Unveiling Learning Mechanisms in the Engineering Design Process: A Sequential Mixed Analysis of Primary School Students’ Cognitive Processes Through the Lens of Engineering Habits of Mind

Unveiling Learning Mechanisms in the Engineering Design Process: A Sequential Mixed Analysis of Primary School Students’ Cognitive Processes Through the Lens of Engineering Habits of Mind | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Unveiling Learning Mechanisms in the Engineering Design Process: A Sequential Mixed Analysis of Primary School Students’ Cognitive Processes Through the Lens of Engineering Habits of Mind Pui Yiu Tam, Gary K.W. Wong, Bixia Chen, Feifei Wang, Liuyufeng Li This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8794720/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background The Engineering Design Process (EDP) is considered as a signature pedagogy for cultivating the Engineering Habits of Mind (EHoMs). However, empirical studies tracing the development of the EHoMs within STEM education are limited. How these habits evolve across EDP stages and the cognitive mechanisms driving this evolution remains a “black box”. This study aims to reveal this process by tracking 63 fifth-grade students during a four-week rubber-band-powered car project. A sequential mixed analysis strategy was employed. Data from online discourse, classroom observations, design workbooks, iterative artifacts, and interviews were integrated and analyzed using Qualitative Content Analysis (QCA), Epistemic Network Analysis (ENA), and Thematic Analysis (TA). STEM education Engineering Design Process Engineering Habits of Mind Primary school Imitate-Fail-Improve pathway Sequential Mixed Analysis Epistemic Network Analysis Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 1. Introduction Science, technology, engineering, and mathematics (STEM) education has been gaining global attention over the past decades. In today’s K–12 classrooms, educators embrace STEM education as an integrative approach that connects these four disciplines, enabling students to apply theoretical knowledge in authentic lessons (Holmlund et al., 2018 ). By engaging in problem-based learning, students can achieve a deep understanding of the subjects and develop higher-order thinking (Hmelo-Silver, 2004 ). Among these fields, engineering offers a body of knowledge focused on designing and developing human-made systems, along with a systematic, constraint-driven approach to problem solving. Engineering leverages principles from science and mathematics along with technological tools to innovate and realize solutions (Honey et al., 2014 ). The process of engineering design is typically iterative rather than linear, permitting constant experimentation and enhancements until the best solution is achieved (English, 2016 ), offering a meaningful and applicable context for STEM learning (Keratithamkul et al., 2020 ). In recent years, educators have increasingly recognized the significance of engineering in STEM education and have actively promoted its integration into K–12 classrooms. The Engineering Design Process (EDP), with its clear structure and stepwise progression in addressing authentic problems, is widely regarded as a key entry point for students into the STEM integration (Wi̇Narno et al., 2020 ). Primary school students can engage with the EDP in ways that are cognitively appropriate and flexible, allowing them to apply problem-solving strategies across varying levels of task complexity (Sung & Kelley, 2023 ). Complementing the EDP, Engineering Habits of Mind (EHoMs) describe the ways of thinking and acting that engineers use when tackling new situations and challenging assignments (Lucas et al., 2014 ). EHoMs are framed as a key component of engineering literacy that can be cultivated in pre-college education and used as clear, observable learning outcomes (Bianchi & Wiskow, 2023 ; Schucker et al., 2024 ). By integrating EHoMs as educational objectives and employing the EDP as the signature pedagogy and practical pathway for cultivating them, this dual approach can advance engineering education from both goal-oriented (fostering the EHoMs) and process-oriented (implementing the EDP) perspectives (Lucas & Hanson, 2016 ). Despite these efforts, enhancing the visibility of engineering within STEM faces ongoing challenges. For instance, empirical research on primary school students’ EHoMs remains limited, and the mechanisms linking engineering practice with the cultivation of these habits are not yet well understood. Additionally, the effectiveness of the EDP as a signature pedagogy lacks fine-grained, process-level evidence. Furthermore, although integration of EHoMs with the EDP from early childhood onward has been advocated (Bianchi & Wiskow, 2023 ), empirical studies tracking students’ EHoM development across EDP stages in authentic STEM contexts are still scarce. To address these gaps, this study examined the development of fifth-grade students’ EHoMs during different stages of the EDP within a rubber-band-powered car project and to explore the underlying mechanisms of this development. Accordingly, this study aims to answer the following research questions: RQ1: What developmental trajectories of Engineering Habits of Mind (EHoMs) do students exhibit across the different stages of the Engineering Design Process (EDP) during STEM activities? RQ2: What mechanisms underlie these developmental trajectories of EHoMs across the EDP stages? This study analyzes cognitive traces based on student on-line discourse, engineering design workbooks, iterative artifacts, interview and classroom observations. To this end, it contributes to the field by: (1) providing a fine-grained, empirical account of malleability and dynamic evolution of EHoMs in upper-primary students throughout the EDP; (2) proposing the “Imitate-Fail-Improve” (IFI) pathway as a mechanism that explains how the EDP scaffold the EHoM development; and (3) delivering practical guidance for setting tiered learning objectives and creating assessments in primary STEM engineering. The work thus offers an integrated, evidence-based foundation for fostering engineering thinking in the early years STEM education. 2. Literature Review 2.1. The EDP as an Entry Point to Integrated STEM Education Given the rapid pace of technological advancement, modern education systems focus on cultivating talent capable of solving real-world problems, viewing such talent as key to success in the 21st-century workforce and society (Binkley et al., 2012 ). Integrated STEM education addresses these demands by promoting interdisciplinary learning and authentic applications, fostering process-oriented thinking to better prepare students for future complex challenges (English, 2016 ). Within the STEM integration, engineering serves as a key interdisciplinary hub, applying propositional scientific, technological, and mathematical knowledge to authentic contexts, thereby encouraging deep learning and addressing the limitations of traditional, transmission-focused teaching methods (Cunningham & Kelly, 2017 ). A core principle of engineering education is the emphasis on “doing” and “understanding design,” with design processes considered fundamental to engineering practice (Moore et al., 2014 ). Engineering design has been defined by the U.S. National Research Council [NRC] (2009) as “a purposeful, iterative process with an explicit goal governed by specifications and constraints” (p. 82). The design process typically includes defining the problem, idea generation, planning, building, testing, reflection, redesign, and communication. This iterative approach fosters “learning through designing” (Crismond & Adams, 2012 ), encouraging experimentation and solution refinement for optimal outcomes. Moreover, effective engineering design often relies on teamwork, as diverse skills are best leveraged collaboratively (Jonassen et al., 2006 ). Research shows that such tasks offer students opportunities to apply STEM knowledge in meaningful, problem-solving contexts, moving beyond rote knowledge acquisition (Lin et al., 2021 ). Prior studies have demonstrated that primary school students can engage in engineering design practices, including modeling under engineering constraints such as constructing paper bridges or earthquake-resistant buildings (English, 2019 ; English et al., 2017 ; English & King, 2015 , 2019 ), thus confirming the applicability of these practices at the primary level. To help scaffold primary students’ engagement in engineering design practices, the Engineering is Elementary (EiE) model of the EDP provides a structured five-step process (see Fig. 1 ): Ask, Imagine, Plan, Create, and Improve (Hester & Cunningham, 2007 ). The steps of the EDP and their corresponding descriptions are as follows: (1) Ask: Defining the problem and constraints; (2) Imagine: Brainstorming and comparing solutions to the problem; (3) Plan: Visualizing the solution and listing the materials; (4) Create: Following the plan to construct the product and testing it; (5) Improve: Evaluating the product, modifying it, and retesting. This model is designed to be accessible to children, enabling them to address engineering challenges in a more immersive and authentic manner (English & King, 2015 ; Lottero-Perdue et al., 2016 ). In a large-scale randomized controlled trial of the EiE program, elementary students who participated in engineering courses based on the EDP performed significantly better in engineering and science knowledge tests than those taught with traditional methods. This improvement was evident across diverse student backgrounds, demonstrating strong inclusiveness (Cunningham et al., 2020 ). By situating learning in real-world, community-based problems, the EDP prompts students to proactively integrate and apply knowledge while designing solutions, boosting engagement and serving as an effective entry point to STEM education (Lai & Cheng, 2023 ). Although the EDP has demonstrated positive effects in educational practice, existing research primarily focuses on improvements in students’ attitudes and content knowledge, wheras assessments of process skills and cognitive development remain relatively limited (Cunningham et al., 2020 ; Wi̇Narno et al., 2020 ). Consequently, even when a curriculum is deemed “effective,” its decisive components and internal mechanisms often remain a “black box” (Cunningham et al., 2020 ; Wi̇Narno et al., 2020 ). Therefore, to further clarify the effectiveness and mechanisms of the EDP, this study investigates the cognitive processes involved in learning activities. It aims to reveal how learning occurs by analyzing learners’ interactions, discourse, and the evolution of their thinking throughout the design process. 2.2. Developing the EHoMs Through the EDP The conceptual foundation of the EHoM lies in the broader framework of Habits of Mind (HoM). Costa and Kallick ( 2000 ) conceptualized HoM as teachable and learnable patterns of intellectual behavior that lead to productive outcomes, advocating for education to focus not just on knowledge transmission but also on fostering students’ lifelong intellectual development (Alhamlan et al., 2017 ). Building on HoM, the development of the EHoM framework seeks to identify the distinctive thought processes of professional engineers, providing guidance for engineering education. The EHoM informs both the thinking and actions of students as they address engineering problems, serving as a key embodiment of engineering literacy (Lucas & Hanson, 2016 ; Schucker et al., 2024 ). This study adopts the UK Royal Academy of Engineering (RAE) EHoM framework (Hanson et al., 2018 ) to investigate students’ cognitive processes during STEM activities. This choice is based on a comparison of two prevalent EHoM frameworks and aligns with the analytical focus of our research. The first framework, from the U.S. NRC (2009), identifies six EHoMs: Systems Thinking, Creativity, Optimism, Collaboration, Communication, and Attention to Ethical Considerations. The second framework, developed by the UK’s RAE through mixed-methods empirical, identifies six descriptors that characterize how engineers think and act when faced with challenges involving the making and improvement of things. These six EHoMs are Systems Thinking, Adapting, Visualizing, Problem Finding, Improving, and Creative Problem-solving (Lucas et al., 2014 ; Lucas & Hanson, 2016 ). A key similarity between the two models is that Systems Thinking is the shared core element. In comparison, the NRC model places greater emphasis on personal and social competencies (e.g., Optimism, Communication, and Ethics), aligning more closely with holistic education, whereas the RAE model focuses more directly on the cognitive processes and strategies inherent to engineering problem-solving. The RAE model (see Fig. 2 ; Hanson et al., 2018 ) was later revised into a multilayered structure. This structure was centered on “making things that work and making things work better” and comprised the original six primary EHoMs and 12 sub-habits. Ethical considerations formed an external shell, encompassing all EHoMs. This revision provided clear, actionable guidance for teaching practice. The international applicability of the RAE EHoM framework has also been officially recognized by Australian national project “Engineering 2035,” which adopts the framework, regarded the defined EHoMs as core engineering thinking, and explicitly incorporates them into the top-level planning for engineering education (Lee et al., 2022 ). Therefore, the RAE framework is adopted for its focus on cognitive patterns, specific sub-habits, actionable teaching practices, and international endorsement, and it specially aligns with our research aim to investigate the cognitive development of young learners through design tasks. To effectively cultivate EHoMs, educators should adopt signature pedagogies that use the design process to bridge knowing and doing through authentic tasks, enabling students to continuously strengthen and internalize EHoMs (Lucas & Hanson, 2016 ). Signature pedagogies have a three-layer structure: the surface level encompasses observable classroom behaviors and activities; the deep level pertains to the underlying educational principles and assumptions of the teaching methods; and the implicit level reflects the professional values and identities conveyed (Shulman, 2005 ). The EDP is recognized as a signature pedagogy for EHoMs (Lucas & Hanson, 2016 ). It features an accessible surface structure (e.g., the EiE model in Fig. 1 ) applicable across educational stages, and a deep structure that emphasizes hands-on action through building and evaluation, aligning with the RAE model’s core focus on “making things that work and making things work better.” Furthermore, by taking students through the process of generating ideas and selecting solutions, the implicit structure of the EDP fosters professional attitudes essential for balancing tensions among different EHoMs. Expanding on this foundation, the RAE has proposed a concrete instructional pathway, the Progressing to be an Engineer (PEng) Cycle framework, which explicitly integrates the EHoMs model with the EDP steps to provide a structured blueprint for school education (Bianchi & Wiskow, 2023 ). Furthermore, a preliminary empirical study of observations in an extracurricular engineering literacy club showed that primary school students can exhibit and develop EHoMs within the EDP (Schucker et al., 2024 ). However, while promising initiatives exist and the importance of incorporating EHoMs into the EDP is acknowledged, two interrelated gaps constrain progress. First, rigorous empirical validation of how EHoMs are integrated within the EDP is limited (Karatas-Aydin & Isiksal-Bostan, 2023 ; Wheeler et al., 2019 ). Second, as noted longitudinally, there is a pressing need to identify effective methods for cultivating EHoMs, especially Systems Thinking, in primary and secondary classrooms (English, et al., 2017 ). Consequently, a process-oriented investigation is required to render visible the cognitive dynamics through which the EDP fosters the EHoMs. 3. Methods This study addresses the research gaps by using EHoMs as an analytical lens to open the “black box” of the EDP as a pathway for enhancing students’ cognitive development. Guided by this lens, multiple sources of qualitative data (i.e., online discussion discourses, engineering design workbooks, artifacts, classroom observations, and interviews) were collected to provide contextualized interpretations of children’s EHoMs. For data analysis, a sequential mixed-methods analytical framework (see Fig. 3 ) was employed to investigate the development of primary school students’ EHoMs across the EDP stages. The analysis first identified quantitative patterns in students’ cognitive behaviors during STEM activities, then qualitatively interpreted the underlying mechanisms shaping those patterns. Through this approach, the study uncovered fine-grained, process-oriented evidence of students’ cognitive and collaborative development during the hands-on, design-based STEM tasks. 3.1. Context This study was conducted in a co-educational primary school in Hong Kong and focused on Grade 5 students, within the context of a larger longitudinal project that followed students from Grade 4 to Grade 6. The teaching and learning activities of the project were co-developed by researchers and teachers to integrate EDP into science education, aligning with Hong Kong’s General Studies Curriculum Guide (Educational Bureau of Hong Kong, 2017 ). Specifically, the project focused on “Strand 3: Science and Technology in Everyday Life” for Key Stage 2 (Primary 4–6, equivalent to Grade 4–6), emphasizing age-appropriate, hands-on learning activities. To keep the activities practical and manageable, the team adapted materials from the EDB Learning and Teaching Resource CD. 3.2. Participants Students from two Grade 5 classes ( N = 63; male = 30, female = 33; mean age = 10 years 7 months) participated in this study. All students had prior experience with a four-week EDP project in Grade 4 where they collaboratively designed a water filter. Students were divided into 15 teams (4–5 members each, with either two boys and two girls, or two boys and three girls), and teachers assigned group leaders to facilitate collaboration. Ethical protocols included parental consent and student assent, with anonymized identifiers (e.g., 5A-G1-S2 for Class 5A, Group 1, Student 2). 3.3. Implementation Over four weeks, students participated in weekly 50-minute project-based STEM sessions that integrated digital collaboration, interdisciplinary content, and hands-on engineering design. Using tablets, students accessed Padlet, a collaborative Web 2.0 platform that allowed real-time sharing during lessons and asynchronous communication beyond class hours. Teachers pre-structured Padlet boards into group sections, enabling the posting of ideas, sketches, images, and videos. This digital approach leveraged students’ prior EDP experience from Primary 4, facilitating a smooth transition into the current project. The interdisciplinary framework of the design challenge connected concepts across the four STEM domains: Science, focusing on energy conversion (e.g., potential to kinetic energy); Technology, emphasizing material selection and processing (e.g., the use of recyclable plastics); Engineering, centering on design constraints and the iterative EDP stages; and Mathematics, involving distance measurement and speed calculations. The project started with a no-contact delivery scenario to frame the design problem. Students were tasked with using household recyclable materials and rubber bands to build a trolley capable of transporting four AA batteries over two meters. Each group collaboratively developed a unified design, while every member constructed an individual prototype based on the shared plan. Guided through the EDP stages, students engaged first in the “Ask” stage, where teachers clarified design requirements through classroom interactive discussion. Students then collaboratively advanced through the “Imagine,” “Plan,” “Create,” and “Improve” stages, using Padlet to brainstorm, provide peer feedback, refine ideas, and document progress. Prototype construction could continue at home, with students uploading photos or videos of their models for teacher feedback. Throughout the four weeks, teachers provided formative, stage-specific guidance to ensure that all student groups maintained alignment with their project goals and engineering constraints. 3.4. Data Collection Reflecting a situative perspective on engineering learning where knowledge is dynamically constructed or reproduced and distributed through design participation (Johri & Olds, 2011 ; Roth, 1996 ), multiple data sources were collected to understand whether and how EDP stages contribute to changes in learners’ EHoMs. First, online collaborative platform data were collected as time-sequenced discourse records, providing a continuous trace of students’ EDP practices. During the EDP activities, students generated 1,088 Padlet posts across fifteen teams in two classes. These posts were compiled into Excel spreadsheets containing both metadata (e.g., post number, section, theme, author, timestamps) and textual content, including embedded images and videos. Student-produced videos were transcribed and integrated with the textual data to form a unified narrative dataset. After removing duplicates and off-task content ( n = 261), a total of 827 posts were retained and categorized according to EDP stages: “imagine & plan” (674), “create” (115), “improve” (36), and “unclassifiable (2).” Second, all four sessions of both Class 5A and Class 5B classroom observation data were gathered as field notes and video records, providing in-situ evidence. Non-intrusive observations conducted throughout the sessions, with researchers documenting field notes on student engagement, problem-solving behaviors, and group dynamics. A front-facing camera was used to record classroom interactions and collaborative work, and these recordings were later cross-referenced with field notes to enhance data validity. Third, formative material artifacts including students’ workbooks and photographs of physical artifacts, were systematically archived as inscription and artifact data to document the iterative refinement of ideas as sketches, testing records, successive prototype versions, and process reflections. Finally, reflective interview data were gathered as learning narrative. Eight semi-structured focus-group interviews (4–5 students each) drew out students’ collaborative experiences, perceptions of the design process, and reflections on learning. All interviews were transcribed verbatim for subsequent analysis. In summary, by integrating process-based cyber discourse records, direct observations, material artifacts, and participants’ narrative reflections, this study established a multidimensional chain of evidence for fine-grained analysis of the students’ EHoM development. 