Noticing student thinking under pressure in primary mathematics and science lessons

Teacher noticing of student thinking represents a key aspect of teacher expertise as it informs teachers’ actions ‘on the fly’ during a lesson. In science and mathematics education, teacher noticing has been conceptualised as being driven by attending to and making sense of student thinking. A range of models of noticing developed in professional fields such as aviation posit that perceptual interaction is also a key driver of noticing. This paper presents an alternate, ecological model of teacher noticing that begins with the premise that teachers have limited capacity to make sense of noteworthy events mid-lesson. Multimodal data collected from a primary science and mathematics lesson, taught by the same teacher, is analysed using the ecological lens. The analysis draws research attention to aspects of teacher/classroom interaction such as attention deployment and the role that classroom environments play in facilitating mid-lesson noticing of student thinking. Differences in teacher noticing between the science and mathematics lesson can be accounted for in terms of variance in environmental structure rather than solely in terms of teacher cognition.


Introduction
Classroom situations cannot be accurately anticipated and scripted in advance, as they tend to be determined by spontaneous events that can develop rapidly and unpredictably. Therefore, teachers must respond flexibly and adequately in-the-moment to promote student learning. Research has identified this ability as a core component of teacher expertise (Berliner, 2001). The construct of teacher noticing has emerged to represent this ability. Schoenfeld (2011) argued that teacher noticing matters because what is noticed impacts on subsequent pedagogical action. In research in mathematics and science education teacher noticing is typically characterised as a process of attending to and making sense of noteworthy classroom events mid-lesson (Chan et al., 2020;Dindyal et al., 2021;Jacobs et al., 2020;. The majority of this body of research has deep roots in a cognitive psychology tradition that focuses on teachers' knowledge base for teaching and the cognitive processes underlying their classroom practice (Santagata et al., 2021). While this approach to the study of teacher noticing remains fruitful, critics have argued that this approach creates a passive account of noticing (Sherin and Star, 2011). A growing body of literature on teacher noticing proposes that a richer account of noticing could be developed by taking into account ecological models of noticing (Jazby, 2019;Scheiner, 2020) and socio-political perspectives (Louie et al., 2021).
Research in other professional contexts has found that noticing and decision making under time pressure are less successful when time-consuming processes of sense making are required (Collyer & Malecki, 1998;Klein, 1997b;Randel et al., 1996). Researchers in the field of Human Factors have developed alternative models of noticing which view noticing under pressure as relying more on perceptual interaction rather than cognitive processes (Gray, 2006;Kirlik, 1995;Klein, 1997b;Smith & Hancock, 1995). This study presents a perception/action cycle model of teacher noticing (Jazby, 2019, which is grounded in the Gibsons' (Gibson, 1969;Gibson, 1979) theory of direct perception and ecological psychology and its application in a range of research fields (Araujo et al., 2006;Kirlik, 1995). The model enables analysis of teacher noticing mid-lesson, where there is limited time for teachers to make sense of noteworthy events. By characterising noticing as a process of seeking out environmental structures that can enable rapid pick up of information mid-lesson, an ecological analysis of teacher noticing shifts the primary focus of analysis away from teacher cognition, so that perception of structures in the teaching environment is given more weight.
The ecological model is applied to classroom data collected from two Year 5, primary lessons-one mathematics lesson and one science lesson-taught by the same teacher in a research classroom in Melbourne, Australia. A key research question which guides the paper is: Which aspects of teacher noticing can be investigated using a perception/action cycle model of noticing? More specifically, the way in which the teacher deploys attention mid-lesson and the environmental structures the teacher perceives as meaningful will be examined.

Conceptualisations of teacher noticing
A wide variety of conceptualisations of teacher noticing are present in the fields of science and mathematics education (Chan et al., 2020;Dindyal et al., 2021;Sherin and Star, 2011;. Much of this research can trace its roots back to Goodwin's (2000) notion of professional vision and Mason's (2002) work on noticing in classroom contexts (Dindyal et al., 2021). Conceptualisations of teacher noticing vary in terms of whether noticing is posited to entail attention, interpretation and/or subsequent action or decision making (Chan et al., 2020;. Many researchers have employed notions of attending to and making sense of as key components of the construct (Dindyal et al., 2021;Jacobs et al., 2020;Sherin & Star, 2011).
These models of teacher noticing posit a relationship between perception and action where a sensory input begins a cognitive process that results in an action or realisation. This approach parallels approaches to the study of perception and action that employ information-processing theory (Lipshitz et al., 2001;Massaro, 1990). Sensory data (information) must be made sense of (processed) so that people can respond to what is happening around them. Uncovering the cognitive processes that enable perception to lead to successful action is of research interest (Massaro, 1990). However, these cognitive processes are complex and may employ multiple subroutines which rely on empirically untestable entities such as mental schema and tacitly held information (Cooke, 1999;Lipshitz & Ben Shaul, 1997;Massaro, 1990). While studies rooted in cognitive psychology draw researchers' attention towards teacherlevel psychological constructs, they often do not give enough weight to factors outside of the teacher for explaining how noticing can occur rapidly under the pressure. Drawing on research from outside of the field of education, this study presents an ecological account of noticing which shifts research attention towards teaching environments so that teacher-level psychological constructs are not the only factors that contributing to teacher noticing.

