Engineering Design and Minecraft Education for Primary Students’ Attitudes Toward Science: Development and Initial Evaluation of a Stakeholder-Informed Framework

Abstract

Positive attitudes toward science are relevant to students’ continuing interest in STEM learning. This study developed and initially evaluated a stakeholder-informed Engineering Design Process–Minecraft Game-Based Learning (EDP-MGBL) framework, with accessibility and participation opportunities treated as instructional design intentions rather than measured inclusion outcomes. Using a sequential two-phase mixed-methods development-and-evaluation design, Phase One drew on semi-structured interviews with Year 5 students, science teachers, curriculum specialists, and educational technology specialists to identify six needs: enjoyment, confidence, perceived relevance, learner agency, opportunities for collaboration, and visualisation of abstract energy concepts. The qualitative findings informed the subsequent design of an eight-week intervention. In Phase Two, 68 Year 5 students in western China participated in a quasi-experimental evaluation. The Attitude Toward Science Scale (ATSS), comprising Scientific Fun Feelings, Scientific Self-Efficacy, Scientific Value Identity, and Scientific Behavioural Intention, was administered at pretest, immediate post-test, and delayed post-test. A significant Group × Time interaction indicated different patterns of attitudinal change between the experimental and comparison groups. The experimental group’s total ATSS mean increased from 83.59 at pretest to 90.59 at post-test and 91.79 at delayed post-test; the comparison group showed comparatively small changes. These results provide preliminary evidence of an association between participation in the EDP-MGBL intervention and more favourable science attitudes. The six stakeholder-informed pathways are untested instructional design propositions, not identified causal mechanisms. Engagement, inclusion, and differential participation were not measured.

1. Introduction

Fostering positive attitudes toward science is an important goal of primary STEM education because students’ enjoyment, confidence, perceived value, and intentions to participate may be associated with later science-related choices (Osborne et al., 2003; Tytler & Osborne, 2012). Reviews and recent research have continued to examine affective outcomes alongside disciplinary learning and engineering performance, while cautioning that these outcomes do not necessarily improve in parallel (Saseendran & Thomas, 2025; Tzeng et al., 2025). Instructional approaches should therefore assess attitudes explicitly rather than infer them from technology use or achievement alone.

Prior science-education scholarship has documented variation in students’ attitudes and concerns about the perceived difficulty, abstraction, and relevance of science (Osborne et al., 2003; Tytler & Osborne, 2012). Such concerns motivate the design of learning activities that connect science concepts with observable problems and meaningful applications (Tytler, 2020). They do not, however, establish that the students recruited for the present intervention were disengaged; no disengagement criterion was used in sampling.

The Engineering Design Process (EDP) offers opportunities to apply scientific knowledge through problem identification, ideation, prototyping, testing, and refinement (Kelley & Knowles, 2016; Moore et al., 2014). Although integrated engineering activities have been associated with various cognitive and affective outcomes, positive attitudes are not guaranteed across contexts (Guzey et al., 2016; Tzeng et al., 2025). The present study therefore evaluates attitudes toward science directly and does not claim that engagement or the processes responsible for changes were measured.

Science attitudes, interest, self-efficacy, participation, and engagement are related but distinguishable constructs. Engagement is often conceptualised in behavioural, emotional, and cognitive dimensions (Fredricks et al., 2004), whereas this study assessed four attitudinal dimensions through the ATSS. Neither initial disengagement nor subsequent re-engagement was operationalised. Phase One captured stakeholder perceptions of instructional barriers; Phase Two examined changes in self-reported attitudes toward science. Statements about learner participation and inclusion in this paper refer to instructional design opportunities, not outcomes established by direct observation or inclusion measures.

Minecraft Education offers a digital setting in which students may construct representations, test ideas, and communicate around design challenges. Reviews identify potential educational applications while emphasising variation in study quality and the need for curriculum alignment and careful evaluation (Di Leo & Traetta, 2025; Slattery et al., 2025). The instructional contribution evaluated here is a combination of Minecraft Education and structured engineering design, not Minecraft as an isolated treatment. The study does not test whether digital activities independently produce attitudinal change.

Stakeholder input can help align educational interventions with the experiences and practical constraints of intended users (Fishman et al., 2013; Penuel et al., 2011). For energy education, stakeholder perspectives are relevant to identifying perceived barriers such as conceptual abstraction, limited opportunities to make design decisions, and the challenge of connecting energy systems with daily life. These needs informed design choices but were not separately evaluated as explanatory mediators.

Accordingly, this study developed and evaluated a stakeholder-informed Engineering Design Process–Minecraft Game-Based Learning (EDP-MGBL) framework for Year 5 energy-related STEM instruction. Phase One translated student, teacher, curriculum-specialist, and educational-technology-specialist perspectives into proposed instructional features. Phase Two assessed differences in science attitudes across three measurement occasions following an eight-week implementation. The resulting design propositions are offered to guide future explanatory research, not as evidence that specific causal pathways, equitable participation, or re-engagement were demonstrated.