3.5. Data Analysis This study adopted a sequential mixed analysis strategy (Teddlie & Tashakkori, 2010 ) to integrate quantitative and qualitative results in two phases to examine both structural patterns and contextual nuances in students’ cognitive processes during collaborative EDP tasks. As shown in the analytical framework in Fig. 3 , Phase 1 investigated stage-specific distribution and relationships between EHoMs and EDP, while Phase 2 provided deeper interpretive insights into emergent patterns. Phase 1 utilized two analytical methods, Qualitative Content Analysis (QCA) and Epistemic Network Analysis (ENA) to address the close-ended RQ1 with quantitative results. QCA integrates quantitative and qualitative paradigms in educational research by categorizing and interpreting textual data to identify patterns and themes (Gläser-Zikuda et al., 2020 ; Mayring, 2015 ). ENA, a discourse analytics tool, quantifies and visualizes dynamic conceptual-procedural connections in learning interactions (e.g., Su et al., 2024 ), offering a systems perspective on complex cognitive processes (Shaffer et al., 2009 ). By applying a deductive coding framework (Mayring, 2015 ) to online discourse data from the Padlet, QCA and ENA jointly identified stage-specific distributions and co-occurrence patterns of EHoM manifestations, revealing trajectories in students’ engineering cognition. A two-stage hierarchical coding procedure was implemented to systematically identify the emergence of students’ EHoM. Each post served as an independent analytic unit, with standardized codes assigned to six primary EHoMs (e.g., Systems Thinking: ST) and their 12 sub-habits (e.g., Connecting: STs1). In Stage 1, Primary EHoM Coding, Hanson et al.’s ( 2018 ) framework (Table 1 ) was applied using binary coding (0/1) to indicate the presence or absence of each primary EHoM. A code of “1” denoted explicit evidence of a primary EHoM (e.g., A = Adapting), while “0” represented its absence. In Stage 2, Sub-EHoM Coding with Dynamic Supplementation, the posts were reanalyzed using twelve sub-EHoM descriptors (see Appendix 1; Hanson et al., 2018 ). A dynamic supplementation rule was applied: if a sub-habit (e.g., Evaluating: As2) appeared in a post that was not initially coded for its parent EHoM (“0”), the corresponding primary EHoM code (e.g., Adapting) was retroactively updated to “1.” This approach ensured that latent EHoM instances were identified, thereby enhancing the validity of the coding process. Multiple codes could be assigned to a single post to account for co-occurring EHoMs. To ensure inter-rater reliability, the first and third authors independently coded all posts. During the calibration phase, two randomly selected post groups from classes 5A and 5B were coded separately. Discrepancies were resolved through iterative discussions, refining the coding standards. Following calibration, the two authors coded the remaining posts. Cohen’s Kappa values of all primary EHoMs between two coders exceeded 0.95, demonstrating near-perfect inter-rater agreement (Landis & Koch, 1977 ). Table 1 Coding Scheme of the Six Primary EHoMs Main EHoM Description Example Systems Thinking (ST) Seeing essential connections between things, seeking out patterns, seeing whole systems and their parts and how they connect, recognizing Interdependencies, synthesizing. “The blades of the propeller are bent forward 30–40 degrees, which allows it to generate airflow when rotating and drive the car forward.” Visualizing (V) A specific form of communication, being able to move from abstract ideas to concrete, manipulating materials, mentally rehearsing alternative design solutions, and communicating them to multiple stakeholders. Student posted a sketch of the car structure to represent the ideas. Adapting (A) Using ingenuity, making something designed for one purpose suitable for another purpose by converting, modifying, transforming, adjusting, changing, re-shaping, redesigning, testing, analyzing, reflecting, rethinking. Student suggested to use 500ml bottle as the car body for it is easier to cut and make. Problem-finding (P) Being curious, deciding what the actual question is, posing a different problem or re-framing the original one, finding out if solutions already exist. “I placed the stick outside and covered it with a drinking straw; otherwise, if it is directly inserted inside the beverage bottle, it will get stuck.” Improving (I) Making things better by experimenting, designing, sketching, guessing, conjecturing, thought-experimenting, prototyping; this necessarily requires resilience, perseverance, reflection and open-mindedness. “In the improved rubber band-powered car, I tied the rubber band directly to the wheels, instead of using wind power.” Creative Problem-solving (CP) Generating ideas and solutions by applying techniques from different traditions, being resourceful, critiquing, giving and receiving feedback, seeing engineering as a “team sport”. Student expressed appreciation to peer’s work, and helped peer to solve the problem. The coded dataset then supported QCA and ENA. Within QCA, using JASP (version 0.18.3), relative frequencies of the primary EHoMs for each EDP stage (“imagine & plan,” “create,” and “improve”) were computed. The EHoM compositional shares were then presented as 100%-normalized, stage-specific bar charts, with each EHoM ordered by descending share to depict the distributional patterns at each stage. In ENA, group mean networks of the six primary EHoMs at each EDP stage were modeled using the ENA Web Tool v1.7.0. Each EHoM functioned as a code, and co-occurrence patterns formed the network edges. Students’ discourses at different EDP stages were treated as stanzas, generating individual adjacency matrices per participant. These were aggregated into group-level matrices representing co-occurrence patterns in high-dimensional space (Shaffer et al., 2016 ; Siebert-Evenstone et al., 2017 ). ENA then applied spherical normalization and singular value decomposition (SVD) to project these networks onto a two-dimensional graph for cross-stage comparison. Together, QCA and ENA provided structural patterns in forms of individual primary EHoM distributions and EHoM interactions across the stages. Building on the structural patterns of EHoM development quantified in Phase 1, Phase 2 explored the contextual factors and cognitive mechanisms underlying these patterns. In Phase 2, Thematic Analysis (TA) was employed to interpret structural patterns in the data, drawing on qualitative evidence from observations, interviews, workbooks, artifacts, and online discourse (extending beyond initial deductive coding to allow new codes to emerge). This phase addressed the open-ended RQ2, which sought to investigate why specific patterns emerged, how students experienced and articulated EHoM, and the contextual influences of artifacts and the EDP. All data sources were imported into NVivo (Version 15.0.0) for organization and analysis. Guided by Braun & Clarke’s ( 2012 ) six-phase analytical approach and using an inductive coding methodology at the TA stage, explanatory themes emerged and facilitated a thick description of the underlying mechanisms, the how and why, that shaped the observed structural patterns. Overall, this study employed a mixed analysis strategy driven by two interlocking rationales, sequential development and complementarity (Onwuegbuzie & Combs, 2011 ). Through sequential development, the Phase 1 quantitative results, serving as structural patterns, directed the Phase 2 qualitative inquiry to make a contextualized interpretation. Through complementarity, the analysis balanced the strengths and weaknesses of each method. QCA provided systematic quantification but lacked insights into the interplay among EHoM elements, as well as depth of meaning. ENA revealed structural connections among EHoM but did not explain why they existed. TA alone provided rich contextual meaning but could not illustrate broader structural patterns or quantify connections. By combining these methods in the mixed analytical framework, this study not only moved beyond revealing patterns to identify underlying mechanisms to address the multifaceted research questions, but also strengthened the validity of the findings. 4. Results 4.1. Students’ Developmental Trajectories Reflected in Structural Patterns To examine students’ EHoM developmental trajectories across the EDP, this study combines QCA and ENA to analyze both relative frequency distributions and epistemic network structures of EHoMs in Padlet discourse. The analysis shows that EHoM development was dynamic and stage-dependent rather than linear. The following sections present changes in relative frequencies of each primary EHoM, chart the evolution of their co-occurrence via ENA, and synthesize these results to delineate developmental trajectories. 4.1.1. Stage-Specific EHoM Relative Frequency Distributions The dominance of different EHoMs changes across the EDP, reflecting the shifting cognitive demands at each stage. QCA quantified the relative frequency of each primary EHoM across the “imagine & plan,” “create,” and “improve” stages of the EDP (see Table 2 ), with Fig. 4 visualizing the within-stage compositions of the EHoMs. In the initial “imagine & plan” stage (see Fig. 4a), CP accounts for 48.19% and serves as the foundation of this stage. A accounts for 25.55% and PF accounts for 16.67%. In comparison, ST and I are much lower, at 1.23% and 0.05%, indicating that both remain rare at this stage. As students moved into “create” stage (see Fig. 4b), the distribution is more balanced, with CP (29.94%), A (27.06%) and V (26.64%) occupying the top three at comparable levels, indicating coordination among multiple EHoMs when translating plans into prototypes. PF and ST are 11.46% and 4.90%, respectively, and I is close to 0.00%. In the “improve” stage (see Fig. 4c), the focus shifts to I, which emerges as central (24.82%). V and A follow at 21.81% each, highlighting the synergy between visual feedback and design adjustment. CP and PF account for 17.29% and 9.01%, respectively. ST is 5.25%, the lowest in this stage. Table 2 Relative Frequency of Students’ Emergent EHoM Across the EDP Stages EDP Stages imagine & plan create improve ST 0.025 0.104 0.194 V 0.165 0.565 0.806 A 0.509 0.574 0.806 PF 0.332 0.243 0.333 I 0.001 0.000 0.917 CP 0.960 0.635 0.639 Overall, these descriptive compositional results reveal the dynamic evolution of focal habits across EDP stages: from CP dominance in “imagine & plan,” to a relatively balanced CP–A–V pattern in “create,” and finally to a marked rise of I in “improve,” forming a multi-habit configuration together with V, A, and CP. PF maintains a steady presence across all three stages, whereas ST remains comparatively infrequent. The percentages reported above are within-stage compositional shares (row-normalized) intended to indicate stage-specific emphases rather than absolute cross-stage quantities. Figure 4. Within-stage Compositions of the EHoMs in Different Stages 4.1.2 Stage-Specific Mean Epistemic Networks of EHoMs Complementing QCA, ENA modeled the co-occurrence patterns of EHoMs, revealing the evolution of their structural relationships across EDP stages. In Fig. 5 , the projection of group network centroids shows a distinct separation by EDP stages. Each point represents a group’s EHoM network centroid, and the projection reveals that group epistemic networks cluster according to EDP stages. Squares indicate the mean centroid for all groups at each stage, and dotted rectangles show the 95% confidence intervals around these means. The first (X/SVD1) and second (Y/SVD2) dimensions explain 40.3% and 21.3% of the variance, respectively, meaning horizontal and vertical separations reflect those proportions of total variance. Differences among EDP stages are evident on both axes: along X, “imagine & plan,” “create,” and “improve” are ordered left to right; along Y, “create” is highest, while “imagine & plan” and “improve” are lower. Mann–Whitney U tests (see Table 3 ) corroborate these patterns: all stages differ significantly along X ( p < .001); along Y, “imagine & plan” and “create” differ significantly ( p < .001), “create” and “improve” differ significantly ( p < .001), whereas “imagine & plan” and “improve” do not differ significantly ( p = .550), indicating that the network structures of EHoMs are stage-specific. Table 3 Mean Epistemic Network Statistical Comparisons Between the EDP Stages EDP Stages X-Axis Y-Axis Mdn U p Mdn U p “imagine & plan” -0.31 10.00 .00* -0.09 13 .00* “create” -0.15 0.20 “imagine & plan” -0.31 0 .00* -0.09 111 .55 “improve” 0.52 -0.20 “create” -0.15 1.00 .00* 0.20 150 .00* “improve” 0.52 -0.20 The corresponding mean networks in Fig. 6 illustrate stage-specific differences in EHoM co-occurrence patterns. In each network, the six primary EHoMs are represented as nodes held at fixed positions to enable cross-network comparisons of connectivity and connection strength; line thickness between nodes indicates the strength of those connections (Shaffer et al., 2016 ; Tan et al., 2022 ). In “imagine & plan” (see Fig. 6 a), nodes A, V, PF, and CP are strongly interconnected; ST maintains weak ties to A, V, PF, and CP, while I remains largely isolated. In “create” (see Fig. 6 b), strong connections persist among A, V, and CP; ST continues to exhibit weak associations with A, V, PF, and CP, and I remains disconnected. A pivotal restructuring occurs in “improve” (see Fig. 6 c). I becomes a central hub, resulting in full connectivity among all six EHoMs. A, V, and I become tightly interconnected. Network-subtraction plots (see Fig. 7 ) further highlight differential connections among EDP stages. In Network 7a (“imagine & plan” vs. “create”), the “imagine & plan” stage dominates PF-A and PF-CP connections, while “create” shows a stronger A-V connection. In Network 7b (“imagine & plan” vs. “improve”), “imagine & plan” maintains stronger PF-CP, PF-A, PF-V, and CP-A ties, while “improve” exhibits enhanced I-CP, I-A, and I-V connectivity. In Network 7c (“create” vs. “improve”), “create” retains a stronger A-V linkage, while “improve” demonstrates significantly stronger I-CP, I-A, and I-V edges. In summary, epistemic networks reveal the dynamic development of EHoMs throughout the EDP. In students’ online discussion discourse, EHoMs frequently co-occurred, and these co-occurrence patterns evolved across the EDP stages. 4.1.3. Synthesis of Developmental Trajectories Drawing on structural patterns, including value distributions (see Table 3 and Fig. 4) and epistemic network structures (see Figs. 5 – 7 ), we identify four dynamic trajectories that characterize the development of students’ EHoMs: (1) CP, A, and V form a stable “core triangle” across all three EDP stages. They occur with steadily high within-stage share (ranking among the top 4 out of 6) across stages and remain densely interconnected within the epistemic networks, serving as the foundation for students’ collaborative cognition of EHoMs. (2) PF acts as a moderately stable element: it persists throughout the entire process and, during the “imagine & plan” stage, forges particularly strong links with the CP–A–V triangle. (3) ST is consistently marginalized, as evidenced by its lowest relative frequency among the six EHoM categories, its weak connections in the epistemic network, and its lack of significant integration into the core networks. (4) Upon entering the “improve” stage, the entire network demonstrates a breakthrough reorganization. Specifically, I evolves from an isolated node into a globally connected hub, enabling all six EHoMs to achieve full connectivity for the first time. Among these new links, the V–I connection emerges as the most pronounced and strongest tie. 4.2. Students’ Developmental Trajectories Interpreted from Explanatory Themes To further interpret students’ EHoM developmental trajectories across the EDP, Thematic Analysis was utilized to identify three key explanatory themes in their cognitive evolution. The following sections outline these themes, showing how students moved from a pragmatic, hands-on focus toward a more analytical and systems-oriented way of thinking. 4.2.1. Adapting Existing Designs: Emergence of CP–A–V with Feasibility-Oriented PF Adapting existing designs emerged as the primary strategy for addressing the engineering task, fostering an early CP–A–V configuration, Creative Problem-solving (CP), Adapting (A), and Visualizing (V). Alongside this triad, Problem Finding (PF) was also present from the outset, but mainly as feasibility-oriented, surface-level identification used to converge on a unified group design. During the “imagine & plan” stage, almost all groups (93.3%) browsed rubber-band-powered car designs via online videos and images, then adapted these ideas for their own projects (e.g., “We can refer to the method in the reference video and attach a carton to the popsicle stick to hold items.” 5A-G1-S1). Students asked option-selection questions (e.g., “Which design do you think is better?” 5A-G4-S4), checked alignments with requirements (e.g., “Are the wheels made from bottle caps? Are bottle caps recyclable?” 5A-G2-S3), and discussed material (e.g., “What material should we use for wheels?” 5B-G1-S2; “Use tape for wheels?” 5B-G4-S1). When evaluating design options, students favored what seemed feasible (e.g., “This design looks easier and more doable.” 5A-G1-S4; “This is the easiest video to follow.” 5A-G4-S2) and intentionally chose simpler solutions. Their material selections reflected the same practical mindset. They prioritized accessibility (e.g., “Plastic bottles are easy to find.” 5A-G1-S1; “The materials are easier to collect.” 5A-G3-S4) and workability (e.g., “The drink carton is easy to cut and shape, so we can choose it.”5A-G4-S2; “A plastic bottle is hard to cut, so I won’t choose it.” 5B-G4-S1), while paying limited attention to mechanical or structural properties. Within this adaptive strategy, PF emerged alongside decision-oriented discussions, focusing mainly on practical issues such as local fit and alignment, and operated in concert with CP and A to carry out feasibility-oriented option screening. This adaptive process was mediated by Visualizing (V) through sketches. As illustrated by 5B–G7 (see Fig. 8 ), students frequently used sketches to visualize ideas and guide group discussions. One member posted a sketch of the car’s components and asked for feedback on where to position the rubber band as a power source (see Fig. 8 a). This prompt led the group to adopt features from an online reference (see Fig. 8 b), refine their design, and develop an updated sketch that clarified the revised concept (see Fig. 8 c). Building on this shared visualization, they then agreed on specific materials and outlined preparation steps. In the subsequent “create” stage, students transformed 2D designs into 3D prototypes. Because the initial sketches lacked precise engineering specifications, construction became an iterative process in which students continuously reconciled conceptual sketches with material constraints and reference solutions. For instance, 5B–G7 compared various beverage containers for the car body and bottle caps for the wheels before building the prototype shown in Fig. 8 d. This material negotiation sustained the CP–A–V triad and further strengthened the interplay between A and V. Overall, the “Adapting Existing Designs” strategy consolidated an early CP–A–V cognitive triad characterized by pragmatic adaptation, visual coordination through 2D sketches, and the translation of conceptual ideas into tangible artifacts through modeling. PF was present chiefly to conduct feasibility-oriented option screening and to convergence on a shared design within the groups. This foundation laid the groundwork for more advanced EHoMs in later EDP stages, where PF progressively deepen into causal diagnosis and, together with Improving (I), enabled evidence-based refinement. 4.2.2. Deficiency-Driven Improvement: Positioning Improving (I) as the Central Hub Synergizing EHoMs As students completed the initial construction, most rubber-band-powered trolleys failed to meet engineering requirements during validation tests. This setback shifted their focus from local-fit construction to a deficiency-driven improvement process, with Improving (I) emerging as the central hub connecting and coordinating EHoMs in the “improve” stage. Under teachers’ guidance, students conducted two formal classroom trials to test whether their trolleys could carry four batteries over a two-meter distance, recording the results in their workbooks. Analysis of 62 workbooks showed that, in the first trial, 17 prototypes (27.42%) succeeded, while 45 (72.58%) fell short of the two-meter benchmark. After iterative modifications, 45 prototypes (72.58%) exceeded the target in the second trial, indicating that despite early deficiencies, the improvements were substantial and advancing in the right direction. Workbook evidence further showed that students moved beyond surface-level observations to diagnose root causes during validation tests. They not only recorded immediate observations, such as “the trolley traveled less than 2 m” (5B-G4-S1) and “its speed was too slow” (5A-G4-S4), but also analyzed the underlying causes of poor performance. As detailed in Appendix 2, four main categories of issues were identified: (1) energy transfer inefficiency (e.g., insufficient winding time, inadequate wind power); (2) friction and mechanical resistance (e.g., high friction between the axle and chassis); (3) structural and stability problems (e.g., the car was overweight, could not travel in a straight line); and (4) component mismatch (e.g., wheels too small, oversized propellers). These insights led students to propose targeted improvements. For example, one student (Workbook, 5B-G2-S2) observed that on smooth surfaces, the trolley’s wheels hardly rotated. He wrapped rubber bands around the rear wheels to increase friction and added extra batteries for weight, improving the trolley’s travel distance from 0.2 m in the first trial to 2 m in the second. In his workbook, he reflected: “Adding weight at the rear makes it go farther.” This example illustrates how the cognitive shift occurred: whereas initial adaptation relied on the CP–A–V triad for rapid prototyping with minimal analysis, prototype deficiencies during the engineering-validation phase triggered deeper engagement with EHoMs. Students began diagnosing structural–environmental interactions (deeper problem finding) and pursued explicit performance goals (i.e., a 2 m travel distance). They entered an improvement-centered, iterative cycle, continuously testing, adjusting, and refining their designs through both analytical reasoning and hands-on modification. Overall, the deficiency-driven improvement process embodies the structural dynamics of the “improve” stage, where Improving (I) acted as a central hub linking EHoMs such as Visualizing, Adapting, Problem Finding, and Creative Problem-solving. Signs of emerging Systems Thinking (ST) also appeared as students recognized and addressed interconnected structural and performance issues through deliberate interventions. 4.2.3. From Knowing to Doing : Failure-Activated Systems Thinking (ST) Systems Thinking remained at a low frequency throughout students’ EDP learning process, with signs of developmental potential in the later stages and beyond. This pattern can be partly explained by a knowing–doing gap: although students understood the theoretical principles behind the trolley’s construction, they rarely applied this knowledge proactively during design. Only when they began analyzing the causes of test failures did this latent understanding become activated, giving rise to the emergence of ST. This gap became particularly evident in relation to energy conversion, a foundational concept in the EDP learning activities. During pre-EDP instruction, students demonstrated an understanding of how energy is stored and transformed in simple converters. Classroom dialogue confirmed their conceptual grasp. For example, one student noted, “The rubber band stores potential energy when twisted, which converts to kinetic energy as it unwinds” (Class Observation, Lesson 1). However, there was no evidence that they applied these principles in their initial design choices, as student discourse indicated they prioritized replicating existing solutions over optimizing them, rarely considering factors such as energy loss minimization. Consequently, without applying energy principles, students diverged into two design paths with distinct performance outcomes. The first, as shown in Fig. 9 , featured direct conversion, in which the rubber band’s potential energy was released directly as kinetic energy to propel the trolley, resulting in high energy conversion efficiency. Thirty-seven students (59.7%) opted for this approach. During the two rounds of formal testing, fifteen students (40.5%) in Round 1 and twenty-nine students (78.4%) in Round 2 passed the 2-meter benchmark traveling distance, consistently exceeding the average rates (27.42% in Round 1 and 72.58% in Round 2). The second propulsion method, as shown in Fig. 10 , involved indirect conversion, where the potential energy stored in the rubber band was first transformed into wind energy via a propeller and then into the trolley’s kinetic energy. Twenty-five students (40.3%) selected this approach. In Round 1, only two indirect cars (8%) reached 2 meters, while twenty-three (92%) failed. Round 2 showed improvement, with sixteen cars (64%) reaching 2 meters. However, overall success of the indirect conversion designs remained below the cohort average. The pronounced struggle with the indirect conversion design prompted students to actively reflect on and analyze the causes of failure, rather than merely imitate, in the later stages of the project. In their workbooks, students noted that the chief flaw of the indirect setup was insufficient propulsive force, and they proposed targeted remedies such as “adding extra rubber bands” (e.g., 5A-G2-S4), “tightening the propeller further”(e.g., 5B-G5-S1), “enlarging the propeller” (e.g., 5A-G3-S2), or, in some cases, “switching to direct rubber-band propulsion” (e.g., 5B-G1-S2). More importantly, focus group interviews revealed that their reasoning had even evolved to consider interactions within the system (e.g., Focus Group, 5A-G2): S3 “I wouldn’t use a propeller again. It’s hard to build, and the car lacks speed and distance. The plastic bottle and other components make it heavy, and determining the correct rubber band length is tricky. If redesigning, I’d use a carton with direct twisting, like other teams did, for better speed and distance.” S1 “We’d adopt G4’s direct-drive design. As S3 said, our design is heavy. The energy converts from potential to wind, then to kinetic, losing power with each conversion. That’s why our car cannot travel as far as the others.” Therefore, through repeated testing and failure analysis, students gradually linked propulsion methods to performance outcomes and closed the knowing–doing gap. They inferred that direct energy conversion reduced energy loss and improved efficiency, and that trolley weight played a crucial role in travel distance. By applying energy principles and comparing design alternatives, they ultimately demonstrated emerging Systems Thinking. 