Teaching is a complex activity. Teachers typically act under time pressure to make instructional decisions in often ambiguous situations, for example, reacting quickly to student misbehavior (Wahl, 1991). Thus, important aspects of a teacher’s expertise are knowing what to be sensitive for in the classroom and how to interpret information, as well as to make instructional decisions based on these interpretations swiftly (Borko et al., 2011; Gaudin & Chaliès, 2015; Jacobs et al., 2010; Kersting et al., 2012; van Es & Sherin, 2008). These instructional decisions include decisions about features of a lesson that should be maintained and features that should be changed in the future, that is, generating teaching alternatives (van Es & Sherin, 2002). Researchers refer to these competencies also as teachers’ professional vision (Goodwin, 1994; Seidel & Stürmer, 2014).

Acquiring professional vision is critical for novice teachers. One of the main concerns of novice teachers is whether they will be able to monitor and manage a classroom (Sadler, 2006). Research shows that teachers’ professional vision can be examined and developed through studying video records of teachers’ own classroom teaching (Kleinknecht & Schneider, 2013; Seidel et al., 2011). There has also been increasing research on the role of the video perspective in teachers’ learning and reflection processes related to professional vision (e.g., Cortina et al., 2018; Theelen et al., 2019). Recent research indicates that mobile eye-tracking (MET) video footage might be a valuable tool for research on teachers’ professional vision in action as well as for teacher training and development, but has yet been scarcely employed (e.g., Cortina et al., 2018).

The goal of the present study was to use MET video footage in combination with think-a-loud protocols to examine differences in expert and novice teachers’ reflection of their regular classroom teaching. Understanding how expert and novice teachers differ in their noticing of classroom events and how they reason about alternatives using this novel technology could be instrumental for the design of teacher education and professional development.

Teachers’ professional vision: Noticing and the generation of alternative teaching strategies

The concept of professional vision was first developed by Charles Goodwin and describes “socially organized ways of seeing and understanding events that are answerable to the distinctive interests of a particular social group” (Goodwin, 1994, p. 606). In research on teaching and teacher education, professional vision has been defined as consisting of two distinct subprocesses: selective attention or noticing and knowledge-based reasoning (Sherin, 2007). Noticing refers to the ability of teachers to direct their attention to relevant events in the classroom (Sherin, 2007; van Es & Sherin, 2002) and is considered a prerequisite for the ability to act adaptively in classroom situations (Berliner, 2001; Kersting et al., 2012; Sherin & van Es, 2009; Weber et al., 2018). The way teachers direct their attention affects their reasoning, which in turn influences further noticing processes (Bromme, 1992; Endsley, 1995; van Es & Sherin, 2002).

Knowledge-based reasoning describes teachers’ cognitive processing of instructional events based on their knowledge about teaching and learning (Borko, 2004; Seidel et al., 2013; Sherin, 2007; van Es & Sherin, 2002). Knowledge-based reasoning comprises the description, explanation, and prediction of classroom situations (Seidel & Stürmer, 2014). In the process of knowledge-based reasoning, teachers draw on their knowledge on the subject matter and the classroom context to understand students’ actions and reactions and connect what happens in the classroom with broader principles of teaching and learning (Sherin, 2007; van Es & Sherin, 2002).

Generating and weighing alternative teaching strategies is one important aspect of teachers’ knowledge-based reasoning (Schwindt, 2008; Seidel et al., 2011). It is assumed that if teachers apply their knowledge, for example, when watching videos of their own or others’ instruction, they come up with and weigh alternative teaching strategies in response to the instruction they observed (e.g., Kersting et al., 2010, 2012). These alternative teaching strategies can be expressed as suggestions for improving instruction, but also as explanations of why the chosen teaching strategy has advantages over other strategies mentioned (Santagata et al., 2007).

Generating and weighing alternative teaching strategies is considered particularly important for effective teaching (Kersting et al., 2012; Kleinknecht & Gröschner, 2016; Santagata & Guarino, 2011). For example, there is evidence that generating alternative teaching strategies is positively related to students’ learning outcomes. Kersting et al. (2010) showed that more effective teachers in terms of student learning gains over a school year are better at generating alternative teaching strategies than less effective teachers. To be able to weigh alternative teaching strategies, teachers first need to identify a critical classroom event. In a second step, they need to access their knowledge of teaching strategies to infer that a different strategy would be advantageous or disadvantageous (Santagata & Guarino, 2011). In the present study, we focus on these two important aspects of teachers’ professional vision: noticing of classroom events and weighing alternative teaching strategies.

Professional vision and expertise

A widely used tool for studying teachers’ professional vision is classroom video (Brophy, 2004; Gaudin & Chaliès, 2015; Goldman et al., 2007). Research on video clubs in teacher professional development––initially stimulated by Sherin and van Es (2005)––has shown that teachers can develop their teaching competencies and professional vision by continuously and collaboratively studying video records of their own or other teachers’ instructional videos (e.g., Kleinknecht & Schneider, 2013; Seidel et al., 2011). It is assumed that teachers improve their awareness of students’ learning processes during classroom instruction through noticing relevant events in the video footage and reasoning about them (Sherin & van Es, 2005).

