:

  1. i)

    Give due visibility to aspects that are meaningful for the members of the classroom themselves. This means that instead of a stable view of the community of the science classroom, where we could try to fit in different categories of discourse to describe interactions, we envision the classroom based on a more fluid view (Green et al. 2020). The ethnographic traditions assume “non-awareness of the other” (Agar 1994), which brings significant analytical potential considering the sociocultural backgrounds of local communities.

  2. ii)

    Analyse science classroom discourse as an amalgam in which translocal meanings are constantly under negotiation (Bloome et al. 2008). This aspect becomes more relevant within a multicultural world, where the science learning process is intertwined with diverse contextual elements that make up human life.

The approach is organised into three articulated levels of analysis (Fig. 1):

Fig. 1
figure 1

Schematic representation of the hourglass approach

Full size image

“Funnelling” refers to learning about the everyday life of participants’ group, seeking to obtain a continuous view of their history. Analytic questions undergo transformation when faced with the characterisation of this history, considering three main axes: (i) the analytical potentiality of events over science lessons; (ii) the significance of the events for the members of the group and (iii) relationships with science classroom discursive practices. A detailed analysis of discursive interactions at such events is the vertex of the hourglass. Finally, new processes deepen the understanding of the interactions within the group, establishing intercontextual relationships in science classroom discursive practices.

The hourglass metaphor is related to the design of the research itself (funnelling > expansion) but that is not all. As the sand is funnelled, the researcher may stop in time and then describe and analyse a “single grain of sand” which passes through the vertex. This grain of sand is not analysed in itself but based on what has occurred previously and what is still going to happen. Thus, an event is analysed considering how it is situated in time and space. Like in the case of a regular hourglass, we can move ahead and backtrack in time, seeking to understand historical relations that have been constructed by the group, and also connections between different contexts.

Ethnography in Education and Science Education: the Bases of the Hourglass Approach

The theoretical and methodological foundations of the hourglass approach are based on the articulation between two fields: ethnography in education and science education.

Ethnography in education gathers elements from different traditions, including ethnography of communication (Gumperz and Hymes 1972), interactional sociolinguistics (Gumperz 1982), critical discourse analysis (Fairclough 1992), interactional ethnography (Castanheira et al. 2001; Erickson 2008), micro-ethnographic discourse analysis (Bloome et al. 2008; Green et al. 2020) and linguistic anthropology (Blommaert 2015). In spite of this theoretical plurality, their assumptions amount to a coordinated effort in an attempt to understand how people create and negotiate the routine activities of their lives by means of discursive interactions. This discourse construction does not unfold in an “aseptic” way but is operated using micro-invocations through discourse that have locally been (in)validated by macro-contextual meanings (Blommaert 2015).

There are other assumptions and analytical practices commonly used in discourse analysis that share similarities with the hourglass approach. One most related practice is looking at discourse events at the different grain levels of micro, meso and macro, establishing links between classroom discourse and larger sociocultural patterns (see Erickson 2008, Franco and Munford 2018). An important difference between the traditions on which the hourglass is based and other traditions in discourse analysis refers to the ways of linking different levels or scales of analysis. One way of establishing these relationships is to understand what happens at the micro-level as if it were embedded in the macro-level. As Erickson (2008) mentions, this top-down approach would lead to ignorance of what is happening at the local level and to overestimation of macro-contextual meanings. Another way would be to give visibility at the micro-level discourse and incorporate elements from macro-level discursive processes (Bloome et al. 2008). One key criticism of this bottom-up approach is that it can concentrate “so intimately on specific characteristics of talk itself, that it ignores global aspects of the ecology of conversation” (Erickson 2008, p. 108).

From an ethnographic point of view, Blommaert (2015) raises the relevance of going beyond these mere unidirectional juxtapositions between different contextual levels in discourse analysis. The hourglass approach follows this advice, organising efforts in building a multidimensional view of the intercontextual relationships. For this, we used the notion of intercontextuality as a theoretical-methodological tool. Intercontextuality “refers to the social construction of relationships among events and contexts” (Bloome et al. 2009, p. 319).