2. Conceptual Framework

2.1. Engineering Design, Game-Based Learning, and Science Attitudes

Engineering design provides learners with opportunities to apply scientific knowledge within iterative processes of problem identification, solution generation, prototyping, testing, evaluation, and refinement. Within integrated STEM education, design-oriented approaches can provide contexts in which scientific knowledge is mobilised to address meaningful problems while simultaneously supporting creativity, inquiry, learner agency, and other cognitive and socio-emotional competencies (Saseendran & Thomas, 2025). Such activities may therefore support affective as well as cognitive dimensions of STEM learning because students encounter science through purposeful problem-solving rather than solely through decontextualised content transmission. However, positive attitudinal outcomes should not be assumed. Recent research examining engineering-focused STEM curricula demonstrates that participation in engineering design does not necessarily produce improvements in students’ attitudes, and affective outcomes may develop differently from disciplinary knowledge or design performance (Tzeng et al., 2025). This reinforces the need for studies that explicitly examine changes in learners’ attitudes alongside cognitive and performance-related outcomes.

Digital game-based environments may complement engineering-design pedagogy by providing interactive settings in which learners can construct, experiment, revise solutions, and collaborate. Minecraft is particularly relevant as an educational environment because its open-ended, construction-oriented structure has been used to support spatial thinking, creativity, engagement, collaboration, and the representation of abstract concepts across diverse curricular contexts (Di Leo & Traetta, 2025; Slattery et al., 2025). Nonetheless, evidence concerning the educational benefits of Minecraft requires cautious interpretation. In their systematic review of 29 intervention studies, Slattery et al. (2025) identified promising findings across academic, cognitive, motivational–affective, and social outcomes, including reported effects in mathematics and science, but found that all included studies were assessed as having a medium or high risk of bias. The authors therefore concluded that promising effects require replication through more rigorous research designs. Similarly, Di Leo and Traetta’s (2025) systematic review of 18 peer-reviewed studies concluded that Minecraft-based educational activities showed potential for supporting spatial thinking, creativity, engagement, and learning of abstract concepts, while emphasising the importance of curriculum alignment, assessment, and teacher preparation for effective implementation. Consequently, positive outcomes reported in Minecraft-based studies should not automatically be attributed to the technological platform itself, because outcomes are likely to depend on how Minecraft is embedded within the broader pedagogical and instructional environment.

The present study addresses this issue by integrating Minecraft Education with an Engineering Design Process (EDP) and grounding the resulting EDP-MGBL framework in stakeholder-identified instructional needs. Rather than positioning Minecraft as an independent technological intervention, the framework treats the platform as a pedagogical resource embedded within a structured engineering-design environment. This positioning is consistent with recent literature indicating that successful technology-supported and design-oriented STEM learning depends on purposeful pedagogical integration, appropriate scaffolding, learner agency, contextualisation, and alignment between instructional activities and intended learning outcomes (Di Leo & Traetta, 2025; Saseendran & Thomas, 2025). Accordingly, the contribution of the present study is not to establish that Minecraft itself improves students’ attitudes toward science, but to develop and provide an initial evaluation of a stakeholder-informed instructional framework in which Minecraft functions as one pedagogical resource within a broader engineering-design learning environment. This distinction is particularly important given current evidence showing both the educational potential of Minecraft-based learning and the methodological limitations of the existing evidence base (Slattery et al., 2025).

2.2. Theoretical Foundations of the EDP-MGBL Framework

The framework draws on integrated STEM education, constructionism, and game-based learning. Integrated STEM education emphasises the meaningful connection of science, technology, engineering, and mathematics through authentic, interdisciplinary challenges that encourage learners to apply knowledge, engage in problem-solving, and develop competencies relevant to contemporary society (English, 2016; Kelley & Knowles, 2016; Tytler, 2020). Inclusion is employed here as a design principle: tasks were intended to offer varied ways to explore, construct, discuss, and revise ideas and to lower perceived participation barriers. The study did not administer measures of accessibility, belonging, equity, or differential participation. The term “inclusive” therefore characterises intended design features, not an empirically established intervention outcome (National Academies of Sciences, Engineering, and Medicine, 2025).

The theory of Constructionism, according to which the process of learners’ cognition enables him/her to understand the material better and make his/her own constructive objects to be shared and tested, has been an important aspect contributing to the framework (Papert, 1980). With the help of the Minecraft Education platform, pupils are not only receivers of information but also creators of digital objects, processes, and products. The process of designing, creating, and improving constructs contributes to pupils’ learning process and enhances pupils’ creativity, technical thinking, and problem-solving ability. From the point of view of Game-Based Learning, education is considered a process taking place due to the presence of various challenges, exploration, experimentation, feedback, and active participation in the educational process. Minecraft Education has created a simulation of a learning platform where pupils can analyze natural phenomena, collaborate with their friends, and implement and test their ideas through a process of online design (Di Leo & Traetta, 2025; Karsenti et al., 2021).

2.3. Research Questions and Study Contributions

These perspectives informed a sequential study in which stakeholders’ accounts guided the development of an engineering-design and Minecraft-based learning environment. The subsequent evaluation focused on attitudes toward science, the outcome measured by the ATSS. Two research questions guided the investigation:

• What instructional needs did stakeholders identify for developing the EDP-MGBL framework?
• How did attitudes toward science change across three assessment occasions for students in the EDP-MGBL and comparison conditions?

The study makes three bounded contributions. First, it documents how six stakeholder-identified needs were translated into design requirements for an eight-week energy-focused intervention. Second, it proposes six theoretically informed explanatory pathways as hypotheses for future testing, without claiming to have established causal mechanisms. Third, it provides an initial quasi-experimental comparison of attitudes toward science, including a delayed assessment that examines persistence within the observed follow-up interval rather than long-term sustainability. The study does not establish improvement in inclusion, re-engagement, or objective participation.