Research that used teachers’ own or other teachers’ teaching videos also provided evidence for differences in expert and novice teachers’ professional vision. Seidel and Prenzel (2008) emphasize that expert teachers (with mostly more than five years of teaching experience) outperformed students in teacher training in their knowledge-based reasoning when viewing authentic classroom videos in standardized assessments. Copeland et al. (1994) let expert and novice teachers watch a video that showed typical interactions between a teacher and students. Experienced teachers, in contrast to novice teachers, saw connections between classroom events and suggested alternative courses of action. In this study, expert teachers were identified based on their extensive teaching experience, their membership in a professional group, and their performance, whereas novice teachers were inexperienced and experienced students in teacher training programs, that is, undergraduate students enrolled in a “preprofessional” teacher program and postgraduate students at the end of a one-year elementary credential program about to start their careers as teachers. In a more recent study by Schäfer and Seidel (2015), participants watched a video in which two pedagogical strategies that were relevant for student learning could be noticed and reasoned on. The results indicated that pre-service teachers struggled to link a noticed classroom event to general pedagogical concepts and their reasoning regarding those classroom events hardly matched experts’ interpretation of the same events. The experts in this study were professionals in the field of teaching and learning with 5 to 10 years of experience in teacher education and systematic classroom observation.

In addition, research shows that expert teachers (i.e., teachers with at least 5 years of teaching experience) seem to construct richer and more meaningful representations of classroom events and focus more strongly on students’ learning than novice teachers do (i.e., students in teacher training or first year classroom teachers; Sabers et al., 1991; Wolff et al., 2015). This was evident, for example, in the fact that expert teachers, as opposed to novice teachers, were able to make predictions about classroom management events and proposed more alternative teaching strategies after viewing several teaching video fragments of a novice teacher, indicating richer knowledge and a deeper understanding of what is happening in the classroom (Wolff et al., 2015).

These differences between expert and novice teachers have been attributed to expert teachers’ more elaborated and coherently organized knowledge structures (Borko & Livingston, 1989; Colton & Sparks-Langer, 1993; Leinhardt & Greeno, 1986; Pauli & Reusser, 2003; Putnam, 1987). It is assumed that these so called “schemata” or “curriculum scripts” enable teachers to rapidly identify meaningful patterns in the classroom and to make informed and flexible instructional decisions (Colton & Sparks-Langer, 1993; Putnam, 1987; see also Lachner et al., 2016 and Wolff et al., 2021).

The ability to notice relevant events (i.e., professional vision) is directly linked to a person’s visual perception. Research has shown that expert and novice teachers differ in their ability to notice relevant events in the classroom (Miller, 2011; Berliner, 2001; Ericsson, 2018; Gegenfurtner et al., 2011). Cognitive psychology uses phenomena such as situational awareness (Endsley, 1995) and cognitive tunneling (Dirkin, 1983) to describe the difference in perception of experts and novices. Situational awareness is defined as comprising three factors: (a) perception of meaningful elements in an environment, (b) comprehension of their meaning, and (c) projection of their status in the near future (Endsley, 1995; see also Miller, 2011). Situational awareness in the classroom could imply, for example, that the teacher notices that a group of students are chatting in the back of the class (perception), realizes that they are not paying attention to the lesson, which disturbs the lesson and could reduce the students’ learning (comprehension), and decides to change his or her position and stand close to the group to stop the distraction (projection). In contrast, cognitive tunneling refers to novices narrowing their attentional field when performing a complex task (Dirkin, 1983). In the classroom context, this could happen, for example, when a novice teacher gives feedback to a particular student and stops paying attention to the rest of the class. Overall, it is assumed that the viewing patterns of experts are attuned to the demands of the situation and thus help to maintain situational awareness, whereas the viewing patterns of novices show less optimal selection from the complexity of the perceptual world (Miller, 2011).

Eye-tracking technology has been used for quite some time to study the visual behavior of experts and novices in a variety of fields (Reingold & Sheridan, 2011) and has more recently been discovered for teaching as well (Beach & McConnel, 2019; Jarodzka et al., 2017). The major advantage of eye-tracking technology is that it can be used to shed light on humans’ initial cognitive information processing (i.e., visual intake, integration, and active search for information; Jarodzka et al., 2021).

