Teachers are expected to create an effective learning environment in their classroom while keeping an eye on classroom disruptions. This requires high monitoring competences as part of their noticing and professional vision. However, novice teachers often struggle dealing with critical situations and show ineffective monitoring behavior. Using dynamic scanpath analyses, this eye-tracking study examined monitoring strategies of novice and expert teachers to understand how classroom disruptions affect teachers’ gaze behavior and what constitutes effective monitoring. The sample consisted of 15 expert and 18 novice teachers. Participants watched a short classroom video in which several minor and one major disruption occurred while their eye movements were recorded with an eye tracker. We analyzed their scanpaths by comparing Shannon entropy values to investigate the predictability of gaze transitions and the stability of their gaze patterns. Compared to novices, experts showed higher transition probabilities in their gaze behavior, both before and after the major disruption. Similarly, experts showed lower entropy scores, indicating more stable gaze behavior, whereas novices with higher entropy scores showed exploratory gaze behavior. In general, experts showed more adaptive monitoring strategies, returning to their routine gaze behavior more quickly than novices after an unexpected disruption. This study enriches the cognitive theory of visual expertise by adding new expertise characteristics of teachers’ dynamic gaze behavior and situation-specific skills in disruptive classroom events. At the same time, the results reveal monitoring strategies that can be addressed in teacher education to enhance professional vision and classroom management. • Expert teachers have more predictable monitoring strategies than novice teachers. • After a classroom disruption, expert teachers' gaze become even more stable. • Novice teachers show more exploratory and unpredictable gaze patterns. • Dynamic scanpath analyses take temporal and contextual factors into account.

Instruction-giving (IG) is central to classroom interaction in Content and Language Integrated Learning (CLIL), yet it is rarely taught explicitly in pre-service CLIL teacher education. In response, the authors redesigned a CLIL practicum course and introduced targeted IG training in the second year. This study (a) described the redesign of an IG-focused training component and (b) examined whether the redesigned course was associated with differences in student teachers’ IG practices across cohorts. Fifty-two pre-service teachers in an eight-semester Primary English Language Teaching program (PELT) in Germany implemented CLIL-based projects with 180 pupils (ages 9–10) across two consecutive cohorts (2024, n = 25; 2025, n = 27). The study compared a baseline course implementation (2024) with a redesigned implementation that included explicit IG training (2025) informed by previous research. Audio-recorded classroom discourse from a randomly selected subsample of STs (n = 20; 10 per cohort) was segmented into CLIL-related IG episodes and analyzed using eight IG principles rated on a four-point scale. Quantitative comparisons were complemented by qualitative classroom discourse analysis and field notes. Mean overall IG scores were higher in 2025 (M = 3.16) than in 2024 (M = 2.79). The largest differences concerned clarity, responding to non-understanding, and strategic L1 use. Differences were smaller for sequencing and multimodality, and modeling did not differ across cohorts in this dataset. The findings suggest that explicit IG training is a teachable component of pre-service teachers’ instructional repertoire. While the cross-cohort design does not support causal claims, consistent descriptive and qualitative differences suggest that explicit IG training holds pedagogical promise for CLIL teacher education. • Instruction-giving (IG) is a teachable component of student teachers' (STs ) instructional repertoire. • Explicit IG training improved STs' clarity, sequencing, multimodality, and strategic L1 use in CLIL classrooms. • Cohort comparison revealed medium-to-large practical gains for the explicitly trained group. • High-quality IG was characterized by micro-sequencing, modeling, and multimodal scaffolding. • Findings underscore the value of embedding explicit IG practice in pre-service CLIL teacher education.