5. Discussion 5.1. Addressing RQ1: Characterizing the Developmental Trajectory of EHoMs The longitudinal analysis revealed that students’ EHoMs developed through a nonlinear and stage-dependent trajectory, with each stage of the EDP eliciting distinct cognitive integrations. Our findings empirically validate the EDP as a signature pedagogy for cultivating primary school students’ EHoMs in STEM activities (Lucas & Hanson, 2016 ). In structured EDP tasks, these habits appeared as clear, observable behaviors, aligning with Bianchi and Wiskow’ s (2023) call to treat EHoMs as visible learning objectives throughout the process to make engineering education more practical for teachers. Through iterative practice in the EDP, students increasingly internalized EHoMs, which became durable habits of mind rather than remaining isolated, one-off skills (Ladion-De Guzmán, 2021). Furthermore, building on earlier research showing that children can develop EHoMs (e.g., Donnelly et al., 2025 ; Karatas-Aydin & Isiksal-Bostan, 2023 ; Schucker et al., 2024 ), our findings show that upper-primary students not only demonstrate continuous EHoM development but also display a hierarchical structure and stage-dependent progression. Specifically, within this energy-converter project, Creative Problem-solving, Adapting, and Visualizing served as foundational habits that persisted across all EDP stages, thereby laying the cognitive groundwork for the subsequent emergence of Improving and Systems Thinking. Additionally, the role of Problem Finding evolved dynamically throughout the EDP. Initially, Problem Finding served primarily as feasibility screening, facilitating quick convergence and establishment of a unified design direction. Later, it transitioned into gap diagnosis, driving students to progress from surface-level observations to an analysis of underlying causes. Among all EHoMs, Improving, which involves refining designs through experimentation, prototyping, and iterative reasoning (Hanson et al., 2018 ), exhibited particularly distinct stages dependency and task specificity. This suggests that the activation of Improving heavily depended on the cognitive accumulation and tangible prototypes developed during earlier stages, which together provided the necessary basis for iterative refinement. Once activated, Improving also synergized with other EHoMs to drive significant overall advancement. The distinctiveness of Improving also aligns with the recognized essential role of the “improve” stage in elementary engineering education, where students have opportunities to critically evaluate and iteratively refine designs using scientific principles and epistemic tools (Kelly & Cunningham, 2019 ). In the observed structural patterns, the rarity of Systems Thinking reflects its role as a higher-order cognitive skill (Frank & Waks, 2000 ). Systems Thinking comprises an integrated set of analytical skills for identifying, understanding, and predicting system behaviors and for designing targeted improvements. As system complexity grows, it demands correspondingly higher levels of systemic reasoning (Arnold & Wade, 2015 ). Since there is a strong connection between having a basic understanding of systems and the subsequent development of Systems Thinking, it is advisable to introduce Systems Thinking instruction as early as primary school (Assaraf & Orion, 2005 ). For primary school students, who often exhibit linear thinking and find it challenging to recognize hidden causal relationships (Nguyen & Santagata, 2021 ), developing Systems Thinking should be a gradual process. Their understanding of systems should be enhanced through consistent engagement with real-world contexts and guided reflection on concrete experiences (Assaraf & Orion, 2010 ). 5.2. Addressing RQ2: Mechanisms of EHoM Growth To further unpack the hierarchical and stage-dependent trajectories of EHoMs, this study identified an “imitate–fail–improve” (IFI) pathway to explain the developmental mechanism of these habits of mind, particularly Systems Thinking, within the EDP. Although children already exhibit basic Systems Thinking (e.g., Donnelly et al., 2025 ; Lippard et al., 2019 ), our findings suggest that students’ Systems Thinking gradually emerges and develops as the knowing–doing gap is progressively bridged through the failure-driven iterative improvement within the EDP. The engineering task began with imitation, which provided a cognitively accessible entry point as students actively sourced existing rubber-band-powered designs from open online platforms. By observing and comparing these designs embodied in physical models, students collaboratively adapted the plans considering material availability and manufacturability. They then translated these visual and structural insights into self-drawn conceptual sketches, a process that interpreted the physical models while integrating personal understanding and imagination. Based on these sketches, students collectively determined material selection and structural implementation for each part of the product, progressively advancing toward the creation of a functional prototype. Throughout the process, by adapting, collaborating, and crafting, students brought their ideas to life and developed a sense of ownership over the final design. Their ideas and creativity were tangibly captured in the sketches they drew and the models they built, which served as visual records of their thinking and practice. This iterative modeling was primarily driven by the synergistic triad of Creative Problem-solving, Adapting, and Visualizing (CP-A-V). Importantly, modeling served not only as a final output but also as an epistemic tool that externalized design thinking and enhanced problem understanding and real-time decision-making (Lammi & Denson, 2017 ). As a result, starting from existing designs reduced cognitive load (Sweller, 2020 ), allowing students to build simple, partially functional prototypes from static components, rather than having to grasp the entire integrated system upfront. Simultaneously, this imitation strategy accommodated the common novice tendency to focus on visible structures rather than system operations and the reasoning behind design choices (Crismond & Adams, 2012 ). Moreover, the prototypes developed from student-sourced exemplars potentially served as improvable models, which also provided scaffolding (Dasgupta, 2019 ). Although the online originals were functionally complete and replicable, students’ imperfect adaptation and construction due to surface understanding of the designs could introduce subsequent defects into the prototypes. Later, through engineering testing, failure acted as a catalyst for deeper exploration. When prototypes underperformed, students transitioned from superficial imitation to diagnostic inquiry. They identified flaws, proposed targeted modifications, and tested solutions while unpacking underlying physical principles. This shift was facilitated by the “improve” stage, which created conditions for productive failure (Kapur, 2008 ), through three key facilitating factors (Johnson et al., 2021 ). (1) Sufficient Opportunities: Students were able to conduct their own tests before the two formal class testing sessions, enabling active, low-stakes experimentation. (2) Fair Comparisons: Vehicle movement distance and speed were recorded and compared under unified engineering requirements. (3) Productive Strategies: Students were encouraged to analyze, diagnose design problems, and actively attempt improvements. As students assessed each optimization’s effectiveness, they engaged more deeply in scientific reasoning, elevating their domain knowledge from “remembering” and “understanding” to “applying” and “evaluating.” Moreover, collaborative learning amplified this process: cross-group critique of failure cases generated productive cognitive conflict, while comparative analysis of divergent designs broadened and deepened students’ systemic understanding through co-constructed knowledge building (Scardamalia & Bereiter, 2014 ). Notably, the evolution of students’ Problem Finding closely aligns with the cognitive patterns commonly observed in novice engineers. To begin with, students’ Problem Finding behavior was oriented toward rapid convergence. Students tended to treat the design task as a well-defined problem and rushed to propose solutions, which corresponds to the “premature commitment” tendency typical of beginners (Crismond & Adams, 2012 ). However, as prototype testing and iterative refinement progressed, students gradually developed the capacity to model dynamic system interactions, which is a hallmark of expert engineering practice (Frank & Waks, 2000 ; Schön, 2017 ). In this constructionist learning environment, the rubber-band-powered trolleys, as physical prototypes, served as shareable artifacts (Papert & Harel, 1991 ) that bridged procedural knowledge with conceptual understanding. The “IFI” pathway thus enabled students to progress from component-focused replication to system-level reasoning, thereby synergistically fostering all EHoMs. 5.3. Implications 5.3.1. Theoretical Implications This study provides nuanced empirical evidence on integrating EHoMs and the EDP in primary school STEM education. The findings support the characterization of the EDP as a “signature pedagogy” in engineering education (Lucas & Hanson, 2016 ) and further clarify the hierarchical and dynamic nature of EHoMs. Specifically, cognitive pillars such as Creative Problem-solving, Adapting, and Visualizing predominate in the initial stages. Problem Finding demonstrates a stable presence throughout the process, whereas higher-order habits of mind, including Improving and Systems Thinking, are primarily activated during the later stages of testing and iteration. As exemplified by the construction of a rubber-band-powered trolley, students initially followed an “adapting existing designs” strategy, combining Creative Problem-solving, Adapting, Visualizing, and Problem Finding to start the project, concretize ideas, tackle early challenges, and complete component-level construction. As they entered the “improve” stage, their focus shifted to the trolley as an integrated system. Through feedback-driven iteration and diagnostic testing, students engaged in purposeful optimization, activating Improving and developing their Systems Thinking progressively. This observed “imitate–fail–improve” (IFI) pathway advances the theoretical understanding of how upper-primary students progressively develop Systems Thinking, enriching the broader cognitive framework for engineering education. 5.3.2. Practical Implications The “IFI” pathway both mirrors real-world engineering practice and aligns with primary students’ cognitive development. On the one hand, in engineering, preliminary research often reveals multiple existing solutions, making adaptive modification the norm in problem solving (Jonassen et al., 2006 ). On the other hand, younger learners, who rely on concrete experiences to grasp abstract ideas, typically build knowledge incrementally through hands-on trial and error with feedback (Firdaus & Rahayu, 2019 ). Accordingly, concrete imitation can be a productive starting point for children’s cognitive development. The “IFI” pathway can be flexibly integrated into STEM instruction to meet diverse needs. For students with limited subject knowledge and emerging systems-thinking skills, early stages can emphasize guided imitation and moderate adaptation of existing designs, providing essential scaffolding for prototyping and turning low-risk failures into opportunities for diagnosis and reflection. For students with stronger foundations, systems-analysis tasks can be introduced before the “create” stage to activate and deepen Systems Thinking, prompting consideration of system boundaries, components, interactions among elements and subsystems, and overall system characteristics and behaviors, thereby fostering richer learning outcomes. Additionally, across different engineering themes, tasks can also embed observable learning-behavior indicators aligned with the EHoMs, and teachers can use ongoing observation and feedback as key elements of formative assessment. 5.4. Limitations While our study provides insights and contributions to the development of EHoM through the EDP, it is important to acknowledge certain limitations that may impact the interpretation and generalizability of the results. First, the exploration of EHoM in this study began only after the “Ask” stage, when students engaged in self-directed collaboration during the subsequent EDP stages. In the “Ask” stage, teachers led the process of defining the engineering problem (e.g., setting constraints, restating goals and introducing materials), ensuring student understanding. Consequently, the absence of student-generated documentation in the “Ask” stage restricts the identification of the structural patterns of EHoM development at this stage and limits comparisons with those in other stages, potentially hindering our holistic understanding of the developmental trajectories of EHoMs across all EDP stages. Second, the engineering task of building a rubber-band-powered trolley has multiple ready-made solutions that are readily accessible online. This situation may lead students to imitate existing solutions rather than demonstrate original thinking. If the task were conducted in an environment without internet access, students might exhibit different patterns of thinking and problem-solving strategies, thereby revealing distinctive characteristics in the development of EHoMs compared to the original context. Third, this study was conducted in a blended learning environment that emphasized online collaboration with limited in-person interaction, which to some extent constrains the external validity of the findings for primarily face-to-face or other types of learning contexts. Last but equally important, the developmental trajectory of EHoM presented may not accurately reflect actual cognitive processes. EHoMs are intended to depict patterns of thinking; however, there may be a disparity between internal thought processes and observable behavior. For example, students may engage in systematic analysis when making decisions, yet this analytical process may not be directly observable in their overt actions. 6. Conclusion This study investigated the developmental trajectories and mechanisms of EHoMs within the EDP among fifth graders. Through a rubber-band-powered trolley project, it traced the emergence of EHoMs and identified their key drivers, yielding two central findings. First, EHoMs develop along a hierarchical trajectory. The foundational CP–A–V triad (Creative Problem-solving, Adapting, Visualizing) underpinned the early “imagine & plan” and “create” stages. PF (Problem Finding) evolved from a shallow to more sophisticated level throughout the entire process, whereas advanced habits like I (Improving) and ST (Systems Thinking) were activated later during the “improve” stage. Second, this development is driven by an “imitate–fail–improve” (IFI) pathway, where imitation served as productive scaffolding, and failure acted as a critical catalyst for a cognitive shift from component-level assembly to system-level reasoning. These findings extend both theory and practice in early engineering education. Theoretically, this study offers fine-grained empirical support for framing the EDP as a signature pedagogy for EHoMs and elucidates a developmental mechanism, the IFI pathway, through which the EHoMs of synergizing and ST emerge from cumulative, reflective practice. Practically, this study provides educators with a recognizable and actionable guidance: instructional activities can be designed to scaffold productive imitation, leverage failure diagnostically, and create structured opportunities for evidence-based improvement, while assessment can incorporate observable indicators aligned with evolving EHoMs. Based on this study, we suggest directions for future work. Future studies should: (1) clarify the contextual boundaries and adaptive mechanism of the IFI pathway by examining whether the IFI sequence remains valid in more open-ended or offline settings and identifying key instructional-contextual factors that support its effective functioning; (2) develop mix-methods approaches that more closely align with cognitive processes to directly capture implicit cognitive development; and (3) uncover the microstructure and interactive relationships among EHoM sub-habits within each primary EHoM (e.g., “Connecting” and “Pattern-making” under Systems Thinking), thereby revealing micro-developmental sequences and co-evolutionary dynamics to refine the educational model of engineering thinking development. Abbreviations EDP Engineering Design Process EHoM(s) Engineering Habit(s) of Mind ENA Epistemic Network Analysis IFI Imitate-Fail-Improve QCA Qualitative Content Analysis TA Thematic Analysis Declarations Competing interests None. Ethics approval and consent to participate This study received ethical approval from the Human Research Ethics Committee (HREC) of the University of Hong Kong (Reference No. EA1803010). All participants and their parents/ guardians signed a consent form prior to participating in the study. Funding This work was supported by the General Research Fund managed by the Research Grant Council under the University Grants Committee in Hong Kong under grant number # 17603619. Author Contribution Authors’ contributionsG.W.: Conceptualization, data curation, methodology, resources, writing-review & editing, project administration, supervision, funding acquisition.P.T.: Conceptualization, data curation, methodology, formal analysis, investigation, writing-original draft, visualization, writing-review & editing.B.C.: Methodology, validation, formal analysis, writing- review & editing.F.W.: Validation, formal analysis, writing- review & editing.L.L.: Validation, visualization, writing- review & editing.All authors read and approved the final manuscript. Data Availability The datasets generated and/ or analyzed during the current study are not publicly available due to ethical restrictions regarding student anonymity and data privacy protocols approved by the Ethics Committee. References Alhamlan, S., Aljasser, H., Almajed, A., Almansour, H., & Alahmad, N. (2017). A systematic review: Using habits of mind to improve student’s thinking in class. Higher Education Studies , 8 (1), 25–35. https://doi.org/10.5539/hes.v8n1p25 Arnold, R. D., & Wade, J. P. (2015). A definition of systems thinking: A systems approach. Procedia Computer Science , 44 , 669–678. https://doi.org/10.1016/j.procs.2015.03.050 Assaraf, O. B., & Orion, N. (2005). Development of system thinking skills in the context of earth system education. Journal of Research in Science Teaching , 42 (5), 518–560. https://doi.org/10.1002/tea.20061 Assaraf, O. B., & Orion, N. (2010). System thinking skills at the elementary school level. Journal of Research in Science Teaching , 47 (5), 540–563. https://doi.org/10.1002/tea.20351 Bianchi, L., & Wiskow, J. (2023). Progressing to be an Engineer – the Approach . (1st ed.) Royal Academy of Engineering. http://raeng.org.uk/media/eegctjlq/rae-ptbae-approach_final-3.pdf Binkley, M., Erstad, O., Herman, J., Raizen, S., Ripley, M., Miller-Ricci, M., & Rumble, M. (2012). Defining twenty-first century skills. In P. Griffin, B. McGaw, & E. Care (Eds.), Assessment and teaching of 21st century skills (pp. 17–66). Springer Netherlands. https://doi.org/10.1007/978-94-007-2324-5_2 Braun, V., & Clarke, V. (2012). Thematic analysis. In H. Cooper, P. M. Camic, D. L. Long, A. T. Panter, D. Rindskopf, & K. J. Sher (Eds.), APA handbook of research methods in psychology, Vol 2. Research designs: Quantitative, qualitative, neuropsychological, and biological (pp. 57–71). American Psychological Association. https://doi.org/10.1037/13620-004 Costa, A. L., & Kallick, B. (2000). Discovering & exploring habits of mind . Association for Supervision and Curriculum Development. Crismond, D. P., & Adams, R. S. (2012). The informed design teaching and learning matrix. Journal of Engineering Education , 101 (4), 738–797. https://doi.org/10.1002/j.2168-9830.2012.tb01127.x Cunningham, C. M., & Kelly, G. J. (2017). Epistemic practices of engineering for education. Science Education , 101 (3), 486–505. https://doi.org/10.1002/sce.21271 Cunningham, C. M., Lachapelle, C. P., Brennan, R. T., Kelly, G. J., Tunis, C. S. A., & Gentry, C. A. (2020). The impact of engineering curriculum design principles on elementary students’ engineering and science learning. Journal of Research in Science Teaching , 57 (3), 423–453. https://doi.org/10.1002/tea.21601 Dasgupta, C. (2019). Improvable models as scaffolds for promoting productive disciplinary engagement in an engineering design activity. Journal of Engineering Education , 108 (3), 394–417. https://doi.org/10.1002/jee.20282 Donnelly, R., Speldewinde, C., & Bridle, H. (2025). Engineering habits of mind in preschool children at Scottish forest nurseries and Australian bush kinders. British Educational Research Journal , 51 (5), 2185–2209. https://doi.org/10.1002/berj.4164 Educational Bureau of Hong Kong (2017). General Studies for Primary Schools Curriculum Guide (Primary 1-Primary 6) . English, L. D. (2016). STEM education K-12: Perspectives on integration. International Journal of STEM Education , 3 (1). Article 3. https://doi.org/10.1186/s40594-016-0036-1 English, L. D. (2019). Learning while designing in a fourth-grade integrated STEM problem. International Journal of Technology and Design Education , 29 (5), 1011–1032. https://doi.org/10.1007/s10798-018-9482-z English, L. D., & King, D. (2019). STEM Integration in Sixth Grade: Desligning and Constructing Paper Bridges. International Journal of Science and Mathematics Education , 17 (5), 863–884. https://doi.org/10.1007/s10763-018-9912-0 English, L. D., King, D., & Smeed, J. (2017). Advancing integrated STEM learning through engineering design: Sixth-grade students’ design and construction of earthquake resistant buildings. The Journal of Educational Research , 110 (3), 255–271. https://doi.org/10.1080/00220671.2016.1264053 English, L. D., & King, D. T. (2015). STEM learning through engineering design: Fourth-grade students’ investigations in aerospace. International Journal of STEM Education , 2 (1). Article 14. https://doi.org/10.1186/s40594-015-0027-7 Firdaus, A. R., & Rahayu, G. D. S. (2019). Effect of STEM-based learning on the cognitive skills improvement. Mimbar Sekolah Dasar , 6 (2), 198–207. https://doi.org/10.17509/mimbar-sd.v6i2.17562 Frank, M., & Waks, S. (2000). Engineering systems thinking: A multifunctional definition. Systemic Practice and Action Research , 14 (3), 361–379. Gläser-Zikuda, M., Hagenauer, G., & Stephan, M. (2020). The potential of qualitative content analysis for empirical educational research. Forum Qualitative Sozialforschung / Forum: Qualitative Social Research , 21 (1). Article 7. https://doi.org/10.17169/FQS-21.1.3443 Hanson, J., Hardman, S., Luke, S., Maunders, P., & Lucas, B. (2018). Engineering the future: Training today’s teachers to develop tomorrow’s engineers . Royal Academy of Engineering. https://doi.org/10.13140/RG.2.2.22598.93765 Hester, K., & Cunningham, C. (2007). Engineering is elementary: An engineering and technology curriculum for children [Paper presentation]. 2007 Annual Conference & Exposition, Honolulu, Hawaii, United States. https://doi.org/10.18260/1-2--1469 Hmelo-Silver, C. E. (2004). Problem-based learning: what and how do students learn? Educational Psychology Review , 16 (3), 235–266. https://doi.org/10.1023/B:EDPR.0000034022.16470.f3 Holmlund, T. D., Lesseig, K., & Slavit, D. (2018). Making sense of STEM education in K-12 contexts. International Journal of STEM Education , 5 (1). https://doi.org/10.1186/s40594-018-0127-2 . Article 32. Honey, M., Pearson, G., & Schweingruber, H. A. (Eds.). (2014). STEM integration in K-12 education: Status, prospects, and an agenda for research . National Academies. Johnson, M., Kelly, G., & Cunningham, C. (2021). Failure and improvement in elementary engineering. Journal of Research in STEM Education , 7 (2), 69–92. https://doi.org/10.51355/jstem.2021.101 Johri, A., & Olds, B. M. (2011). Situated engineering learning: Bridging engineering education research and the learning sciences. Journal of Engineering Education , 100 (1), 151–185. https://doi.org/10.1002/j.2168-9830.2011.tb00007.x Jonassen, D., Strobel, J., & Lee, C. B. (2006). Everyday problem solving in engineering: Lessons for engineering educators. Journal of Engineering Education , 95 (2), 139–151. https://doi.org/10.1002/j.2168-9830.2006.tb00885.x Kapur, M. (2008). Productive failure. Cognition and Instruction , 26 (3), 379–424. https://doi.org/10.1080/07370000802212669 Karatas-Aydin, F. I., & Isiksal-Bostan, M. (2023). Engineering-based modelling experiences of elementary gifted students: An example of bridge construction. Thinking Skills and Creativity , 47 ., Article 101237. https://doi.org/10.1016/j.tsc.2023.101237 Kelly, G. J., & Cunningham, C. M. (2019). Epistemic tools in engineering design for K-12 education. Science Education , 103 (4), 1080–1111. https://doi.org/10.1002/sce.21513 Keratithamkul, K., Kim, J. N., & Roehrig, G. H. (2020). Cultural competence or deficit-based view? A qualitative approach to understanding middle school students’ experience with culturally framed engineering. International Journal of STEM Education , 7 (1). Article 26. https://doi.org/10.1186/s40594-020-00224-5 Ladion-De Guzman, M. G. (2021). Nurturing engineering habits of mind in the 21st century learners. International Forum Journal , 24 (2), 5–20. https://doi.org/10.63201/plqc6617 Lai, M. Y., & Cheng, E. (2023). Bringing engineering into primary science classrooms using engineering design and community of practice approach—An evaluation of STEM × Play program. Frontiers in Education , 8 . https://doi.org/10.3389/feduc.2023.1290857 Lammi, M. D., & Denson, C. D. (2017). Modeling as an engineering habit of mind and practice. Advances in Engineering Education , 6 (1). Landis, J. R., & Koch, G. G. (1977). The Measurement of observer agreement for categorical data. Biometrics , 33 (1), 159–174. https://doi.org/10.2307/2529310 Lee, P., Crosthwaite, C., Reidsema, C., Burnett, I., Foley, B., Hargreaves, D., King, R., Lamborn, J., Symes, M., & Wilson, J. (2022). Preparing engineers for 2035: Transforming Australia’s engineering education for emerging roles and expectations. In C. Hong & W. W. K. Ma (Eds.), Applied degree education and the future of learning (pp. 29–52). Springer Nature Singapore. https://doi.org/10.1007/978-981-16-9812-5_2 Lin, K. Y., Wu, Y. T., Hsu, Y. T., & Williams, P. J. (2021). Effects of infusing the engineering design process into STEM project-based learning to develop preservice technology teachers’ engineering design thinking. International Journal of STEM Education , 8 (1). Article 1. https://doi.org/10.1186/s40594-020-00258-9 Lippard, C. N., Lamm, M. H., Tank, K. M., & Choi, J. Y. (2019). Pre-engineering thinking and the engineering habits of mind in preschool classroom. Early Childhood Education Journal , 47 (2), 187–198. https://doi.org/10.1007/s10643-018-0898-6 Lottero-Perdue, P., Bowditch, M., Kagan, M., Robinson-Cheek, L., Webb, T., Meller, M., & Nosek, T. (2016). An engineering design process for early childhood: Trying to engineer an egg package. Science and Children , 54 (3), 70–77. Lucas, B., & Hanson, J. (2016). Thinking like an engineer: Using engineering habits of mind and signature pedagogies to redesign engineering education. International Journal of