Major research themes in teacher noticing include, in particular, mathematics and mathematics pedagogical topics such as students’ mathematical thinking (Jacobs et al., 2010) or ways of dealing with representations (Dreher and Kuntze, 2015), and general pedagogical topics, such as classroom management (Gold and Holodynski, 2017; Weber et al., 2018). Recent studies exploring teachers’ noticing have considered dealing with heterogeneity (Keppens et al., 2021) and sociopolitical dimensions, such as ethnicity or socioeconomic background (Shah and Coles, 2020; Louie et al., 2021), as well as strength-oriented noticing (Scheiner, 2023).

Based on a cognitive–psychological approach, teacher noticing is widely understood as paying attention to or perceiving classroom events, interpreting these events, and—in some conceptualizations—employing decision-making processes to act/react based on the interpretations (Sherin et al., 2011; Dindyal et al., 2021; König et al., 2022b). For example, Jacobs et al. (2010) focused in their seminal work on children’s mathematical understanding, distinguishing three facets of teachers’ noticing: attending to children’s mathematical strategies, interpreting their understanding, and deciding how to respond. Although the facets of teachers’ noticing support each other and may give the impression that noticing in the classroom is a linear or even chronological process (van Es, 2011), from a theoretical perspective, they are understood as deeply “interrelated and cyclical” (Sherin et al., 2011, p. 5), operating through complex interactions (Kaiser et al., 2015; Santagata and Yeh, 2016; Thomas et al., 2021). This circular modeling of in-the-moment teachers’ noticing is often expressed as perception ↔ interpretation ↔ decision-making (Santagata and Yeh, 2016; Thomas et al., 2021).

However, little empirical attention has been paid to these relationships and their nature, despite knowledge of them being crucial for comprehending teachers’ noticing and its effects in the classroom (Thomas et al., 2021). Some studies have claimed that perception and interpretation, and interpretation and decision-making, are more closely related than perception and decision-making, providing some insight into the internal structure of the construct teacher noticing (Thomas et al., 2021; Bastian et al., 2022), but less is known about the development of the facets’ relationships over time and how they influence and change each other. This deficiency may be explained by the lack of quantitative pretest–posttest studies on teacher noticing (König et al., 2022b).

One exception is a study by Jong et al. (2021), who examined changes in PSTs’ noticing skills over the course of one semester. They reported a significant effect of perception at Time 1 (T1) on perception at Time 2 (T2, β=0.27) and similar results for interpretation (β=0.35), but no such association between decision-making at T1 and T2. In addition, only one cross-lagged path was significant (the one from interpretation at T1 to perception at T2 [β=0.16]), indicating an effect of interpretation on perceptual skills. These results were consistent with a study by Superfine et al. (2017), which demonstrated that it may be better to foster interpretation skills before perceptual skills. In our study, we further analyzed the cross-lagged, reciprocal effects of the three facets of teacher noticing to provide a clearer picture of their relationships.

Competence can be conceptualized as a set of cognitive skills and abilities needed to successfully cope with the demands of professional situations, including the motivational, volitional, and social willingness to apply them (Weinert, 2001). Following this conceptualization, competence research at first particularly focused on cognitive dispositions of teacher such as their professional knowledge (Kunter et al., 2013). To consider the situation-specificity of competence, conceptualizations have been further developed in recent years based on complementary components, such as teacher noticing (Kaiser et al., 2017; Metsäpelto et al., 2021). For example, Blömeke et al. (2015a) conceptualized competence as a continuum between dispositions (e.g., knowledge) and actual classroom performance, and they included perceiving, interpreting, and decision-making as mediating skills to emphasize the situation-specific aspects of competence. These skills, which we describe as teacher noticing, contextualize knowledge and affects and link them to teachers’ actual performance. Thus, they also connect theory learned in initial teacher education to practice in the classroom.