Teacher noticing and cognition under pressure
Researchers who have studied noticing and decision making in naturistic settings, outside of laboratory conditions, have also cited potential short comings of an information-processing approach in dynamic, complex settings (Endsley, 2012;Klein, 1997b;Lipshitz & Ben Shaul, 1997;Lipshitz et al., 2001)-particularly when action must be taken under pressure. Tolcott (1992) summarised much of the research relevant to military contexts and concluded that noticing and decision making can be considered as being under pressure when a person has limited time, high workload, ambiguous and incomplete information and the presence of stressors such as noise and competing attentional demands. Under these conditions, 'naturalistic' models of noticing and decision making are required (Tolcott, 1992), and constructs of 'naturalistic decision making' (Lipshitz et al., 2001) and 'situation awareness' (Endsley, 2012) were developed throughout the 1990s. Unlike models that have been developed in laboratory experiments, naturalistic models are predicated on the assumption that, in the field and under pressure, time is limited and information is incomplete (Lipshitz et al., 2001).
When noticing occurs under pressure, task performance may be negatively impacted by conscious, sense-making processes (Randel et al., 1996). Fast-paced recognition of environmental cues that prompt workable courses of action has been noted as a hallmark of expert noticing in a range of professional fields (Klein, 1997a;Lipshitz et al., 2001). Randel et al. (1996) found that novice technicians in naval training exercises tended to weight up options and engage in deliberative cognitive processing, and this was associated with less success in exercises. Similar findings in a range of settings (Lipshitz et al., 2001) suggest that, with experience, task performance may become less cognitively demanding as there is less sense-making required, but still perceptually demanding as attention to key environmental cues enables rapid response to a changing environment. Gibson (1979) argued that behaviour can be better explained by analysing how a person perceptually interacts with their environment rather than by analysing their cognitive processes. According to Gibson (1979), the world serves as its own best model and perception of meaning is direct (Chemero, 2009). A teacher, entering a familiar environment, such as a classroom, will recognise familiar structures as having meaning directly. Perception of these meaningful entities also entails possible courses of action, so action can be taken rapidly in response to what is perceived, without engaging in time-consuming cognitive processing. Gibson's (1979) use of the term 'ecological' differs from the use of the term in a range of current theoretical frameworks that are used in education research. Chevallard (2019), for example, uses the term ecology to describe the wider conditions that may constrain the use of a mathematical practice (Barquero et al., 2013). Gibson's (1979) use of the term was less far reaching. In response to cognitive psychology which viewed behaviour as something that was guided by what happened in a person's head, and behaviourism, which viewed behaviour as being triggered by external stimulus, J. J. Gibson argued that behaviour was neither purely in the head nor purely triggered by the environment. Instead, there was an ecology in which the actor and the environment interacted (Chemero, 2009). This ecology does not include aspects such as larger cultural factors but does present an account of a person's psychological experience which is not purely based on what is happening within a person's body or mind.
The Gibsons' theory of direct perception has been used in a variety of fields, but most who use the theory posit that cognition also plays an important role in behaviour (Chemero, 2009;Gray, 2006;Kirlik, 1995). One approach has been to posit that some behaviour can be perceptually guided, while other behaviour is cognitively guided (Gray, 2006). Kirlik (1995) suggested that much behaviour can be explained as occurring as a rapid cycle of perception and action, but cognition can be drawn upon when perception/action solutions fail. Gray (2006) noted that while researchers have found merit in conceptualising in-the-moment noticing and decision making as being guided by an ecological and a cognitive system, evidence suggests that the interaction between each system may be more complex than Kirlik's model suggests. However, modelling skilled behaviour as cycles of perception/action has occurred in a range of fields (Araujo et al., 2006;Kirlik, 1995;Smith & Hancock, 1995), and this analysis has shifted researchers' attention to aspects of noticing that are not given significant weight in information-processing models (Araujo et al., 2006;Passos et al., 2008).
Drawing on analysis of noticing and decision making in a range of different sports, Araujo et al. (2006) have argued that noticing is both a technical and a tactical skill. Passos et al. (2008) argued that an ecological approach switches the unit of analysis to 'performer-environment interaction' so that the constraints and limitations that the environment places on performance are considered. In dynamic situations, a performer has limited control over events; hence, they need to be tactical and environmental factors may inhibit their capacity to enact successful task performance. When applied to teacher noticing, adopting a 'teacher-classroom interaction' unit of analysis shifts focus from noticing as a teacher action towards a phenomenon in which a teacher acts within environmental constraints (Jazby, 2019).
In summary, if teacher noticing mid-lesson is conceptualised as an in-the-moment phenomenon which is subject to time pressure, then attending to and making sense of noteworthy classroom events may give too much weight to cognitive processes as the proximal mechanisms that drive noticing. The ecological framework that guides the development of the perception/action cycle model of noticing presented in the next section views noticing as a product of 'teacher-classroom interaction' rather than an activity that is purely within the control of the teacher. This interaction is perceptually demanding, requiring limited attentional resources be tactically deployed in classrooms which James (1890) famously described as 'blooming, buzzing confusions'. With limited time available to make sense of environmental inputs, the model presumes that teachers will encounter familiar environmental structures which they will see as being meaningful, at a glance, with time-consuming conscious meaning making cognitive processes only required when perceptual information is insufficient.