3. Methodology

3.1. Research Design

This study employed a sequential two-phase mixed-methods development-and-evaluation design to develop and initially evaluate a stakeholder-informed Engineering Design Process–Minecraft Game-Based Learning (EDP-MGBL) framework. The design addressed two complementary research questions: the identification of contextual instructional needs and the evaluation of changes in students’ attitudes toward science. Phase One involved semi-structured interviews with students, science teachers, curriculum specialists, and educational technology specialists. The qualitative findings informed the development of the EDP-MGBL framework by translating stakeholder-identified needs into instructional design features. Phase Two involved a quasi-experimental evaluation of the resulting framework using two intact Year 5 classes. Students’ attitudes toward science were assessed at three measurement occasions: pretest, immediate post-test, and delayed post-test. The integration of qualitative and quantitative components occurred primarily through a building approach, whereby the qualitative findings informed the intervention, subsequently evaluated quantitatively (Creswell & Plano Clark, 2018; Fishman et al., 2013). Although the study drew on educational design research principles, it did not involve repeated cycles of classroom implementation, evaluation, and redesign (McKenney & Reeves, 2019, 2025). Accordingly, the study is characterised as a two-phase development-and-evaluation investigation rather than a full educational design research study.

3.2. Phase One: Stakeholder-Informed Framework Development

3.2.1. Rationale for Stakeholder Involvement

The first step toward the development of STEM programs usually involves the use of theoretical models or curriculum goals as reference points. While these approaches make it possible to create ideas, they are often unsuccessful in solving the problems faced by students and educators in real classrooms. In recent studies on STEM education, the importance of engaging stakeholders in the instructional process was noted, as students and teachers often have experience in the given context that cannot be derived from theory alone (Fishman et al., 2013; Penuel et al., 2011). Therefore, the development of the EDP-MGBL framework was based on the input from different parties involved in the process. Students participated because they are the main beneficiaries of the program and could provide firsthand information about the challenges they face, their interests and motivation, and the reasons they feel unmotivated to learn science. Science teachers were involved in the process because of their experience with the curriculum and teaching practice. The experts in curriculum and instructional design have shared their insights regarding curriculum alignment, pedagogical quality, inclusivity, and instructional coherence. The aim of incorporating various groups of stakeholders was to guarantee the framework meets real educational needs while sustaining theoretical integrity and curriculum applicability, as well as being accessible for different types of learners in an inclusive STEM environment (National Academies of Sciences, Engineering, and Medicine, 2025).

3.2.2. Participants of Phase One

Phase One involved 10 Year 5 students, three primary science teachers, two curriculum specialists, and two educational technology specialists. Participants were selected purposively to obtain information-rich perspectives relevant to science learning, curriculum requirements, and technology-supported STEM instruction. This selection logic is consistent with criterion-based purposive sampling approaches (Palinkas et al., 2015). The stakeholder groups contributed different types of information for identifying needs and refining the proposed intervention. Recruitment criteria and contextual descriptors should correspond to the original recruitment records.

The 10 Year 5 students represented the intended learner population for the intervention. They were selected because they could provide first-hand accounts of their experiences of primary science learning, including perceived difficulties with energy concepts, confidence in completing science-related tasks, opportunities for participation and collaboration, and experiences of technology-supported learning. The students were drawn from the broader primary-school context in western China in which the study was conducted. Year 5 was particularly relevant because students at this level were approximately 11 years old, had already experienced several years of formal science education, were generally familiar with basic computer operation, and encountered systematic energy-related content within the primary science curriculum.

The three primary science teachers contributed classroom perspectives on Year 5 science content, student experiences, instructional approaches, and practical implementation constraints. Their accounts complemented student perspectives in identifying potential design requirements.

The two curriculum specialists were selected on the basis of their professional knowledge of primary science curriculum requirements, learning objectives, content appropriateness, pedagogical alignment, and inclusive STEM education. Their contribution focused on examining whether the proposed learning activities were consistent with curriculum expectations and whether the content and instructional structure were appropriate for Year 5 learners. In the framework-development process, their recommendations were used to strengthen curriculum alignment and instructional coherence.

The two educational technology specialists were selected because of their expertise in technology-supported learning and the pedagogical application of digital tools. Their role was to examine the technical and pedagogical suitability of Minecraft Education, including usability, accessibility, student participation, and the alignment between the technological activities and the intended learning objectives. Their feedback was also used to inform the design of Minecraft-based engineering activities and collaborative learning tasks.

3.2.3. Development of the EDP-MGBL Framework

Framework development proceeded through four stages. First, stakeholder interview data were analysed to identify recurrent barriers and instructional needs relating to science learning. Second, the identified needs were compared across stakeholder groups to distinguish recurring concerns from group-specific perspectives. Third, each recurring need was translated into an instructional design requirement. For example, stakeholder interest in enjoyable and hands-on science experiences informed the use of mission-based Minecraft engineering challenges, low confidence informed progressively scaffolded design activities, insufficient real-life relevance informed authentic energy-related problems, teacher-directed learning informed student-led design decision-making, stakeholder interest in purposeful collaborative learning informed team-based engineering activities, and difficulty visualising energy processes informed interactive digital modelling and simulation. Fourth, the proposed activities were reviewed by curriculum and educational technology specialists for curricular alignment, pedagogical appropriateness, technological feasibility, and accessibility. Feedback from these reviews was used to refine the framework before implementation in Phase Two.