The TEDS-M research program includes situation-specific skills in the measurement of teachers’ professional competencies by incorporating teacher noticing as an integral part in the competence framework since the TEDS-Follow-Up (TEDS-FU) study (Kaiser et al., 2017). Based on the model by Blömeke et al. (2015a), we conceptualize teacher noticing as consisting of three facets: (1) perception of specific classroom events, (2) interpretation of the perceived events, and (3) decision-making, that is, the preparation of responses to student actions or alternative instructional strategies (Kaiser et al., 2015). This can be classified as an analytical and cognitive–psychological conceptualization of teacher noticing (König et al., 2022b). A cognitive-psychological approach was chosen to model and measure the cognitive processes involved in teacher noticing and to compare the skills of groups of teachers as well as the development of individual teachers over time (König et al., 2022b). The approach allows for a standardized and feasible operationalization of the noticing facets and is thus commonly applied in the psychometric measurement of teacher noticing (Weyers et al., 2023b).

Based on studies from the expertise research (Berliner, 1988; Stigler and Miller, 2018), the first facet of teacher noticing in our conceptualization involves perceptual processes with no or only minimal interpretation (Bastian et al., 2022), encompassing the observation of clearly discernable incidents that happen in classroom. The second facet, interpretation, involves the analysis of observed events based on an individual’s knowledge, experiences, and beliefs, excluding the consideration of instructional responses to these events. This, in turn, is part of the third facet of teacher noticing that includes the development of possible continuations of the lesson, responses to student behavior, and alternative approaches to observed teacher actions (Yang et al., 2021). At the content level, our conceptualization, used in this study, does not focus only on one topic, such as children’s mathematical understanding, but encompasses a broad field of classroom situations and features that are relevant to high-quality mathematics education from general and mathematics pedagogical perspectives, such as perceiving effective classroom management, analyzing the use of different representations, interpreting students’ mathematical thinking, or making decisions about teaching and learning processes (Kaiser et al., 2015; Yang et al., 2021).

Over the past 20 years, a growing body of research has considered teaching internships and their impact on all stakeholders, particularly PSTs (see reviews by Lawson et al., 2015; Ulrich et al., 2020). Early studies revealed the crucial role of teaching internships in the self-perceived personal and professional development of PSTs (e.g., Caires and Almeida, 2005), and more recent literature has demonstrated the positive impact of teaching internships on the beliefs, self-efficacy, and motivation of PSTs (Ng et al., 2010; Seifert and Schaper, 2018; García-Lázaro et al., 2022). Mentors foster the development of their mentees and play a crucial role in how PSTs use and perceive their teaching internships (Hudson and Millwater, 2008; König et al., 2018b; Festner et al., 2020; García-Lázaro et al., 2022). In addition, research has suggested that professional competencies, which, in this context, are often distinguished as teaching, educating, assessing, and innovating, are expanded during teaching internships (Gröschner et al., 2013; Seifert et al., 2018).

However, these studies have usually been based solely on PSTs’ self-assessments of their own competencies and have not applied objective tests to measure PSTs’ professional competencies (Ulrich et al., 2020). Consequently, evidence of competence gains is limited, although there are some exceptions. Recent studies within the Learning to Practice project have used knowledge tests to assess PSTs’ general pedagogical knowledge, have combined and linked these with self-assessment measures, and demonstrated improvements in complex decision-related knowledge among student teachers from three different universities. Reflection and theory–practice links were shown to facilitate this improvement (König et al., 2018a). Furthermore, the incorporation of reflection activities and the establishment of links between theoretical concepts and school practice have the potential to enhance knowledge development, as shown by Schlag and Glock (2019). They conducted an analysis of classroom management knowledge using standardized tests and highlighted the significance of practice activities in the acquisition of classroom management knowledge. In summary, although teaching internships have been examined in several studies, research gaps remain regarding the investigation of teachers’ competence development using standardized measures (Ulrich et al., 2020).

To date, situation-specific facets of competence, such as teacher noticing, have scarcely been investigated in the teaching internship context (Ulrich et al., 2020). However, in a more general context, comparisons between teachers with different durations of experience have emphasized the critical role of teaching experiences in the development of teachers’ noticing (e.g., Jacobs et al., 2010; Gold and Holodynski, 2017; Yang et al., 2021; Bastian et al., 2022). In these studies, PSTs performed worse than in-service teachers on standardized tests of teachers’ noticing (Bastian et al., 2022). Thus, practical experiences in schools (e.g., teaching internships) should facilitate the development of PSTs’ teacher noticing. This hypothesis was confirmed in a study by Stürmer et al. (2013), who investigated a five-month teaching internship at a German university, focusing on perceptual and interpretive facets of teachers’ general pedagogical noticing conceptualized as professional vision. They demonstrated increases (η² = 0.09, corresponding to an effect size of d = 0.63; Lenhard and Lenhard, 2016-2022) in both facets using the video-based test Observer (Seidel and Stürmer, 2014).

