Adolescents of color are particularly poised to experience the mental health crisis partly due to the absence of a clear-cut solution that prepares them to cope with ethnic/racial discrimination. One resilience-promoting factor in minoritized adolescents' lives is cultural socialization (i.e., the beliefs, practices, and worldviews that youth receive about their ethnic/racial group's heritage, history, and values), but the role of cultural socialization in relation to adolescents' resilience in the face of ethnic/racial discrimination is sporadic with extant studies documenting mixed results. Prior studies are likely limited by their focus on cultural socialization from parents relative to school adults and the larger school context. Following ethnic/racial discrimination, school-based cultural socialization may reduce youth's anticipation of discrimination, trust in others from different ethnic/racial groups, and rejection sensitivity. To test our theories, the present study used two daily diaries: Study 1 followed 134 African American adolescents over a 14-day period (N diaries = 1,494), and Study 2 followed 159 Asian American and Latinx adolescents over a 30-day period (N diaries = 3,458). In both studies, on days when ethnic/racial discrimination occurred, adolescents reported greater negative affect. This daily effect of ethnic-racial discrimination on negative affect was exacerbated on days when adolescents received less school-based cultural socialization but weaker on days when adolescents received more school-based cultural socialization. The present studies underscore how school adults foster youth's resilience in the context of ethnic/racial adversity.

Undergraduates enrolled in large, active learning courses must self-regulate their learning (self-regulated learning [SRL]) by appraising tasks, making plans, setting goals, and enacting and monitoring strategies. SRL researchers have relied on self-report and learner-mediated methods during academic tasks studied in laboratories and now collect digital event data when learners engage with technology-based tools in classrooms. Inferring SRL processes from digital events and testing their validity is challenging. We aligned digital and verbal SRL event data to validate digital events as traces of SRL and used them to predict achievement in lab and course settings. In Study 1, we sampled a learning task from a biology course into a laboratory setting. Enrolled students (N = 48) completed the lesson using digital resources (e.g., online textbook, course site) while thinking aloud weeks before it was taught in class. Analyses confirmed that 10 digital events reliably co-occurred >= 70% of the time with verbalized task definition and strategy use macroprocesses. Some digital events co-occurred with multiple verbalized SRL macroprocesses. Variance in occurrence of validated digital events was limited in lab sessions, and they explained statistically nonsignificant variance in learners' performance on lesson quizzes. In Study 2, lesson-specific digital event data from learners (N = 307) enrolled in the course (but not in Study 1) predicted performance on lesson-specific exam items, final exams, and course grades. Validated digital events also predicted final exam and course grades in the next semester (N = 432). Digital events can be validated to reflect SRL processes and scaled to explain achievement in naturalistic undergraduate education settings.

In educational settings, automated program repair techniques serve as a feedback mechanism to guide students working on their programming assignments. Recent work has investigated using large language models (LLMs) for program repair. In this area, most of the attention has been focused on using proprietary systems accessible through APIs. However, the limited access and control over these systems remain a block to their adoption and usage in education. The present work studies the repairing capabilities of open large language models. In particular, we focus on a recent family of generative models, which, on top of standard left-to-right program synthesis, can also predict missing spans of code at any position in a program. We experiment with one of these models on four programming datasets and show that we can obtain good repair performance even without additional training. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

This paper evaluates an automatically extracted domain model from textbooks and applies learning curve analysis to assess its ability to represent students’ knowledge and learning. Results show that extracted concepts are meaningful knowledge components with varying granularity, depending on textbook authors’ perspectives. The evaluation demonstrates the acceptable quality of the extracted domain model in knowledge modeling. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

While AI literacy is regarded as an essential competency to become a citizen in a rapidly changing society, it is challenging for people without computer science (CS) backgrounds to develop a sufficient level of AI competency. The main goal of this research is to examine the impact of the flipped learning approach to equip non-CS major students who intend to pursue careers in AI-related fields with basic AI literacy. Among various learner-centered methods, flipped learning was chosen as the main pedagogical frame to design an AI literacy curriculum. The participants were 80 adult learners who enrolled in the AI education program in Korea. The control group (N = 40) was taught in traditional instructor-centered methods whereas the experimental group (N = 40) was taught with a flipped learning method. Our research results indicate that AI literacy education with flipped learning improves the learning achievements of both CS majors and non-majors, especially effective for higher-order problem-solving skills. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