Engineering Pedagogy (iJEP) , 6 (2), 4–13. https://doi.org/10.3991/ijep.v6i2.5366 Lucas, B., Hanson, J., & Claxton, G. (2014). Thinking like an engineer: Implications for the education system . Royal Academy of Engineering . https://winchester.elsevierpure.com/en/publications/thinking-like-an-engineer-implications-for-the-education-system-3/ Mayring, P. (2015). Qualitative content analysis: Theoretical background and procedures. In A. Bikner-Ahsbahs, C. Knipping, & N. Presmeg (Eds.), Approaches to qualitative research in mathematics education (pp. 365–380). Springer Netherlands. https://doi.org/10.1007/978-94-017-9181-6_13 Moore, T. J., Glancy, A. W., Tank, K. M., Kersten, J. A., Smith, K. A., & Stohlmann, M. S. (2014). A framework for quality K-12 engineering education: Research and development. Journal of Pre-College Engineering Education Research (J-PEER) , 4 (1). Article 2. https://doi.org/10.7771/2157-9288.1069 National Research Council. (2009). Engineering in K-12 education: Understanding the status and improving the prospects . National Academies. https://doi.org/10.17226/12635 Nguyen, H., & Santagata, R. (2021). Impact of computer modeling on learning and teaching systems thinking. Journal of Research in Science Teaching , 58 (5), 661–688. https://doi.org/10.1002/tea.21674 Onwuegbuzie, A. J., & Combs, J. P. (2011). Data analysis in mixed research: A primer. International Journal of Education , 3 (1). https://doi.org/10.5296/ije.v3i1.618 Papert, S., & Harel, I. (1991). Situating constructionism. constructionism , 36 (2), 1–11. Roth, W. M. (1996). Art and artifact of children’s designing: A situated cognition perspective. Journal of the Learning Sciences , 5 (2), 129–166. https://doi.org/10.1207/s15327809jls0502_2 Scardamalia, M., & Bereiter, C. (2014). Knowledge building and knowledge creation: Theory, pedagogy, and technology. In R. K. Sawyer (Ed.), The Cambridge handbook of the learning sciences (2nd ed., pp. 397–417). Cambridge University Press. https://doi.org/10.1017/CBO9781139519526.025 Schön, D. A. (2017). The reflective practitioner: How professionals think in action . Routledge. https://doi.org/10.4324/9781315237473 Schucker, K., Mcvee, M. B., Jarmark, C. J., & Shanahan, L. E. (2024). Exploring engineering habits of mind and disciplinary literacy connections in an elementary afterschool engineering literacies club. Pedagogies: An International Journal , 1–22. https://doi.org/10.1080/1554480X.2022.2134871 Shaffer, D. W., Collier, W., & Ruis, A. R. (2016). A tutorial on epistemic network analysis: Analyzing the structure of connections in cognitive, social, and interaction data. Journal of Learning Analytics , 3 (3), 9–45. https://doi.org/10.18608/jla.2016.33.3 Shaffer, D. W., Hatfield, D., Svarovsky, G. N., Nash, P., Nulty, A., Bagley, E., Frank, K., Rupp, A. A., & Mislevy, R. (2009). Epistemic network analysis: A prototype for 21st-century assessment of learning. International Journal of Learning and Media , 1 (2), 33–53. https://doi.org/10.1162/ijlm.2009.0013 Shulman, L. S. (2005). Signature pedagogies in the professions. Daedalus , 134 (3), 52–59. https://doi.org/10.1162/0011526054622015 Siebert-Evenstone, A. L., Irgens, A., Collier, G., Swiecki, W., Ruis, Z., A. R., & Shaffer, W., D (2017). In search of conversational grain size: Modeling semantic structure using moving stanza windows. Journal of Learning Analytics , 4 (3), 123–139. https://doi.org/10.18608/jla.2017.43.7 Su, Y. S., Wang, S., & Liu, X. (2024). Using epistemic network analysis to explore primary school students’ computational thinking in pair programming learning. Journal of Educational Computing Research , 62 (2), 559–593. https://doi.org/10.1177/07356331231210560 Sung, E., & Kelley, T. R. (2023). Elementary students’ engineering design process: How young students solve engineering problems. International Journal of Science and Mathematics Education , 21 (5), 1615–1638. https://doi.org/10.1007/s10763-022-10317-y Sweller, J. (2020). Cognitive load theory and educational technology. Educational Technology Research and Development , 68 (1), 1–16. https://doi.org/10.1007/s11423-019-09701-3 Tan, Y., Hinojosa, C., Marquart, C., Ruis, A. R., & Shaffer, D. W. (2022). Epistemic network analysis visualization. In B. Wasson & S. Zörgő (Eds.), Advances in quantitative ethnography (Vol. 1522, pp. 129–143). Springer International Publishing. https://doi.org/10.1007/978-3-030-93859-8_9 Teddlie, C., & Tashakkori, A. (2010). Foundations of mixed methods research: Integrating quantitative and qualitative approaches in the social and behavioral sciences . Sage. Wheeler, L. B., Navy, S. L., Maeng, J. L., & Whitworth, B. A. (2019). Development and validation of the Classroom Observation Protocol for Engineering Design (COPED). Journal of Research in Science Teaching , 56 (9), 1285–1305. https://doi.org/10.1002/tea.21557 Wi̇Narno, N., Rusdi̇Ana, D., Samsudi̇N, A., Susi̇Lowati̇, E., Ahmad, N., & Afi̇Fah, R. M. A. (2020). The steps of the engineering design process (EDP) in science education: A systematic literature review. Journal for the Education of Gifted Young Scientists , 8 (4), 1345–1360. https://doi.org/10.17478/jegys.766201 Additional Declarations No competing interests reported. Supplementary Files Appendix1.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8794720","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":600138912,"identity":"78f64c38-5091-45c3-82d9-a5ceb02352d9","order_by":0,"name":"Pui Yiu Tam","email":"","orcid":"","institution":"University of Hong Kong","correspondingAuthor":false,"prefix":"","firstName":"Pui","middleName":"Yiu","lastName":"Tam","suffix":""},{"id":600138913,"identity":"68bcd20a-1482-49e5-9166-5c70afa5ea6c","order_by":1,"name":"Gary K.W. Wong","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAtElEQVRIiWNgGAWjYJACxgYGBjkDHgYGZoYDJGgxJl1L4gaitRgc7z38cmZbbfp2njMGzAVniNFy5lya5ca247k7e3sMmGfcIEKL2Y0cM8OH247lbjjPY8DM84EELekGpGgxfrhxW02CwVmgw3iIcZj9mTNmjDP/HTDccOZYwWEeYrwv2d5j/LHnTJ28wZnkjY95jhGhBQjYJBgYDoNZB4jTAIxAoJ/riFU8CkbBKBgFIxEAANkJP+FZi3Q3AAAAAElFTkSuQmCC","orcid":"","institution":"University of Hong Kong","correspondingAuthor":true,"prefix":"","firstName":"Gary","middleName":"K.W.","lastName":"Wong","suffix":""},{"id":600138914,"identity":"a56abec2-1138-4cd4-ba7e-031172189e89","order_by":2,"name":"Bixia Chen","email":"","orcid":"","institution":"University of Hong Kong","correspondingAuthor":false,"prefix":"","firstName":"Bixia","middleName":"","lastName":"Chen","suffix":""},{"id":600138915,"identity":"1877cccb-85d1-4a35-a178-e9eae1331417","order_by":3,"name":"Feifei Wang","email":"","orcid":"","institution":"University of Hong Kong","correspondingAuthor":false,"prefix":"","firstName":"Feifei","middleName":"","lastName":"Wang","suffix":""},{"id":600138916,"identity":"e1ae0e34-4edc-4380-812d-fbb7722a2d86","order_by":4,"name":"Liuyufeng Li","email":"","orcid":"","institution":"University of Hong Kong","correspondingAuthor":false,"prefix":"","firstName":"Liuyufeng","middleName":"","lastName":"Li","suffix":""}],"badges":[],"createdAt":"2026-02-05 09:08:51","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8794720/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8794720/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":104101988,"identity":"f6255f4d-dec1-41a3-8872-f70962addad8","added_by":"auto","created_at":"2026-03-06 20:07:19","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":17070,"visible":true,"origin":"","legend":"\u003cp\u003eEngineering design process (\u003cem\u003eHester \u0026amp; Cunningham, 2007\u003c/em\u003e)\u003c/p\u003e","description":"","filename":"Picture1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8794720/v1/bc10ceaceb4528f89ed82468.jpg"},{"id":104101993,"identity":"870e3d21-b24a-4861-928c-cf4c094a2469","added_by":"auto","created_at":"2026-03-06 20:07:19","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":40519,"visible":true,"origin":"","legend":"\u003cp\u003eEHoM from the royal academy of engineering (\u003cem\u003eHanson et al., 2018\u003c/em\u003e)\u003c/p\u003e","description":"","filename":"Picture2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8794720/v1/e0566ba4f43244c0c04b5ae3.jpg"},{"id":104101989,"identity":"5c3de858-84d7-450b-9495-6bd95e9ba0d2","added_by":"auto","created_at":"2026-03-06 20:07:19","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":74538,"visible":true,"origin":"","legend":"\u003cp\u003eAnalytical Framework of this Study\u003c/p\u003e","description":"","filename":"Picture3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8794720/v1/1aafa8b6ae5fbdbf4e9bac74.jpg"},{"id":104403338,"identity":"38eff1d1-200f-4906-a7a2-329f65f0d729","added_by":"auto","created_at":"2026-03-11 12:18:05","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":133908,"visible":true,"origin":"","legend":"\u003cp\u003eWithin-stage Compositions of the EHoMs in Different Stages\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eNote:\u003c/strong\u003e The percentagesof EHoMs for each stage sum to 100%\u003c/p\u003e","description":"","filename":"Picture4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8794720/v1/dcd14f21b864cfddf0194580.jpg"},{"id":104404196,"identity":"0766f199-3561-4307-8b03-4961b3c69e63","added_by":"auto","created_at":"2026-03-11 12:19:48","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":41053,"visible":true,"origin":"","legend":"\u003cp\u003eGroups’ Epistemic Networks Cluster at each EDP Stage\u003c/p\u003e","description":"","filename":"Picture5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8794720/v1/60b17c37dabf8e30e3d74b46.jpg"},{"id":104403611,"identity":"5b03a010-6753-466a-b537-96a992d4f9ce","added_by":"auto","created_at":"2026-03-11 12:18:41","extension":"jpeg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":176800,"visible":true,"origin":"","legend":"\u003cp\u003eStage-Specific EHoM Epistemic Mean Networks\u003c/p\u003e","description":"","filename":"floatimage8.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8794720/v1/996b99ce3ba5b687782fe1fe.jpeg"},{"id":104403387,"identity":"d44b388f-ddc7-41ee-a68e-1af038010df2","added_by":"auto","created_at":"2026-03-11 12:18:13","extension":"jpeg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":182043,"visible":true,"origin":"","legend":"\u003cp\u003eNetwork Subtraction Between the EDP Stages\u003c/p\u003e","description":"","filename":"floatimage9.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8794720/v1/b5de52ba602906b2fc5804b3.jpeg"},{"id":104101995,"identity":"a9f0fe2e-39d1-48af-bc12-960185816423","added_by":"auto","created_at":"2026-03-06 20:07:19","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":308910,"visible":true,"origin":"","legend":"\u003cp\u003eFrom Conceptual Sketch to Prototype (Padlet Post - 5B G7)\u003c/p\u003e","description":"","filename":"floatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-8794720/v1/10267e3a63e63225af84a49b.png"},{"id":104101998,"identity":"dd4395ae-37a9-438c-a9c2-c4508857c940","added_by":"auto","created_at":"2026-03-06 20:07:19","extension":"jpeg","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":576893,"visible":true,"origin":"","legend":"\u003cp\u003eDesign Using Direct Conversion (Prototype, 5A-G4-S4)\u003c/p\u003e","description":"","filename":"floatimage11.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8794720/v1/1aa1b32d76de0072d38f0669.jpeg"},{"id":104403362,"identity":"87b9a87f-b245-4501-8e89-49ebb825f9ea","added_by":"auto","created_at":"2026-03-11 12:18:09","extension":"png","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":380722,"visible":true,"origin":"","legend":"\u003cp\u003eDesign Using Indirect Conversion (Prototype, 5A-G2-S2)\u003c/p\u003e","description":"","filename":"floatimage12.png","url":"https://assets-eu.researchsquare.com/files/rs-8794720/v1/f7f8b38df031b2e01594e995.png"},{"id":104408811,"identity":"2de9582a-48a1-4ed7-9567-b2ce676f0156","added_by":"auto","created_at":"2026-03-11 12:43:29","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3123100,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8794720/v1/a2d7ce2e-a886-4b82-bd18-c50f5ff9a069.pdf"},{"id":104403267,"identity":"e4950fc3-4fba-4801-a116-b0f3462a53c3","added_by":"auto","created_at":"2026-03-11 12:17:52","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":21030,"visible":true,"origin":"","legend":"","description":"","filename":"Appendix1.docx","url":"https://assets-eu.researchsquare.com/files/rs-8794720/v1/933b6395668af59c89572b89.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Unveiling Learning Mechanisms in the Engineering Design Process: A Sequential Mixed Analysis of Primary School Students’ Cognitive Processes Through the Lens of Engineering Habits of Mind","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eScience, technology, engineering, and mathematics (STEM) education has been gaining global attention over the past decades. In today\u0026rsquo;s K\u0026ndash;12 classrooms, educators embrace STEM education as an integrative approach that connects these four disciplines, enabling students to apply theoretical knowledge in authentic lessons (Holmlund et al., \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). By engaging in problem-based learning, students can achieve a deep understanding of the subjects and develop higher-order thinking (Hmelo-Silver, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2004\u003c/span\u003e). Among these fields, engineering offers a body of knowledge focused on designing and developing human-made systems, along with a systematic, constraint-driven approach to problem solving. Engineering leverages principles from science and mathematics along with technological tools to innovate and realize solutions (Honey et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). The process of engineering design is typically iterative rather than linear, permitting constant experimentation and enhancements until the best solution is achieved (English, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2016\u003c/span\u003e), offering a meaningful and applicable context for STEM learning (Keratithamkul et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIn recent years, educators have increasingly recognized the significance of engineering in STEM education and have actively promoted its integration into K\u0026ndash;12 classrooms. The Engineering Design Process (EDP), with its clear structure and stepwise progression in addressing authentic problems, is widely regarded as a key entry point for students into the STEM integration (Wi̇Narno et al., \u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Primary school students can engage with the EDP in ways that are cognitively appropriate and flexible, allowing them to apply problem-solving strategies across varying levels of task complexity (Sung \u0026amp; Kelley, \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Complementing the EDP, Engineering Habits of Mind (EHoMs) describe the ways of thinking and acting that engineers use when tackling new situations and challenging assignments (Lucas et al., \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). EHoMs are framed as a key component of engineering literacy that can be cultivated in pre-college education and used as clear, observable learning outcomes (Bianchi \u0026amp; Wiskow, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Schucker et al., \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). By integrating EHoMs as educational objectives and employing the EDP as the signature pedagogy and practical pathway for cultivating them, this dual approach can advance engineering education from both goal-oriented (fostering the EHoMs) and process-oriented (implementing the EDP) perspectives (Lucas \u0026amp; Hanson, \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2016\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eDespite these efforts, enhancing the visibility of engineering within STEM faces ongoing challenges. For instance, empirical research on primary school students\u0026rsquo; EHoMs remains limited, and the mechanisms linking engineering practice with the cultivation of these habits are not yet well understood. Additionally, the effectiveness of the EDP as a signature pedagogy lacks fine-grained, process-level evidence. Furthermore, although integration of EHoMs with the EDP from early childhood onward has been advocated (Bianchi \u0026amp; Wiskow, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), empirical studies tracking students\u0026rsquo; EHoM development across EDP stages in authentic STEM contexts are still scarce.\u003c/p\u003e \u003cp\u003eTo address these gaps, this study examined the development of fifth-grade students\u0026rsquo; EHoMs during different stages of the EDP within a rubber-band-powered car project and to explore the underlying mechanisms of this development. Accordingly, this study aims to answer the following research questions:\u003c/p\u003e \u003cp\u003eRQ1: What developmental trajectories of Engineering Habits of Mind (EHoMs) do students exhibit across the different stages of the Engineering Design Process (EDP) during STEM activities?\u003c/p\u003e \u003cp\u003eRQ2: What mechanisms underlie these developmental trajectories of EHoMs across the EDP stages?\u003c/p\u003e \u003cp\u003eThis study analyzes cognitive traces based on student on-line discourse, engineering design workbooks, iterative artifacts, interview and classroom observations. To this end, it contributes to the field by: (1) providing a fine-grained, empirical account of malleability and dynamic evolution of EHoMs in upper-primary students throughout the EDP; (2) proposing the \u0026ldquo;Imitate-Fail-Improve\u0026rdquo; (IFI) pathway as a mechanism that explains how the EDP scaffold the EHoM development; and (3) delivering practical guidance for setting tiered learning objectives and creating assessments in primary STEM engineering. The work thus offers an integrated, evidence-based foundation for fostering engineering thinking in the early years STEM education.\u003c/p\u003e"},{"header":"2. Literature Review","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1. The EDP as an Entry Point to Integrated STEM Education\u003c/h2\u003e \u003cp\u003eGiven the rapid pace of technological advancement, modern education systems focus on cultivating talent capable of solving real-world problems, viewing such talent as key to success in the 21st-century workforce and society (Binkley et al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). Integrated STEM education addresses these demands by promoting interdisciplinary learning and authentic applications, fostering process-oriented thinking to better prepare students for future complex challenges (English, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). Within the STEM integration, engineering serves as a key interdisciplinary hub, applying propositional scientific, technological, and mathematical knowledge to authentic contexts, thereby encouraging deep learning and addressing the limitations of traditional, transmission-focused teaching methods (Cunningham \u0026amp; Kelly, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). A core principle of engineering education is the emphasis on \u0026ldquo;doing\u0026rdquo; and \u0026ldquo;understanding design,\u0026rdquo; with design processes considered fundamental to engineering practice (Moore et al., \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2014\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eEngineering design has been defined by the U.S. National Research Council [NRC] (2009) as \u0026ldquo;a purposeful, iterative process with an explicit goal governed by specifications and constraints\u0026rdquo; (p. 82). The design process typically includes defining the problem, idea generation, planning, building, testing, reflection, redesign, and communication. This iterative approach fosters \u0026ldquo;learning through designing\u0026rdquo; (Crismond \u0026amp; Adams, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2012\u003c/span\u003e), encouraging experimentation and solution refinement for optimal outcomes. Moreover, effective engineering design often relies on teamwork, as diverse skills are best leveraged collaboratively (Jonassen et al., \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2006\u003c/span\u003e). Research shows that such tasks offer students opportunities to apply STEM knowledge in meaningful, problem-solving contexts, moving beyond rote knowledge acquisition (Lin et al., \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003cp\u003ePrior studies have demonstrated that primary school students can engage in engineering design practices, including modeling under engineering constraints such as constructing paper bridges or earthquake-resistant buildings (English, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; English et al., \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; English \u0026amp; King, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2015\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2019\u003c/span\u003e), thus confirming the applicability of these practices at the primary level. To help scaffold primary students\u0026rsquo; engagement in engineering design practices, the Engineering is Elementary (EiE) model of the EDP provides a structured five-step process (see Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e): Ask, Imagine, Plan, Create, and Improve (Hester \u0026amp; Cunningham, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2007\u003c/span\u003e). The steps of the EDP and their corresponding descriptions are as follows: (1) Ask: Defining the problem and constraints; (2) Imagine: Brainstorming and comparing solutions to the problem; (3) Plan: Visualizing the solution and listing the materials; (4) Create: Following the plan to construct the product and testing it; (5) Improve: Evaluating the product, modifying it, and retesting. This model is designed to be accessible to children, enabling them to address engineering challenges in a more immersive and authentic manner (English \u0026amp; King, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Lottero-Perdue et al., \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). In a large-scale randomized controlled trial of the EiE program, elementary students who participated in engineering courses based on the EDP performed significantly better in engineering and science knowledge tests than those taught with traditional methods. This improvement was evident across diverse student backgrounds, demonstrating strong inclusiveness (Cunningham et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). By situating learning in real-world, community-based problems, the EDP prompts students to proactively integrate and apply knowledge while designing solutions, boosting engagement and serving as an effective entry point to STEM education (Lai \u0026amp; Cheng, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eAlthough the EDP has demonstrated positive effects in educational practice, existing research primarily focuses on improvements in students\u0026rsquo; attitudes and content knowledge, wheras assessments of process skills and cognitive development remain relatively limited (Cunningham et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Wi̇Narno et al., \u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Consequently, even when a curriculum is deemed \u0026ldquo;effective,\u0026rdquo; its decisive components and internal mechanisms often remain a \u0026ldquo;black box\u0026rdquo; (Cunningham et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Wi̇Narno et al., \u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Therefore, to further clarify the effectiveness and mechanisms of the EDP, this study investigates the cognitive processes involved in learning activities. It aims to reveal how learning occurs by analyzing learners\u0026rsquo; interactions, discourse, and the evolution of their thinking throughout the design process.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2. Developing the EHoMs Through the EDP\u003c/h2\u003e \u003cp\u003eThe conceptual foundation of the EHoM lies in the broader framework of Habits of Mind (HoM). Costa and Kallick (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2000\u003c/span\u003e) conceptualized HoM as teachable and learnable patterns of intellectual behavior that lead to productive outcomes, advocating for education to focus not just on knowledge transmission but also on fostering students\u0026rsquo; lifelong intellectual development (Alhamlan et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). Building on HoM, the development of the EHoM framework seeks to identify the distinctive thought processes of professional engineers, providing guidance for engineering education. The EHoM informs both the thinking and actions of students as they address engineering problems, serving as a key embodiment of engineering literacy (Lucas \u0026amp; Hanson, \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Schucker et al., \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThis study adopts the UK Royal Academy of Engineering (RAE) EHoM framework (Hanson et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) to investigate students\u0026rsquo; cognitive processes during STEM activities. This choice is based on a comparison of two prevalent EHoM frameworks and aligns with the analytical focus of our research. The first framework, from the U.S. NRC (2009), identifies six EHoMs: Systems Thinking, Creativity, Optimism, Collaboration, Communication, and Attention to Ethical Considerations. The second framework, developed by the UK\u0026rsquo;s RAE through mixed-methods empirical, identifies six descriptors that characterize how engineers think and act when faced with challenges involving the making and improvement of things. These six EHoMs are Systems Thinking, Adapting, Visualizing, Problem Finding, Improving, and Creative Problem-solving (Lucas et al., \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Lucas \u0026amp; Hanson, \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). A key similarity between the two models is that Systems Thinking is the shared core element. In comparison, the NRC model places greater emphasis on personal and social competencies (e.g., Optimism, Communication, and Ethics), aligning more closely with holistic education, whereas the RAE model focuses more directly on the cognitive processes and strategies inherent to engineering problem-solving. The RAE model (see Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e; Hanson et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) was later revised into a multilayered structure. This structure was centered on \u0026ldquo;making things that work and making things work better\u0026rdquo; and comprised the original six primary EHoMs and 12 sub-habits. Ethical considerations formed an external shell, encompassing all EHoMs. This revision provided clear, actionable guidance for teaching practice. The international applicability of the RAE EHoM framework has also been officially recognized by Australian national project \u0026ldquo;Engineering 2035,\u0026rdquo; which adopts the framework, regarded the defined EHoMs as core engineering thinking, and explicitly incorporates them into the top-level planning for engineering education (Lee et al., \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Therefore, the RAE framework is adopted for its focus on cognitive patterns, specific sub-habits, actionable teaching practices, and international endorsement, and it specially aligns with our research aim to investigate the cognitive development of young learners through design tasks.