These findings were replicated by Mertens and Gräsel (2018) for a holistic understanding of professional vision, which included perceptual and interpretive aspects, based on five-month teaching internships via another university. The increase in teachers’ noticing skills had a large effect size (d = 0.79) and was validated using a control group (Mertens et al., 2018; Mertens and Gräsel, 2018). Further evidence suggests that teaching internships benefit PSTs with weak teacher noticing skills, helping them to develop these skills (Stürmer et al., 2013; Orschulik, 2020).

In a similar study exploring the effects of two short-term (seven-week) practical activities at a German university, Weber et al. (2018) reported no significant development in the classroom management and general pedagogical knowledge-related facets of holistic professional vision for PSTs who participated in the practical activities without accompanying institutionalized reflection and feedback. In contrast, a group of PSTs who received peer and expert feedback on their videotaped lessons and gave feedback themselves significantly improved their skills, with a large effect size of d = 1.10 (Weber et al., 2018). Focused accompanying intervention to link theory and practice seems to support the development of teachers’ noticing skills (Stürmer et al., 2013; Scholten and Orschulik, 2022).

However, until now the development of decision-making skills during teaching internships has not been investigated, and subject-specific insights in the noticing of mathematics teachers are lacking. Moreover, the influence of process variables (i.e., PSTs’ individual teaching experiences and opportunities to learn within the internship) on the development of teacher noticing has not been examined before but are needed to assess benefits of teaching internships as an important part of initial teacher education and to provide first explanations for the development. Overall, previous results have been based solely on data collected at single universities, thus limiting their generalizability. Our study addressed this broad research gap.

Based on current state-of-the-art research and the accompanying research desiderata, we aimed to investigate the reciprocal influences of the facets of teachers’ noticing over the course of teaching internships and the influences of teaching internships and their process variables on the development of PSTs’ teacher noticing. Therefore, we considered the following research questions:

In this section, we briefly introduce teacher education in Germany and describe the characteristics of teaching internships to place the study in its context, to facilitate the understanding of our research design, methodology, and findings, and to enable researchers to connect our findings to the practical experiences in teacher education in their countries, as teaching internships have been identified as a crucial part of teacher education worldwide (Cohen et al., 2013; Lawson et al., 2015; Cabaroğlu and Öz, 2023).

In general, teacher education in Germany consists of three main phases, the first two constitute initial teacher education: (1) higher education at universities and pedagogical universities, (2) practical teacher training (known as preparatory service or induction) through specific state-based teacher seminars, and (3) elective professional development courses, which accompany professional practice as an in-service teacher following the first two phases of teacher education (Drahmann, 2020; Eckhardt, 2021).

In Germany, the exact design and organization of teacher education depends on the federal state (Bundesland) in which it is conducted, although shared norms and principles exist (Cortina and Thames, 2013). Several different types of schooling have been implemented for German secondary education (the focus of this paper): teaching at academic-track schools (Gymnasium), teaching at non-academic track schools, teaching at intermediate forms which combine academic and non-academic track education, teaching at vocational schools and special needs teaching (Cortina and Thames, 2013; Drahmann, 2020). For each type, PSTs must study two school subjects, with the exception of special needs education. University study is divided into four areas: subject matter, subject-related pedagogy, general pedagogy, and practical activities in school (Drahmann, 2020).

Due to the identified gap between theory and practice in teaching and teacher education (König and Rothland, 2018), connecting these practical activities with academic opportunities to learn has become increasingly important recently. Hence, school practical studies have been reorganized and reshaped in the wake of the Bologna reforms, that is, the transformation from a traditional state examination system to a bachelor-master-system comparable within Europe (Schubarth et al., 2012; Drahmann, 2020; Terhart, 2021). They are usually spread over several semesters of university teacher education depending on the federal state and aim to achieve several goals: professional orientation, competence improvement by linking theory and practice, and the development of teaching abilities (König, 2019).