Theoretical framework: a perception/action cycle model of noticing
Noticing in the blooming, buzzing confusion of the classroom is-at times-posited to be subject to the sources of pressure listed by Tolcott (1992): time limitations; high workload; incomplete and ambiguous sources of information; and potentially noise and competing attentional demands. The demands of teaching under pressure may limit the degree to which a teacher can utilise cognitively intensive information processing mid-lesson. Hence, the psychological experience of noticing student thinking in-the-moment may not involve any conscious experience of being able to stop and make sense of environmental inputs, which is the basis of most studies of teacher noticing (Cross Francis et al., 2021;Jacobs et al., 2010).
If a teacher cannot stop and make sense of environmental inputs mid-lesson, ecological psychology provides an alternate account of the psychological experience of teacher noticing. Rather than engaging in sense making, a teacher may notice and act by perceptually interacting with the classroom environment in a perception/action cycle. By tactically deploying their attention (Araujo et al., 2006) which, according to Gibson (1979), includes moving your body through the environment, the teacher can place themselves so that key environmental structures can be perceived. In order to illustrate how the psychological experience of noticing mid-lesson is argued to unfold, an example is presented and discussed as the elements of an ecological analysis of noticing are explained. Consider a teacher asking a student questions and the student responds. When some questions are answered, the teacher notices that the student lacks confidence while other questions are answered confidently. If a cognitivist model of noticing is employed to analyse this example, the student responses are an input, the teacher makes sense of these mentally and we may make inferences about the teacher's knowledge, thought processes, beliefs or emotions (Cross Francis et al., 2021;Jacobs et al., 2010). Gibson (1972Gibson ( /2002 argued that while sense-making processes and knowledge activation may occur neurologically as we watch and listen to a student give their answer, these are not the dominant features of the psychological experience of taking action in-the-moment. We are able to gauge a student's confidence in their reply by looking and listening to our environment, which contains the student. Hence, the key unit of analysis of teacher noticing is teacher-classroom interaction rather than the teacher's sense-making processes. If we can ascertain what a teacher was attending to mid-lesson, we can start to build an account of their experience if we can identify the environmental structures that were salient to them. Meaningful environmental structures (MES) are a key component of an ecological model and require that meaning is considered quite differently than in cognitivist models of noticing. MES are meaningful to the person who perceives them. Ecological psychology begins with an assumption that the psychological experience of perception is not devoid of meaning. When a person sees an object in their environment, such as a student's facial expression when asked a question, the person may perceive the student's expression and vocal tone as meaningful at a glance, without having to consciously experience a sense-making process. An experienced teacher who knows the student who is answering may attend to certain environmental structures-the slope of the student's mouth; an upwards inflection at the end of a sentence-and be able to pick-up whether the student is confident or not confident in their response. These structures are meaningful to that specific teacher at that specific time and are the proximal mechanism by which the teacher may pick up meaning. From the teacher's perspective, seeing a student's facial expression and hearing the student's tone as they respond enables the teacher to pick up information without engaging in what Chemero (2009) referred to as 'mental gymnastics'. Complex neurological processes may be occurring, but the teacher's psychological experience is posited to involve looking and listening for meaningful environmental structures which, when perceived, enables the teacher to pick up information that can they guide action. Environmental structures are not meaningful in their own right, but as a person perceptually interacts with environmental structures, they may perceive some structures as being meaningful. Multiple teachers might listen and watch the same student provide a reply to a question (hence they may all attend to the same environmental structure), but the meaning perceived via attention to the structure by each teacher could vary. Gibson (1969) argued that a key element of learning how to perform a task in a particular environment involves perceptual learning where, over time, a person becomes better able to attend to variations in environmental structure so that MES can be recognised more efficiently. A teacher who has taught the student who is responding to the question for a long period may be able to pick up on a particular aspect of the student's body language-a slight tilt of the head perhaps-and perception of this MES would lead to pick up of different information than a teacher who had not encountered the student before (Jazby, 2019).
Within the ecology of the teaching environment, only some environmental structures will be MES that enable pick up of information to guide action. By identifying which structures a teacher perceived as being MES within a teaching episode, an ecological analysis can develop an account of the teacher's psychological experience of the teaching episode (Jazby, 2019;Kirlik, 1995). While cognitivist accounts of the psychological experience of teacher noticing focus on teacher thinking, knowledge and goals (Cross Francis et al., 2021;Jacobs et al., 2010), identification of MES shifts some of the weight of analysis onto the environment in which noticing occurred. If a teacher noticed something noteworthy, then there had to be some element of the environmental structures that were attended to that made the noteworthy event noticeable. Noticing is not purely the product of the teacher's mental processes but is also dependant on there being sufficient structure in the environment that enables MES to be perceived. By analysing the MES that a teacher perceived mid-lesson, we may be able to identify which structures enabled a teacher to pick up information relating to student thinking and these structures can then be analysed in order to ascertain the degree to which a MES could enable information pick up.
Environmental entities that enable information pick up relating to student thinking require variance in their structure so that meaning can be perceived at a glance. In the example of a student answering a question, a teacher looking at the student as they answer would see a range of structures such as the position of the student's mouth, eyebrows, hair, clothing colour and posture. As the student answers questions, some of these structures may be invariant in that they don't change (e.g. clothing colour), or may change without providing information relating to student thinking (e.g. hair may move as the student replies). Other structures would be considered variant structures that could be used to pick up information relating to student thinking. For example, in Jazby's (2019) previous trial of ecological task analysis, teachers described posture and movements of a student's mouth and eyebrows as enabling them to pick up when a student was confused.
By comparing MES that are similar but enabled different information to be picked up, such as comparing a confused facial expression to a facial expression that was perceived as meaning that something was understood, the invariant and variant structural elements of each MES can be identified and described. If the environment does not contain variant environmental structure, then information pick up via attentional interaction may become limited or impossible. Hence a knowledgeable teacher, who is able to make sense of student responses (Jacobs et al., 2010) and who has exemplary goals mid-lesson (Schoenfeld, 2011), may still not be able to notice student thinking if the environment does not contain variance and invariance in environmental structure. By adopting an ecological framing, the environment is given a greater role in noticing mid-lesson.
In summary, the key unit of analysis in the perception/action cycle model of teacher noticing is teacher-classroom interaction. The position of the teacher in Fig. 1 highlights the focus on teacher action and perception, but this is considered in relation to the encompassing entity of the classroom environment. Within that environment, the teacher has limited perceptual resources that they can deploy. When the teacher enters the classroom environment, this contains environmental structures which the teacher may or may not find meaningful. The teacher cannot attend to all potentially meaningful sources of environmental structure and inevitably, the way in which the teacher positions themselves in the classroom will lead to the occlusion of some sources of environmental structure (represented by the grey, left-hand region in Fig. 1).
Teachers may deploy their attention tacitly, but how attention is deployed affects which regions of the classroom are perceptually accessible (the right-hand region in Fig. 1). Rather than waiting to encounter an event to attend to, employing an ecological framework assumes the teacher to be actively trying to locate structures that can enable information pick up. Hence, attention deployment will shift during the lesson as the teacher tries to locate information that can guide action. Teachers can then take action based on the information that has been picked up via attention to environmental structures. There is a range of actions that could be taken. For example, Fig. 1 shows two possible teacher actions that could continue a cycle of perception/ action. The first potential action that a teacher could take is that they shift position and deploy attention in order to locate further meaningful environmental structures. This could lead to further information pick up, and further teacher action that is perceptually guided. Another possibility is that a teacher could take action to change their environment by giving students directions, for example. Both actions-shifting attention deployment and giving students directions-may increase the likelihood that the teacher will encounter environmental structures that enable information pick up that can guide action with minimal need to interrupt this perception/action cycle with cognitive sense-making processes. Teacher action can be fluid as perceptual interaction guides behaviour as long as meaningful environmental structures are perceptually accessible to the teacher.
The teacher may also have a role in creating the conditions in which MES which contain variance that enables pick up of information relating to student thinking would be more likely to be produced. The teacher would also need to deploy their attention so that they are in the right position at the right time to encounter these structures mid-lesson, hence how a teacher deploys their attention would influence their capacity to notice student thinking mid-lesson. Given time pressure, the teacher needs to find a workable course of action quickly and as information is picked up, the teacher can take action-including actions which impact on attention deployment (moving to another part of the room, for example) and environmental structure (directing students to draw a picture, for example). Thus, perception guides action, and these actions affect perception and environmental structure so that new MES can be located, and further information is picked up in an ongoing cycle throughout the lesson.

