Tina Seufert

Solving Bayesian problems poses many challenges, such as identifying relevant numerical information, classifying and translating it into mathematical formula language, and forming a mental representation. This triggers research on how to support the solving of Bayesian problems. The facilitating effect of using numerical data in frequency format instead of probabilities is well documented, as is the facilitating effect of given visualizations of statistical data. Accordingly, this study examines the differences, in learning success and cognitive load, between the formula, the 2 × 2 table, and the unit square. The visualizations are additionally explained in a descriptive way and created by the participants themselves. The results confirm the hypothesis of the study that learning success is significantly higher when using the unit square and the 2 × 2 table than when using the formula. A contrasting pattern emerged for passive and active load. Significant differences between the unit square and the 2 × 2 table could not be found for learning success and passive and active load. Consequently, the visualization of Bayesian problems, which are explicitly explained and created by the participants, increase solution performance and reduce the effort that the solution of a task requires from the learners.

This paper complements retrospective approaches to researching informal learning in the workplace with experience sampling. Since (conscious) informal learning is becoming increasingly important for successfully keeping pace with rapid changes in working environments, a clear understanding of the construct and its precise measurement are essential. This is opposed by the error‐proneness of retrospective measurements. The aim of this study was therefore to better understand how employees learn informally by situationally measuring their use of informal learning activities (e.g., searching the internet). Three hundred sixty four full‐time employees with desk jobs in various professions completed short online questionnaires twice a day for 8 days about their current work task and the informal learning activities used. A Multilevel CFA with good model fit revealed the same three factors at the situation and person level: learning from oneself, learning from others, and learning from media. Thinking/reflection, professional exchange/co‐working with colleagues, and sharing knowledge were most frequently reported. Learning activities with media were used comparatively little. These findings underscore the metacognitive role of reflection in informal learning, challenge previous models of informal learning that underrepresent situational antecedents and learning from media, and emphasize the importance of including multilevel data in research. Our research provides human resource development professionals and supervisors with a foundation for promoting informal learning and the society with basic knowledge for familiarizing students with the need for informal learning. This ability to adapt quickly to changing knowledge requirements is crucial to the competitiveness of many industries.

Research on self-regulated learning (SRL) is undergoing a methodological and conceptual transformation from static, retrospective measures toward dynamic, multimodal, and temporally sensitive analyses. This discussion paper synthesizes and extends the contributions of a Topical Collection devoted to multimodal approaches in SRL research. It examines how diverse studies conceptualize and operationalize SRL as a complex interplay of cognitive, metacognitive, affective, and motivational (CAMM) processes. The Self-regulated learning, Multimodal data, and Analysis Grid (SMA Grid) serves as a shared framework for classifying and integrating different modalities and analytical designs. Across the reviewed contributions, a general shift from unimodal toward integrated multimodal approaches is evident, though motivational and affective dimensions remain underrepresented. The paper argues for expanding existing frameworks—particularly SMA and CAMM—toward explanatory models that account for social, contextual, and resource-based factors shaping regulatory processes. It also highlights persistent challenges in aligning data richness with theoretical depth, especially regarding temporal modeling and causal inference. A central concern is the translation of multimodal diagnostics into actionable pedagogical support, an area still underdeveloped despite the rise of AI-based analytics. Building on concepts such as cognitive load theory and resource-based perspectives, the paper proposes that SRL should be understood as a function of the dynamic balance between learners’ resources, task demands, and instructional context. Ultimately, it calls for a more integrated, theory-driven, and practice-oriented research agenda that connects analysis with support, and measurement with meaningful educational intervention.

Introduction: The multimedia principle states that learning with text and pictures is better than learning with text only. It might also depend on the context, in particular on the material the neighbour is learning with - which might look much more interesting.Methods: Therefore, in our experiments (n = 48; 47) we analysed whether learning outcomes, situational interest and extraneous load depend on the learning material (text versus text with pictures) and the context (the neighbour learns with either the same or different material). In the second study, we additionally analysed the influence of the learner’s affective state. In particular, we analysed with concurrent hypotheses whether affect is a consequence of material and context, whether effects of material and context on learning outcomes and interest are mediated by negative affect, or whether negative affect prior to learning is a moderator of material and context effects.Results: The multimedia effect was replicated for all dependent variables. This is particularly interesting because in classical multimedia design studies the effect has only been shown for learning outcomes. The fact that the addition of pictures also has a positive effect on situational interest and on the experienced extraneous load is of additional interest. The interaction between material and context was significant for extraneous load (Exp.1) and for situational interest (Exp.2). The role of affect could not be clarified.Discussion: Overall, it seems relevant to consider not only the learning material, but also the context in which the material is provided. Even if the effects of context are not consistent for all learning parameters, situational interest was strongly influenced by context when inequality was particularly salient.