Intercontextuality is discursively constructed in the classroom by teachers and students (Bloome et al. 2009). The classroom members circulate in different space-time dimensions, every day, and live through distinct experiences in their families and communities; they may be spatially close to, or distant from, their groups of origin and construct their groups of friends within the groups they participate in, using the Internet with access to a wide range of information, and also strike up relationships through social networks (Blommaert and Backus 2013). In the science classroom, such macro-contextual repertoires have translocal meanings that cross each other in the micro-contextual process of lessons; they are dynamic and interactive (Bloome et al. 2008; Franco and Munford 2018).

The other theoretical-methodological basis of the hourglass approach comes from the academic field of science education. In order to analyse science classroom discourse practices in each hourglass phase, we established articulations between ethnographic perspective and science education constructs through the notion of science learning opportunities (Rex 2006).

As proposed by Rex (2006), learning opportunities are “social events in which a person or persons are positioned to adopt and take up a set of social and cultural practices associated with academic domains” (Rex 2006, p. 165). We established relationships between these conceptions and Duschl’s (2008) proposal of a science education curriculum in three-part harmony: a balance between conceptual, epistemic and social domains of scientific knowledge (Franco and Munford 2020).

The conceptual domain involves opportunities for students to learn scientific explanations about the natural world and use the body of knowledge that represents these explanations (Duschl 2008; Stroupe 2015). The epistemic domain is related to opportunities that students have to use epistemic criteria that the scientific community uses to construct knowledge. These criteria allow students to recognise what they know and why they are convinced that they know it (Stroupe 2015). The social domain involves opportunities for understanding “processes and contexts that shape how knowledge is communicated, represented, argued, and debated” (Duschl 2008, p. 277). This domain can be characterised by the routines from which students develop, criticise and apply ideas to build knowledge (Stroupe 2015). The hourglass approach used these discussions to characterise how students positioned themselves to adopt a set of practices from the conceptual, epistemic and social domains of scientific knowledge.

Data Used to Illustrate the Hourglass Approach

To illustrate the hourglass approach, we used data regarding a Brazilian class group that we followed through its first 3 years of elementary school, between 2012 and 2014. These 27 students entered the school through a publicly open random draw, meaning that there were students from different regions of the city, coming from a range of different early childhood education institutions; in addition, they were also diverse in terms of social and economic level and ethnicity, as shown in Table 1.Footnote 1

Table 1 Demographic features of the group of children investigated
Full size table

The school is run by the Brazilian Federal Government and it is located in a metropolitan area. The teacher we have followed, Karina, taught Science and Portuguese. Despite her vast experience as a teacher and plenty of expertise in the language area, she had little contact with science as a curriculum component. The lessons were guided by inquiry-based science teaching (e.g. Pedaste et al. 2015), based on work carried out jointly between the teacher and the research team. The subject matter of the science lessons over 3 years that we followed the group is represented in Table 2.

Table 2 Science lessons in the first 3 years of primary school, in the class, are investigated in the study
Full size table

The immersion in the class everyday life throughout 3 years took place based on participant observation of science lessons (Spradley 1980), and the records produced videos, field notes, photographs and materials from various activities. The organisation of this extensive database involved the work of our research team in a cloud file storage service. All data was organised in an extensive database and summarised in a large spreadsheet that gathered information about the collected material: general information about each lesson (day, month, year, activities, observers, comments) and the encoded identification of each file and its respective links directed to cloud storage (video files, photos, digitalised activities, digitalised field notes and other artefacts).

The Upper Part of the Hourglass

In the upper part of the hourglass, the grains of sand represent the events that unfolded in the group’s science lessons. Moving down the hourglass, we observe a funnelling towards the vertex. This process reflects procedures that aim to select an ever-smaller number of events, to make certain features of the group more evident.Footnote 2

This analysis involves some questions that are wider in scope. “What is happening here?” is the key question for studies guided by ethnography (Goffman 1974). The researcher constructs answers through different representations of history. Descriptive charts of the lessons, as well as event maps and timelines, are all useful tools in this process (Green et al. 2020). As an example, we now present an excerpt of a long timeline (Fig. 2).