In introductory programming courses, automated repair tools (ARTs) are used to provide feedback to students struggling with debugging. Most successful ARTs take advantage of context-specific educational data to construct repairs to students’ buggy codes. Recent work in student program repair using large language models (LLMs) has also started to utilize such data. An underexplored area in this field is the use of ARTs in combination with LLMs. In this paper, we propose to transfer the repairing capabilities of existing ARTs to open large language models by finetuning LLMs on ART corrections to buggy codes. We experiment with this approach using three large datasets of Python programs written by novices. Our results suggest that a finetuned LLM provides more reliable and higher-quality repairs than the repair tool used for finetuning the model. This opens venues for further deploying and using educational LLM-based repair techniques. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Hand-raising signals students’ willingness to participate actively in the classroom discourse. It has been linked to academic achievement and cognitive engagement of students and constitutes an observable indicator of behavioral engagement. However, due to the large amount of effort involved in manual hand-raising annotation by human observers, research on this phenomenon, enabling teachers to understand and foster active classroom participation, is still scarce. An automated detection approach of hand-raising events in classroom videos can offer a time- and cost-effective substitute for manual coding. From a technical perspective, the main challenges for automated detection in the classroom setting are diverse camera angles and student occlusions. In this work, we propose utilizing and further extending a novel view-invariant, occlusion-robust machine learning approach with long short-term memory networks for hand-raising detection in classroom videos based on body pose estimation. We employed a dataset stemming from 36 real-world classroom videos, capturing 127 students from grades 5 to 12 and 2442 manually annotated authentic hand-raising events. Our temporal model trained on body pose embeddings achieved an F1 score of 0.76. When employing this approach for the automated annotation of hand-raising instances, a mean absolute error of 3.76 for the number of detected hand-raisings per student, per lesson was achieved. We demonstrate its application by investigating the relationship between hand-raising events and self-reported cognitive engagement, situational interest, and involvement using manually annotated and automatically detected hand-raising instances. Furthermore, we discuss the potential of our approach to enable future large-scale research on student participation, as well as privacy-preserving data collection in the classroom context. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Compelling evidence, from multiple levels of schooling, suggests that teachers' knowledge and beliefs about knowledge, knowing, and learning (i.e., epistemologies) play a strong role in shaping their approaches to teaching and learning. Given the importance of epistemologies in science teaching, we as researchers must pay careful attention to how we model them in our work. That is, we must work to explicitly and cogently develop theoretical models of epistemology that account for the learning phenomena we observe in classrooms and other settings. Here, we use interpretation of instructor interview data to explore the constraints and affordances of two models of epistemology common in chemistry and science education scholarship: epistemological beliefs and epistemological resources. Epistemological beliefs are typically assumed to be stable across time and place and to lie somewhere on a continuum from instructor-centered (worse) to student-centered (better). By contrast, a resources model of epistemology contends that one's view on knowledge and knowing is compiled in-the-moment from small-grain units of cognition called resources. Thus, one's epistemology may change one moment to the next. Further, the resources model explicitly rejects the notion that there is one best epistemology, instead positing that different epistemologies are useful in different contexts. Using both epistemological models to infer instructors' epistemologies from dialogue about their approaches to teaching and learning, we demonstrate that how one models epistemology impacts the kind of analyses possible as well as reasonable implications for supporting instructor learning. Adoption of a beliefs model enables claims about which instructors have better or worse beliefs and suggests the value of interventions aimed at shifting toward better beliefs. By contrast, modeling epistemology as in situ activation of resources enables us to explain observed instability in instructors' views on knowing and learning, surface and describe potentially productive epistemological resources, and consider instructor learning as refining valuable intuition rather than fixing wrong beliefs.

Teachers' mathematical knowledge has important consequences for the quality of the learning environment they create for their students to learn mathematics. Yet relatively little is known about how teachers reason proportionally, despite the fact that proportional reasoning is foundational for several mathematics concepts and that ratios and proportional relationships constitute a major component of the middle school mathematics curriculum. In this study, we investigated how teachers reasoned proportionally on a nonroutine ratio task and the extent to which their proportional reasoning was able to predict their overall understanding of the relevant concepts: ratios and proportional relationships. Using data collected from 238 US mathematics teachers, we found that teachers' proportional reasoning could be grouped into four categories: incorrect, additive, relative, and proportional reasoning. Our results also indicated that teachers' overall knowledge of ratios and proportional relationships aligned with the way they reasoned proportionally, meaning that teachers who used incorrect reasoning on a separate task received the lowest scores on average on the ratios and proportional relationships measure, whereas those who reasoned proportionally had the highest mean scores on average. Implications of the study include the need to shift attention to the way teachers reason in relation to the two elements of proportional reasoning (covariance and invariance) to capture the nuances in their understanding of ratios and proportional relationships.