Eye-tracking studies indicated that there is evidence of expertise in teachers’ visual behavior. For example, studies showed that expert teachers (i.e., recognized teachers with more than seven years of teaching experience) fixated on more areas, revisited them, and had more fixations on areas where relevant information was available (i.e., areas showing students and classroom activity) than novice teachers (i.e., students in teacher training), whereas novice teachers tended to skip areas in their visual field (Wolff et al., 2016). Furthermore, students in teacher training failed more often to identify relevant classroom events in standardized video sequences and consequently missed the cues for teacher action, whereas recognized teachers with at least 10 years of teaching experience paid attention to the event while simultaneously maintaining a broad overview of the classroom (van den Bogert et al., 2014; see also Wyss et al., 2021). In addition, expert teachers needed less time to comprehend the event as they had shorter fixation durations than novice teachers when a classroom event occurred (van den Bogert et al., 2014). Using MET technology, Cortina et al., (2015) found that first year classroom teachers who were rated as being highly responsive to individual students were less able to divide their attention evenly across the whole class, whereas experienced teachers did not show this trade-off: They managed to be both responsive to individual students and attentive to the whole class.

However, an eye-tracking study that examined the professional vision of experienced physical education teachers with at least five years of teaching experience compared to that of first- and last-year students in teacher training found that the professional vision of experienced teachers and last-year students was equally developed and only less developed in first-year students. Commenting on 12 slides from a gymnastics lesson after viewing them for a few seconds, the three groups did not differ in the number of events reported or in the number and duration of the fixations. But last-year students and experienced teachers correctly reported more critical events (e.g., student misbehavior or the teacher giving a poor demonstration) than first-year students did (Behets, 1996).

In sum, eye-tracking and video studies in the field of expertise research provide evidence that experts rather than novices are able to rapidly perceive relevant classroom events and ignore unimportant events, while at the same time paying attention to the whole classroom. This is also reflected in their eye movements. Further, experts are better able to link classroom events to their experience and knowledge, are more likely to suggest alternative teaching strategies when reflecting on lessons, and to make predictions about subsequent classroom events.

Importantly, although MET video provides the opportunity to evaluate teachers’ professional vision in their gaze patterns, the present study focuses on analyzing teachers’ professional vision expressed in their reflections on their MET videos rather than in their eye-tracking metrics. Eye-tracking is used in this study to illustrate the eye movements of experts and novices in their teaching videos.

Effects of different types of classroom videos on the assessment of teachers’ professional vision

Different types of classroom video are used to investigate teachers’ professional vision, differing in the perspective (i.e., filming from the student perspective or the teacher perspective; e.g., Borko et al., 2008; Snoeyink, 2010), agency (i.e., filming the teacher’s own lesson or the lesson of another teacher; e.g., Gold et al., 2020; Krammer et al., 2006; Zhang et al., 2011), as well as in the technology used (e.g., stationary video, 360° cameras with virtual reality, or MET; e.g., Cortina et al., 2018; Theelen et al., 2019).

The use of teachers’ own versus others’ classroom videos

Whether to use teachers’ own teaching videos or recordings of other teachers depends on the learning goals (Blomberg et al., 2014). A video of other teachers is considered helpful for developing a critical stance toward instruction, whereas self-recordings give teachers the opportunity to critically reflect and hone their own skill sets (Gaudin & Chaliès, 2015; Seidel et al., 2011). Empirical evidence supports the notion that self-video compared to other-teacher video increases pre-service and in-service teachers’ sense of authenticity in the professional learning process that is perceived as motivating (Borko et al., 2008; Brouwer, 2012; Kleinknecht & Schneider, 2013; Rosaen et al., 2008; Seidel et al., 2011).

However, studies also demonstrated that in-service teachers tend to comment less critically, reflect on events less profoundly, and identify fewer alternative teaching strategies when watching stationary videos of their own teaching than when watching other-video (Kleinknecht & Poschinski, 2014; Kleinknecht & Schneider, 2013; Seidel et al., 2011). Seidel et al. (2011) illustrate that teachers who watched their own teaching might be more prone to apply self-protection mechanisms that impair their critical reflection and articulation than teachers who watch other-video. In addition, Kleinknecht and Schneider (2013) suggested that teachers who reflect on their own videos may be less able to develop alternative teaching strategies than teachers who reflect on videos of other teachers because they are too accustomed to their own teaching practices. Furthermore, two studies reported that teachers who saw and reflected on their own teaching practices reported less negative (i.e., disappointment with teaching actions) and more positive emotions (e.g., enjoyment, interest, well-being) than teachers who saw and reflected on other-video (Kleinknecht & Poschinski, 2014; Kleinknecht & Schneider, 2013).

The researchers reasoned that teachers watching their own video might have perceived a discrepancy between their actual teaching practice and their teaching self-concept. The perceived discrepancy might have threatened their teacher identity and induced shame or guilt (see also Tracy & Robins, 2004). As a consequence, teachers could have been less open about their emotions to protect their self-worth (Kleinknecht & Poschinski, 2014; Kleinknecht & Schneider, 2013). To conclude, previous studies indicated that reflecting on their own videos potentially impairs teachers’ reflexive learning processes in video-based interventions. However, these studies did not examine the effect of teacher expertise.