Fig. 2
figure 2

Timeline of the group’s activities in some lessons that occurred in the month of November, in the 1st grade

Full size image

In the top column, broad narrative descriptions present information on the science teaching/learning context, considering conceptual, epistemic and social domains of scientific knowledge (Duschl 2008). In the bottom column, there are comments about elements from different contexts (Bloome et al. 2008) as evoked by the participants.

Through this representation, we have identified points of intersections between discursive practices and intercontextual meanings in the science classroom. For example, on 8 November, students had to observe three stick bugs in the classroom and, also using other knowledge about the insects, decide the sex of each one. A conceptual topic of science, sexual dimorphism, was being discussed based on observational criteria and required argumentation between peers, relevant practices from epistemic and social domains of scientific knowledge (Duschl 2008). Based on field notes, and sometimes videos,Footnote 3 we identified conversations in which students evoked translocal meanings throughout their engagement in science classroom discursive practices, such as gender issues, family experiences and out-of-school sources of information.

We chose this excerpt as it is the case explored in more detail to illustrate the use of the hourglass approach. The complete timeline, over 3 years, allowed us to identify different intercontextual references/resources: the community of students, their family units and networks of friends, Internet and social networks, out-of-school sources of information, issues such as those of gender, religion, ethnicity and social class. We selected all those lessons in which these different sociocultural contexts seemed to be related to discourse practices in the science classroom.

At the next analytical level, the hourglass leads to a new database cut, from the cut generated in the previous level. The metaphor of the hourglass funnelling is inspired by a telling case (Mitchell 1984) logic. This involves looking for an event, or set of related events, in which “particular circumstances surrounding a case serve to make previously obscure theoretical relationships suddenly apparent” (Mitchell 1984, p. 239). According to Agar (1994), moments of unfulfilled expectations are rich points for this type of analysis. Educational ethnographers have taken over this proposal based on the notion of frame clash (Green et al. 2013).

Frame clashes are situations in which the ordinary flow of everyday life is disrupted, by conflicts or breaches of expectations between the participants or conflicts between the researcher’s knowledge/expectations and the group’s sociocultural knowledge (Agar 1994; Green et al. 2013). The frame clash analysis has the potential to make “visible the cultural knowledge and meanings of the group and individuals-within-the group that differ from those of the ethnographer” (Green et al. 2013, p. 127).

In the analyses that we now present, we returned to all lessons in which there was evidence of links between science learning opportunities and different contexts evoked by the participants (e.g. Figure 2) to identify frame clashes. In all, we identified 23 frame clash events over time (Fig. 3).

Fig. 3
figure 3

Set of events in which we identified frame clashes

Full size image

At this third level of analysis, we move to a new data funnelling in the hourglass. We constructed mesoscale descriptions of each frame clash in a historical perspective (Castanheira et al. 2001) in order to select one of them to analyse in detail. We have contrasted the frame clashes over time in order to mapping how each of them has assumed different social consequences in the history of the class (Fig. 4).

Fig. 4
figure 4

Excerpts from the historical analysis of two frame clashes

Full size image

An important first decision at this level of analysis was to turn our attention to those frame clashes more related to gender issues. These issues were quite expressive over the science lessons, not only because they appeared more frequently than other conflicts but also were those that were more closely related to students’ engagement in practices of conceptual, epistemic and social domains of scientific knowledge. This led to new queries and framings that gave visibility to this specific context within the social life of the group.

Among the “gender frame clashes”, we also make a historically informed decision. We distinguished between those events evoked over time as relevant resources in ongoing discussions, from others that have not been mobilised by participants or which appear as patterns that occur repeatedly throughout the history of the group. In doing so, we come across three gender frame clashes (28/6/2012, 11/8/2012, 3/9/2014) with greater potential. Among them, we selected the “stick bug” frame clash (11/8/2012) for the hourglass vertex.