Method
This paper reports a case study (Yin, 2014) of Ava, who taught two lessons, one focusing on mathematics and one on science. Ava is a generalist teacher who is teaching across all subject areas. Prior to her employment at a research partner school, she had been teaching Year 5/6 classes in another school for 5 years. The same group of 24 Year 5/6 students (10-12 years old) participated in both lessons as this was Ava's regular class.
Task selection for each lesson occurred via collaboration between the teacher and the research team. The teacher provided the research team with their planning documents for the upcoming 10-week teaching term. The research team then presented the teacher with suggested tasks that aligned with the teacher's term plan, while also being tasks that could be presented in one session in the research centre. The research team presented the teacher with tasks that matched her learning goals for term that also involved using materials that students could manipulate in a way that could be visible when filmed. The teacher then selected each task and the research team ensured that the materials required were available and tested whether the materials would be visible on-camera when recorded in the research centre. The plan for each lesson involved a lesson overview and suggested progression for the lesson, but there was no script for the teacher to follow, and the teacher was able to adapt the lesson plan as she saw fit. For the mathematics lesson, the painted cube task was selected. A version of the task developed by the reSolve: Maths by inquiry project (Australian Academy of Science, 2018) was used and this included materials and worksheets for student use. The task requires students to explore number patterns by examining the number of smaller cubes with varying number of faces covered in paints when different sized cubes are dipped in paint. This task focuses on developing algebraic reasoning through exploring a variety of number patterns that can be described spatially, numerically, or symbolically (see Appendix for task instructions). In science, the school was engaging in a chemistry unit which included investigating states of matter at the time of data collection. The science lesson involved predicting, observing and explaining what happened when icing sugar was mixed with vinegar in a bottle which had a balloon on top (the icing sugar would dissolve); then conducting the same experiment when the bottle was filled with vinegar and bicarb soda was added (a reaction would occur that would fill the balloon and bottle with gas). Students were given several resources, including a white board, markers, toothpicks and playdough, to express their ideas about these experiments.
The lessons and data collection took place at a custom-built facility of a research centre X at University Y with multiple cameras that enabled video and audio from each student group to be collected. Two tracking cameras followed Ava's movement during each lesson, and the students worked at six tables. Each student worktable was allocated a camera and microphone. This produced multiple, synchronised video records of each lesson. Ava participated in interviews held straight after each lesson, as well as a post-lesson discussion approximately one month after the lessons.