In all federal states, PSTs undertake some extensive, long-term (i.e., lasting several months) practical activities in schools in the master’s phase of their university teacher education to link theory with practice, which is often referred to as a teaching internship or (long-term) teaching practicum (Gröschner et al., 2015; Ulrich et al., 2020) and is, again, the focus of this paper. During their teaching internships, PSTs attend a school of their teaching type and participate in almost all aspects of daily school life. In particular, they are required to observe a certain number of lessons taught by in-service teachers and to teach lessons themselves under supervision. Mentors (i.e., in-service teachers at the school) support the PSTs in practical matters during their internships, while teacher educators from the university, and in some federal states, from the state-run teacher training seminars of the second phase of teacher education, facilitate the theory–practice linkage.

To address the aforementioned lack of studies based on multiple universities, the sample consisted of 175 secondary mathematics education PSTs from six universities (Hamburg, Cologne, Münster, Paderborn, Vechta, and Würzburg) in four German federal states who participated in an online survey before and after their (long-term) teaching internships between spring 2019 and fall 2021. At T1, 313 PSTs participated in the survey. The panel sample decreased with a panel attrition of 44% due to lack of participation at T2. Since the process variable data were collected at T2, these individuals were missing not only the T2 ability scores but also important predictor variables and were therefore excluded from the final panel sample. Data imputation was not possible because the missing predictor variables would have led to a circular process. Teaching internships were conducted at the master’s level for all universities except Würzburg. Since initial teacher education in Bavaria is organized as a state examination with no division between bachelor’s and master’s programs, PSTs from Würzburg participated in the survey during their 5th or 6th semesters of study. The demographic sample statistics are presented in Table 1.

Deploying a pretest–posttest design, we administered a survey to the participants via an online platform before their teaching internships (at T1) and after they completed their internships (at T2) in the frame of the TEDS-Validate-Transfer research project. Each questionnaire took approximately 90 min to complete. At T1, we obtained demographic data and measured teachers’ noticing using a video-based instrument. At T2, the latter measure was repeated and supplemented with an assessment of process variables related to teaching internships: learning time, teaching practice activities, and mentor support. PSTs received a small financial reward for their participation. Data were collected and processed according to the requirements of the General Data Protection Regulation.

To address the first research question concerning reciprocal effects of the three facets of teacher noticing, we computed autocorrelations and an autoregressive manifest model with a cross-lagged panel design using the WLEs with Mplus 6.8 software (Muthén and Muthén, 1998-2021). Although the teacher education programs and, in particular, the teaching internships at all six universities in our sample shared key components, they were not, of course, completely similar. To account for differences and impacts of teacher educators, teaching type and modules and thus the stratified nature of the sample, we specified a stratification variable combining university and teaching types using the “type = complex” option.

We approached the second research question which addressed changes in PST’s noticing skills using paired sample t-tests. We calculated the Cohen’s d effect sizes for a within-subjects design (Lakens, 2013) to estimate the impact of teaching internships on teacher noticing and compare our results with previous findings.

Again, we used multiple regression analyses to investigate the third research question, considering the stratified sample. We conducted several analyses (for each set of subscales separately) to examine the effects of process variables on the ability scores for each teacher noticing facet at T2. Regression analyses were controlled for high school diploma grade, semester, and dichotomized teaching type (academic-track and vocational school vs. non-academic-track and special needs education), as well as the ability scores for the teacher noticing facets at T1.

To answer our first research question, we conducted cross-lagged panel analyses. The stability of the teacher noticing facets was investigated using autocorrelations. Positive medium to strong autocorrelations for perception (r = 0.50, p < 0.001), interpretation (r = 0.62, p < 0.001) and decision-making (r = 0.52, p < 0.001) indicated stable parts of these constructs, but also emphasized intrapersonal variation and, thus, change during the teaching internships⁴. We applied a cross-lagged panel model to assess the predictive quality of the three facets at T1 for the facets at T2 (see Figure 5). The means and standard deviations for all ability scores are presented in Table 4, and the correlations between all facets of teachers’ noticing are presented in Table 5. The correlations suggested possible cross-lagged effects between the three facets. Measures of reliability have already been discussed (see Table 2).