As the world has rapidly transitioned to virtual platforms for education due to the Covid-19 pandemic, a new phenomenon known as “Zoom fatigue” has emerged. This term refers to the exhaustion associated with the extensive use of virtual communication platforms. In this context, the present study investigates the effects of different virtual class settings on cognitive load and fatigue. Four interventions were examined: self-view on vs. self-view off, focus-view function vs. grid-view function, activating a virtual background vs. natural background, and active participation vs. no active participation. The results suggest that turning off self-view can significantly reduce both cognitive load and fatigue. However, the effects of focus-view function vs. grid-view function did not reach significance for either cognitive load or fatigue. Similarly, using a virtual background vs. a natural background did not significantly affect fatigue, but it did influence cognitive load, with higher values for virtual backgrounds. Lastly, active participation in classes was associated with decreased fatigue and lower extraneous cognitive load compared to no active participation. The findings highlight the importance of considering both technological aspects (e.g., self-view, background settings) and pedagogical approaches (e.g., active participation) in online teaching to promote a more engaging and less fatiguing learning environment.

In learning journals, prompts were shown to increase self-regulated learning processes effectively. As studies on effects of long-term prompting are sparse, this study investigates the effects of prompting cognitive and metacognitive self-regulation strategies short-term and long-term in learning journals on learners’ strategy use, self-efficacy, and learning outcome. Therefore, 74 university students kept a weekly learning journal as follow-up course work over a period of eight weeks. All students’ learning journals included prompts for a short-term period, half of the students were prompted long-term. While self-efficacy was assessed via self-reports, strategy use was measured with self-reports and qualitative data from the learning journals. Learning outcomes were assessed via course exams. Short-term prompting increased self-reported cognitive and metacognitive strategy use, and the quantity of cognitive strategy use. Yet, it did not affect self-efficacy, which predicted the learning outcome. Irrespective whether prompting continued or not, self-reported cognitive and metacognitive strategy use, and self-efficacy decreased. Qualitative data indicate that the quantity of learners’ cognitive strategy use kept stable irrespective of the condition. The results indicate that short-term prompting activates cognitive and metacognitive strategy use. Long-term prompting in learning journals had no effect on strategy use, self-efficacy, and performance. Future research should investigate possible enhancers of long-term prompting like feedback, adaptive prompts or additional support.

Background and Aims: Endoscopy simulators are primarily designed to provide training in interventions performed during procedures. Peri-interventional tasks such as checking patient data, filling out forms for team time-out, patient monitoring, and performing sedation are often not covered. This study assesses the face, content, and construct validity of the ViGaTu (Virtual Gastro Tutor) immersive virtual reality (VR) simulator in teaching these skills. Methods: 71 nurses and physicians were invited to take part in VR training. The participants experienced an immersive VR simulation of an endoscopy procedure, including setting up the endoscopic devices, checking sign-in and team time-out forms, placing monitoring devices, and performing sedation. The actions performed by the participants and their timing were continuously recorded. Face and content validity, as well as the System Usability Scale (SUS), were then assessed. Results: 43 physicians and 28 nurses from 43 centers took a mean of 27.8 min (standard deviation ± 14.42 min) to complete the simulation. Seventy-five percent of the items for assessing face validity were rated as realistic, and 60% of items assessing content validity and usefulness of the simulation for different learning goals were rated as useful by the participants (four out of five on a Likert scale). The SUS score was 70, demonstrating a high degree of usability. With regard to construct validity, experienced endoscopy staff were significantly faster in setting up the endoscope tower and instruments than beginners. Conclusions: This multicenter study presents a new type of interdisciplinary endoscopy training system featuring peri-interventional tasks and sedation in an immersive VR environment.

Self-regulated learning depends on task difficulty and on learners’ resources and cognitive load, as described by an inverted U-shaped relationship in Seufert’s (2018) model: for easy tasks, resources are high and load is low, so there is no need to regulate, whereas for difficult tasks, load is too high and resources are too low to regulate. Only at moderate task difficulty do learners regulate, as resources and load are in equilibrium. The purpose of this study is to validate this model, i.e., the inverted U-shaped relationship between task difficulty and self-regulatory activities, as well as learner resources and cognitive load as mediators. In the within-subject study, 67 participants reported their cognitive and metacognitive strategy use for four exams of varying difficulty. For each exam task difficulty, cognitive load, and available resources (such as prior knowledge, interest, etc.) were assessed. Multilevel analysis revealed an inverted U-shaped relationship between task difficulty and the use of cognitive strategies. For metacognitive strategies, only a linear relationship was found. Increasing cognitive load mediated these relationship patterns. For learner resources we found a competitive mediation, indicating that further mediators could be relevant. In future investigations a broader range of task difficulty should be examined.

Solving Bayesian problems poses many challenges, such as identifying relevant numerical information, classifying, and translating it into mathematical formula language, and forming a mental representation. This triggers research on how to support the solving of Bayesian problems. The facilitating effect of using numerical data in frequency format instead of probabilities is well documented, as is the facilitating effect of given visualizations of statistical data. The present study not only compares the visualizations of the 2 × 2 table and the unit square, but also focuses on the results obtained from the self-creation of these visualizations by the participants. Since it has not yet been investigated whether the better correspondence between external and internal visualization also has an effect on cognitive load when solving Bayesian tasks, passive and active cognitive load are additionally measured. Due to the analog character and the proportional representation of the numerical information by the unit square, it is assumed that the passive cognitive load is lower when using the unit square as visualization than when using the 2 × 2 table. The opposite is true for active cognitive load.