Method of data analysis
Multi-camera data collection enabled Ava's position in the classroom to be identified as even if she shifted her position away from one camera, she would move into shot for another camera. Her gaze direction could be ascertained by reviewing video footage from multiple sources. Knowing where Ava was looking at any given time is not sufficient data to then ascertain what she saw as being meaningful. Post-lesson interview data, both from the interview immediately after each lesson and in the later interview where vignettes selected by the research team were used as a recall prompt, were used to ascertain the meaning Ava ascribed to particular environmental structures she focused on. For example, camera data revealed a point in the science lesson where Ava focused her gaze on a student's playdough model. Using the video recording of this situation as a stimulus, Ava recalled that she could tell that the student was thinking about how there are tiny things moving in a liquid by looking at the student's playdough representation. Hence, the playdough model was coded as a MES, and Ava's description in the post-lesson interviews of what the environmental structure of the playdough meant to her was used to identify the meaning that she ascribed to the environmental structure.
Identification of MES required the coordination of video and post-lesson interview data as described above. Ava pointed to and described various environmental entities during the post-lesson interviews. Primarily, in both lessons, she pointed to and described student created representations as having meaning that related to student thinking. In the mathematics lesson, students worked with wooden cubes (multi-base arithmetic blocks) and a worksheet which contained a table (see Fig. 2), and a space to write patterns they had recognised as they worked through the task (see Fig. 3). In the science lessons, students either drew representations or used playdough and toothpicks to model representations (see Fig. 4). Referring to one of the student's drawings of a bottle shape with wavy lines in the science lesson, Ava remarked that the student was thinking about "the gas going up" when the bicarb soda was added to vinegar. This environmental structure was coded as being meaningful to Ava as she described it as having a meaning that was related to student thinking. It is not presumed that the student's thinking was as Ava inferred, as the analysis is concerned with developing an account of Ava's psychological experience during the lesson. Hence, the coding is designed to capture the meaning that Ava perceived mid-lesson, but her perception of meaning may be inaccurate. Having   Given the time constraints of post-lesson interviews, Ava was able to describe a few key episodes in enough detail for MES to be identified. Two episodes of noticing from each lesson could be analysed in detail as these episodes contained enough description of MES in the post-lesson interviews to enable analysis. Hence, these episodes are presented as vignettes in the findings section.
Information pick up at a glance is theorised to be facilitated by variance and invariance in environmental structure (Gibson, 1979). Hence, by comparing images of the environmental structures that Ava saw as being meaningful, identification of differences between each student representation was used to identify what was similar (or invariant) or different (or variant) about each structure. Figure 2 provides an example from the maths lesson. Ava perceived the left image in Fig. 2 as meaning that the students 'didn't really get it', while the image on the right meant that students had 'got it'. Invariant elements of both environmental structures were then identified by the research team: both structures are tables which are organised the same way and parts of each table have had numbers entered into them. What varies between each environmental structure is the numbers that have been entered into the table, the number of cells that have been filled in and the number of entries that have been crossed out by students. Hence, these variant elements were coded as the elements of these MES that enabled different meaning to be picked up at a glance by Ava. In the science lesson, student representations (either drawn or made of playdough and tooth picks) were compared by the research team so that the variant and invariant elements of structures that Ava perceived as having different meanings could be identified. Noticing requires that environmental structures contain variance and invariance if meaning is to be picked up at a glance. Hence the research team also evaluated the degree to which MES provided Ava with sources of environmental structure that enabled the pick up of information relating to student thinking.