3.3. Phase Two: Evaluation of the EDP-MGBL Framework

3.3.1. Quasi-Experimental Design

Phase Two employed a non-equivalent control group design with three repeated measurements to evaluate changes in students’ attitudes toward science following implementation of the EDP-MGBL framework. The intervention was conducted over eight weeks with two intact Year 5 classes, comprising an experimental group (n = 34) and a comparison group (n = 34). The experimental group participated in engineering-design activities supported by Minecraft Education, whereas the comparison group received conventional science instruction covering the same curricular content. Both groups completed the Attitude Toward Science Scale (ATSS) before the intervention, immediately after the intervention, and at a delayed post-test. The principal quantitative comparison concerned differences in attitudinal change across the three measurement occasions between the two groups. Because the groups consisted of intact classes from different schools and participants were not randomly assigned, potential school-level, teacher-related, and contextual differences could not be fully controlled. Therefore, the design supports a preliminary comparison of attitudinal changes but does not establish definitive causal effects of the intervention (Capili & Anastasi, 2024).

3.3.2. Participants and Context

The evaluative phase involved 68 Year 5 students from two public primary schools located in an urban area of Sichuan Province, western China. The students were approximately 11 years old and were drawn from two intact Grade 5 classes. One class consisting of 34 students was assigned to experimental conditions, while another class of 34 students from a second public primary school served as the comparison group. The use of intact classes was intended to preserve the existing classroom structure and minimise disruption to normal school organisation.

The two classes were selected purposively to ensure a broadly comparable instructional context. According to the sampling criteria reported in the study, students in both classes had similar educational backgrounds, followed the same teaching schedule, possessed comparable levels of computer skills, and had access to computers through their respective school computer laboratories. These characteristics were considered important because the intervention required students to participate in STEM learning activities based on Minecraft Education.

Purposive sampling was also used because Year 5 represented a contextually appropriate stage for the intervention. Students at this level had already experienced several years of formal science education, were generally familiar with basic computer operation, and encountered systematic learning of energy-related topics within the primary science curriculum. Year 5 was also described as an important developmental stage for the formation of attitudes toward science, logical thinking, and creativity.

The activities were targeted at inclusive STEM learning by providing chances to explore, collaborate, and solve problems, as well as to participate in learning activities actively. In the process of the activities, pupils had to find problems, generate ideas, choose solutions, create their prototypes, and continually improve their designs (Kelley & Knowles, 2016; Moore et al., 2014).

3.3.3. Intervention Structure and Learning Activities

The eight-week EDP-MGBL intervention addressed energy-related problems through an Ask–Imagine–Plan–Create–Improve sequence. Across sessions, learners examined energy-system problems, suggested solutions, constructed and tested Minecraft representations, and revised designs. Table 1 identifies the reported weekly topics and tasks and offers operational descriptions synthesised from the intervention account. These descriptions are summaries of the instructional structure, not verbatim copies of teacher-issued prompts; detailed lesson scripts, digital files, rubrics, and fidelity documentation are not reproduced.

Table 1. Eight-Week Structure of the EDP-MGBL Intervention.

Week	STEM Focus	EDP Activities	Minecraft Activity and Indicative Task Direction
1	Energy concepts	Ask–Imagine	Explore energy systems and identify design problems Indicative direction: Inspect a model energy system; identify what it does and propose one energy-related problem the design should solve.
2	Electrical energy	Ask–Plan	Examine and repair Redstone-based circuits Indicative direction: Inspect an incomplete Redstone electrical circuit; identify a fault, propose a connection change, and test whether the revised circuit functions.
3	Renewable energy	Imagine–Plan	Design a renewable-energy solution Indicative direction: Choose a renewable-energy problem; sketch or describe a Minecraft design and explain the energy source and proposed operation.
4	Solar energy	Create	Construct and test a solar-energy system Indicative direction: Build a model solar-energy system; run a test and explain what the representation does and does not show about solar power.
5	Hydropower	Create–Improve	Repair or redesign a hydropower system Indicative direction: Inspect a hydropower model; identify a problem and revise the representation to explain energy transfer or system function.
6	Energy efficiency	Plan–Create	Construct an energy-efficient building Indicative direction: Design an energy-efficient building; choose features intended to reduce energy demand and justify these choices.
7	Sustainable community	Create–Improve	Develop an integrated sustainable-energy solution Indicative direction: Combine selected renewable-energy and efficiency features into a Minecraft community design and explain key trade-offs.
8	Design refinement	Test–Improve–Reflect	Test, present, critique, and improve designs Indicative direction: Test the integrated design, describe a weakness, revise a feature, and present the reasoning for the improvement.

Note. EDP = Engineering Design Process; MGBL = Minecraft game-based learning.

The comparison group continued to receive conventional science instruction on the same curricular content during the corresponding eight-week period. Instruction followed the school’s existing teaching approach and curriculum rather than the EDP-MGBL activities. Students in the comparison condition therefore did not participate in the structured Minecraft Education-based engineering-design activities implemented with the intervention group. Both conditions followed the same curricular topic and were implemented over the same study period, while differing principally in the instructional approach used to engage students with the energy content.