The historical analysis indicated its relevance not only between 1st-grade lessons but also in the 2nd and 3rd grade. Furthermore, another criterion was relevant in this decision: the specificities of the science learning opportunities in this frame clash. Unlike other frame clashes, gender in “stick bug” frame clash was not a pattern related to boys/girls’ participation in science lessons. It was a conflict directly intertwined with a scientific matter under construction. Evidently, boys/girls’ participation in science lessons is an extremely relevant concern. However, the gender issues appeared during the stick bug frame clash in an uncommon way within an absolutely common situation. Children were mobilising gender to legitimise scientific knowledge.

The selection of the stick bug frame clash led us to look at the set of lessons in which the group discussed the stick bug biology (Table 3). In the first observations of the insects, there was a relative consensus that the bigger one would be the “father”, the medium-sized one the “mother” and the smaller one the “son”, as noted in Karla’s record (Fig. 5). However, the idea that the bigger insect could be the female became more accepted as the lessons progressed, creating a conflict between students. An apparently consensual knowledge started to be questioned, generating an increase in uncertainty among the participants.

Table 3
Full size table
Fig. 5
figure 5

The “stick bug” frame clash

Full size image

The increase in uncertainty and conflict gave visibility to translocal contexts that were being raised by the pupils while they established criteria for the sexing of the insect. During the lessons, the participants supported their views with arguments such as being stronger, being calmer, passive, more nervous and angrier, or social roles, like who takes care of the family, who eats more and who defends the family. In this way, the local activity flow about studying sexual dimorphism, and the related epistemic and social practices, was operating, in this local community, through negotiations about what it means to “be a man” and “be a woman”.

Figure 6 is the result of a mapping of events in which the “sex of the stick bug” had visibility in the ongoing discussions. In all these events, it was possible to identify evocations related to gender issues intertwined with the sexual definition of insects. That is relevant because it indicates that the stick bug frame clash became a “collective memory” within the history of the group. In other words, we identified narratives about past events that were acknowledged by the members of the classroom and used as resources in future discussions over 3 years (Bloome et al. 2009).

Fig. 6
figure 6

Timeline “the sex of the stick bug” events

Full size image

In short, the hourglass funnelling is developed through three levels of analysis. First, a panoramic view of the group’s track record in science lessons, highlighting links between science teaching/learning context and other sociocultural contexts. Second, we identified frame clashes as rich points for further analysis. And third, we make decisions about which frame clash to explore in detail, considering two criteria: the historical relevance of the events and their relationships with science learning opportunities.

The Vertex of the Hourglass

The vertex of the hourglass means analysing a “single grain of sand” which passes through the vertex. As the sand is funnelled, the researcher may stop at any time and then analyse an event or set of related events that become anchors for subsequent analyses.

The vertex analysis is carried out through the transcription of talk in message units. Green and Wallat (1981) define the message unit as the smallest unit of meaning within the analysis of a conversation. To establish the borders and limits between each unit, we use signs that Gumperz (1982) calls contextualisation cues and includes verbal, non-verbal and prosodic signals, such as changes in intonation when speaking; rhythm; emphasis; speed; pauses; gestures and eye contact. This type of analysis does not aim to understand the internal state of people or their internal intentions. Aiming to interpret the meanings shared in a group, it focuses on consequences of what is said, analysing how people act and react to each other (Bloome et al. 2008).Footnote 4 From the science education perspective, the vertex shows ways in which participants engage in practices from conceptual, epistemic and social domains of scientific knowledge through discourse (Duschl 2008; Rex 2006), as illustrated in Chart 1, considering the interactional unit 1.

Chart 1 Interactional unit 1
Full size table

In this paper, we select an anchor event to conduct what we called “vertex analysis”. This event is part of a discussion that took place on 22 November, during a discussion about the stick bug’s sexual identification. We organised the event in three interactional units (Bloome et al. 2008). Here we shall only present the first interactional unit,Footnote 5 to provide sufficient elements for understanding the process of analysis (for other related events, see Franco and Munford 2020).

We have selected this particular event as this is a scenario marked by greater uncertainty with regard to the sex of the stick bug in lesson 7 (22 November), different from previous and posterior events. Until then, most students agreed that the largest stick bug would be the male (8 to 19 November). At later events, most students agreed that the larger of the two stick bugs would be the female (from 26 November onwards). This change shows the relevance of this event for the group, especially in the discursive construction of science learning opportunities. In addition, disagreements and the different arguments became much more visible in students’ discourse at this event.