Deployment of attention
Ava deployed her attention in similar ways both lessons. She walked between tables and would stop next to one table. In most cases she would look at what students were doing or tilt her head so that she could see the representations that students were making. She would then start talking with the students. When Ava focused on a specific table group, she turned her back on other tables so that what was occurring outside of her focus group was occluded from her view. In both lessons, Ava engaged in a total of 37 interactions with a single table group during the body of the lesson. These interactions ranged from 4 to 97 s in the science lesson, and 5 to 172 s in the mathematics lesson. The average length of an interaction was slightly longer in the mathematics lesson (43 s) compared to the science lesson (34 s). Overall, the pattern of attention deployment in both lessons was remarkably similar.
We will now present examples of information pick up by Ava during the two lessons, related to two vignettes in each of these lessons, followed by an analysis.

Identification of MES and information pick up in mathematics vignette 1
Ava could pick up mid-lesson that the pair of students she interacted with in vignette 1 were getting incorrect answers to the problem. She walked over to the pair and quietly looked at their worksheet for approximately 3 s before saying, "you guys are so confused, aren't you?". This suggests that the students' filled in worksheet was the source of environmental structure that enabled her pick-up of this information. She then explained to students how to correct their work. Approximately three minutes later, both students exclaimed, "Oh. I get it now" and looked upwards and smiled then stopped her explanation and began to move away from the pair. This suggests that students' facial expressions and utterances enabled the pick-up of the information that students understood the teacher's explanation. Table 1 provides a summary of the information picked up during this cycle of interaction and the sources of environmental structure that were perceived by Ava as MES.

Identification of MES and information pick up in mathematics vignette 2
The second vignette was chosen by Ava in the post-lesson interview as a point in the lesson which was noteworthy. In the vignette, two students are explaining their thinking to the class. Ava engaged in a cycle of interaction with the pair that consisted of four interactions during the body of the lesson, prior to selecting the students to lead the whole-class discussion. In each interaction, she moved behind the students and watched what they were doing for an average of 12 s before she spoke to them. During this time, her gaze was fixed on the worksheet the students were working on, suggesting that this was the environmental structure that she was attending to. In three of the four interactions (in one interaction she just watched the students), she asked a question after watching the students filling out the worksheet. This led to discussion between herself and the students. She also pointed to cells of the table on the worksheet and moved cubes on the desk.
Data from both post-lesson interviews were used to ascertain what Ava picked up via this interaction. In relation to these two students, she stated in the interview that, "[she] was impressed with their reasoning" and she commented on their ability to find a pattern. She also inferred that the students were using a formula for some parts of the problem but counting strategies for others. Also, a miscalculation prevented the pair from recognising one of the patterns. Table 2 shows what environmental structure was attended to in relation to each piece of information picked up midlesson in relation to the two students that were the focus of vignette 2.

Analysis of environmental structures-mathematics lesson
Ava used student worksheets, student facial expressions, student utterances and student manipulation of materials to pick up information mid-lesson. In the mathematics lesson, she attended to student worksheets in particular. Figure 2 shows two different student worksheets that Ava attended to. The main sources of variation between the two representations were whether the correct numbers had been entered into the table and the degree to which answers had been crossed out by the students. There is no variation in the structures that could enable the teacher to pick up information relating to the mathematical thinking that was employed to arrive at the response written in the table. Hence, the structure of the table cannot enable pick up of information relating to student thinking beyond whether a student has the correct numerical value in each cell and the number of cells filled out at different points in the lesson could enable the teacher to be aware of the speed with which each group was completing the task. After completing the table, students were instructed to describe any patterns they saw in the painted cube task. Figure 3 shows different student responses to this instruction. The image on the left was common in that most student worksheets provided limited information regarding patterns that students had employed mid-lesson. While the second page of the worksheet (see Fig. 3) prompted students to search for patterns based on the table, most students did not reach this point. The image on the right shows what was written by the pair of students who featured in mathematics clip 2. Unlike most of the environmental structures that were produced by students in response to the instruction, the students featured in vignette 2 provided some explanation for the patterns that they had discovered and included a diagram. Compared to most of the class, the sheer volume of writing in this section is a key source of variation in structure. This pair was then chosen by Ava to explain what they had done in the end of lesson whole-class discussion. In the post-lesson interview, Ava remarked that she was impressed with this pair's reasoning. Ava's response suggests Table 2 Information pick up in mathematics vignette 2 Information picked up mid-lesson Source of environmental structure that enabled pick up Student reasoning is good; the student has used patterns for some parts of the problem Correct responses on student worksheet; student articulates a pattern in conversation with teacher Student has made a miscalculation Incorrect response is written on worksheet; student tells class about a doubling pattern which is incorrect The student was using a formula for initial problems Correct response on worksheet The student was using a counting-based strategy for later problems Student manipulation of the cubes provided that the environmental structures produced in the right-hand image enabled her to pick up more information about student thinking mid-lesson.