Advantages of MET versus other video technology

MET technology could be a viable solution to this problem as it can be used to assess aspects of pre-service teachers’ and in-service teachers’ perception and cognition that are not accessible to conscious thought (Gaudin & Chaliès, 2015). MET measures what direction teachers look in the classroom and what their gaze is focused on. Teachers usually wear the measuring device (i.e., the eye-tracker) as a pair of glasses which allows them to move freely through the classroom. Video footage of the precise focus of a teacher’s attention make studying teachers’ situational awareness and thought processes during teaching very feasible (Cortina et al., 2015, 2018), which is often difficult in other video formats (e.g., stationary video footage; Brophy, 2004; Shepherd & Hannafin, 2008).

In contrast to the video formats predominantly used in teacher education and teacher professional development, MET video footage has also the advantage that it rarely shows the teachers themselves. MET hence allows teachers to “re-live” their teaching from their own viewpoint instead from the perspective of a classroom observer (Stigler & Miller, 2018). For example, there is evidence that viewing their own MET videos after instruction facilitates a shift in perspective such that teachers focused substantially more on the students and less on themselves in spontaneous think-aloud commentaries (Cortina et al., 2018).

Today’s MET technology has evolved over the last 20 years to the point that the teacher now can move around fully unencumbered by the equipment; even the students quickly ignore the device, as it looks very similar to regular glasses (Stigler & Miller, 2018). According to participating teachers, the impairment of the lessons due to MET technology is minimal (Miller, 2011). Thus, MET has the potential to provide insight into teachers’ professional vision in action (Cortina et al., 2015). However, there are only a few studies that have used MET to study professional vision in teacher education and teacher professional development (see for example, Cortina et al., 2015, 2017, 2018).

The present study

The present study addresses the research gap outlined above by examining differences in expert and novice teachers’ professional vision using MET video feedback. Expert and novice teachers’ professional vision is examined based on the commentary on their own MET videos. The main goal of this study was to further our understanding of the difference in expert and novice teachers’ noticing and weighing of alternative teaching strategies when reflecting on their own teaching using MET video feedback. The following research questions are addressed:

  1. (1)

    What differences can be found in expert and novice teachers’ noticing when reflecting on their own MET video footage?

    Based on the theoretical assumption of expert teachers’ superior knowledge organization in curriculum scripts (Putnam, 1987; Wolff et al., 2021) and previous empirical research findings (e.g., van den Bogert et al., 2014; Wolff et al., 2015; Wyss et al., 2021), we predicted that expert teachers would notice more classroom events than novice teachers (Hypothesis 1). That is, we expect that when expert teachers watch and comment on their MET teaching videos, they are more likely than novices to report that they discovered something relevant to them in the video.

    Furthermore, we explored whether expert and novice teachers differ in how they perceive the events they noticed (i.e., whether they perceived the noticed classroom event as negative, positive, or neutral).

  2. (2)

    What differences can be found in expert and novice teachers’ weighing of alternative teaching strategies when reflecting on their own MET video footage?

    Based on previous findings on expert teachers’ superior knowledge-based reasoning compared to novices (e.g., Copeland et al., 1994; Wolff et al., 2015), we expected that expert teachers would suggest more alternative teaching strategies than novice teachers when reflecting on their own MET video footage (Hypothesis 2).

    Furthermore, expert and novice teachers may mention alternative teaching strategies for different reasons in their reflection process. Either they might mention them because they recognize that an alternative course of action would have been better in a particular teaching situation, they might mention alternative teaching strategies neutrally as part of their reflection process, or they might mention alternative courses of action to explain why their actions were in their eyes superior compared to alternatives. This distinction in terms of weighing alternative courses of action has not yet been investigated and will be explored in the present study.

Method

Sample

The present study is based on a subsample of a larger MET study in which 52 classroom teachers from 26 schools participated (see Cortina et al., 2015). The schools were located in southeast Michigan, United States, and included schools in affluent as well as economically challenged neighborhoods. Half of the participants were experienced teachers; the other half were student teachers who had recently completed their university teacher training. The experienced teachers served as mentors to the student teachers during the academic year. Mentors and mentees taught the same class on separate occasions. Among these 52 classroom teachers, 45 teachers volunteered to participate in an additional meeting with the research team to watch their MET recording, 40 teachers approved using the tape recordings of this session for further anonymous analyses. Three of those recordings could not be used for the current study (low battery of the recording device, microphone covered by clothing that muffled the recording, device failure for the MET recording). There is no association between data missing and expertise status and other demographics. In total, data from 37 classroom teachers could be included in this study (see Table 1 for sample and teacher characteristics). Experienced teachers had, on average, 10.7 years more teaching experience (varying between 4 and 25 years) than the novice teachers. Although years of experience alone do not necessarily translate into being an expert teacher, research indicates that it takes between 4.5 and 7 years of teaching experience to develop expertise as a teacher (Berliner, 2004). Teachers taught students in grades 1 to 11 in 16 different elementary, middle, and high schools. Participation in the study was uncompensated.