The teacher proposed students’ engagement in practices related to conceptual, epistemic and social domains within scientific knowledge (Duschl 2008). It relates to the conceptual domain because the concept of sexual dimorphism suggests there is no universal morphological standard that would establish who/what is male and female. For this reason, it would not be possible to categorise the largest as necessarily being the male, as had been the case in previous lessons. It relates to the epistemic domain because the instructional expectation in this event was that students used data that could sustain their conclusions as they established causal relationships. It relates to the social domain because there was a need to reach a consensus with grounds, while also considering the opinions of the peers.

Mauricio was moving away from this proposal. On looking at the terrarium, Mauricio was not identifying male and female, as he says (L23). He was in fact seeing one larger and one smaller animal. As the larger animal would [already] be the male; hence, the male would eat more. The confirmation of this fact backs up the idea but does not correspond to the cause. The deviation was made clear by the reaction of the teacher. She did not accept this argument without first trying to establish the relationships present in Mauricio’s talk (L16–22, L24–26). The teacher tried to shed light on possible connections that would be necessary for the student to acknowledge the proposal.

Breno, in turn, gave an alternative interpretation for the stick bugs’ eating habits. Mauricio’s observation (L5) was used considering a relationship between form and function in the insect (L41). The mother would need to eat more and not the male. This suggests a point of articulation between the conceptual, epistemic and social domains of scientific knowledge: based on the knowledge that it is the female that shall “have babies”, the student linked this piece of data to the conclusion, on articulating Mauricio’s contributions and defend his own point of view.

On the continuation of the event (interactional unit 2), the teacher highlighted the disagreement and, once again, asked the students for their opinions. Mauricio did not agree with Breno’s point. However, he did not present new arguments, and he talked in a way that indicated that obviously, the male was the biggest animal. To not acknowledge the possibility that the female is bigger than the male indicates distancing from instructional expectation in relation to the scientific concept under construction. In face of Mauricio’s reaction, Breno restated his position.

In the final part of the event (interactional unit 3), Mauricio stopped completely to provide arguments, and only emphasises that his idea was very important. The student moved away from social norms that helped to form the development, criticism and use of knowledge within the group. The student sought to devalue the opinion of his colleague, instead of pressing on with the work, based on arguments, giving due value to what peers have contributed. This example illustrates the analysis involved in the vertex of the hourglass.

Expansion: the Lower Part of the Hourglass

The lower part of the hourglass involves interpreting how the discursive practices in the science classroom analysed in the vertex phase were linked with other classroom events and translocal contexts. Following the hourglass metaphor, this means that the vertex grain of sand is not in itself analysed in this new phase. The analysis is based on what has passed and what is still to pass. The researcher can move forward and backward in time, seeking to understand historical relations (Green et al. 2013). In addition, expanding means interpreting the vertex interactions beyond the “classroom walls”, establishing relationships between science classroom discourse and translocal meanings from different sociocultural contexts. In short, expansion is an analysis of discursive interactions through intercontextual lenses.

An important aspect of the expansion is the choice of theoretical frameworks in order to analyse these intercontextual relationships. In the case used in this article, we turn to Butler’s discussions of gender, which gave us a framework based on which we can interpret these relationships. However, other elements of sociocultural backgrounds could have been the object of the hourglass vertex (e.g. religion, ethnicity, family experiences), and this would require the use of other theories and concepts. This indicates a dynamic and challenging aspect imposed by the ethnographic perspective. It was not our initial interest to discuss gender. The funnelling analysis indicated gender as a sociocultural element with important consequences for the discursive practices in this science classroom.

During the interactions on 22 November and beyond, the class accepted Breno’s arguments. However, one of his colleagues was not convinced. The vertex of the hourglass gave us some cues about what events should be resumed and how to pursue an understanding of participants’ meaning making that explains their positioning in the frame clash. Why does Mauricio not accept Breno’s argument?