3.4. Instrumentation

The Attitude Toward Science Scale (ATSS; Guo & Liu, 2022) was used to assess science attitudes among Year 5 students. Its 20 items cover four dimensions: Scientific Fun Feelings (items 1–8), Scientific Self-Efficacy (items 9–12), Scientific Value Identity (items 13–16), and Scientific Behavioural Intention (items 17–20). These dimensions concern enjoyment, perceived competence, perceived value, and intended future participation, respectively. Importantly, the instrument does not directly record classroom engagement, observed participation, accessibility, or equitable inclusion.

All items were rated using a five-point Likert scale, ranging from 1 = Strongly Disagree to 5 = Strongly Agree, with the intermediate categories of Disagree, Unsure, and Agree. Accordingly, the total ATSS score could range from 20 to 100, with higher scores indicating a more favourable attitude toward science. The possible score range for Scientific Fun Feelings was 8–40, whereas each of the other three four-item dimensions ranged from 4–20. The ATSS was selected because it was developed and validated specifically for Chinese primary-school students. Prior large-scale use involved more than 60,000 Year 5 students from 277 schools in Zhengzhou City, Henan Province, providing evidence that the instrument is appropriate for the age and educational context of the present study (Guo & Liu, 2022). Because the norm population consisted of Chinese Year 5 students, the original Chinese-language version of the scale was directly adopted for the present study rather than translated from another language.

Reported scale-validation indices in the study materials included (CFI = 0.992, TLI = 0.991, RMSEA = 0.076), factor loadings above 0.40, and reliability coefficients above 0.90 for the large-scale validation; these specific psychometric details should be checked against the primary validation report. The present study reported a 50-student pilot, with overall Cronbach’s alpha = 0.80, and assessed internal consistency at all three study occasions (See Section 4.3.1). Interpretation of coefficient alpha should consider the scale context rather than treating 0.70 as an absolute validity threshold; the frequently used 0.70 guideline is associated with Nunnally and Bernstein (1994). Appendix A presents the reported dimensional structure of the ATSS together with an adapted English-language item set. The ATSS was administered at pretest, immediately after the eight-week intervention, and at a delayed post-test. The design supports inference about change in self-reported science attitudes during the study period only.

3.5. Data Collection

The semi-structured interviews were conducted during the needs phase regarding the stakeholders’ experiences, expectations, and perceptions related to different science-related activities, education in STEM, and challenges encountered while participating in such activities. This method was chosen as it provided consistency while retaining flexibility (Creswell & Creswell, 2018). The students were interviewed on various issues such as their knowledge of energy concepts, difficulties they faced during the process of studying science, what they think about studying science, applying technologies, attitudes towards gamified learning, and reasons why students might quit studying science. These discussions helped understand students’ needs better. Regarding the teachers, the interviews were aimed at exploring questions concerning issues related to teaching difficulties, involvement of the students, implementation of the curriculum, teaching methods, inclusiveness, and possibilities of integrating STEM learning. Teachers analysed their opinions about barriers to engagement.

Independent interviews were held with each participant so that they could express their ideas and opinions freely and deeply. They were recorded on tape. In this case, we received qualitative data for analysis. The data identified stakeholders’ needs and provided support for developing the EDP-MGBL framework (Braun & Clarke, 2006, 2021). In addition to students and teachers, curriculum experts and educational technology specialists were invited, as important stakeholders, to provide expert opinions on curriculum programming, pedagogy, technology, and inclusiveness in learning. Data from educational curriculum professionals were generated through semi-structured interviews conducted with the experts and framework evaluation sessions with the professionals. The experts checked whether the learning objectives were achieved and whether the curriculum met the requirements of inclusiveness in the teaching process.

The qualitative findings are reported using pseudonymous participant identifiers, with S referring to students and ST referring to science teachers. Representative excerpts are presented in English to illustrate the six stakeholder-informed themes identified during the framework-development phase. The original interview language, the accuracy of the English wording, and the correspondence between each quotation and its source recording or transcript have been independently verified using the primary study records.

Experts working in educational technology took part in the same process but paid special attention to the implementation of Minecraft Education and other technological aspects into the project. They evaluated the technical side of the process, ease of use and participation in the process, accessibility of the learning process, and whether the objectives of learning in the process of using Minecraft’s tools were achieved.

The evaluation from experts was processed along with the results gathered from learners and the educators; as a result, instead of using expert opinions in the verification process, they were included in the creation of the EDP-MGBL framework. The recommendations of the curriculum specialists contributed to the improvement of the curriculum, while the suggestions made by the specialist in educational technology were used to create the Minecraft exercises and group learning activities. Taking this into account, the involvement of educational experts in the process of creating the framework increased its trustworthiness and validity because the process was based not only on the requirements of the students and the teachers but also on the knowledge in the field of curriculum development and STEM education (Fishman et al., 2013; Penuel et al., 2011).

3.6. Data Analysis

Qualitative and quantitative data were analysed separately according to the objectives of each phase of the study. Phase One interview data were analysed using reflexive thematic analysis, whereas Phase Two ATSS data were analysed using descriptive and inferential statistics.