Table 1 Sample and teacher characteristics reported separately for experts and novices and for the total sample
Full size table

Equipment

Teachers wore the ASL Mobile Eye Tracker, a completely self-contained eye-tracking system (ASL mobile Eye, version 2009: https://www.asleyetracking.com). The system consisted of a pair of glasses containing infrared recording device for the eye-tracking and a small digital camera that captured the teacher’s field of vision (optics: 640 × 480 px, fixed focal range, alternate 30 Hz recording). As shown in Fig. 1, the MET device recorded the teacher’s field of view and fixation in the classroom (depicted by the circle). A cable transmitted the data to a recording unit that the teacher wore in a fanny pack around the hips. Two minutes prior to the lesson, a 5-point system calibration was performed in 5–10-m distance. Two stationary cameras provided additional video footage.

Fig. 1
figure 1

Still shot from a MET video commented on by a teacher in the present study. The circle indicates the fixation of the teacher in the classroom

Full size image

Procedure

For each participating teacher, a regular class period was recorded (ranging from 35 to 55 min). Two to ten days after the recordings (depending on teacher availability), expert and novice teachers were asked to freely comment while watching their own MET videos in the research lab using video-stimulated recall (Calderhead, 1981). The recording consisted of the footage of the visual field camera and a superimposed moving red circle indicating the fixation point. Occasionally, the recording started later into the lesson or stopped earlier due to technical issues during recording. A trained research assistant supported teachers’ think-aloud process, but was instructed not to make any directive comments. This was facilitated by the seating arrangement (assistant sat behind the teacher) and the use of headphones. Support was sometimes necessary when teachers did not express their thoughts spontaneously at the beginning of the session (e.g., the research assistant asked generic questions such as “So, what’s going on right now?” or reminded them of the instruction to freely articulate what is going through their minds). The think-aloud protocols were recorded, transcribed verbatim, and analyzed using qualitative content analysis (Braun & Clarke, 2006).

For the content analysis, we developed a two-dimensional, disjunct category system (Table 2) that included teachers’ noticing (“Noticing”) and their weighing of alternative teaching strategies (“Alternatives”). For the coding, clauses were the defining unit. Every clause was rated on the respective categories. In the first step of the coding process, coders decided whether teachers noticed a relevant classroom event and/or proposed an alternative teaching strategy in their statements that went beyond what the teachers said they actually used with their students in the classroom videos. In a second step, for the category "Noticing", the coders rated whether the teachers perceived the noticed classroom event as negative, positive, or neutral. For the category "Alternatives", coders rated the mentioned alternative teaching strategy as negative if the teacher indicated that this would have been a better alternative in the given situation, as positive if the teacher framed it as a worse alternative to their actual teaching, and as neutral if the teacher did not compare it to their actual behavior in class.

Table 2 Category system for content analysis
Full size table

The transcripts were coded by the first and second author. To measure interrater reliability, assessed by the percentage of agreement, a subsample of 10 transcripts was used. Deviations in the coding were resolved through discussion until consensus was reached. For the dimension "Noticing", the range of the percentage of agreement was between 74.6 and 97.4% prior to consensus discussion. For the dimension "Alternatives", the percentage of agreement ranged between 93.8 and 100% before discussion. The remaining 27 transcripts were then coded by the first author only.

Data analysis

We investigated whether experts and novice teachers differed in the number of events noticed and in the number of alternative teaching strategies mentioned using t-tests. To examine whether experts and novices differed in the valence of their noticing and weighing of alternative teaching strategies, we applied multinomial loglinear modelling. This approach tests for significant associations between Noticing/Alternatives and Expertise and allowed us to include relevant control variables in the same analysis. Because teachers differed on several characteristics, such as subject taught, grade level taught, minority status, and gender, that might influence teachers’ professional vision (e.g., Miller & Zhou 2007; Blomberg et al., 2011; Jarodzka et al., 2021), we controlled for these characteristics by including them as control variables in the analyses.

A loglinear analysis is a direct generalization of the common cross tabulation and its associated χ2 statistic to more than two categorical variables (Green, 1988). It allows, for example, to test statistically whether the strength of the association between the categorical variables "Noticing" and "Expertise" is moderated by the school subject or by teachers’ gender. The goal of loglinear modelling is to identify the most parsimonious model with respect to the number of effects included to reproducing the observed frequencies for all level combinations of the variables. A good model fit is indicated by an insignificant χ2 statistic. Significance tests for specific effects are based on iterative χ2 difference tests for a model that compares the expected and observed cell frequencies. If the discrepancy is large, the χ2 values is large and leads to the rejection of the null hypothesis of the current model and additional effects are included, making the model less parsimonious. The interpretation of effects is based on the crosstabulation of the variables.