The auto-lagged path coefficients showed significant moderate predictive power for perception (β = 0.36, p < 0.001) and decision-making (β = 0.36, p < 0.001), and a strong effect for interpretation (β = 0.57, p < 0.001). The cross-lagged path coefficients demonstrated that only interpretation had an effect on the other facets, since both paths from interpretation at T1 to perception at T2 (β = 0.30, p = 0.001) and decision-making at T2 (β = 0.28, p = 0.002) were significant, but not the reverse cross-lagged paths. This indicated a prominent role for interpretation in the development of all three facets. No cross-lagged effect was observed between perception and decision-making. For all three facets, the model explained a significant amount of the variance.

To answer the second research question, we compared the measurements of individual teachers’ noticing skills before and after the teaching internships using t-tests to explore the effects of the practical activities (see Table 4). Significant increases were found for all three facets of teachers’ noticing over the course of the teaching internships. The effect sizes, expressed as Cohen’s d, revealed small effects for all three facets, with a higher effect size for interpretation.

The third research question was answered using multiple regression analyses to investigate the effects of the teaching internship process variables on the development of PSTs’ teacher noticing. For each scale (i.e., learning time, teaching practice activities, and mentor support) and for each noticing facet a model was calculated using all subscales of the scale as predictors. The results of these analyses are illustrated in Tables 6–8. They showed the significant influences of process variables for only a few subscales. A significant proportion of the variance was explained.

For learning time, there was only a positive effect of lesson follow-up on decision-making (β = 0.19, p = 0.006), while lesson preparation negatively influenced decision-making (β = −0.21, p = 0.039). For the teaching practice activities, linking theories to specific situations significantly predicted perception and interpretation, with positive regression coefficients of β = 0.19, p = 0.014 and β = 0.25, p < 0.001, respectively. In contrast, teaching (i.e., performing situations) showed a negative effect on perception (β = −0.26, p = 0.001) but no effects on interpretation and decision-making.

Emotional support provided by PSTs’ mentors had a significant positive effect on decision-making (β = 0.27, p = 0.002). In contrast, instrumental support, such as providing teaching materials, negatively affected interpretation (β = −0.14, p = 0.048) and decision-making skills (β = −0.15, p = 0.046).

The limitations of this study should be considered. First, the results presented herein were derived from convenience sampling. Consequently, generalizations should only be made with caution. Group effects of the six-university sample were controlled for in the analyses using stratification variables. However, this approach may not have accounted for all effects of the different locations and may potentially have overlooked additional group-specific effects. In addition, for organizational reasons, it was not possible to establish control groups to evaluate the influence of the teaching internships against a group that did not undertake internships and to control for memory effects. However, this only influenced the overall effect of the internships on the three facets of teachers’ noticing for the second research question.

Moreover, the design of the study included only two measurement time points, which provided a rather rough representation of the development of teachers’ noticing over the course of the teaching internships. Research with more measurements at shorter time intervals is needed to gain more insight into the development of teachers’ noticing and, in particular, into the optimal length of an internship, since saturation effects may occur during this activity.

Process variables (i.e., learning time, teaching practice activities, and mentor support) were measured only through self-report instruments, which may have created bias in the dataset. Third-party assessments by mentors or researchers, for example, should be used to complement our findings. Furthermore, our test instrument assessed only a subset of teachers’ professional competencies; thus, the development of other competence facets may have been overlooked, and more comprehensive survey designs are needed to consider knowledge, beliefs, and situation-specific and/or performance-related facets together and to obtain a deeper understanding of the effects of teaching internships on PSTs.

Finally, some parts of the data collection took place during the COVID-19 pandemic. This influenced the organization of the teaching internships in ways that differed among schools. Our items for the internship process variables specifically included participation in online and distance learning and teaching, which provided some control for the changed conditions. Nevertheless, the pandemic may have had some unknown effects on our study.