3.6.1. Qualitative Data Analysis

In Phase One, interview recordings were transcribed verbatim and analysed using reflexive thematic analysis (RTA), following the approach developed by Braun and Clarke (2021) and further illustrated in contemporary applications of the method (Byrne, 2022). Reflexive thematic analysis conceptualises analysis as an active and interpretive process in which themes are developed through the researcher’s sustained engagement with the dataset rather than treated as objective entities that simply emerge from the data (Braun & Clarke, 2021; Byrne, 2022). Analysis proceeded recursively through familiarisation with the dataset, initial coding, development of candidate themes, review of themes against coded extracts and the complete dataset, refinement and naming of themes, and production of the analytic narrative. These phases were treated as iterative rather than strictly sequential, allowing movement between coding, theme development, and interpretation as analytic understanding developed (Braun & Clarke, 2021; Byrne, 2022).

Coding was primarily inductive while remaining sensitised to the study’s focus on barriers, instructional needs, confidence, participation, collaboration, technology use, and understanding of energy concepts. This approach allowed patterns of meaning to be developed from participants’ accounts while recognising that qualitative analysis is inevitably shaped by the research questions, theoretical assumptions, and interpretive positioning of the researchers (Braun & Clarke, 2021). Data from students, teachers, curriculum specialists, and educational technology specialists were initially examined within stakeholder groups before comparisons were made across groups to identify areas of convergence, divergence, and complementary perspectives.

Throughout the analysis, the research team maintained analytic notes documenting coding decisions, theme revisions, interpretive reflections, and links between stakeholder evidence and subsequent framework-design decisions. The thematic analysis was conducted using the original-language transcripts, while English translations were prepared subsequently for reporting representative quotations. The language in which the interview data were transcribed and coded, and the procedures through which the English-language excerpts were produced, gained confirmation from the primary research records.

Reflexive discussions were used to critically examine how the researchers’ prior interests and experiences in STEM education, engineering design, and educational technology might shape their engagement with and interpretation of the data. Such reflexive engagement is consistent with RTA, which acknowledges researchers’ subjectivity as an integral analytic resource rather than viewing it solely as a source of bias to be eliminated (Braun & Clarke, 2021; Byrne, 2022).

Consistent with this methodological orientation, credibility was pursued through transparent documentation of the analytic process, repeated and sustained engagement with the transcripts, recursive development and review of themes, comparison of perspectives across stakeholder groups, and the use of participant quotations to demonstrate how the interpretations were grounded in the dataset. Inter-coder agreement or coding consensus was not used as an indicator of analytic reliability, because contemporary RTA does not conceptualise coding as a process in which independent coders are expected to arrive at a single objectively correct interpretation. Instead, researcher reflexivity, theoretical coherence, depth of engagement with the data, and transparency of the analytic process were emphasised as indicators of methodological quality (Braun & Clarke, 2021; Byrne, 2022).

3.6.2. Quantitative Data Analysis

During the second phase, the quantitative data gathered from the Attitude Toward Science Scale were inspected for missing data, errors in data entry, outliers, and the assumptions necessary for parametric tests. Summary statistics, including means and standard deviations, were computed for the total score on the ATSS and the four dimensions of Scientific Fun Feelings, Scientific Self-Efficacy, Scientific Value Identity, and Scientific Behavioural Intention in the pre-test and immediate and delayed post-tests. Reliability of measurement was assessed using Cronbach’s alpha at each measurement point of both the total scale and subscales. Effectiveness of the intervention was analysed using a mixed-design ANOVA with group (EDP-MGBL vs. comparison) between participants and time (pre-test vs. post-test vs. delayed post-test) within participants as independent variables. The group-by-time interaction was used as the main test for effectiveness, since it shows whether changes occurred over time in the experimental group differ from those taking place in the comparison group.

3.6.3. Integration of Qualitative and Quantitative Findings

Integration occurred through a building sequence at the interface between Phase One and Phase Two. The researchers translated each stakeholder-informed theme into a design feature (See Section 4.2) used the resulting features to structure the intervention, and subsequently evaluated science attitudes with the ATSS. This is an integration of qualitative evidence into intervention development, not triangulation of qualitative and quantitative measurements of the same outcome. No post-intervention interview or observation dataset was collected to explain the ATSS patterns. Accordingly, the linkage between design features and changes in science attitudes should be regarded as theoretically plausible, not empirically identified mediation or process evidence (Creswell & Plano Clark, 2018).

3.7. Ethical Procedures and Informed Consent

Before the commencement of recruitment and data acquisition, a thorough briefing and workshop were conducted for the parents and/or legal representatives, teachers, as well as pupil volunteers. During the meeting, the researchers explained the study’s aims and procedures, activities to be conducted, timelines, confidentiality conditions, and the participants’ rights. The participants learned that their educational activities might bring a chance to learn STEM subjects, engineering design, and problem-solving procedures, as well as the use of the Minecraft Education app. Nonetheless, their participation was totally optional, and there were no guarantees of any benefits.

The researchers made sure to differentiate clearly between educational programme participation and research involvement. Thus, individuals who were part of the programme or the workshop were not obliged to participate in the research study, finish the questionnaire, or have a formal interview. Participants were given the option of partaking in the educational programme without being obliged to complete the research study. They were notified that they had the right not to answer any questions if they wished to do so and that they could choose to withdraw from the research at any time. The researchers also obtained informed consent from participants before conducting the study. In case the participant was a fifth grader, parenting consent was obtained from their parents or guardians beforehand.