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eTo effectively cultivate EHoMs, educators should adopt signature pedagogies that use the design process to bridge knowing and doing through authentic tasks, enabling students to continuously strengthen and internalize EHoMs (Lucas \u0026amp; Hanson, \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). Signature pedagogies have a three-layer structure: the surface level encompasses observable classroom behaviors and activities; the deep level pertains to the underlying educational principles and assumptions of the teaching methods; and the implicit level reflects the professional values and identities conveyed (Shulman, \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2005\u003c/span\u003e). The EDP is recognized as a signature pedagogy for EHoMs (Lucas \u0026amp; Hanson, \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). It features an accessible surface structure (e.g., the EiE model in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e) applicable across educational stages, and a deep structure that emphasizes hands-on action through building and evaluation, aligning with the RAE model\u0026rsquo;s core focus on \u0026ldquo;making things that work and making things work better.\u0026rdquo; Furthermore, by taking students through the process of generating ideas and selecting solutions, the implicit structure of the EDP fosters professional attitudes essential for balancing tensions among different EHoMs. Expanding on this foundation, the RAE has proposed a concrete instructional pathway, the Progressing to be an Engineer (PEng) Cycle framework, which explicitly integrates the EHoMs model with the EDP steps to provide a structured blueprint for school education (Bianchi \u0026amp; Wiskow, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Furthermore, a preliminary empirical study of observations in an extracurricular engineering literacy club showed that primary school students can exhibit and develop EHoMs within the EDP (Schucker et al., \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eHowever, while promising initiatives exist and the importance of incorporating EHoMs into the EDP is acknowledged, two interrelated gaps constrain progress. First, rigorous empirical validation of how EHoMs are integrated within the EDP is limited (Karatas-Aydin \u0026amp; Isiksal-Bostan, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Wheeler et al., \u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Second, as noted longitudinally, there is a pressing need to identify effective methods for cultivating EHoMs, especially Systems Thinking, in primary and secondary classrooms (English, et al., \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). Consequently, a process-oriented investigation is required to render visible the cognitive dynamics through which the EDP fosters the EHoMs.\u003c/p\u003e \u003c/div\u003e"},{"header":"3. Methods","content":"\u003cp\u003eThis study addresses the research gaps by using EHoMs as an analytical lens to open the \u0026ldquo;black box\u0026rdquo; of the EDP as a pathway for enhancing students\u0026rsquo; cognitive development. Guided by this lens, multiple sources of qualitative data (i.e., online discussion discourses, engineering design workbooks, artifacts, classroom observations, and interviews) were collected to provide contextualized interpretations of children\u0026rsquo;s EHoMs. For data analysis, a sequential mixed-methods analytical framework (see Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e) was employed to investigate the development of primary school students\u0026rsquo; EHoMs across the EDP stages. The analysis first identified quantitative patterns in students\u0026rsquo; cognitive behaviors during STEM activities, then qualitatively interpreted the underlying mechanisms shaping those patterns. Through this approach, the study uncovered fine-grained, process-oriented evidence of students\u0026rsquo; cognitive and collaborative development during the hands-on, design-based STEM tasks.\u003c/p\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e3.1. Context\u003c/h2\u003e \u003cp\u003eThis study was conducted in a co-educational primary school in Hong Kong and focused on Grade 5 students, within the context of a larger longitudinal project that followed students from Grade 4 to Grade 6. The teaching and learning activities of the project were co-developed by researchers and teachers to integrate EDP into science education, aligning with Hong Kong\u0026rsquo;s General Studies Curriculum Guide (Educational Bureau of Hong Kong, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). Specifically, the project focused on \u0026ldquo;Strand 3: Science and Technology in Everyday Life\u0026rdquo; for Key Stage 2 (Primary 4\u0026ndash;6, equivalent to Grade 4\u0026ndash;6), emphasizing age-appropriate, hands-on learning activities. To keep the activities practical and manageable, the team adapted materials from the EDB Learning and Teaching Resource CD.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e3.2. Participants\u003c/h2\u003e \u003cp\u003eStudents from two Grade 5 classes (\u003cem\u003eN\u003c/em\u003e\u0026thinsp;=\u0026thinsp;63; male\u0026thinsp;=\u0026thinsp;30, female\u0026thinsp;=\u0026thinsp;33; mean age\u0026thinsp;=\u0026thinsp;10 years 7 months) participated in this study. All students had prior experience with a four-week EDP project in Grade 4 where they collaboratively designed a water filter. Students were divided into 15 teams (4\u0026ndash;5 members each, with either two boys and two girls, or two boys and three girls), and teachers assigned group leaders to facilitate collaboration. Ethical protocols included parental consent and student assent, with anonymized identifiers (e.g., 5A-G1-S2 for Class 5A, Group 1, Student 2).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e3.3. Implementation\u003c/h2\u003e \u003cp\u003eOver four weeks, students participated in weekly 50-minute project-based STEM sessions that integrated digital collaboration, interdisciplinary content, and hands-on engineering design. Using tablets, students accessed Padlet, a collaborative Web 2.0 platform that allowed real-time sharing during lessons and asynchronous communication beyond class hours. Teachers pre-structured Padlet boards into group sections, enabling the posting of ideas, sketches, images, and videos. This digital approach leveraged students\u0026rsquo; prior EDP experience from Primary 4, facilitating a smooth transition into the current project.\u003c/p\u003e \u003cp\u003eThe interdisciplinary framework of the design challenge connected concepts across the four STEM domains: Science, focusing on energy conversion (e.g., potential to kinetic energy); Technology, emphasizing material selection and processing (e.g., the use of recyclable plastics); Engineering, centering on design constraints and the iterative EDP stages; and Mathematics, involving distance measurement and speed calculations. The project started with a no-contact delivery scenario to frame the design problem. Students were tasked with using household recyclable materials and rubber bands to build a trolley capable of transporting four AA batteries over two meters. Each group collaboratively developed a unified design, while every member constructed an individual prototype based on the shared plan. Guided through the EDP stages, students engaged first in the \u0026ldquo;Ask\u0026rdquo; stage, where teachers clarified design requirements through classroom interactive discussion. Students then collaboratively advanced through the \u0026ldquo;Imagine,\u0026rdquo; \u0026ldquo;Plan,\u0026rdquo; \u0026ldquo;Create,\u0026rdquo; and \u0026ldquo;Improve\u0026rdquo; stages, using Padlet to brainstorm, provide peer feedback, refine ideas, and document progress. Prototype construction could continue at home, with students uploading photos or videos of their models for teacher feedback. Throughout the four weeks, teachers provided formative, stage-specific guidance to ensure that all student groups maintained alignment with their project goals and engineering constraints.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e3.4. Data Collection\u003c/h2\u003e \u003cp\u003eReflecting a situative perspective on engineering learning where knowledge is dynamically constructed or reproduced and distributed through design participation (Johri \u0026amp; Olds, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2011\u003c/span\u003e; Roth, \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e1996\u003c/span\u003e), multiple data sources were collected to understand whether and how EDP stages contribute to changes in learners\u0026rsquo; EHoMs. First, online collaborative platform data were collected as time-sequenced discourse records, providing a continuous trace of students\u0026rsquo; EDP practices. During the EDP activities, students generated 1,088 Padlet posts across fifteen teams in two classes. These posts were compiled into Excel spreadsheets containing both metadata (e.g., post number, section, theme, author, timestamps) and textual content, including embedded images and videos. Student-produced videos were transcribed and integrated with the textual data to form a unified narrative dataset. After removing duplicates and off-task content (\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;261), a total of 827 posts were retained and categorized according to EDP stages: \u0026ldquo;imagine \u0026amp; plan\u0026rdquo; (674), \u0026ldquo;create\u0026rdquo; (115), \u0026ldquo;improve\u0026rdquo; (36), and \u0026ldquo;unclassifiable (2).\u0026rdquo; Second, all four sessions of both Class 5A and Class 5B classroom observation data were gathered as field notes and video records, providing in-situ evidence. Non-intrusive observations conducted throughout the sessions, with researchers documenting field notes on student engagement, problem-solving behaviors, and group dynamics. A front-facing camera was used to record classroom interactions and collaborative work, and these recordings were later cross-referenced with field notes to enhance data validity. Third, formative material artifacts including students\u0026rsquo; workbooks and photographs of physical artifacts, were systematically archived as inscription and artifact data to document the iterative refinement of ideas as sketches, testing records, successive prototype versions, and process reflections. Finally, reflective interview data were gathered as learning narrative. Eight semi-structured focus-group interviews (4\u0026ndash;5 students each) drew out students\u0026rsquo; collaborative experiences, perceptions of the design process, and reflections on learning. All interviews were transcribed verbatim for subsequent analysis. In summary, by integrating process-based cyber discourse records, direct observations, material artifacts, and participants\u0026rsquo; narrative reflections, this study established a multidimensional chain of evidence for fine-grained analysis of the students\u0026rsquo; EHoM development.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003e3.5. Data Analysis\u003c/h2\u003e \u003cp\u003eThis study adopted a sequential mixed analysis strategy (Teddlie \u0026amp; Tashakkori, \u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e2010\u003c/span\u003e) to integrate quantitative and qualitative results in two phases to examine both structural patterns and contextual nuances in students\u0026rsquo; cognitive processes during collaborative EDP tasks. As shown in the analytical framework in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, Phase 1 investigated stage-specific distribution and relationships between EHoMs and EDP, while Phase 2 provided deeper interpretive insights into emergent patterns.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003ePhase 1 utilized two analytical methods, Qualitative Content Analysis (QCA) and Epistemic Network Analysis (ENA) to address the close-ended RQ1 with quantitative results. QCA integrates quantitative and qualitative paradigms in educational research by categorizing and interpreting textual data to identify patterns and themes (Gl\u0026auml;ser-Zikuda et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Mayring, \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). ENA, a discourse analytics tool, quantifies and visualizes dynamic conceptual-procedural connections in learning interactions (e.g., Su et al., \u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), offering a systems perspective on complex cognitive processes (Shaffer et al., \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e2009\u003c/span\u003e). By applying a deductive coding framework (Mayring, \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2015\u003c/span\u003e) to online discourse data from the Padlet, QCA and ENA jointly identified stage-specific distributions and co-occurrence patterns of EHoM manifestations, revealing trajectories in students\u0026rsquo; engineering cognition.\u003c/p\u003e \u003cp\u003eA two-stage hierarchical coding procedure was implemented to systematically identify the emergence of students\u0026rsquo; EHoM. Each post served as an independent analytic unit, with standardized codes assigned to six primary EHoMs (e.g., Systems Thinking: ST) and their 12 sub-habits (e.g., Connecting: STs1). In Stage 1, Primary EHoM Coding, Hanson et al.\u0026rsquo;s (\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) framework (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e) was applied using binary coding (0/1) to indicate the presence or absence of each primary EHoM. A code of \u0026ldquo;1\u0026rdquo; denoted explicit evidence of a primary EHoM (e.g., A\u0026thinsp;=\u0026thinsp;Adapting), while \u0026ldquo;0\u0026rdquo; represented its absence. In Stage 2, Sub-EHoM Coding with Dynamic Supplementation, the posts were reanalyzed using twelve sub-EHoM descriptors (see Appendix 1; Hanson et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). A dynamic supplementation rule was applied: if a sub-habit (e.g., Evaluating: As2) appeared in a post that was not initially coded for its parent EHoM (\u0026ldquo;0\u0026rdquo;), the corresponding primary EHoM code (e.g., Adapting) was retroactively updated to \u0026ldquo;1.\u0026rdquo; This approach ensured that latent EHoM instances were identified, thereby enhancing the validity of the coding process. Multiple codes could be assigned to a single post to account for co-occurring EHoMs. To ensure inter-rater reliability, the first and third authors independently coded all posts. During the calibration phase, two randomly selected post groups from classes 5A and 5B were coded separately. Discrepancies were resolved through iterative discussions, refining the coding standards. Following calibration, the two authors coded the remaining posts. Cohen\u0026rsquo;s Kappa values of all primary EHoMs between two coders exceeded 0.95, demonstrating near-perfect inter-rater agreement (Landis \u0026amp; Koch, \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e1977\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eCoding Scheme of the Six Primary EHoMs\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMain EHoM\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDescription\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eExample\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSystems Thinking (ST)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSeeing essential connections between things, seeking out patterns, seeing whole systems and their parts and how they connect, recognizing Interdependencies, synthesizing.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026ldquo;The blades of the propeller are bent forward 30\u0026ndash;40 degrees, which allows it to generate airflow when rotating and drive the car forward.\u0026rdquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVisualizing (V)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eA specific form of communication, being able to move from abstract ideas to concrete, manipulating materials, mentally rehearsing alternative design solutions, and communicating them to multiple stakeholders.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eStudent posted a sketch of the car structure to represent the ideas.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAdapting (A)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eUsing ingenuity, making something designed for one purpose suitable for another purpose by converting, modifying, transforming, adjusting, changing, re-shaping, redesigning, testing, analyzing, reflecting, rethinking.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eStudent suggested to use 500ml bottle as the car body for it is easier to cut and make.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eProblem-finding (P)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eBeing curious, deciding what the actual question is, posing a different problem or re-framing the original one, finding out if solutions already exist.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026ldquo;I placed the stick outside and covered it with a drinking straw; otherwise, if it is directly inserted inside the beverage bottle, it will get stuck.\u0026rdquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eImproving (I)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMaking things better by experimenting, designing, sketching, guessing, conjecturing, thought-experimenting, prototyping; this necessarily requires resilience, perseverance, reflection and open-mindedness.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026ldquo;In the improved rubber band-powered car, I tied the rubber band directly to the wheels, instead of using wind power.\u0026rdquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCreative Problem-solving (CP)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGenerating ideas and solutions by applying techniques from different traditions, being resourceful, critiquing, giving and receiving feedback, seeing engineering as a \u0026ldquo;team sport\u0026rdquo;.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eStudent expressed appreciation to peer\u0026rsquo;s work, and helped peer to solve the problem.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThe coded dataset then supported QCA and ENA. Within QCA, using JASP (version 0.18.3), relative frequencies of the primary EHoMs for each EDP stage (\u0026ldquo;imagine \u0026amp; plan,\u0026rdquo; \u0026ldquo;create,\u0026rdquo; and \u0026ldquo;improve\u0026rdquo;) were computed. The EHoM compositional shares were then presented as 100%-normalized, stage-specific bar charts, with each EHoM ordered by descending share to depict the distributional patterns at each stage. In ENA, group mean networks of the six primary EHoMs at each EDP stage were modeled using the ENA Web Tool v1.7.0. Each EHoM functioned as a code, and co-occurrence patterns formed the network edges. Students\u0026rsquo; discourses at different EDP stages were treated as stanzas, generating individual adjacency matrices per participant. These were aggregated into group-level matrices representing co-occurrence patterns in high-dimensional space (Shaffer et al., \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Siebert-Evenstone et al., \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). ENA then applied spherical normalization and singular value decomposition (SVD) to project these networks onto a two-dimensional graph for cross-stage comparison. Together, QCA and ENA provided structural patterns in forms of individual primary EHoM distributions and EHoM interactions across the stages.\u003c/p\u003e \u003cp\u003eBuilding on the structural patterns of EHoM development quantified in Phase 1, Phase 2 explored the contextual factors and cognitive mechanisms underlying these patterns. In Phase 2, Thematic Analysis (TA) was employed to interpret structural patterns in the data, drawing on qualitative evidence from observations, interviews, workbooks, artifacts, and online discourse (extending beyond initial deductive coding to allow new codes to emerge). This phase addressed the open-ended RQ2, which sought to investigate why specific patterns emerged, how students experienced and articulated EHoM, and the contextual influences of artifacts and the EDP. All data sources were imported into NVivo (Version 15.0.0) for organization and analysis. Guided by Braun \u0026amp; Clarke\u0026rsquo;s (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2012\u003c/span\u003e) six-phase analytical approach and using an inductive coding methodology at the TA stage, explanatory themes emerged and facilitated a thick description of the underlying mechanisms, the how and why, that shaped the observed structural patterns.\u003c/p\u003e \u003cp\u003eOverall, this study employed a mixed analysis strategy driven by two interlocking rationales, sequential development and complementarity (Onwuegbuzie \u0026amp; Combs, \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2011\u003c/span\u003e). Through sequential development, the Phase 1 quantitative results, serving as structural patterns, directed the Phase 2 qualitative inquiry to make a contextualized interpretation. Through complementarity, the analysis balanced the strengths and weaknesses of each method. QCA provided systematic quantification but lacked insights into the interplay among EHoM elements, as well as depth of meaning. ENA revealed structural connections among EHoM but did not explain why they existed. TA alone provided rich contextual meaning but could not illustrate broader structural patterns or quantify connections. By combining these methods in the mixed analytical framework, this study not only moved beyond revealing patterns to identify underlying mechanisms to address the multifaceted research questions, but also strengthened the validity of the findings.\u003c/p\u003e \u003c/div\u003e"},{"header":"4. Results","content":"\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\n \u003ch2\u003e4.1. Students\u0026rsquo; Developmental Trajectories Reflected in Structural Patterns\u003c/h2\u003e\n \u003cp\u003eTo examine students\u0026rsquo; EHoM developmental trajectories across the EDP, this study combines QCA and ENA to analyze both relative frequency distributions and epistemic network structures of EHoMs in Padlet discourse. The analysis shows that EHoM development was dynamic and stage-dependent rather than linear. The following sections present changes in relative frequencies of each primary EHoM, chart the evolution of their co-occurrence via ENA, and synthesize these results to delineate developmental trajectories.\u003c/p\u003e\n \u003cdiv id=\"Sec13\" class=\"Section3\"\u003e\n \u003ch2\u003e4.1.1. Stage-Specific EHoM Relative Frequency Distributions\u003c/h2\u003e\n \u003cp\u003eThe dominance of different EHoMs changes across the EDP, reflecting the shifting cognitive demands at each stage. QCA quantified the relative frequency of each primary EHoM across the \u0026ldquo;imagine \u0026amp; plan,\u0026rdquo; \u0026ldquo;create,\u0026rdquo; and \u0026ldquo;improve\u0026rdquo; stages of the EDP (see Table\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e), with Fig. 4 visualizing the within-stage compositions of the EHoMs. In the initial \u0026ldquo;imagine \u0026amp; plan\u0026rdquo; stage (see Fig. 4a), CP accounts for 48.19% and serves as the foundation of this stage. A accounts for 25.55% and PF accounts for 16.67%. In comparison, ST and I are much lower, at 1.23% and 0.05%, indicating that both remain rare at this stage. As students moved into \u0026ldquo;create\u0026rdquo; stage (see Fig. 4b), the distribution is more balanced, with CP (29.94%), A (27.06%) and V (26.64%) occupying the top three at comparable levels, indicating coordination among multiple EHoMs when translating plans into prototypes. PF and ST are 11.46% and 4.90%, respectively, and I is close to 0.00%. In the \u0026ldquo;improve\u0026rdquo; stage (see Fig. 4c), the focus shifts to I, which emerges as central (24.82%). V and A follow at 21.81% each, highlighting the synergy between visual feedback and design adjustment. CP and PF account for 17.29% and 9.01%, respectively. ST is 5.25%, the lowest in this stage.