Identification of MES and information pick up in science vignette 1
Prior to selecting the first pair of students to present to the class in the science lesson, Ava interacted with the pair three times. Each interaction began with an average of seven seconds of watching what the students were doing. In this period, Ava's gaze shifted between the representation that one student was making with playdough and the whiteboard the other student was writing on. This suggests that each of these environmental structures were potential sources of information pick up. Ava then asked the students to explain what their playdough model represented, and she paid particular attention to words like 'thinner' and 'moving' as evidenced by her recall during the post-lesson interview that, "I was like 'thinner'; good word", and "[one student] picks up that the liquid … they're moving all the time". Hence, she was able to infer the student was thinking about viscosity and the movement of matter in a liquid. Table 3 summarises the connection between environmental structures and information pick up in relation to this student pairing.

Identification of MES and information pick up in science vignette 2
Prior to selecting the second pair of students to present in the science lesson, Ava interacted with the pair 8 times in the lesson body. Two of these interactions were initiated by the students asking the teacher to come over so that they could explain what was happening in the experiment to her. In one interaction, Ava happened to be passing by the pair when one student said, "it's all dissolved into little tiny particles". Ava changed the direction she was walking in as the student said this and asked the student to explain what they'd meant. In the remaining 6 interactions, Ava began each interaction by watching the students for an average of 4 s. Her gaze moved between the playdough representations the students had made, what they were writing on their whiteboard and also watching the students conducting the experiment. From these interactions, Ava was able to pick up that these students recognised that gas was produced and rose into the balloon when bicarb was mixed with vinegar, and that they thought that icing sugar had dissolved when mixed with vinegar. Ava stated in her interview that these students' thinking was "more shallow" because "they were just staying at a surface level" without trying to explain what was happening (Table 4).

Analysis of environmental structures-science lesson
Student created representations of what occurred during the science experiments varied in their structure in terms of whether they were micro-or macro-representations, referring to particle models or real-world phenomena, respectively (De Jong et al., 2005). The left image in Fig. 4 shows a typical macro representation. This representation shows three stages to the first experiment (mixing icing sugar and vinegar) and the drawings show what was visible during the experiment. In contrast, the image on the right shows a micro-representation in which students attempted to represent what was happening at the level of (invisible) particles during this experiment. There is no bottle in the micro-representation. In the post-lesson interview, Ava commented that the students who created the micro-image in Fig. 4 were thinking about viscosity and the movement of particles in a liquid. During the lesson, she had asked the students to explain their representation and the students had explained that the balls represent atoms. She then asked why they were joined together by the toothpicks and the students replied that they weren't joined, but they needed space between each ball because they were moving. Ava then asked whether the second experiment (i.e. the mixing of bicarb and vinegar) would have the same representation and the student stated that it won't be different because it would be 'thinner'. Hence, while perception of the student created representation did not enable Ava to pick up what students were thinking at a glance, questioning the students about the representation led the students to mention 'movement' and a 'thinner' solution and these utterances enabled information pick up.
The macro-representation shown on the left of Fig. 4 was associated with what Ava referred to as "more shallow thinking". Attention to this structure enabled Ava to pick up that students were describing what happened in the experiments, but they were not explaining why it happened. Ava also asked students to explain what their macro-representations meant and noted that students did not use as precise language to describe what happened as children who had created micro-representations. Using the representation as a basis for discussion, Ava was able to pick up that students were thinking about a gas being produced (i.e. experiment 2) or a solid being dissolved (i.e. experiment 1).