We ran separate models for Noticing (three levels: negative, positive vs. neutral) and Alternatives (three levels: negative, positive vs. neutral) in combination with the variables Expertise (two levels: expert vs. novice), Gender (two levels: female vs. male), Minority Status (two levels: white vs. minority status), Subject (Math vs. other), and Grade Level (elementary vs. middle/high school). Note that for Subject, Grade Level, and Minority Status, the originally more differentiated categories had to be simplified to reduce the number of empty cells in the model.

We started for both analyses with a model that contains no interaction effects (i.e., a model that assumes that none of the variables affect the distribution to the three levels of Noticing or Alternatives; Model I), followed by a model with all two-way interaction effects (Model II; Noticing/Alternatives × Subject, Noticing/Alternatives × Grade Level, Noticing/Alternatives × Minority Status, Noticing/Alternatives × Gender, Noticing/Alternatives × Expertise). If this model remained significant, all higher order interactions involving the key variable Expertise were included one at a time to inspect fit improvement. In a last step (Model III), two-way interaction effects were eliminated one at a time to test if this significantly lowered the model fit. The final model (Model IV) is the model with the fewest number of effects necessary to produce an insignificant χ2 statistic. Then, all significant effects were analyzed separately as cross-tabulations to identify those cells in the frequency table that contribute substantially to the significant effect. For simplicity, the standardized residual (SR) is used for this purpose. Deviations of ± 2 SR are used for interpretation because, analogous to a z value, as indicate of a significant deviation.

Results

Contrary to our expectations, we found that expert and novice teachers did not differ significantly in the number of noticed classroom events (MDiff = −0.56, 95% CI [–14.18, 13.06], p = .933; Hypothesis 1) or in the number of teaching alternatives they mentioned (MDiff = −1.01, 95% CI [–5.05, 3.04], p = .617; Hypothesis 2) while watching their own MET video footage. However, the loglinear analysis indicated that expert and novice teachers differed in how they perceive the events they noticed (i.e., whether they perceived the noticed classroom event as negative or positive) and why they mentioned alternative teaching strategies (Table 3). Table 3 shows the model testing sequence from the loglinear analysis for Noticing and Alternatives. For both variables, the overall null hypothesis that no effect exists was rejected (significant χ2 for Model I in both cases). Including all two-way interactions yielded a good model fit (insignificant χ2) for Noticing and Alternatives (Model II). However, as Models IIIa-e show, none of the two-way interaction effects between Noticing or Alternatives and Subject, Grade Level, Minority Status, and Gender were essential except for Expertise. This means that even after controlling for these variables, Expertise was associated with a difference in the distribution across the three levels of Noticing as well as Alternatives. From Model II we can infer that no higher order interaction effects needed to be considered because a two-way interaction model already produced a good fit. The loglinear models suggested that none of the additional control variables needed to be considered for the interpretation of the crosstabs Noticing by Expertise and Alternatives by Expertise.

Table 3 Loglinear model history
Full size table

Novice teachers noticed more negative classroom events (43.6%) than positive classroom events (26.1%), whereas the opposite was true for expert teachers with 18.8% negative and 47% positive commentary (Table 4). As the standardized residuals indicate, there was no difference between expert and novice teachers regarding neutral commentary on noticed classroom events.

Table 4 Crosstabulation noticing by expertise
Full size table

Novice teachers mentioned more alternative teaching strategies with which they would improve their own teaching (81.4%), whereas expert teachers mentioned alternative teaching strategies with which they would improve their own teaching (42.7%) and which they consider inferior to their own instruction (41.8%) with about the same frequency (Table 5). Neutral commentary on alternative teaching strategies was as common for novice teachers as it was for expert teachers.

Table 5 Crosstabulation alternatives by expertise
Full size table

Discussion

Classroom videos have become increasingly popular to study teachers’ professional vision (Brophy, 2004; Gaudin & Chaliès, 2015; Goldman et al., 2007). The purpose of this study was to use MET video feedback to examine differences in expert and novice teachers’ professional vision. More specifically, we compared expert and novice teachers’ noticing and weighing of alternative teaching strategies while reflecting on their own MET video footage.

There are three key findings of the present research. First, overall, our findings point to expertise differences in teachers’ professional vision when reflecting on their own MET video footage. These differences in expertise only show in the valence of the comments, but not in their number. This might indicate that the novice teachers in our sample have already gained some professional vision allowing them to identify classroom events that are significant to them with heightened sensitivity (see also Behets, 1996) and to re-interpret the classroom situation and generate alternative courses of action (e.g., Santagata & Guarino, 2011). However, this finding deviates from previous research on teachers’ professional vision (e.g., Seidel & Prenzel, 2008; Schäfer & Seidel, 2015; van den Bogert et al., 2014; Wolff et al., 2015; Wyss et al., 2021) which indicated that expert teachers exceeded novice teachers in their noticing and knowledge-based reasoning abilities. However, these studies differ from our study in three ways: (1) they examined expertise effects in standardized teaching situations, (2) they used stationary video recordings, and (3) they used video sequences of other teachers to do this. In contrast, the present study examined expertise effects by asking teachers to reflect on their own authentic teaching videos recorded with MET technology. While the total amount of commentary did not differ between expert and novice teachers, it is likely that there are more critical events that lend themselves to commentary (i.e., noticing of critical events and generating alternative teaching strategies) in the footage of the novice teachers. Based on the analysis of the stationary video recording from the same study, for example, Ebright et al., (2021) found that student misbehavior during class could be identified about 1.5 times as often in the recordings for student teachers compared to the recording when the same class was taught by the experienced teachers. Arguably, many misbehaviors qualify as noticeable events.