\u003c/p\u003e\n \u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u0026nbsp;\u003ctable id=\"Tab2\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eRelative Frequency of Students\u0026rsquo; Emergent EHoM Across the EDP Stages\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" rowspan=\"2\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\" colspan=\"3\"\u003e\n \u003cp\u003eEDP Stages\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eimagine \u0026amp; plan\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ecreate\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eimprove\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eST\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.025\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.104\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.194\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eV\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.165\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.565\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.806\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.509\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.574\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.806\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.332\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.243\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.333\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.917\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eCP\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.960\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.635\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.639\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n \u003cp\u003eOverall, these descriptive compositional results reveal the dynamic evolution of focal habits across EDP stages: from CP dominance in \u0026ldquo;imagine \u0026amp; plan,\u0026rdquo; to a relatively balanced CP\u0026ndash;A\u0026ndash;V pattern in \u0026ldquo;create,\u0026rdquo; and finally to a marked rise of I in \u0026ldquo;improve,\u0026rdquo; forming a multi-habit configuration together with V, A, and CP. PF maintains a steady presence across all three stages, whereas ST remains comparatively infrequent. The percentages reported above are within-stage compositional shares (row-normalized) intended to indicate stage-specific emphases rather than absolute cross-stage quantities.\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eFigure 4.\u003c/strong\u003e Within-stage Compositions of the EHoMs in Different Stages\u003c/p\u003e\n \u003c/div\u003e\n \u003cdiv id=\"Sec14\" class=\"Section3\"\u003e\n \u003ch2\u003e4.1.2 Stage-Specific Mean Epistemic Networks of EHoMs\u003c/h2\u003e\n \u003cp\u003eComplementing QCA, ENA modeled the co-occurrence patterns of EHoMs, revealing the evolution of their structural relationships across EDP stages. In Fig. \u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003e, the projection of group network centroids shows a distinct separation by EDP stages. Each point represents a group\u0026rsquo;s EHoM network centroid, and the projection reveals that group epistemic networks cluster according to EDP stages. Squares indicate the mean centroid for all groups at each stage, and dotted rectangles show the 95% confidence intervals around these means. The first (X/SVD1) and second (Y/SVD2) dimensions explain 40.3% and 21.3% of the variance, respectively, meaning horizontal and vertical separations reflect those proportions of total variance. Differences among EDP stages are evident on both axes: along X, \u0026ldquo;imagine \u0026amp; plan,\u0026rdquo; \u0026ldquo;create,\u0026rdquo; and \u0026ldquo;improve\u0026rdquo; are ordered left to right; along Y, \u0026ldquo;create\u0026rdquo; is highest, while \u0026ldquo;imagine \u0026amp; plan\u0026rdquo; and \u0026ldquo;improve\u0026rdquo; are lower. Mann\u0026ndash;Whitney U tests (see Table \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e) corroborate these patterns: all stages differ significantly along X (\u003cem\u003ep\u003c/em\u003e \u0026lt; .001); along Y, \u0026ldquo;imagine \u0026amp; plan\u0026rdquo; and \u0026ldquo;create\u0026rdquo; differ significantly (\u003cem\u003ep\u003c/em\u003e \u0026lt; .001), \u0026ldquo;create\u0026rdquo; and \u0026ldquo;improve\u0026rdquo; differ significantly (\u003cem\u003ep\u003c/em\u003e \u0026lt; .001), whereas \u0026ldquo;imagine \u0026amp; plan\u0026rdquo; and \u0026ldquo;improve\u0026rdquo; do not differ significantly (\u003cem\u003ep =\u003c/em\u003e .550), indicating that the network structures of EHoMs are stage-specific.\u0026nbsp;\u003c/p\u003e\n \u003ctable id=\"Tab3\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eMean Epistemic Network Statistical Comparisons Between the EDP Stages\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003eEDP Stages\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"3\"\u003e\n \u003cp\u003eX-Axis\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"3\"\u003e\n \u003cp\u003eY-Axis\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMdn\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003eU\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003ep\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMdn\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003eU\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003ep\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026ldquo;imagine \u0026amp; plan\u0026rdquo;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-0.31\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003e10.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003e.00*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-0.09\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003e.00*\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026ldquo;create\u0026rdquo;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-0.15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.20\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026ldquo;imagine \u0026amp; plan\u0026rdquo;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-0.31\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003e.00*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-0.09\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003e111\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003e.55\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026ldquo;improve\u0026rdquo;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.52\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-0.20\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026ldquo;create\u0026rdquo;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-0.15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003e1.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003e.00*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.20\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003e.00*\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026ldquo;improve\u0026rdquo;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.52\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-0.20\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003cp\u003e\u003c/p\u003e\n \u003cp\u003eThe corresponding mean networks in Fig. \u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003e illustrate stage-specific differences in EHoM co-occurrence patterns. In each network, the six primary EHoMs are represented as nodes held at fixed positions to enable cross-network comparisons of connectivity and connection strength; line thickness between nodes indicates the strength of those connections (Shaffer et al., \u003cspan class=\"CitationRef\"\u003e2016\u003c/span\u003e; Tan et al., \u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e). In \u0026ldquo;imagine \u0026amp; plan\u0026rdquo; (see Fig. \u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003ea), nodes A, V, PF, and CP are strongly interconnected; ST maintains weak ties to A, V, PF, and CP, while I remains largely isolated. In \u0026ldquo;create\u0026rdquo; (see Fig. \u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003eb), strong connections persist among A, V, and CP; ST continues to exhibit weak associations with A, V, PF, and CP, and I remains disconnected. A pivotal restructuring occurs in \u0026ldquo;improve\u0026rdquo; (see Fig. \u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003ec). I becomes a central hub, resulting in full connectivity among all six EHoMs. A, V, and I become tightly interconnected. Network-subtraction plots (see Fig. \u003cspan class=\"InternalRef\"\u003e7\u003c/span\u003e) further highlight differential connections among EDP stages. In Network 7a (\u0026ldquo;imagine \u0026amp; plan\u0026rdquo; vs. \u0026ldquo;create\u0026rdquo;), the \u0026ldquo;imagine \u0026amp; plan\u0026rdquo; stage dominates PF-A and PF-CP connections, while \u0026ldquo;create\u0026rdquo; shows a stronger A-V connection. In Network 7b (\u0026ldquo;imagine \u0026amp; plan\u0026rdquo; vs. \u0026ldquo;improve\u0026rdquo;), \u0026ldquo;imagine \u0026amp; plan\u0026rdquo; maintains stronger PF-CP, PF-A, PF-V, and CP-A ties, while \u0026ldquo;improve\u0026rdquo; exhibits enhanced I-CP, I-A, and I-V connectivity. In Network 7c (\u0026ldquo;create\u0026rdquo; vs. \u0026ldquo;improve\u0026rdquo;), \u0026ldquo;create\u0026rdquo; retains a stronger A-V linkage, while \u0026ldquo;improve\u0026rdquo; demonstrates significantly stronger I-CP, I-A, and I-V edges. In summary, epistemic networks reveal the dynamic development of EHoMs throughout the EDP. In students\u0026rsquo; online discussion discourse, EHoMs frequently co-occurred, and these co-occurrence patterns evolved across the EDP stages.\u003c/p\u003e\n \u003c/div\u003e\n \u003cdiv id=\"Sec15\" class=\"Section3\"\u003e\n \u003ch2\u003e4.1.3. Synthesis of Developmental Trajectories\u003c/h2\u003e\n \u003cp\u003eDrawing on structural patterns, including value distributions (see Table \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e and Fig. 4) and epistemic network structures (see Figs. \u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003e\u0026ndash;\u003cspan class=\"InternalRef\"\u003e7\u003c/span\u003e), we identify four dynamic trajectories that characterize the development of students\u0026rsquo; EHoMs: (1) CP, A, and V form a stable \u0026ldquo;core triangle\u0026rdquo; across all three EDP stages. They occur with steadily high within-stage share (ranking among the top 4 out of 6) across stages and remain densely interconnected within the epistemic networks, serving as the foundation for students\u0026rsquo; collaborative cognition of EHoMs. (2) PF acts as a moderately stable element: it persists throughout the entire process and, during the \u0026ldquo;imagine \u0026amp; plan\u0026rdquo; stage, forges particularly strong links with the CP\u0026ndash;A\u0026ndash;V triangle. (3) ST is consistently marginalized, as evidenced by its lowest relative frequency among the six EHoM categories, its weak connections in the epistemic network, and its lack of significant integration into the core networks. (4) Upon entering the \u0026ldquo;improve\u0026rdquo; stage, the entire network demonstrates a breakthrough reorganization. Specifically, I evolves from an isolated node into a globally connected hub, enabling all six EHoMs to achieve full connectivity for the first time. Among these new links, the V\u0026ndash;I connection emerges as the most pronounced and strongest tie.\u003c/p\u003e\n \u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec16\" class=\"Section2\"\u003e\n \u003ch2\u003e4.2. Students\u0026rsquo; Developmental Trajectories Interpreted from Explanatory Themes\u003c/h2\u003e\n \u003cp\u003eTo further interpret students\u0026rsquo; EHoM developmental trajectories across the EDP, Thematic Analysis was utilized to identify three key explanatory themes in their cognitive evolution. The following sections outline these themes, showing how students moved from a pragmatic, hands-on focus toward a more analytical and systems-oriented way of thinking.\u003c/p\u003e\n \u003cdiv id=\"Sec17\" class=\"Section3\"\u003e\n \u003ch2\u003e4.2.1. Adapting Existing Designs: Emergence of CP\u0026ndash;A\u0026ndash;V with Feasibility-Oriented PF\u003c/h2\u003e\n \u003cp\u003eAdapting existing designs emerged as the primary strategy for addressing the engineering task, fostering an early CP\u0026ndash;A\u0026ndash;V configuration, Creative Problem-solving (CP), Adapting (A), and Visualizing (V). Alongside this triad, Problem Finding (PF) was also present from the outset, but mainly as feasibility-oriented, surface-level identification used to converge on a unified group design.\u003c/p\u003e\n \u003cp\u003eDuring the \u0026ldquo;imagine \u0026amp; plan\u0026rdquo; stage, almost all groups (93.3%) browsed rubber-band-powered car designs via online videos and images, then adapted these ideas for their own projects (e.g., \u0026ldquo;We can refer to the method in the reference video and attach a carton to the popsicle stick to hold items.\u0026rdquo; 5A-G1-S1). Students asked option-selection questions (e.g., \u0026ldquo;Which design do you think is better?\u0026rdquo; 5A-G4-S4), checked alignments with requirements (e.g., \u0026ldquo;Are the wheels made from bottle caps? Are bottle caps recyclable?\u0026rdquo; 5A-G2-S3), and discussed material (e.g., \u0026ldquo;What material should we use for wheels?\u0026rdquo; 5B-G1-S2; \u0026ldquo;Use tape for wheels?\u0026rdquo; 5B-G4-S1). When evaluating design options, students favored what seemed feasible (e.g., \u0026ldquo;This design looks easier and more doable.\u0026rdquo; 5A-G1-S4; \u0026ldquo;This is the easiest video to follow.\u0026rdquo; 5A-G4-S2) and intentionally chose simpler solutions. Their material selections reflected the same practical mindset. They prioritized accessibility (e.g., \u0026ldquo;Plastic bottles are easy to find.\u0026rdquo; 5A-G1-S1; \u0026ldquo;The materials are easier to collect.\u0026rdquo; 5A-G3-S4) and workability (e.g., \u0026ldquo;The drink carton is easy to cut and shape, so we can choose it.\u0026rdquo;5A-G4-S2; \u0026ldquo;A plastic bottle is hard to cut, so I won\u0026rsquo;t choose it.\u0026rdquo; 5B-G4-S1), while paying limited attention to mechanical or structural properties. Within this adaptive strategy, PF emerged alongside decision-oriented discussions, focusing mainly on practical issues such as local fit and alignment, and operated in concert with CP and A to carry out feasibility-oriented option screening.\u003c/p\u003e\n \u003cp\u003eThis adaptive process was mediated by Visualizing (V) through sketches. As illustrated by 5B\u0026ndash;G7 (see Fig. \u003cspan class=\"InternalRef\"\u003e8\u003c/span\u003e), students frequently used sketches to visualize ideas and guide group discussions. One member posted a sketch of the car\u0026rsquo;s components and asked for feedback on where to position the rubber band as a power source (see Fig. \u003cspan class=\"InternalRef\"\u003e8\u003c/span\u003ea). This prompt led the group to adopt features from an online reference (see Fig. \u003cspan class=\"InternalRef\"\u003e8\u003c/span\u003eb), refine their design, and develop an updated sketch that clarified the revised concept (see Fig. \u003cspan class=\"InternalRef\"\u003e8\u003c/span\u003ec). Building on this shared visualization, they then agreed on specific materials and outlined preparation steps.\u003c/p\u003e\n \u003cp\u003eIn the subsequent \u0026ldquo;create\u0026rdquo; stage, students transformed 2D designs into 3D prototypes. Because the initial sketches lacked precise engineering specifications, construction became an iterative process in which students continuously reconciled conceptual sketches with material constraints and reference solutions. For instance, 5B\u0026ndash;G7 compared various beverage containers for the car body and bottle caps for the wheels before building the prototype shown in Fig. \u003cspan class=\"InternalRef\"\u003e8\u003c/span\u003ed. This material negotiation sustained the CP\u0026ndash;A\u0026ndash;V triad and further strengthened the interplay between A and V.\u003c/p\u003e\n \u003cp\u003eOverall, the \u0026ldquo;Adapting Existing Designs\u0026rdquo; strategy consolidated an early CP\u0026ndash;A\u0026ndash;V cognitive triad characterized by pragmatic adaptation, visual coordination through 2D sketches, and the translation of conceptual ideas into tangible artifacts through modeling. PF was present chiefly to conduct feasibility-oriented option screening and to convergence on a shared design within the groups. This foundation laid the groundwork for more advanced EHoMs in later EDP stages, where PF progressively deepen into causal diagnosis and, together with Improving (I), enabled evidence-based refinement.\u003c/p\u003e\n \u003c/div\u003e\n \u003cdiv id=\"Sec18\" class=\"Section3\"\u003e\n \u003ch2\u003e4.2.2. Deficiency-Driven Improvement: Positioning Improving (I) as the Central Hub Synergizing EHoMs\u003c/h2\u003e\n \u003cp\u003eAs students completed the initial construction, most rubber-band-powered trolleys failed to meet engineering requirements during validation tests. This setback shifted their focus from local-fit construction to a deficiency-driven improvement process, with Improving (I) emerging as the central hub connecting and coordinating EHoMs in the \u0026ldquo;improve\u0026rdquo; stage.\u003c/p\u003e\n \u003cp\u003eUnder teachers\u0026rsquo; guidance, students conducted two formal classroom trials to test whether their trolleys could carry four batteries over a two-meter distance, recording the results in their workbooks. Analysis of 62 workbooks showed that, in the first trial, 17 prototypes (27.42%) succeeded, while 45 (72.58%) fell short of the two-meter benchmark. After iterative modifications, 45 prototypes (72.58%) exceeded the target in the second trial, indicating that despite early deficiencies, the improvements were substantial and advancing in the right direction.\u003c/p\u003e\n \u003cp\u003eWorkbook evidence further showed that students moved beyond surface-level observations to diagnose root causes during validation tests. They not only recorded immediate observations, such as \u0026ldquo;the trolley traveled less than 2 m\u0026rdquo; (5B-G4-S1) and \u0026ldquo;its speed was too slow\u0026rdquo; (5A-G4-S4), but also analyzed the underlying causes of poor performance. As detailed in Appendix 2, four main categories of issues were identified: (1) energy transfer inefficiency (e.g., insufficient winding time, inadequate wind power); (2) friction and mechanical resistance (e.g., high friction between the axle and chassis); (3) structural and stability problems (e.g., the car was overweight, could not travel in a straight line); and (4) component mismatch (e.g., wheels too small, oversized propellers). These insights led students to propose targeted improvements. For example, one student (Workbook, 5B-G2-S2) observed that on smooth surfaces, the trolley\u0026rsquo;s wheels hardly rotated. He wrapped rubber bands around the rear wheels to increase friction and added extra batteries for weight, improving the trolley\u0026rsquo;s travel distance from 0.2 m in the first trial to 2 m in the second. In his workbook, he reflected: \u0026ldquo;Adding weight at the rear makes it go farther.\u0026rdquo; This example illustrates how the cognitive shift occurred: whereas initial adaptation relied on the CP\u0026ndash;A\u0026ndash;V triad for rapid prototyping with minimal analysis, prototype deficiencies during the engineering-validation phase triggered deeper engagement with EHoMs. Students began diagnosing structural\u0026ndash;environmental interactions (deeper problem finding) and pursued explicit performance goals (i.e., a 2 m travel distance). They entered an improvement-centered, iterative cycle, continuously testing, adjusting, and refining their designs through both analytical reasoning and hands-on modification.\u003c/p\u003e\n \u003cp\u003eOverall, the deficiency-driven improvement process embodies the structural dynamics of the \u0026ldquo;improve\u0026rdquo; stage, where Improving (I) acted as a central hub linking EHoMs such as Visualizing, Adapting, Problem Finding, and Creative Problem-solving. Signs of emerging Systems Thinking (ST) also appeared as students recognized and addressed interconnected structural and performance issues through deliberate interventions.\u003c/p\u003e\n \u003c/div\u003e\n \u003cdiv id=\"Sec19\" class=\"Section3\"\u003e\n \u003ch2\u003e4.2.3. \u003cstrong\u003eFrom Knowing to Doing\u003c/strong\u003e: \u003cstrong\u003eFailure-Activated Systems Thinking (ST)\u003c/strong\u003e\u003c/h2\u003e\n \u003cp\u003eSystems Thinking remained at a low frequency throughout students\u0026rsquo; EDP learning process, with signs of developmental potential in the later stages and beyond. This pattern can be partly explained by a knowing\u0026ndash;doing gap: although students understood the theoretical principles behind the trolley\u0026rsquo;s construction, they rarely applied this knowledge proactively during design. Only when they began analyzing the causes of test failures did this latent understanding become activated, giving rise to the emergence of ST.\u003c/p\u003e\n \u003cp\u003eThis gap became particularly evident in relation to energy conversion, a foundational concept in the EDP learning activities. During pre-EDP instruction, students demonstrated an understanding of how energy is stored and transformed in simple converters. Classroom dialogue confirmed their conceptual grasp. For example, one student noted, \u0026ldquo;The rubber band stores potential energy when twisted, which converts to kinetic energy as it unwinds\u0026rdquo; (Class Observation, Lesson 1). However, there was no evidence that they applied these principles in their initial design choices, as student discourse indicated they prioritized replicating existing solutions over optimizing them, rarely considering factors such as energy loss minimization.\u003c/p\u003e\n \u003cp\u003eConsequently, without applying energy principles, students diverged into two design paths with distinct performance outcomes. The first, as shown in Fig. \u003cspan class=\"InternalRef\"\u003e9\u003c/span\u003e, featured direct conversion, in which the rubber band\u0026rsquo;s potential energy was released directly as kinetic energy to propel the trolley, resulting in high energy conversion efficiency. Thirty-seven students (59.7%) opted for this approach. During the two rounds of formal testing, fifteen students (40.5%) in Round 1 and twenty-nine students (78.4%) in Round 2 passed the 2-meter benchmark traveling distance, consistently exceeding the average rates (27.42% in Round 1 and 72.58% in Round 2). The second propulsion method, as shown in Fig. \u003cspan class=\"InternalRef\"\u003e10\u003c/span\u003e, involved indirect conversion, where the potential energy stored in the rubber band was first transformed into wind energy via a propeller and then into the trolley\u0026rsquo;s kinetic energy. Twenty-five students (40.3%) selected this approach. In Round 1, only two indirect cars (8%) reached 2 meters, while twenty-three (92%) failed. Round 2 showed improvement, with sixteen cars (64%) reaching 2 meters. However, overall success of the indirect conversion designs remained below the cohort average.\u003c/p\u003e\n \u003cp\u003eThe pronounced struggle with the indirect conversion design prompted students to actively reflect on and analyze the causes of failure, rather than merely imitate, in the later stages of the project. In their workbooks, students noted that the chief flaw of the indirect setup was insufficient propulsive force, and they proposed targeted remedies such as \u0026ldquo;adding extra rubber bands\u0026rdquo; (e.g., 5A-G2-S4), \u0026ldquo;tightening the propeller further\u0026rdquo;(e.g., 5B-G5-S1), \u0026ldquo;enlarging the propeller\u0026rdquo; (e.g., 5A-G3-S2), or, in some cases, \u0026ldquo;switching to direct rubber-band propulsion\u0026rdquo; (e.g., 5B-G1-S2). More importantly, focus group interviews revealed that their reasoning had even evolved to consider interactions within the system (e.g., Focus Group, 5A-G2):\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eS3\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cem\u003e\u0026ldquo;I wouldn\u0026rsquo;t use a propeller again. It\u0026rsquo;s hard to build, and the car lacks speed and distance. The plastic bottle and other components make it heavy, and determining the correct rubber band length is tricky. If redesigning, I\u0026rsquo;d use a carton with direct twisting, like other teams did, for better speed and distance.\u0026rdquo;\u003c/em\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eS1\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cem\u003e\u0026ldquo;We\u0026rsquo;d adopt G4\u0026rsquo;s direct-drive design. As S3 said, our design is heavy. The energy converts from potential to wind, then to kinetic, losing power with each conversion. That\u0026rsquo;s why our car cannot travel as far as the others.\u0026rdquo;\u003c/em\u003e\u003c/p\u003e\n \u003cp\u003eTherefore, through repeated testing and failure analysis, students gradually linked propulsion methods to performance outcomes and closed the knowing\u0026ndash;doing gap. They inferred that direct energy conversion reduced energy loss and improved efficiency, and that trolley weight played a crucial role in travel distance. By applying energy principles and comparing design alternatives, they ultimately demonstrated emerging Systems Thinking.