Backward mapping analysis shows that the group had already experienced related practices throughout the school year and that Mauricio himself was involved in such practices. In the first semester, for example the group was studying growth of plants. On 25 June, Mauricio took part in the discussions about where the violets should be planted. Mauricio’s criterion to make his decision came out of a link between data, conclusion and prediction: the connection between sunlight and the survival of the plant gave grounds for making the decision about where the plant should be planted.

In the case of the stick bug, the problem was not just Mauricio’s limitations in engaging in argumentation. There was something else in play ever since the first contact of the students with the stick bug: translocal meanings of gender. At the anchor event, on 22 November, Mauricio gave an active character to the male, a sense of masculinity which, legitimated by the standards regarding gender (Butler 1993), took on the status of a biological characteristic. The male, a spender of intrinsic energy, “naturally” eats more and is bigger for this reason.

The backward/forward mapping suggests that Mauricio had taken on an important role in defending the view that the male from all the species would be larger, whenever there was any discussion about the sexual identification of insects. On 8 November, on seeing the different sizes of insects, there was a doubt: Who is who here? Mauricio and most of his colleagues had no doubts. The class upheld naturalised expressions of gender, stressing the normFootnote 6 (Butler 1993): a nuclear family following the model of a fertile heterosexual couple, with the hierarchy between father and mother being a guide for sexual categorisation for most of the children. Even so, in this same lesson, some children went against the norm. For example, Livia said that the larger insect would be the mother and the other two would be the offspring. Her view is that there was no father within the terrarium. Livia’s position, despite not having received support from the group, shows the possibility of doubt, something which started to recur during the events. This doubt is an example of established links between the “here and now” of group interactions and the translocal meanings of gender.

In Fig. 7, we represent the complexity of this process. The curves in the central line to the left suggest movements to reinforce the norm, while those to the right suggest contradiction/challenging the norm. The increase as observed in the occurrence of right-leaning curves over time represents growing tension in negotiations, generated by the progressive movement towards challenging the norm throughout the events. In Fig. 7, there are some quotations in brackets, selected as representatives for each of the events indicating a tension between these two movements of reinforcing/challenging the gender norm. In the corners on each side, we map relationships between each classroom event and different spatio-temporal dimensions. All these dimensions (e.g. naturalised notions of masculinity and femininity, data gathered in the classroom or at home, science textbooks, family experiences, collective memories, different data sources) constitute a social amalgam in which learning about sexual dimorphism took place.

Fig. 7
figure 7

Representation of the process to expand analysis of interactions

Full size image

In this broader scenario that considers a set of events, as well as the meanings of different dimensions, we rely on the notion of intercontextuality to interpret how Mauricio had difficulty in engaging in science classroom discursive practices that were coherent with what the group has constructed in a previous event (i.e. in the vertex). This seemed as if Mauricio’s answer had already been given: the largest one was the male. This evidences that it was difficult for the student to perceive the proposal made by the teacher or accept Breno’s argument. His commitment was with the norm and not with looking for the “best” answer in a science lesson. Therefore, the discussion about the sex of insects was much wider in scope, something that extended beyond the classroom walls. In this case, the gender norm was a meaningful sociocultural element closely related to engagement in practices from conceptual, epistemic and social domains of scientific knowledge.

Synthesis

In sum, the hourglass is a synthesis of criteria used in ethnography in education considering specific aspects of the science classroom. Figure 8 summarises the proposal. In the boxes, there are examples of theoretical and methodological tools that we used in each analytical phase. We are not arguing that these are the only tools one should use. However, it is important to note that in each hourglass phase, there is a concern with this articulation between ethnographic assumptions, discourse analysis and science classroom.

Fig. 8
figure 8

The hourglass approach

Full size image

Diversity and Equity Issues

Our analytical metaphor organises, in an instructive way, contributions from ethnographic traditions to the development of analysis concerned with science classroom diversity. The perspective of the participants can help us understand how science classroom discourse is constructed in classrooms with increasingly less predictable and stable communities. The hourglass also indicates how elements from different contexts are intertwined with science learning opportunities. The example we used to illustrate the proposal unfolds important aspects of this phenomenon: gender issues, a significant component of the participants’ sociocultural backgrounds, played an important role in their engagement in practices of conceptual, epistemic and social domains of scientific knowledge.

