Although generative AI is increasingly integrated into K–12 education, prior research has emphasized post-intervention outcomes rather than how students interact with AI or how post-intervention competencies support human–AI collaboration. This mixed-methods study examined phase-based changes in students’ AI interactions, pre–post changes in AI dispositions, prompt engineering skills, and human–AI collaboration competencies, and predictors of post-intervention collaboration competencies. Sixty-nine eighth-grade students participated in a five-day STEM–AI curriculum using ChatGPT. Data from student-generated prompts, pre–post surveys, and competency tests were analyzed through content analysis, repeated-measures MANOVA, and multiple regression analyses. Results indicated that students’ AI interactions evolved from exploratory use toward argumentation and metacognitive monitoring. Students showed significant improvements in AI dispositions, prompt engineering skills, and human–AI collaboration competencies. Ethical awareness, particularly accountability and privacy, emerged as a significant predictor of post-intervention collaboration competencies. These findings suggest that generative AI can support higher-order thinking as a collaborative partner and that the development of human–AI collaboration competencies depends more on ethical awareness than on prompt engineering skills alone.

Abstract: This study explores the integration of Augmented Reality (AR) and collaborative activities to leverage abstract Computational Thinking (CT) concepts accessible to young students. The instructional design follows Plan, Act, Reflect (PAR) cycles that consist of three types of collaborative activities: Hands-on, AR-integrated, and self-directed robot programming activities. Findings highlight the importance of scaffolding in helping young learners, particularly those with low spatial ability, grasp directional concepts. Role-based collaboration proved effective in fostering engagement and problem-solving skills, though challenges emerged in the AR-based activity. This study contributes to immersive learning by demonstrating practical application of AR technology into K-12 classrooms.

The purpose of this study was to address case-based instruction in entrepreneurship education from multiple instructors’ perspectives and describe their experiences with this approach as it relates to gender diversity. This study answers the call for more substantive research focused on issues related to entrepreneurship education, including an examination of specific teaching methods and an in-depth exploration of gender and entrepreneurship education. The findings suggest the role that case-based instruction can play in efforts to advance gender equality as part of the diversity imperative in education. Instructors considered gender diversity in case selection but found it to be a challenge. Moreover, the descriptions of the entrepreneurs and their venture behaviors in the cases were gendered, which might have resulted in unintended discriminatory gender lessons. While none of the women were portrayed in a specifically negative light, underlying assumptions about gender roles infused the cases. While the instructors considered gender diversity in case selection, they did not necessarily consider gender stereotypes, which is problematic because gender stereotypes are cognitive schemas that influence the ways in which individuals make sense of their social world and may discourage some women’s aspirations to become business owners.

This study examined the effects of embodied learning experiences on students' understanding of computational thinking (CT) concepts and their ability to solve CT problems. In a mixed-reality learning environment, students mapped CT concepts, such as sequencing and loops, onto their bodily movements. These movements were later applied to robot programming tasks, where students used the same CT concepts in a different modality. By explicitly connecting embodied actions with programming tasks, the intervention aimed to enhance students' comprehension and transfer of CT skills. Forty-four first- and second-grade students participated in the study. The results showed significant improvements in students' CT competency and positive attitudes toward CT. Additionally, an analysis of robot programming performance identified common errors and revealed how students employed embodied strategies to overcome challenges. The effects of embodied learning and the impact of embodied learning strategies were discussed.

This study explores the impact of embodied learning activities, implemented in both mixed-reality and unplugged contexts, along with robot programming, on early elementary students' computational thinking (CT) knowledge, competence, and attitudes. A total of 67 first-grade students from a rural elementary school participated in the quasi-experimental study, which compared how these different modes of embodied learning influenced CT skill acquisition. The findings indicate that embodied learning activities significantly enhanced students' CT knowledge, regardless of whether they were conducted in a mixed-reality or unplugged setting. Additionally, robot programming, used as a transfer task, improved students' CT competence, demonstrating the benefits of applying learned skills in new contexts. In addition, students in the mixed-reality condition reported higher satisfaction with their learning experiences.

This study examined how embodied learning experiences support students to grasp abstract CT concepts in early primary education. Forty-seven students were recruited from first- and second-grade classrooms. They had five embodied activities that simulated robot programming tasks requiring students to (1) figure out a route from a starting point to a destination, (2) convert the spatial information into codes, and (3) perform spatial movements according to the codes. The results revealed that students' CT and spatial reasoning skills significantly improved after the embodied learning. No gender differences were confirmed regarding learning outcomes and attitudes. The students' ages were closely related to their learning outcomes. The effects of embodied learning, differences between genders, and cognitive development of students were discussed.