Another alternative explanation for our finding that expert and novice teachers did not differ in the number of noticed classroom events could be that novices are less able to distinguish between important and unimportant events than experts. Thus, it is possible that novice teachers noticed a similar number of classroom events when watching their MET videos as experts, but the novices’ noticed events were not important, whereas experts only mentioned important classroom events in their reflection of their own teaching. This is supported by previous research that indicated that expert teachers identified relevant classroom events and ignored unimportant events in standardized teaching situations (e.g., Behets, 1996; Carter et al., 1988; van den Bogert et al., 2014; Wolff et al., 2016; Wyss et al., 2021).

To conclude, our findings that expert and novice teachers did not differ in the number of noticed classroom events and generated alternative teaching strategies in their own MET videos does not imply that novice teachers have an equally good professional vision as expert teachers. It is more likely that novice teachers noticed as many classroom events in their teaching as experts because they reported more irrelevant events or that there is simply more happening in novices’ teaching that lends itself to (negative) commentary and suggestions for improvement compared with experts’ teaching.

Second, the present study extended previous research on teachers’ professional vision by investigating expertise differences in the valence of teachers’ noticing and weighing of alternative teaching strategies. Regarding noticing, previous studies have focused on the ability of experts and novices to identify critical events in (standardized) classroom situations (e.g., Schäfer & Seidel, 2015; van den Bogert et al., 2014; Wolff et al., 2016; Wyss et al., 2021). However, it is not clear to what extent the quality of teacher noticing differs between experts and novices. The present study tackled this research gap by exploring the extent to which expert and novice teachers differed in their noticing of negative, positive, and neutral aspects when commenting on their own MET videos. Our results showed that novice teachers were more critical of their own teaching practice than expert teachers. Novice teachers noticed more negative than positive aspects in their teaching videos, whereas the opposite was true for expert teachers. That novice teachers noticed more negative classroom events in their MET videos could indicate that there were more critical events to note in novices’ lessons than in experts’ lessons, as discussed above (Ebright et al., 2021). The finding that expert teachers noticed more positive than negative aspects in their MET videos suggests that experts were more satisfied with their own teaching.

Moreover, our study expands on previous studies that focused on comparing expert and novice teachers’ generation of alternative teaching strategies when reflecting on (standardized) classroom situations (i.e., a video or slides of other teachers’ lesson; e.g., Copeland et al., 1994; Wolff et al., 2015) by exploring differences in experts’ and novices’ generation and weighing of alternative teaching strategies when reflecting on their own MET teaching videos. We found that novice teachers were particularly more critical when they considered possible alternative teaching strategies than expert teachers, indicated by the high percentage (over 80%) of alternatives they would have preferred over the behavior they displayed in situations they considered worth reflecting on. Importantly, expert teachers were still often critical of their own teaching behavior as they were dismissing potential instructional alternatives in merely half on the cases. Thus, despite their richer teaching experience, expert teachers were still open to entertaining the idea of optimizing their behavior in specific situations when reflecting on their own MET video footage.

Third, the present study also allows conclusions regarding possible effects of MET video feedback on teachers’ noticing and weighing of alternative teaching strategies. We found that MET video directs the attention of the teacher in their self-reflection of their teaching to events that they most likely missed while teaching but noticed in the review process with MET video footage. Furthermore, we showed that both expert and novice teachers were able to reflect on and identify alternatives to their teaching practices when watching their own MET videos. Thus, our results suggest that MET video feedback might help expert and novice teachers to advance their professional vision with regard to their noticing and weighing of alternative teaching strategies. Until now, only one study, which focused on pre-service teachers, suggested that MET video feedback could be fruitful in this endeavor (Stigler & Miller, 2018). Watching their lessons “through their own eyes” with MET video feedback offers teachers a new approach to reflect on their own teaching.

Practical implications

We maintain that MET video feedback holds promise to improve video feedback for novice teachers in various settings. The second author used MET video feedback with great success in a classroom setting with novice teachers who taught regular middle and high school classes. In the think-aloud phase, the instructor listened and took note when a “Noticing” or “Alternative” comment was a candidate for individual or (if the teacher was comfortable with it) group discussion. The MET combined with think-aloud commentary lends itself to a discourse that is led by the novice teachers themselves which help avoiding the instructor to simply impart knowledge about arguable better ways to handle a given situation, alternatives are explored in the discussion and––ideally––also reflected with respect to their feasibility at the actual stage of training. These reflections and discussions allow novice teachers to gradually acquire better strategies for teaching in a process that feels authentically their own without lecture. Of course, in-service teachers could also benefit from such courses in teacher professional development settings. This technique is consistent with the deliberate practice approach in which expertise can be acquired through a goal-oriented, focused, intentional, and structured effort by combining effective teachers, immediate feedback, and opportunity for repetitions and practice (e.g., Ericsson, 2018).