Discussion
This study sought to ascertain which aspects of teacher noticing could be investigated when an ecological perception/action cycle model was used to frame teaching interactions. Through this lens, teacher noticing under pressure is not characterised by sense-making mental processes (Jacobs et al., 2010) as there is little time for cognitively intensive analysis mid-lesson (Lipshitz et al., 2001;Salmon et al., 2008;Tolcott, 1992). Instead, teacher noting entails fast-paced recognition of environmental cues so the psychological experience of teaching entails deploying attention tactically so that you can 'see' what to do. By interacting with environmental structures, structures which are perceived as being meaningful (MES) may be encountered, and attention to these structures enables the teacher to pick up information to guide action. Hence, teacher noticing is not purely supported by teacher-level factors such as knowledge, goals and orientations. The environment also needs to contain structures that can enable pick up of information relating to student thinking and teachers need to deploy their attention so that they're in the right place at the right time to perceive these (Jazby, 2021). Ava employed similar patterns of attention deployment in both lessons and did not significantly vary where her attention was directed when the content area of each lesson was different. The similarity in the way in which Ava deployed her attention in each lesson suggests a pattern of behaviour that is specific to teaching. Ava 'checked in' with every table group as the students worked. When checking in, groups outside of her immediate vicinity were occluded from her view. This analysis highlights the attentional limitations of working in classroom environments which contain more potential sources of MES than can be attended to by an individual teacher.
The degree to which Ava noticed student thinking varied between each lesson. In the mathematics lesson, she mainly noticed whether students were correct or incorrect with pick up of information regarding the thinking strategies that students may have used. In contrast, she noticed particular science concepts students were employing in the science lesson. When noticing is viewed as a perception/action cycle these differences may be driven by environmental structure. In the mathematics lesson, students worked through a worksheet where they entered numbers into a table and wrote sentences which described patterns they had noticed. Analysis of these structures demonstrates that they are limited in enabling information pickup relating to student thinking beyond whether a student is correct or incorrect and the pace that students are moving through the task. In contrast, the student representations that were produced during the science lesson enabled pick up of information relating to student thinking about the behaviour of matter, at a glance, due to variation in structure between student representations.
This ecological account of teacher noticing gives more weight to environmental structure as a proximal mechanism that influences what a teacher can notice than when noticing is characterised as cognitive processing (Jacobs et al., 2010). The key structures that Ava attended to included student-created representations that students created in response to the instructions they were given. Hence, noticing could be affected by changing the instructions for each task. Jazby and Widjaja (2019) suggested an alternative design to the painted cube task that could increase the likelihood that students would create representations that support the pick-up of information relating to student thinking. While task design has been subject to a wide range of research approaches, Sherin and Star (2011) noted that models of teacher noticing tended to conceptualise noticing as occurring once a noteworthy event was encountered by the teacher rather than starting when a task is designed. Choy (2016) argued that noticing of student thinking is influenced by teacher planning that occurs prior to the lesson. The perception/action cycle model of noticing presented in this study provides an analytical framework that creates an inherent link between task design and noticing. Task design which supports information pick up that relates to student thinking is posited to facilitate improvements in teacher noticing. Hence the perception/action cycle model presented strengthens the construct of teacher noticing by providing a clearer connection between task design and mid-lesson noticing. By analysing the degree to which environmental structures created by students facilitate information pick-up, structures that support information pick up could be explicitly built into tasks so that noticing of student thinking becomes more obvious.

Conclusion
When viewed through an ecological lens, noticing student thinking mid-lesson is subject to time pressure in which a teacher has limited recourse to employ timeconsuming cognitive processing. Instead, perceptual interaction can be used to guide action quickly, so that workable inferences and decisions can be made in the moment. Ava is posited to be engaged in a perception/action cycle which entails minimal mid-lesson sense making. When viewed as a perception/action cycle, noticing is attentionally demanding as it requires teachers to strategically deploy their attention so that they are in the right place at the right time to pick up information that can guide action. This ecological analysis of Ava's noticing produces an account that differs from cognitivist accounts of teacher noticing such as Jacobs et al. (2010), Blömeke et al. (2015) and Kaiser et al. (2017) which posit that noticing is primarily cognitively demanding (Scheiner, 2020). It is likely that noticing student thinking mid-lesson entails both attentional and cognitive demands, but most approaches to studying teacher noticing primarily focus on cognitive processing (Santagata et al., 2021;Scheiner, 2020;Sherin & Star, 2011). The perception/action cycle model presented in this study enables the attentional demands of noticing to be analysed so that the way in which teachers deploy their attention and the environmental structures they attend to can be identified and enrich analysis of noticing. Provided that there are sufficient MES present in the classroom, the teacher could conduct a lesson by employing a perception/action cycle, but they could also employ more demanding cognitive processing if required. Should environmental pressures ease (which frees up limited resources for cognitively intensive processing) or novel environmental structures be encountered (which do not provide sufficient information pick up), the teacher might have need of sense-making processes. Hence, while the perception/action cycle model shifts the weight of analysis of teacher noticing away from primarily focusing on attending to and making sense of, it is designed to complement an information-processing approach as some instances of noticing may require sense making and not be perceptually guided (Gray, 2006).
The small-scale trail of the data collection technique and analytical framework presented in this study suggests that analysis of perceptual interaction can produce an account of teacher noticing under pressure that addresses elements of noticing that are absent from the accounts of noticing informed by the work of Goodwin (2000) and Mason (2002). The scale of the present study is not sufficient to determine whether analysis of attention deployment, MES, information pick up and variance in environmental structure could produce generalisable results that can help develop our understanding of teacher noticing.
The ecological analysis presented in this study explores an approach to analysing teacher noticing that shifts beyond solely relying on a process of a teacher attending to and making sense of noteworthy events. Improving teacher capacity to attend to and make sense of noteworthy events could assist teachers when noticing occurs with less pressure, such as when reviewing classroom events (Kilic, 2018;Luna & Sherin, 2017). It is hoped that this account enables researchers to investigate teacher noticing under pressure by bringing perspectives on noticing that have been developed in professions other than teaching.