This study aims to investigate the effects of competition through a digital leaderboard in gamified online discussions on learners’ behavioral and cognitive engagement in learning. Twenty-three graduate students pursuing master’s degree in instructional technology in a public university in Indonesia were involved in a five-week quasi-experiment (N = 23; Group A = 12 Group B = 11). All discussion activities were performed on a gamification platform designed specifically to facilitate online discussions. Several game elements were used on this platform, including points, badges, quizzes, and leaderboards. The discussion topics encompassed instructional design case studies in varied contexts (K-12, higher education, and industry). In this study, we measured behavioral engagement by tallying the number of posts and earned points from participants. Additionally, cognitive engagement was evaluated by analyzing the cognitive level evident in participants’ posts. A framework entitled the levels of Knowledge construction was used for analyzing the quality of the posts. After a five-week implementation period, our investigation revealed that incorporating competition through leaderboards did not yield significant differences in learners’ behavioral engagement, measured by two metrics: the number of posts (Z = − 0.346, p = .729) and earned points (Z = -1.283, p = .200). Regarding cognitive engagement in online asynchronous discussions, no evidence supported the effectiveness of competition in gamified online discussions across various levels of knowledge construction, including initiation (Z = -1.826, p = .068), development (Z = -1.604, p = .109), and construction (Z = -1.289, p = .197). This study also scrutinized limitations and critical issues, offering essential recommendations to guide future research.

Successful problem-based learning (PBL) often requires students to collectively regulate their learning processes as a group and engage in socially shared regulation of learning (SSRL). This paper focuses on how facilitators supported SSRL in the context of middle-school game-based PBL. Using conversation analysis, this study analyzed text-based chat messages of facilitators and students collected during gameplay. The analysis revealed direct modeling strategies such as performing regulative processes, promoting group awareness, and dealing with contingency as well as indirect strategies including prompting questions and acknowledgment of regulation, and the patterns of how facilitation faded to yield responsibilities to students to regulate their own learning. The findings will inform researchers and practitioners to design prompts and develop technological tools such as adaptive scaffolding to support SSRL in PBL or other collaborative inquiry processes.

Collaborative game-based learning environments have significant potential for creating effective and engaging group learning experiences. These environments offer rich interactions between small groups of students by embedding collaborative problem solving within immersive virtual worlds. Students often share information, ask questions, negotiate, and construct explanations between themselves towards solving a common goal. However, students sometimes disengage from the learning activities, and due to the nature of collaboration, their disengagement can propagate and negatively impact others within the group. From a teacher's perspective, it can be challenging to identify disengaged students within different groups in a classroom as they need to spend a significant amount of time orchestrating the classroom. Prior work has explored automated frameworks for identifying behavioral disengagement. However, most prior work relies on a single modality for identifying disengagement. In this work, we investigate the effects of using multiple modalities to detect disengagement behaviors of students in a collaborative game-based learning environment. For that, we utilized facial video recordings and group chat messages of 26 middle school students while they were interacting with Crystal Island: EcoJourneys, a game-based learning environment for ecosystem science. Our study shows that the predictive accuracy of a unimodal model heavily relies on the modality of the ground truth, whereas multimodal models surpass the unimodal models, trading resources for accuracy. Our findings can benefit future researchers in designing behavioral engagement detection frameworks for assisting teachers in using collaborative game-based learning within their classrooms.

The primary goals of this research were to investigate the development of computational thinking (CT) skills among elementary students and to identify areas for improvement in their CT practices. Empirical investigations, accomplished in a learner-centered, problem-based learning curriculum for sixth-graders, sought to examine student proficiency in CT practices by analyzing programs developed by students. A total of 30 students participated in this study and eleven Scratch projects were analyzed. Results revealed that students showed high competence in Event, Parallelism, Sequence, and Design while low competence in Loop & Operator. Examination of the flow of their programs revealed how effectively students utilized CT concepts in CT practices. As areas for improvement, researchers identified an extremely fine-grained programming (EFGP) approach and a lack of abstraction.

This study used an explanatory mixed methods research design to examine the scaffolding strategies of middle school teachers during a design problem-based learning (PBL) unit. Both quantitative and qualitative techniques were used in data collection and analysis, and the findings were integrated for interpretation. Our analysis of classroom observations revealed that teachers relied heavily on soft scaffolding throughout the five days of instruction. In addition, both classroom observations and semi-structured teacher interviews revealed that the teachers' primary goal for scaffolding was to facilitate students' cognitive structuring. Qualitative data collected through teacher interviews were used to further explore how the teachers made sense of scaffolding. Suggestions are made for incorporating distributed scaffolding in design PBL makerspaces which tend to be highly complex and place a heavy cognitive load on learners.

Collaborative game-based learning environments offer significant promise for creating effective and engaging group learning experiences. These environments enable small groups of students to work together toward a common goal by sharing information, asking questions, and constructing explanations. However, students periodically disengage from the learning process, which negatively affects their learning, and the impacts are more severe in collaborative learning environments as disengagement can propagate, affecting participation across the group. Here, we introduce a multimodal behavioral disengagement detection framework that uses facial expression analysis in conjunction with natural language analyses of group chat. We evaluate the framework with students interacting with a collaborative game-based learning environment for middle school science education. The multimodal behavioral disengagement detection framework integrating both facial expression and group chat modalities achieves higher levels of predictive accuracy than those of baseline unimodal models.