\u003c/p\u003e\n \u003c/div\u003e\n\u003c/div\u003e"},{"header":"5. Discussion","content":"\u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003e5.1. Addressing RQ1: Characterizing the Developmental Trajectory of EHoMs\u003c/h2\u003e \u003cp\u003eThe longitudinal analysis revealed that students\u0026rsquo; EHoMs developed through a nonlinear and stage-dependent trajectory, with each stage of the EDP eliciting distinct cognitive integrations. Our findings empirically validate the EDP as a signature pedagogy for cultivating primary school students\u0026rsquo; EHoMs in STEM activities (Lucas \u0026amp; Hanson, \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). In structured EDP tasks, these habits appeared as clear, observable behaviors, aligning with Bianchi and Wiskow\u0026rsquo; s (2023) call to treat EHoMs as visible learning objectives throughout the process to make engineering education more practical for teachers. Through iterative practice in the EDP, students increasingly internalized EHoMs, which became durable habits of mind rather than remaining isolated, one-off skills (Ladion-De Guzm\u0026aacute;n, 2021).\u003c/p\u003e \u003cp\u003eFurthermore, building on earlier research showing that children can develop EHoMs (e.g., Donnelly et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; Karatas-Aydin \u0026amp; Isiksal-Bostan, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Schucker et al., \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), our findings show that upper-primary students not only demonstrate continuous EHoM development but also display a hierarchical structure and stage-dependent progression. Specifically, within this energy-converter project, Creative Problem-solving, Adapting, and Visualizing served as foundational habits that persisted across all EDP stages, thereby laying the cognitive groundwork for the subsequent emergence of Improving and Systems Thinking. Additionally, the role of Problem Finding evolved dynamically throughout the EDP. Initially, Problem Finding served primarily as feasibility screening, facilitating quick convergence and establishment of a unified design direction. Later, it transitioned into gap diagnosis, driving students to progress from surface-level observations to an analysis of underlying causes. Among all EHoMs, Improving, which involves refining designs through experimentation, prototyping, and iterative reasoning (Hanson et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2018\u003c/span\u003e), exhibited particularly distinct stages dependency and task specificity. This suggests that the activation of Improving heavily depended on the cognitive accumulation and tangible prototypes developed during earlier stages, which together provided the necessary basis for iterative refinement. Once activated, Improving also synergized with other EHoMs to drive significant overall advancement. The distinctiveness of Improving also aligns with the recognized essential role of the \u0026ldquo;improve\u0026rdquo; stage in elementary engineering education, where students have opportunities to critically evaluate and iteratively refine designs using scientific principles and epistemic tools (Kelly \u0026amp; Cunningham, \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2019\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIn the observed structural patterns, the rarity of Systems Thinking reflects its role as a higher-order cognitive skill (Frank \u0026amp; Waks, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2000\u003c/span\u003e). Systems Thinking comprises an integrated set of analytical skills for identifying, understanding, and predicting system behaviors and for designing targeted improvements. As system complexity grows, it demands correspondingly higher levels of systemic reasoning (Arnold \u0026amp; Wade, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). Since there is a strong connection between having a basic understanding of systems and the subsequent development of Systems Thinking, it is advisable to introduce Systems Thinking instruction as early as primary school (Assaraf \u0026amp; Orion, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2005\u003c/span\u003e). For primary school students, who often exhibit linear thinking and find it challenging to recognize hidden causal relationships (Nguyen \u0026amp; Santagata, \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), developing Systems Thinking should be a gradual process. Their understanding of systems should be enhanced through consistent engagement with real-world contexts and guided reflection on concrete experiences (Assaraf \u0026amp; Orion, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2010\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003e5.2. Addressing RQ2: Mechanisms of EHoM Growth\u003c/h2\u003e \u003cp\u003eTo further unpack the hierarchical and stage-dependent trajectories of EHoMs, this study identified an \u0026ldquo;imitate\u0026ndash;fail\u0026ndash;improve\u0026rdquo; (IFI) pathway to explain the developmental mechanism of these habits of mind, particularly Systems Thinking, within the EDP. Although children already exhibit basic Systems Thinking (e.g., Donnelly et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; Lippard et al., \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2019\u003c/span\u003e), our findings suggest that students\u0026rsquo; Systems Thinking gradually emerges and develops as the knowing\u0026ndash;doing gap is progressively bridged through the failure-driven iterative improvement within the EDP.\u003c/p\u003e \u003cp\u003eThe engineering task began with imitation, which provided a cognitively accessible entry point as students actively sourced existing rubber-band-powered designs from open online platforms. By observing and comparing these designs embodied in physical models, students collaboratively adapted the plans considering material availability and manufacturability. They then translated these visual and structural insights into self-drawn conceptual sketches, a process that interpreted the physical models while integrating personal understanding and imagination. Based on these sketches, students collectively determined material selection and structural implementation for each part of the product, progressively advancing toward the creation of a functional prototype.\u003c/p\u003e \u003cp\u003eThroughout the process, by adapting, collaborating, and crafting, students brought their ideas to life and developed a sense of ownership over the final design. Their ideas and creativity were tangibly captured in the sketches they drew and the models they built, which served as visual records of their thinking and practice. This iterative modeling was primarily driven by the synergistic triad of Creative Problem-solving, Adapting, and Visualizing (CP-A-V). Importantly, modeling served not only as a final output but also as an epistemic tool that externalized design thinking and enhanced problem understanding and real-time decision-making (Lammi \u0026amp; Denson, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). As a result, starting from existing designs reduced cognitive load (Sweller, \u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), allowing students to build simple, partially functional prototypes from static components, rather than having to grasp the entire integrated system upfront. Simultaneously, this imitation strategy accommodated the common novice tendency to focus on visible structures rather than system operations and the reasoning behind design choices (Crismond \u0026amp; Adams, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). Moreover, the prototypes developed from student-sourced exemplars potentially served as improvable models, which also provided scaffolding (Dasgupta, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Although the online originals were functionally complete and replicable, students\u0026rsquo; imperfect adaptation and construction due to surface understanding of the designs could introduce subsequent defects into the prototypes.\u003c/p\u003e \u003cp\u003eLater, through engineering testing, failure acted as a catalyst for deeper exploration. When prototypes underperformed, students transitioned from superficial imitation to diagnostic inquiry. They identified flaws, proposed targeted modifications, and tested solutions while unpacking underlying physical principles. This shift was facilitated by the \u0026ldquo;improve\u0026rdquo; stage, which created conditions for productive failure (Kapur, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2008\u003c/span\u003e), through three key facilitating factors (Johnson et al., \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). (1) Sufficient Opportunities: Students were able to conduct their own tests before the two formal class testing sessions, enabling active, low-stakes experimentation. (2) Fair Comparisons: Vehicle movement distance and speed were recorded and compared under unified engineering requirements. (3) Productive Strategies: Students were encouraged to analyze, diagnose design problems, and actively attempt improvements. As students assessed each optimization\u0026rsquo;s effectiveness, they engaged more deeply in scientific reasoning, elevating their domain knowledge from \u0026ldquo;remembering\u0026rdquo; and \u0026ldquo;understanding\u0026rdquo; to \u0026ldquo;applying\u0026rdquo; and \u0026ldquo;evaluating.\u0026rdquo; Moreover, collaborative learning amplified this process: cross-group critique of failure cases generated productive cognitive conflict, while comparative analysis of divergent designs broadened and deepened students\u0026rsquo; systemic understanding through co-constructed knowledge building (Scardamalia \u0026amp; Bereiter, \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2014\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eNotably, the evolution of students\u0026rsquo; Problem Finding closely aligns with the cognitive patterns commonly observed in novice engineers. To begin with, students\u0026rsquo; Problem Finding behavior was oriented toward rapid convergence. Students tended to treat the design task as a well-defined problem and rushed to propose solutions, which corresponds to the \u0026ldquo;premature commitment\u0026rdquo; tendency typical of beginners (Crismond \u0026amp; Adams, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). However, as prototype testing and iterative refinement progressed, students gradually developed the capacity to model dynamic system interactions, which is a hallmark of expert engineering practice (Frank \u0026amp; Waks, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2000\u003c/span\u003e; Sch\u0026ouml;n, \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). In this constructionist learning environment, the rubber-band-powered trolleys, as physical prototypes, served as shareable artifacts (Papert \u0026amp; Harel, \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e1991\u003c/span\u003e) that bridged procedural knowledge with conceptual understanding. The \u0026ldquo;IFI\u0026rdquo; pathway thus enabled students to progress from component-focused replication to system-level reasoning, thereby synergistically fostering all EHoMs.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec23\" class=\"Section2\"\u003e \u003ch2\u003e5.3. Implications\u003c/h2\u003e \u003cdiv id=\"Sec24\" class=\"Section3\"\u003e \u003ch2\u003e5.3.1. Theoretical Implications\u003c/h2\u003e \u003cp\u003eThis study provides nuanced empirical evidence on integrating EHoMs and the EDP in primary school STEM education. The findings support the characterization of the EDP as a \u0026ldquo;signature pedagogy\u0026rdquo; in engineering education (Lucas \u0026amp; Hanson, \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2016\u003c/span\u003e) and further clarify the hierarchical and dynamic nature of EHoMs. Specifically, cognitive pillars such as Creative Problem-solving, Adapting, and Visualizing predominate in the initial stages. Problem Finding demonstrates a stable presence throughout the process, whereas higher-order habits of mind, including Improving and Systems Thinking, are primarily activated during the later stages of testing and iteration. As exemplified by the construction of a rubber-band-powered trolley, students initially followed an \u0026ldquo;adapting existing designs\u0026rdquo; strategy, combining Creative Problem-solving, Adapting, Visualizing, and Problem Finding to start the project, concretize ideas, tackle early challenges, and complete component-level construction. As they entered the \u0026ldquo;improve\u0026rdquo; stage, their focus shifted to the trolley as an integrated system. Through feedback-driven iteration and diagnostic testing, students engaged in purposeful optimization, activating Improving and developing their Systems Thinking progressively. This observed \u0026ldquo;imitate\u0026ndash;fail\u0026ndash;improve\u0026rdquo; (IFI) pathway advances the theoretical understanding of how upper-primary students progressively develop Systems Thinking, enriching the broader cognitive framework for engineering education.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec25\" class=\"Section3\"\u003e \u003ch2\u003e5.3.2. Practical Implications\u003c/h2\u003e \u003cp\u003eThe \u0026ldquo;IFI\u0026rdquo; pathway both mirrors real-world engineering practice and aligns with primary students\u0026rsquo; cognitive development. On the one hand, in engineering, preliminary research often reveals multiple existing solutions, making adaptive modification the norm in problem solving (Jonassen et al., \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2006\u003c/span\u003e). On the other hand, younger learners, who rely on concrete experiences to grasp abstract ideas, typically build knowledge incrementally through hands-on trial and error with feedback (Firdaus \u0026amp; Rahayu, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Accordingly, concrete imitation can be a productive starting point for children\u0026rsquo;s cognitive development.\u003c/p\u003e \u003cp\u003eThe \u0026ldquo;IFI\u0026rdquo; pathway can be flexibly integrated into STEM instruction to meet diverse needs. For students with limited subject knowledge and emerging systems-thinking skills, early stages can emphasize guided imitation and moderate adaptation of existing designs, providing essential scaffolding for prototyping and turning low-risk failures into opportunities for diagnosis and reflection. For students with stronger foundations, systems-analysis tasks can be introduced before the \u0026ldquo;create\u0026rdquo; stage to activate and deepen Systems Thinking, prompting consideration of system boundaries, components, interactions among elements and subsystems, and overall system characteristics and behaviors, thereby fostering richer learning outcomes. Additionally, across different engineering themes, tasks can also embed observable learning-behavior indicators aligned with the EHoMs, and teachers can use ongoing observation and feedback as key elements of formative assessment.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec26\" class=\"Section2\"\u003e \u003ch2\u003e5.4. Limitations\u003c/h2\u003e \u003cp\u003eWhile our study provides insights and contributions to the development of EHoM through the EDP, it is important to acknowledge certain limitations that may impact the interpretation and generalizability of the results. First, the exploration of EHoM in this study began only after the \u0026ldquo;Ask\u0026rdquo; stage, when students engaged in self-directed collaboration during the subsequent EDP stages. In the \u0026ldquo;Ask\u0026rdquo; stage, teachers led the process of defining the engineering problem (e.g., setting constraints, restating goals and introducing materials), ensuring student understanding. Consequently, the absence of student-generated documentation in the \u0026ldquo;Ask\u0026rdquo; stage restricts the identification of the structural patterns of EHoM development at this stage and limits comparisons with those in other stages, potentially hindering our holistic understanding of the developmental trajectories of EHoMs across all EDP stages. Second, the engineering task of building a rubber-band-powered trolley has multiple ready-made solutions that are readily accessible online. This situation may lead students to imitate existing solutions rather than demonstrate original thinking. If the task were conducted in an environment without internet access, students might exhibit different patterns of thinking and problem-solving strategies, thereby revealing distinctive characteristics in the development of EHoMs compared to the original context. Third, this study was conducted in a blended learning environment that emphasized online collaboration with limited in-person interaction, which to some extent constrains the external validity of the findings for primarily face-to-face or other types of learning contexts. Last but equally important, the developmental trajectory of EHoM presented may not accurately reflect actual cognitive processes. EHoMs are intended to depict patterns of thinking; however, there may be a disparity between internal thought processes and observable behavior. For example, students may engage in systematic analysis when making decisions, yet this analytical process may not be directly observable in their overt actions.\u003c/p\u003e \u003c/div\u003e"},{"header":"6. Conclusion","content":"\u003cp\u003eThis study investigated the developmental trajectories and mechanisms of EHoMs within the EDP among fifth graders. Through a rubber-band-powered trolley project, it traced the emergence of EHoMs and identified their key drivers, yielding two central findings. First, EHoMs develop along a hierarchical trajectory. The foundational CP\u0026ndash;A\u0026ndash;V triad (Creative Problem-solving, Adapting, Visualizing) underpinned the early \u0026ldquo;imagine \u0026amp; plan\u0026rdquo; and \u0026ldquo;create\u0026rdquo; stages. PF (Problem Finding) evolved from a shallow to more sophisticated level throughout the entire process, whereas advanced habits like I (Improving) and ST (Systems Thinking) were activated later during the \u0026ldquo;improve\u0026rdquo; stage. Second, this development is driven by an \u0026ldquo;imitate\u0026ndash;fail\u0026ndash;improve\u0026rdquo; (IFI) pathway, where imitation served as productive scaffolding, and failure acted as a critical catalyst for a cognitive shift from component-level assembly to system-level reasoning.\u003c/p\u003e \u003cp\u003eThese findings extend both theory and practice in early engineering education. Theoretically, this study offers fine-grained empirical support for framing the EDP as a signature pedagogy for EHoMs and elucidates a developmental mechanism, the IFI pathway, through which the EHoMs of synergizing and ST emerge from cumulative, reflective practice. Practically, this study provides educators with a recognizable and actionable guidance: instructional activities can be designed to scaffold productive imitation, leverage failure diagnostically, and create structured opportunities for evidence-based improvement, while assessment can incorporate observable indicators aligned with evolving EHoMs.\u003c/p\u003e \u003cp\u003eBased on this study, we suggest directions for future work. Future studies should: (1) clarify the contextual boundaries and adaptive mechanism of the IFI pathway by examining whether the IFI sequence remains valid in more open-ended or offline settings and identifying key instructional-contextual factors that support its effective functioning; (2) develop mix-methods approaches that more closely align with cognitive processes to directly capture implicit cognitive development; and (3) uncover the microstructure and interactive relationships among EHoM sub-habits within each primary EHoM (e.g., \u0026ldquo;Connecting\u0026rdquo; and \u0026ldquo;Pattern-making\u0026rdquo; under Systems Thinking), thereby revealing micro-developmental sequences and co-evolutionary dynamics to refine the educational model of engineering thinking development.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cdiv class=\"DefinitionList\"\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eEDP\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eEngineering Design Process\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eEHoM(s)\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eEngineering Habit(s) of Mind\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eENA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eEpistemic Network Analysis\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eIFI\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eImitate-Fail-Improve\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eQCA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eQualitative Content Analysis\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eTA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eThematic Analysis\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e \u003ch2\u003eCompeting interests\u003c/h2\u003e \u003cp\u003eNone.\u003c/p\u003e \u003c/p\u003e\u003cp\u003e \u003ch2\u003eEthics approval and consent to participate\u003c/h2\u003e \u003cp\u003e This study received ethical approval from the Human Research Ethics Committee (HREC) of the University of Hong Kong (Reference No. EA1803010). All participants and their parents/ guardians signed a consent form prior to participating in the study.\u003c/p\u003e \u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e \u003cp\u003e This work was supported by the General Research Fund managed by the Research Grant Council under the University Grants Committee in Hong Kong under grant number # 17603619.\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eAuthors\u0026rsquo; contributionsG.W.: Conceptualization, data curation, methodology, resources, writing-review \u0026amp; editing, project administration, supervision, funding acquisition.P.T.: Conceptualization, data curation, methodology, formal analysis, investigation, writing-original draft, visualization, writing-review \u0026amp; editing.B.C.: Methodology, validation, formal analysis, writing- review \u0026amp; editing.F.W.: Validation, formal analysis, writing- review \u0026amp; editing.L.L.: Validation, visualization, writing- review \u0026amp; editing.All authors read and approved the final manuscript.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eThe datasets generated and/ or analyzed during the current study are not publicly available due to ethical restrictions regarding student anonymity and data privacy protocols approved by the Ethics Committee.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAlhamlan, S., Aljasser, H., Almajed, A., Almansour, H., \u0026amp; Alahmad, N. (2017). A systematic review: Using habits of mind to improve student\u0026rsquo;s thinking in class. \u003cem\u003eHigher Education Studies\u003c/em\u003e, \u003cem\u003e8\u003c/em\u003e(1), 25\u0026ndash;35. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.5539/hes.v8n1p25\u003c/span\u003e\u003cspan address=\"10.5539/hes.v8n1p25\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eArnold, R. D., \u0026amp; Wade, J. P. (2015). A definition of systems thinking: A systems approach. \u003cem\u003eProcedia Computer Science\u003c/em\u003e, \u003cem\u003e44\u003c/em\u003e, 669\u0026ndash;678. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.procs.2015.03.050\u003c/span\u003e\u003cspan address=\"10.1016/j.procs.2015.03.050\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAssaraf, O. B., \u0026amp; Orion, N. (2005). Development of system thinking skills in the context of earth system education. \u003cem\u003eJournal of Research in Science Teaching\u003c/em\u003e, \u003cem\u003e42\u003c/em\u003e(5), 518\u0026ndash;560. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/tea.20061\u003c/span\u003e\u003cspan address=\"10.1002/tea.20061\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAssaraf, O. B., \u0026amp; Orion, N. (2010). System thinking skills at the elementary school level. \u003cem\u003eJournal of Research in Science Teaching\u003c/em\u003e, \u003cem\u003e47\u003c/em\u003e(5), 540\u0026ndash;563. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/tea.20351\u003c/span\u003e\u003cspan address=\"10.1002/tea.20351\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBianchi, L., \u0026amp; Wiskow, J. (2023). \u003cem\u003eProgressing to be an Engineer \u0026ndash; the Approach\u003c/em\u003e. (1st ed.) Royal Academy of Engineering. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://raeng.org.uk/media/eegctjlq/rae-ptbae-approach_final-3.pdf\u003c/span\u003e\u003cspan address=\"http://raeng.org.uk/media/eegctjlq/rae-ptbae-approach_final-3.pdf\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBinkley, M., Erstad, O., Herman, J., Raizen, S., Ripley, M., Miller-Ricci, M., \u0026amp; Rumble, M. (2012). Defining twenty-first century skills. In P. Griffin, B. McGaw, \u0026amp; E. Care (Eds.), \u003cem\u003eAssessment and teaching of 21st century skills\u003c/em\u003e (pp. 17\u0026ndash;66). Springer Netherlands. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/978-94-007-2324-5_2\u003c/span\u003e\u003cspan address=\"10.1007/978-94-007-2324-5_2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBraun, V., \u0026amp; Clarke, V. (2012). Thematic analysis. In H. Cooper, P. M. Camic, D. L. Long, A. T. Panter, D. Rindskopf, \u0026amp; K. J. Sher (Eds.), \u003cem\u003eAPA handbook of research methods in psychology, Vol 2. Research designs: Quantitative, qualitative, neuropsychological, and biological\u003c/em\u003e (pp. 57\u0026ndash;71). American Psychological Association. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1037/13620-004\u003c/span\u003e\u003cspan address=\"10.1037/13620-004\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCosta, A. L., \u0026amp; Kallick, B. (2000). \u003cem\u003eDiscovering \u0026amp; exploring habits of mind\u003c/em\u003e. Association for Supervision and Curriculum Development.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCrismond, D. P., \u0026amp; Adams, R. S. (2012). The informed design teaching and learning matrix. \u003cem\u003eJournal of Engineering Education\u003c/em\u003e, \u003cem\u003e101\u003c/em\u003e(4), 738\u0026ndash;797. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/j.2168-9830.2012.tb01127.x\u003c/span\u003e\u003cspan address=\"10.1002/j.2168-9830.2012.tb01127.