However, gender was not the basis from which the hourglass explored the nature of scientific discourse in the classroom. The hourglass focus is on intercontextual relationships. Gender gained visibility throughout our example because it had expressive consequences in the construction of science learning opportunities in the participants’ group. In other communities, the hourglass may indicate the role of other contextual elements depending on each sociocultural practice observed.

Similarly, in the final stages of the analysis, intercontextual relations also have a central role. Selecting events that have gender as an expressive contextual element did not mean analysing it from unidirectional models of discourse analysis (Bloome et al. 2008). Figure 7, for example is a zoom-in on the “expansion” of Fig. 8. Two central aspects are presented in Fig. 7: the relationships between events in the classroom and the different contextual elements that shape a social amalgam, from different spatio-temporal dimensions. In this way, gender, or any other contextual element, is understood in the light of diverse sociocultural repertoires converging and intersecting over science discursive practices.

The use of intercontextuality throughout the hourglass indicates possibilities of mediation so that elements of translocal meanings that permeate scientific knowledge are identified by the participants themselves, promoting equity in science education. Science teachers through intercontextual lenses promote a more complex view of science among students. This can support equity through a critical analysis of the ways of doing and speaking science in the classroom and even the use of scientific concepts themselves, generally taught as truth (Shea and Sandoval 2020).

In science education research, these relationships are normally established based on the assumption that there exists a context “external” to the classroom that goes into this space, making students replicate (or not) social hegemonic patterns (see Brickhouse 2011). The hourglass approach indicates possibilities of moving beyond the unidirectional contrast between what’s “inside” and what’s “outside” the science classroom to avoid the risk of developing limited interpretations of students’ and teachers’ agency in face of the “outside” world. Macro-contextual elements, actually, are “inside” the classroom, because classroom members constantly negotiate these elements while talking science.

Methodological Implications

The traditions we rely on provide some strategies to face practical challenges in working with video analysis that are related to the extensive volume of data and the complexity of event selection. Transcriptions at different levels (macro, meso, micro) are very useful. Word-by-word transcriptions are not always central to the analysis. The proposal is to identify a smaller number of events which evidence certain aspects of the social life of a group and which are significant for its members themselves. Only those events considered most significant are transcribed word-by-word.

The smaller number of events that are transcribed this way do not cast aside the importance of longer-term data collection. On the contrary, it would hardly be possible to obtain historical data able to evidence the role of one or few events in a classroom over a short period of data collection. We would hardly identify the frame clash analysed in this paper as something that had such relevant consequences in these conditions. So, our interpretations of micro-events were only possible with the macro- and meso-analysis over time.

However, acknowledging that longer periods of participant observation are key to analysis does not imply that the hourglass approach can only be adopted in research with 3 years or more years. There are examples of research guided by ethnography in education with shorter periods of time, such as 1 year or even less (see Kelly and Green 2019). What is the key in using this approach is to follow ethnographic principles to analyse everyday life in the science classroom, as discussed previously in this paper.

Data collection over time can also create practical problems and limitations. One of them refers to the massive investment of time in prolonged monitoring of classroom interactions. Video files and other artefacts generate large volumes of data and must be carefully organised and coded so that there is no risk of material loss or difficulties in their subsequent analysis. Another limitation is that funnelling process leads the researcher to a progressive selection of events that has implications. When making a selection in the database, the next level of analysis depends on the data selected at the previous level and so on. Thus, the analyses will always be limited by the selected data.

Finally, there are issues related to the concept of validity. Different from other approaches to the analyses of interaction carried out in science classrooms, in our approach, the validity is not based on the degree of compatibility in categorisations between different researchers. In ethnographic rationale, trustworthiness is constructed based on the transparency of the research processes. The hourglass approach follows this tradition by making visible processes that are different yet related: the details in the documentation of the types of records, description of how such records were transformed into data, the construction of visual representations of the relations between questions, data and analysis (Skuskauskaite 2019). Researchers that chose this kind of analysis must be aware of such challenges and differences.