As artificial intelligence (AI) technologies develop and even become ubiquitous, their promise for supporting teaching and learning affords new possibilities. To support teachers and teaching, AI technologies can be designed in ways that are human-centered, amplifying the teacher’s cognitive capacity and expertise, particularly with respect to supporting ambitious learning practices such as problem-based and inquiry learning. In particular, AI technologies have immense potential to support classroom orchestration – how teachers organize classroom activity across individual, small group, and whole-class scales. Effective classroom orchestration is key to supporting STEM inquiry, productive disciplinary engagement, and collaboration. However, given the practical challenges of classroom activity, effective orchestration can be difficult to achieve. Teacher orchestration systems supported by AI can assist teachers in facilitating classroom activity by offering them information about pedagogical tools (e.g., curriculum materials, technologies) and suggestions for strategies, tracking student activity in real time, and seamlessly completing managerial tasks to “free up” the teacher for supporting authentic inquiry. In this chapter, we outline design considerations that emerged as we designed such an orchestration assistant, an AI-supported teacher orchestration system. We first consider the theoretical possibilities currently available for supporting pedagogy with AI. Next, we discuss how those possibilities might be enacted, complicated, or transformed in the context of real classroom activity. We ground our discussion in design considerations from our orchestration assistant project. We conclude with some open questions about such AI-supported systems for teachers.

Scaffolding is one of the critical features in a problem-based learning environment to address challenges associated with problem solving. While transfer of responsibility is considered as an ultimate goal in scaffolding that is adaptive and contingent, it is rarely studied and practiced. Thus, the purpose of this study was to inform a deeper understanding of one middle school teacher's manner of soft scaffolding, which refers to just-in-time and contingent support, through teacher-student interaction to examine how transfer of responsibility was achieved. We investigated one middle school teacher's forms of scaffolding during a problem of food systems and supply chains related multiple aspects of sustainability and social justice issues. Using conversation analysis, three discursive patterns in scaffolding emerged: (1) shifting patterns of turn-taking organization; (2) leaving room for the students to take responsibility by giving extended wait time; and (3) extending the discussion with different examples. The paper concludes with implications for PBL teachers and researchers.

Problem-based learning (PBL) has been widely incorporated in STEM classrooms. Unfortunately, its effectiveness for foreign language teaching is less explored. This design case describes the design and implementation process of a Chinese PBL unit in a US elementary school along with the design considerations of teaching dilemma-centered instruction. We provide detailed accounts of our process of developing this PBL curriculum, learning materials, and the two rounds of implementations. We also reflected on the design process and examined the design dilemmas faced by this interdisciplinary design team. Findings show multiple design tensions, which include balancing the language and PBL teaching goals, balancing L1 and L2 use, and the communication challenges within a collaborative design project.

The purpose of this study was to explore the impact a specific socio-scientific inquiry (SSI) unit has on student achievement and attitudes over the course of two years with two different implementations of the unit. Specifically, this study addressed the following research question: What impact does participation in an SSI unit have on student achievement and attitudes? A total of 212 students in 10 different biology classes participated in the study. Data collected included student pre- and post-test results from an assessment on genetics content, results from the state-delivered end of course assessment in biology, student focus group interviews, and results from a student attitudinal survey. Student achievement results (as measured by gains from pre-test to post-test) demonstrated statistically significant increases for all students participating in the SSI unit, and were even more pronounced for students with a low level of prior knowledge of genetics content. In addition, end-of-course assessment results for students participating in the unit were significantly higher than the passing scores set by the state. Finally, student attitudes towards SSI unit activities were highly positive, and students indicated that the technology resources available to them facilitated their completion of unit activities. These findings suggest that engagement in inquiry-based instruction that incorporates the SSI model may have a positive effect on students’ knowledge of science content regardless of their level of prior knowledge and that well-implemented SSI can positively address the challenges faced by students engaged in SSI instruction.

This study investigated how a computer science (CS) problem-based curriculum impacted elementary students' CS learning and attitudes. Four sixth-grade teachers and 200 of their students participated in the study. Researchers developed a CS curriculum in collaboration with the teachers, which consisted of two main units: (1) an introduction to block-based coding and (2) a problem-based learning (PBL) applied coding project. Overall, students significantly improved their knowledge of CT concepts after the introductory block-based coding lessons and retained that knowledge after completing the PBL activities approximately three months later. Results suggest that Event and Parallelism were challenging concepts for most of the students, whereas Loop and Sequence were easily grasped by most of the students. Further analysis based on prior knowledge levels revealed that the high-prior knowledge (HK) group outperformed the low-prior knowledge (LK) group on every measure. However, LK narrowed the gap of CT concepts after the introductory block-based coding lessons. Students also communicated relatively positive attitudes towards CS at the conclusion of the PBL unit. These results provide support for further exploring the integration of inquiry-oriented instructional strategies such as PBL to support CS instruction.