Purpose This paper aims to build on work that has demonstrated the value of play or game-based learning environments and to further unpack how different kinds of play activities can support learning of academic concepts. To do so, this paper explores how students learn complex science concepts through collective embodied play by comparing two forms of play labeled as Inquiry Play and Game Play. Design/methodology/approach This study builds off of previous research that uses the Science Through Technology Enhanced Play (STEP) technology platform (Authors et al., 2015). STEP is a mixed reality platform that allows learners to playfully explore science phenomena, such as the rules of particle behavior in solid, liquid and gas, through collective embodied activity. A combination of interaction analysis and qualitative coding of teacher and student interactions are used to examine patterns in the learning processes during embodied play activities. Findings Both forms of play led to similar learning gains. However, Inquiry Play promoted more emergent, flexible modeling of underlying mechanisms while Game Play oriented students more towards “winning”. Originality/value By contrasting play environments, this paper provides new insights into how different features of play activities, as well as how teachers orient their students according to these different features, support students’ learning in collective activity. As a result, these findings can provide insights into the design of future play-based learning environments that are intended to support the learning of academic concepts.

Both professional and classroom‐based scientific communities develop and test explanatory models of the natural world. For students to take up models as tools for sensemaking, practice must be agentive (where students use and revise models for specific purposes) and conceptually productive (where students make progress on their ideas). In this paper, we explore principles to support agentive and conceptually productive modeling. One is that models can “do work”; that is, participate in students' sensemaking by offering resources, making gaps visible, or pushing back on modelers' understandings. A second is that working across, and seeking to align, multiple models—what we explain as interlocking models—supports models to do work. A third is that modeling activity can support fine‐grained conceptual progress. We detail how we used these ideas to guide and refine the design of a fifth‐grade investigation into the conservation of matter across phase change. We identify four ways that models participated in students' sensemaking as they interlocked: by providing contradictions, constraints, representational surplus, and gaps for students to engage with. We discuss how designing for models to be co‐participants in sense‐making and to interlock can provide productive paths forward for curriculum designers, researchers, and teachers.

Purpose This paper aims to show how collective embodiment with physical objects (i.e. props) support young children’s learning through the construction of liminal blends that merge physical, virtual and conceptual resources in a mixed-reality (MR) environment.. Design/methodology/approach Building on Science through Technology Enhanced Play (STEP), we apply the Learning in Embodied Activity Framework to further explore how liminal blends can help us understand learning within MR environments. Twenty-two students from a mixed first- and second-grade classroom participated in a seven-part activity sequence in the STEP environment. The authors applied interaction analysis to analyze how student’s actions performed with the physical objects helped them to construct liminal blends that allowed key concepts to be made visible and shared for collective sensemaking. Findings The authors found that conceptually productive liminal blends occurred when students constructed connections between the resources in the MR environment and coordinated their embodiment with props to represent new understandings. Originality/value This study concludes with the implications for how the design of MR environment and teachers’ facilitation in MR environment supports students in constructing liminal blends and their understanding of complex science phenomena.