The results of the present study did show, however, that novice teachers’ evaluative reflections on their own teaching remained negatively tinted which may require counteractive measures (Sherin & van Es, 2009). Previous research shows that the perception of failure in teaching reduces teachers’ self-efficacy (i.e., teachers’ belief that they can influence how well students learn, including those who may be difficult or unmotivated; Guskey & Passaro, 1994). A lowered teaching self-efficacy is likely to contribute to teachers’ expectation that they will also fail in future teaching situations (Bandura, 1997; Tschannen-Moran & Hoy, 2007). Thus, if MET video feedback is used in teacher education as a tool to self-reflect on prospective teachers’ own instruction, it might be advisable to train prospective teachers to also identify the positive aspects of their teaching. For example, providers of feedback could remind novice teachers to also focus on the positive aspects or ask them whether there are different perspectives on the critical classroom situation (Kleinknecht & Gröschner, 2016).

Limitations and future research

Although this study significantly expands the body of knowledge on differences between expert and novice teachers with regard to their noticing and weighing of alternative teaching strategies using modern MET technology, we acknowledge the limitations of our results. For example, the time between the MET recordings and the stimulated recall sessions was not constant between teachers (2–10 days). The varying time span could have an impact on participants’ recall and thus on their comments in the think-aloud protocols. It has been discussed in the literature that a delay between an event and the recall of an event leads to a loss in the accuracy of the stimulated recall (e.g., Gass, 2001), but also that it may enhance recall (e.g., Spellman & Bjork, 1992). Note, however, that the recall stimulus (MET footage) is very strong and teachers never reported difficulty remembering the lesson.

Furthermore, it is also worth noticing that the perceived quality of the recorded instruction rated by experts varied and occasionally crossed between expert and novice teachers who were teaching the same class on different occasions (i.e., the quality rating of the novice teacher exceeded the rating for the expert, see Cortina et al., 2015). While we did not find significant interactions with the school subject, grade level, gender, or minority status of the teachers observed, the power of the statistical tests for higher order interactions was low. Those variables all have the potential to moderate the reported expert–novice differences.

Moreover, expert and novice teachers in our sample taught a wide range of grade levels (grades 1–11). Teaching younger and older students inevitably requires different instructional practices. While older students are often able to work rather independently on tasks in class, younger students need more individual attention and guidance from the teacher. In addition, depending on (subject-specific) classroom activities, seating arrangements may also differ between grade levels. For example, carpet seating close to the teacher during book reading is common in grade 1, while this is no longer practiced in higher grades. These differences in instructional practices between grade levels could influence teachers’ field of vision as well as the classroom events in general that teachers comment on while watching their MET videos. However, it is important to note that the loglinear analyses showed that the different grade levels taught did not affect the results of this study.

As it is true for field studies in schools in general, it is safe to assume that the classrooms observed were a positive selection with respect to (dis-)functionality of the classroom; teachers of very unruly classroom are less likely to have volunteered to participate in the study.

In terms of future research, it would be useful to extend the current findings by investigating the development of teachers’ professional vision and expertise more closely. In the present study, we compared noticing and weighing of alternative teaching strategies in two expertise groups: expert and novice teachers. However, examining expertise development with three or more groups (e.g., first-year students, advanced student teachers, first-year teachers, and expert teachers) could elucidate the role of transitions (i.e., from educational or academic contexts into professional contexts) in expertise development as well as enable researchers to examine nonlinear trajectories of expertise development (Boshuizen et al., 2020; for examples see, Behets, 1996; Boshuizen et al., 1995; Boshuizen & Schmidt, 1992; Carter et al., 1988; Copeland et al., 1994; Sabers et al., 1991).

Conclusion

Using teachers’ MET videos of their own teaching, the present study confirmed that expert and novice teachers differed in how they perceived their own instructional practices, reflecting differences in expertise. While expert and novice teachers did not differ in the number of noticed classroom events or generated alternative teaching strategies, differences between expert and novice teachers were apparent in the valence of their commentary. Novices noticed more negative than positive events in their teaching videos, whereas the opposite was true for experts. Furthermore, expert teachers proposed equal numbers of alternative teaching strategies that they perceived as superior and inferior to their own teaching, whereas novices were much more critical towards their teaching practices and explained mainly how they could improve their teaching using a different approach. We conclude that MET video feedback offers a natural way to broaden expert and novice teachers’ reflection on their teaching practices.