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCunningham, C. M., \u0026amp; Kelly, G. J. (2017). Epistemic practices of engineering for education. \u003cem\u003eScience Education\u003c/em\u003e, \u003cem\u003e101\u003c/em\u003e(3), 486\u0026ndash;505. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/sce.21271\u003c/span\u003e\u003cspan address=\"10.1002/sce.21271\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCunningham, C. M., Lachapelle, C. P., Brennan, R. T., Kelly, G. J., Tunis, C. S. A., \u0026amp; Gentry, C. A. (2020). The impact of engineering curriculum design principles on elementary students\u0026rsquo; engineering and science learning. \u003cem\u003eJournal of Research in Science Teaching\u003c/em\u003e, \u003cem\u003e57\u003c/em\u003e(3), 423\u0026ndash;453. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/tea.21601\u003c/span\u003e\u003cspan address=\"10.1002/tea.21601\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDasgupta, C. (2019). Improvable models as scaffolds for promoting productive disciplinary engagement in an engineering design activity. \u003cem\u003eJournal of Engineering Education\u003c/em\u003e, \u003cem\u003e108\u003c/em\u003e(3), 394\u0026ndash;417. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/jee.20282\u003c/span\u003e\u003cspan address=\"10.1002/jee.20282\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDonnelly, R., Speldewinde, C., \u0026amp; Bridle, H. (2025). Engineering habits of mind in preschool children at Scottish forest nurseries and Australian bush kinders. \u003cem\u003eBritish Educational Research Journal\u003c/em\u003e, \u003cem\u003e51\u003c/em\u003e(5), 2185\u0026ndash;2209. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/berj.4164\u003c/span\u003e\u003cspan address=\"10.1002/berj.4164\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEducational Bureau of Hong Kong (2017). \u003cem\u003eGeneral Studies for Primary Schools Curriculum Guide (Primary 1-Primary 6)\u003c/em\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEnglish, L. D. (2016). STEM education K-12: Perspectives on integration. \u003cem\u003eInternational Journal of STEM Education\u003c/em\u003e, \u003cem\u003e3\u003c/em\u003e(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003eArticle 3. https://doi.org/10.1186/s40594-016-0036-1\u003c/span\u003e\u003cspan address=\"Article 3. 10.1186/s40594-016-0036-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEnglish, L. D. (2019). Learning while designing in a fourth-grade integrated STEM problem. \u003cem\u003eInternational Journal of Technology and Design Education\u003c/em\u003e, \u003cem\u003e29\u003c/em\u003e(5), 1011\u0026ndash;1032. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10798-018-9482-z\u003c/span\u003e\u003cspan address=\"10.1007/s10798-018-9482-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEnglish, L. D., \u0026amp; King, D. (2019). STEM Integration in Sixth Grade: Desligning and Constructing Paper Bridges. \u003cem\u003eInternational Journal of Science and Mathematics Education\u003c/em\u003e, \u003cem\u003e17\u003c/em\u003e(5), 863\u0026ndash;884. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10763-018-9912-0\u003c/span\u003e\u003cspan address=\"10.1007/s10763-018-9912-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEnglish, L. D., King, D., \u0026amp; Smeed, J. (2017). Advancing integrated STEM learning through engineering design: Sixth-grade students\u0026rsquo; design and construction of earthquake resistant buildings. \u003cem\u003eThe Journal of Educational Research\u003c/em\u003e, \u003cem\u003e110\u003c/em\u003e(3), 255\u0026ndash;271. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/00220671.2016.1264053\u003c/span\u003e\u003cspan address=\"10.1080/00220671.2016.1264053\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEnglish, L. D., \u0026amp; King, D. T. (2015). STEM learning through engineering design: Fourth-grade students\u0026rsquo; investigations in aerospace. \u003cem\u003eInternational Journal of STEM Education\u003c/em\u003e, \u003cem\u003e2\u003c/em\u003e(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003eArticle 14. https://doi.org/10.1186/s40594-015-0027-7\u003c/span\u003e\u003cspan address=\"Article 14. 10.1186/s40594-015-0027-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFirdaus, A. R., \u0026amp; Rahayu, G. D. S. (2019). Effect of STEM-based learning on the cognitive skills improvement. \u003cem\u003eMimbar Sekolah Dasar\u003c/em\u003e, \u003cem\u003e6\u003c/em\u003e(2), 198\u0026ndash;207. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.17509/mimbar-sd.v6i2.17562\u003c/span\u003e\u003cspan address=\"10.17509/mimbar-sd.v6i2.17562\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFrank, M., \u0026amp; Waks, S. (2000). Engineering systems thinking: A multifunctional definition. \u003cem\u003eSystemic Practice and Action Research\u003c/em\u003e, \u003cem\u003e14\u003c/em\u003e(3), 361\u0026ndash;379.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGl\u0026auml;ser-Zikuda, M., Hagenauer, G., \u0026amp; Stephan, M. (2020). The potential of qualitative content analysis for empirical educational research. \u003cem\u003eForum Qualitative Sozialforschung / Forum: Qualitative Social Research\u003c/em\u003e, \u003cem\u003e21\u003c/em\u003e(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003eArticle 7. https://doi.org/10.17169/FQS-21.1.3443\u003c/span\u003e\u003cspan address=\"Article 7. 10.17169/FQS-21.1.3443\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHanson, J., Hardman, S., Luke, S., Maunders, P., \u0026amp; Lucas, B. (2018). \u003cem\u003eEngineering the future: Training today\u0026rsquo;s teachers to develop tomorrow\u0026rsquo;s engineers\u003c/em\u003e. Royal Academy of Engineering. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.13140/RG.2.2.22598.93765\u003c/span\u003e\u003cspan address=\"10.13140/RG.2.2.22598.93765\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHester, K., \u0026amp; Cunningham, C. (2007). \u003cem\u003eEngineering is elementary: An engineering and technology curriculum for children\u003c/em\u003e [Paper presentation]. 2007 Annual Conference \u0026amp; Exposition, Honolulu, Hawaii, United States. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.18260/1-2--1469\u003c/span\u003e\u003cspan address=\"10.18260/1-2--1469\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHmelo-Silver, C. E. (2004). Problem-based learning: what and how do students learn? \u003cem\u003eEducational Psychology Review\u003c/em\u003e, \u003cem\u003e16\u003c/em\u003e(3), 235\u0026ndash;266. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1023/B:EDPR.0000034022.16470.f3\u003c/span\u003e\u003cspan address=\"10.1023/B:EDPR.0000034022.16470.f3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHolmlund, T. D., Lesseig, K., \u0026amp; Slavit, D. (2018). Making sense of STEM education in K-12 contexts. \u003cem\u003eInternational Journal of STEM Education\u003c/em\u003e, \u003cem\u003e5\u003c/em\u003e(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s40594-018-0127-2\u003c/span\u003e\u003cspan address=\"10.1186/s40594-018-0127-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Article 32.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHoney, M., Pearson, G., \u0026amp; Schweingruber, H. A. (Eds.). (2014). \u003cem\u003eSTEM integration in K-12 education: Status, prospects, and an agenda for research\u003c/em\u003e. National Academies.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJohnson, M., Kelly, G., \u0026amp; Cunningham, C. (2021). Failure and improvement in elementary engineering. \u003cem\u003eJournal of Research in STEM Education\u003c/em\u003e, \u003cem\u003e7\u003c/em\u003e(2), 69\u0026ndash;92. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.51355/jstem.2021.101\u003c/span\u003e\u003cspan address=\"10.51355/jstem.2021.101\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJohri, A., \u0026amp; Olds, B. M. (2011). Situated engineering learning: Bridging engineering education research and the learning sciences. \u003cem\u003eJournal of Engineering Education\u003c/em\u003e, \u003cem\u003e100\u003c/em\u003e(1), 151\u0026ndash;185. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/j.2168-9830.2011.tb00007.x\u003c/span\u003e\u003cspan address=\"10.1002/j.2168-9830.2011.tb00007.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJonassen, D., Strobel, J., \u0026amp; Lee, C. B. (2006). Everyday problem solving in engineering: Lessons for engineering educators. \u003cem\u003eJournal of Engineering Education\u003c/em\u003e, \u003cem\u003e95\u003c/em\u003e(2), 139\u0026ndash;151. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/j.2168-9830.2006.tb00885.x\u003c/span\u003e\u003cspan address=\"10.1002/j.2168-9830.2006.tb00885.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKapur, M. (2008). Productive failure. \u003cem\u003eCognition and Instruction\u003c/em\u003e, \u003cem\u003e26\u003c/em\u003e(3), 379\u0026ndash;424. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/07370000802212669\u003c/span\u003e\u003cspan address=\"10.1080/07370000802212669\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKaratas-Aydin, F. I., \u0026amp; Isiksal-Bostan, M. (2023). Engineering-based modelling experiences of elementary gifted students: An example of bridge construction. \u003cem\u003eThinking Skills and Creativity\u003c/em\u003e, \u003cem\u003e47\u003c/em\u003e., Article 101237. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.tsc.2023.101237\u003c/span\u003e\u003cspan address=\"10.1016/j.tsc.2023.101237\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKelly, G. J., \u0026amp; Cunningham, C. M. (2019). Epistemic tools in engineering design for K-12 education. \u003cem\u003eScience Education\u003c/em\u003e, \u003cem\u003e103\u003c/em\u003e(4), 1080\u0026ndash;1111. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/sce.21513\u003c/span\u003e\u003cspan address=\"10.1002/sce.21513\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKeratithamkul, K., Kim, J. N., \u0026amp; Roehrig, G. H. (2020). Cultural competence or deficit-based view? A qualitative approach to understanding middle school students\u0026rsquo; experience with culturally framed engineering. \u003cem\u003eInternational Journal of STEM Education\u003c/em\u003e, \u003cem\u003e7\u003c/em\u003e(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003eArticle 26. https://doi.org/10.1186/s40594-020-00224-5\u003c/span\u003e\u003cspan address=\"Article 26. 10.1186/s40594-020-00224-5\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLadion-De Guzman, M. G. (2021). Nurturing engineering habits of mind in the 21st century learners. \u003cem\u003eInternational Forum Journal\u003c/em\u003e, \u003cem\u003e24\u003c/em\u003e(2), 5\u0026ndash;20. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.63201/plqc6617\u003c/span\u003e\u003cspan address=\"10.63201/plqc6617\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLai, M. Y., \u0026amp; Cheng, E. (2023). Bringing engineering into primary science classrooms using engineering design and community of practice approach\u0026mdash;An evaluation of STEM \u0026times; Play program. \u003cem\u003eFrontiers in Education\u003c/em\u003e, \u003cem\u003e8\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/feduc.2023.1290857\u003c/span\u003e\u003cspan address=\"10.3389/feduc.2023.1290857\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLammi, M. D., \u0026amp; Denson, C. D. (2017). Modeling as an engineering habit of mind and practice. \u003cem\u003eAdvances in Engineering Education\u003c/em\u003e, \u003cem\u003e6\u003c/em\u003e(1).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLandis, J. R., \u0026amp; Koch, G. G. (1977). The Measurement of observer agreement for categorical data. \u003cem\u003eBiometrics\u003c/em\u003e, \u003cem\u003e33\u003c/em\u003e(1), 159\u0026ndash;174. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2307/2529310\u003c/span\u003e\u003cspan address=\"10.2307/2529310\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLee, P., Crosthwaite, C., Reidsema, C., Burnett, I., Foley, B., Hargreaves, D., King, R., Lamborn, J., Symes, M., \u0026amp; Wilson, J. (2022). Preparing engineers for 2035: Transforming Australia\u0026rsquo;s engineering education for emerging roles and expectations. In C. Hong \u0026amp; W. W. K. Ma (Eds.), \u003cem\u003eApplied degree education and the future of learning\u003c/em\u003e (pp. 29\u0026ndash;52). Springer Nature Singapore. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/978-981-16-9812-5_2\u003c/span\u003e\u003cspan address=\"10.1007/978-981-16-9812-5_2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLin, K. Y., Wu, Y. T., Hsu, Y. T., \u0026amp; Williams, P. J. (2021). Effects of infusing the engineering design process into STEM project-based learning to develop preservice technology teachers\u0026rsquo; engineering design thinking. \u003cem\u003eInternational Journal of STEM Education\u003c/em\u003e, \u003cem\u003e8\u003c/em\u003e(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003eArticle 1. https://doi.org/10.1186/s40594-020-00258-9\u003c/span\u003e\u003cspan address=\"Article 1. 10.1186/s40594-020-00258-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLippard, C. N., Lamm, M. H., Tank, K. M., \u0026amp; Choi, J. Y. (2019). Pre-engineering thinking and the engineering habits of mind in preschool classroom. \u003cem\u003eEarly Childhood Education Journal\u003c/em\u003e, \u003cem\u003e47\u003c/em\u003e(2), 187\u0026ndash;198. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10643-018-0898-6\u003c/span\u003e\u003cspan address=\"10.1007/s10643-018-0898-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLottero-Perdue, P., Bowditch, M., Kagan, M., Robinson-Cheek, L., Webb, T., Meller, M., \u0026amp; Nosek, T. (2016). An engineering design process for early childhood: Trying to engineer an egg package. \u003cem\u003eScience and Children\u003c/em\u003e, \u003cem\u003e54\u003c/em\u003e(3), 70\u0026ndash;77.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLucas, B., \u0026amp; Hanson, J. (2016). Thinking like an engineer: Using engineering habits of mind and signature pedagogies to redesign engineering education. \u003cem\u003eInternational Journal of Engineering Pedagogy (iJEP)\u003c/em\u003e, \u003cem\u003e6\u003c/em\u003e(2), 4\u0026ndash;13. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3991/ijep.v6i2.5366\u003c/span\u003e\u003cspan address=\"10.3991/ijep.v6i2.5366\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLucas, B., Hanson, J., \u0026amp; Claxton, G. (2014). \u003cem\u003eThinking like an engineer: Implications for the education system\u003c/em\u003e. \u003cem\u003eRoyal Academy of Engineering\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://winchester.elsevierpure.com/en/publications/thinking-like-an-engineer-implications-for-the-education-system-3/\u003c/span\u003e\u003cspan address=\"https://winchester.elsevierpure.com/en/publications/thinking-like-an-engineer-implications-for-the-education-system-3/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMayring, P. (2015). Qualitative content analysis: Theoretical background and procedures. In A. Bikner-Ahsbahs, C. Knipping, \u0026amp; N. Presmeg (Eds.), \u003cem\u003eApproaches to qualitative research in mathematics education\u003c/em\u003e (pp. 365\u0026ndash;380). Springer Netherlands. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/978-94-017-9181-6_13\u003c/span\u003e\u003cspan address=\"10.1007/978-94-017-9181-6_13\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMoore, T. J., Glancy, A. W., Tank, K. M., Kersten, J. A., Smith, K. A., \u0026amp; Stohlmann, M. S. (2014). A framework for quality K-12 engineering education: Research and development. \u003cem\u003eJournal of Pre-College Engineering Education Research (J-PEER)\u003c/em\u003e, \u003cem\u003e4\u003c/em\u003e(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003eArticle 2. https://doi.org/10.7771/2157-9288.1069\u003c/span\u003e\u003cspan address=\"Article 2. 10.7771/2157-9288.1069\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNational Research Council. (2009). \u003cem\u003eEngineering in K-12 education: Understanding the status and improving the prospects\u003c/em\u003e. National Academies. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.17226/12635\u003c/span\u003e\u003cspan address=\"10.17226/12635\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNguyen, H., \u0026amp; Santagata, R. (2021). Impact of computer modeling on learning and teaching systems thinking. \u003cem\u003eJournal of Research in Science Teaching\u003c/em\u003e, \u003cem\u003e58\u003c/em\u003e(5), 661\u0026ndash;688. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/tea.21674\u003c/span\u003e\u003cspan address=\"10.1002/tea.21674\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOnwuegbuzie, A. J., \u0026amp; Combs, J. P. (2011). Data analysis in mixed research: A primer. \u003cem\u003eInternational Journal of Education\u003c/em\u003e, \u003cem\u003e3\u003c/em\u003e(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.5296/ije.v3i1.618\u003c/span\u003e\u003cspan address=\"10.5296/ije.v3i1.618\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePapert, S., \u0026amp; Harel, I. (1991). Situating constructionism. \u003cem\u003econstructionism\u003c/em\u003e, \u003cem\u003e36\u003c/em\u003e(2), 1\u0026ndash;11.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRoth, W. M. (1996). Art and artifact of children\u0026rsquo;s designing: A situated cognition perspective. \u003cem\u003eJournal of the Learning Sciences\u003c/em\u003e, \u003cem\u003e5\u003c/em\u003e(2), 129\u0026ndash;166. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1207/s15327809jls0502_2\u003c/span\u003e\u003cspan address=\"10.1207/s15327809jls0502_2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eScardamalia, M., \u0026amp; Bereiter, C. (2014). Knowledge building and knowledge creation: Theory, pedagogy, and technology. In R. K. Sawyer (Ed.), \u003cem\u003eThe Cambridge handbook of the learning sciences\u003c/em\u003e (2nd ed., pp. 397\u0026ndash;417). Cambridge University Press. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1017/CBO9781139519526.025\u003c/span\u003e\u003cspan address=\"10.1017/CBO9781139519526.025\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSch\u0026ouml;n, D. A. (2017). \u003cem\u003eThe reflective practitioner: How professionals think in action\u003c/em\u003e. Routledge. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.4324/9781315237473\u003c/span\u003e\u003cspan address=\"10.4324/9781315237473\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchucker, K., Mcvee, M. B., Jarmark, C. J., \u0026amp; Shanahan, L. E. (2024). Exploring engineering habits of mind and disciplinary literacy connections in an elementary afterschool engineering literacies club. \u003cem\u003ePedagogies: An International Journal\u003c/em\u003e, 1\u0026ndash;22. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/1554480X.2022.2134871\u003c/span\u003e\u003cspan address=\"10.1080/1554480X.2022.2134871\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShaffer, D. W., Collier, W., \u0026amp; Ruis, A. R. (2016). A tutorial on epistemic network analysis: Analyzing the structure of connections in cognitive, social, and interaction data. \u003cem\u003eJournal of Learning Analytics\u003c/em\u003e, \u003cem\u003e3\u003c/em\u003e(3), 9\u0026ndash;45. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.18608/jla.2016.33.3\u003c/span\u003e\u003cspan address=\"10.18608/jla.2016.33.3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShaffer, D. W., Hatfield, D., Svarovsky, G. N., Nash, P., Nulty, A., Bagley, E., Frank, K., Rupp, A. A., \u0026amp; Mislevy, R. (2009). Epistemic network analysis: A prototype for 21st-century assessment of learning. \u003cem\u003eInternational Journal of Learning and Media\u003c/em\u003e, \u003cem\u003e1\u003c/em\u003e(2), 33\u0026ndash;53. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1162/ijlm.2009.0013\u003c/span\u003e\u003cspan address=\"10.1162/ijlm.2009.0013\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShulman, L. S. (2005). Signature pedagogies in the professions. \u003cem\u003eDaedalus\u003c/em\u003e, \u003cem\u003e134\u003c/em\u003e(3), 52\u0026ndash;59. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1162/0011526054622015\u003c/span\u003e\u003cspan address=\"10.1162/0011526054622015\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSiebert-Evenstone, A. L., Irgens, A., Collier, G., Swiecki, W., Ruis, Z., A. R., \u0026amp; Shaffer, W., D (2017). In search of conversational grain size: Modeling semantic structure using moving stanza windows. \u003cem\u003eJournal of Learning Analytics\u003c/em\u003e, \u003cem\u003e4\u003c/em\u003e(3), 123\u0026ndash;139. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.18608/jla.2017.43.7\u003c/span\u003e\u003cspan address=\"10.18608/jla.2017.43.7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSu, Y. S., Wang, S., \u0026amp; Liu, X. (2024). Using epistemic network analysis to explore primary school students\u0026rsquo; computational thinking in pair programming learning. \u003cem\u003eJournal of Educational Computing Research\u003c/em\u003e, \u003cem\u003e62\u003c/em\u003e(2), 559\u0026ndash;593. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/07356331231210560\u003c/span\u003e\u003cspan address=\"10.1177/07356331231210560\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSung, E., \u0026amp; Kelley, T. R. (2023). Elementary students\u0026rsquo; engineering design process: How young students solve engineering problems. \u003cem\u003eInternational Journal of Science and Mathematics Education\u003c/em\u003e, \u003cem\u003e21\u003c/em\u003e(5), 1615\u0026ndash;1638. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10763-022-10317-y\u003c/span\u003e\u003cspan address=\"10.1007/s10763-022-10317-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSweller, J. (2020). Cognitive load theory and educational technology. \u003cem\u003eEducational Technology Research and Development\u003c/em\u003e, \u003cem\u003e68\u003c/em\u003e(1), 1\u0026ndash;16. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s11423-019-09701-3\u003c/span\u003e\u003cspan address=\"10.1007/s11423-019-09701-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTan, Y., Hinojosa, C., Marquart, C., Ruis, A. R., \u0026amp; Shaffer, D. W. (2022). Epistemic network analysis visualization. In B. Wasson \u0026amp; S. Z\u0026ouml;rgő (Eds.), \u003cem\u003eAdvances in quantitative ethnography\u003c/em\u003e (Vol. 1522, pp. 129\u0026ndash;143). Springer International Publishing. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/978-3-030-93859-8_9\u003c/span\u003e\u003cspan address=\"10.1007/978-3-030-93859-8_9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTeddlie, C., \u0026amp; Tashakkori, A. (2010). \u003cem\u003eFoundations of mixed methods research: Integrating quantitative and qualitative approaches in the social and behavioral sciences\u003c/em\u003e. Sage.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWheeler, L. B., Navy, S. L., Maeng, J. L., \u0026amp; Whitworth, B. A. (2019). Development and validation of the Classroom Observation Protocol for Engineering Design (COPED). \u003cem\u003eJournal of Research in Science Teaching\u003c/em\u003e, \u003cem\u003e56\u003c/em\u003e(9), 1285\u0026ndash;1305. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/tea.21557\u003c/span\u003e\u003cspan address=\"10.1002/tea.21557\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWi̇Narno, N., Rusdi̇Ana, D., Samsudi̇N, A., Susi̇Lowati̇, E., Ahmad, N., \u0026amp; Afi̇Fah, R. M. A. (2020). The steps of the engineering design process (EDP) in science education: A systematic literature review. \u003cem\u003eJournal for the Education of Gifted Young Scientists\u003c/em\u003e, \u003cem\u003e8\u003c/em\u003e(4), 1345\u0026ndash;1360. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.17478/jegys.766201\u003c/span\u003e\u003cspan address=\"10.17478/jegys.766201\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"STEM education, Engineering Design Process, Engineering Habits of Mind, Primary school, Imitate-Fail-Improve pathway, Sequential Mixed Analysis, Epistemic Network Analysis","lastPublishedDoi":"10.21203/rs.3.rs-8794720/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8794720/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eThe Engineering Design Process (EDP) is considered as a signature pedagogy for cultivating the Engineering Habits of Mind (EHoMs). However, empirical studies tracing the development of the EHoMs within STEM education are limited. How these habits evolve across EDP stages and the cognitive mechanisms driving this evolution remains a \u0026ldquo;black box\u0026rdquo;. This study aims to reveal this process by tracking 63 fifth-grade students during a four-week rubber-band-powered car project. A sequential mixed analysis strategy was employed. Data from online discourse, classroom observations, design workbooks, iterative artifacts, and interviews were integrated and analyzed using Qualitative Content Analysis (QCA), Epistemic Network Analysis (ENA), and Thematic Analysis (TA).\u003c/p\u003e","manuscriptTitle":"Unveiling Learning Mechanisms in the Engineering Design Process: A Sequential Mixed Analysis of Primary School Students’ Cognitive Processes Through the Lens of Engineering Habits of Mind","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-03-06 20:07:14","doi":"10.21203/rs.3.rs-8794720/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"2237bdf2-d6c3-4496-bafb-f97b2cb7644c","owner":[],"postedDate":"March 6th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-03-06T20:07:14+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-06 20:07:14","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8794720","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8794720","identity":"rs-8794720","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}