4. Results

Phase One generated six stakeholder-informed instructional themes that were used to develop the EDP-MGBL framework. The themes represent analytically distinct concerns: enjoyment (affective experience), confidence (perceived capability), relevance (everyday meaning), agency (control over design decisions), collaboration (opportunities for joint reasoning), and visualisation (representing abstract energy processes). Representative quotations are presented only once under the theme they most directly illustrate. Phase Two reports a separate quantitative evaluation of attitudes toward science.

4.1. Stakeholder Needs Informing the Development of the EDP-MGBL Framework

Phase One stakeholder accounts were organised into six analytically distinct themes: the affective appeal of science activities, confidence in performing tasks, connections to everyday applications, opportunities for learner decision-making, collaborative knowledge construction, and the visualisation of abstract energy concepts. The themes were distinguished by their principal analytic focus, although stakeholders could discuss more than one instructional concern. The selected quotations illustrate the interpretation of interview accounts and informed the instructional design of the EDP-MGBL framework; they do not demonstrate changes resulting from the subsequent intervention.

4.1.1. Theme 1: Affective Appeal of Hands-On Science Activities

Theme 1 concerns stakeholders’ accounts of what makes science learning enjoyable and attention-catching. It addresses affective responses to instructional activities rather than students’ capacity to complete tasks (Theme 2) or their control over instructional decisions (Theme 4). One science teacher described students’ immediate interest in experiments:

“When students hear of an experiment, they immediately show interest. That’s why I want to include more experimental activities.”

(ST1)

The teacher’s observation suggests that hands-on experimentation was perceived as an attractive instructional format. It does not establish that all students disliked science or that disengagement had been measured. The framework consequently incorporated mission-based Minecraft engineering activities as a design response intended to create opportunities for enjoyable exploration; changes in enjoyment were examined subsequently as part of the attitude scale, not established by these interview data.

4.1.2. Theme 2: Limited Confidence Constrained Students’ Participation in Science Learning

Theme 2 captures perceived difficulties in undertaking science tasks independently, especially tasks requiring idea generation, experimentation, or original design. Unlike Theme 1, the central issue is not whether students found activities interesting but whether they believed they could complete them successfully. Two students described difficulties with working independently:

“I could not handle the project on my own and the teacher said that my project was unoriginal.”

(S4)

“For me, doing experiments is very hard and I cannot do it by myself.”

(S5)

Both excerpts refer directly to perceived ability and task independence. Their relevance to self-efficacy differs from the teachers’ comments about enjoyment or group discussion. These accounts informed progressively scaffolded engineering tasks, opportunities for testing, and revision of unsuccessful designs. They do not establish that the intervention subsequently caused improvements in students’ perceived competence.

4.1.3. Theme 3: Interest in Practical Applications of Energy Concepts

Theme 3 concerns how the relevance of science was expressed through questions about recognisable technologies and their applications. It is distinct from Theme 6, which focuses on the representational difficulty of understanding processes that cannot readily be observed. A student expressed curiosity about the uses of solar energy:

“I am very interested in the solar energy. I would like to know how it works and what products can be made using the solar energy.”

(S2)

A science teacher also emphasised the importance of drawing on existing knowledge when teaching unfamiliar content:

“I will draw on what learners already know to enhance their understanding.”

(ST2)

These accounts do not suggest that students regarded science as unimportant; rather, they indicate an instructional opportunity to link interest in energy technologies with practical applications. They informed the use of authentic solar-energy, hydropower, and sustainable-design problems as settings for applying science knowledge in the EDP-MGBL framework.

4.1.4. Theme 4: Limited Learner Agency in Teacher-Directed Science Lessons

Theme 4 addresses instructional decision-making: who determines the learning sequence, proposes solutions, and decides how tasks are carried out. It is analytically separate from Theme 1 (the emotional appeal of activities) and Theme 5 (opportunities for peer discussion). A student described a teacher-led sequence:

“In science classes, teachers usually give lectures first and afterward let us do the tasks.”

(S3)

The excerpt concerns the sequencing of teaching and student tasks. It suggests a teacher-directed approach without proving that learners were wholly passive or lacked interest. This interpretation informed the framework’s emphasis on student-generated design options, selection among alternatives, testing, and revision. Learner agency itself was not separately measured as an intervention outcome.

4.1.5. Theme 5: Stakeholder-Identified Need for Collaborative Knowledge Construction

Theme 5 concerns stakeholder interest in learning experiences where peers explain, discuss, and develop ideas together. It is distinct from Theme 2, which concerns confidence in completing tasks independently, and Theme 4, which concerns who directs the learning process. One teacher specifically mentioned group discussion:

“Students prefer student-centered approaches which include hands-on experiences and discussions in groups.”

(ST3)

This quotation identifies the perceived educational value of group discussion. By itself, however, it does not document how often group work occurred or whether existing classroom collaboration was insufficient. Accordingly, the theme is conservatively framed as a stakeholder-identified design need rather than a demonstrated deficit in classroom participation. Team-based engineering tasks were included to create opportunities for shared explanation, negotiation, and design refinement; the quality of collaboration was not measured in Phase Two.

4.1.6. Theme 6: Energy Concepts Were Difficult to Visualize and Consequently Difficult to Understand

Theme 6 concerns students’ difficulty forming concrete representations of energy processes that they could not directly observe or experience. Whereas Theme 3 focuses on the practical relevance of energy technologies, Theme 6 concerns understanding how unfamiliar energy phenomena operate. One student described the distance between energy content and lived experience:

“Some energy knowledge that I have is outside of my real experience like knowledge of nuclear energy or wind energy.”

(S1)

A science teacher similarly identified the abstract character of the topic and students’ limited experiential background:

“The knowledge points are too abstract, and students’ life experience is very limited.”

(ST2)

These excerpts suggest a representational challenge in learning about energy rather than an absence of curiosity or scientific value. They informed the use of interactive Minecraft and Redstone models to make selected circuits and energy-system relationships more accessible to exploration. The hypothesised contribution of such models to understanding was not independently tested as a causal pathway.

4.2. Resulting EDP-MGBL Framework

The framework-development process produced six instructional design features informed by the stakeholder needs identified in Phase One. Table 2 presents the relationships between these needs, their corresponding instructional features, and the hypothesised functions of those features.

Table 2. Stakeholder-Identified Needs, EDP-MGBL Design Features, and Hypothesised Design Propositions.

Stakeholder Need	Design Feature	Hypothesised Explanatory Function
Need for enjoyable science learning experiences	Minecraft-based engineering missions	Potential enjoyment-related function
Reported difficulty with independent science tasks	Scaffolded engineering activities	Potential perceived-competence function
Interest in everyday energy applications	Authentic energy projects	Potential relevance-related function
Teacher-directed task sequencing	Student-led design challenges	Potential agency-related function
Stakeholder-identified need for purposeful group discussion	Team-based EDP activities	Potential collaborative-learning function
Difficulty representing unfamiliar energy processes	Minecraft simulations	Potential visual-representation function

The resulting EDP-MGBL framework combined an engineering-design sequence (Ask–Imagine–Plan–Create–Improve) with energy-related Minecraft tasks over eight weeks. The intervention was organised around the six stakeholder-derived design needs presented in Table 2. This is a design mapping, not a post-intervention qualitative explanation of the quantitative results. The six design features were incorporated into a combined intervention rather than implemented as separate experimental components.

The framework was intended to increase opportunities for enjoyable science learning, and scaffolded tasks were intended to support confidence through achievable design, testing, and revision steps. These aims informed lesson planning; changes in enjoyment and self-efficacy were assessed only as components of the attitudinal measure, not as experimentally isolated pathways.

Authentic tasks such as solar-energy design, hydropower-system repair, and sustainable-community planning were intended to connect energy concepts with practical applications. Learners were also invited to identify problems, propose alternatives, and appraise solutions. This describes the instructional opportunities provided, not directly measured growth in perceived relevance or learner agency.

Team tasks were designed to invite idea sharing, division of roles, explanation, and joint design revision. Minecraft representations, including Redstone circuits and renewable-energy systems, were intended to help make energy concepts more visible and manipulable. Collaboration and visualisation were proposed design features; the present study did not directly observe changes in either construct.

The framework integrates science (energy concepts), technology (Minecraft Education and Redstone representations), engineering (iterative design), and mathematics (measurement and comparison during design). Its inclusion-oriented purpose was to provide varied ways for students to engage with the activities. Because accessibility, belonging, equitable participation, and outcomes for learner groups were not evaluated, the study cannot conclude that the intervention achieved inclusive outcomes.

4.3. Effects of the EDP-MGBL Framework on Attitudes Toward Science

4.3.1. Internal Consistency Reliability of the ATSS

Cronbach’s alpha was calculated for the ATSS at pretest, post-test, and delayed post-test (Table 3). The total-scale estimates were 0.913, 0.924, and 0.932, respectively; subscale estimates ranged from 0.792 to 0.891 across the three occasions. All estimates exceeded 0.70, a commonly cited heuristic associated with Nunnally and Bernstein (1994), although alpha alone does not establish dimensional validity or the absence of measurement bias.

Table 3. Internal Consistency Reliability of the ATSS Across Measurement Occasions.

ATSS Dimension	Items	Pretest α	Post-Test α	Delayed Post-Test α
Scientific Fun Feelings	8	0.864	0.882	0.891
Scientific Self-Efficacy	4	0.821	0.843	0.851
Scientific Value Identity	4	0.803	0.824	0.836
Scientific Behavioural Intention	4	0.792	0.814	0.826
Total ATSS	20	0.913	0.924	0.932

4.3.2. Descriptive Changes in Attitudes Toward Science

The second research question examined the effects of the EDP-MGBL framework on students’ attitudes toward science. Table 4 presents the mean scores and standard deviations for students’ overall attitudes toward science across the three measurement occasions. At pretest, the experimental group recorded a mean score of 83.59 (SD = 5.720), while the comparison group obtained a mean score of 82.62 (SD = 5.680). The relatively similar pretest means indicate that the two groups began the study with broadly comparable levels of attitude toward science.

Table 4. Descriptive Statistics for Overall Attitude Toward Science Scores.

Assessment	Experimental Mean Score	Experimental SD	Comparison Mean Score	Comparison SD
Pretest	83.59	5.720	82.62	5.680
Post-test	90.59	5.612	83.62	5.724
Delayed post-test	91.79	5.381	83.62	5.484

Following the intervention, the experimental group showed a marked increase in its mean ATSS score to 90.59 (SD = 5.612), whereas the comparison group recorded only a small increase to 83.62 (SD = 5.724). At the delayed post-test, the experimental group maintained the higher score, with a mean of 91.79 (SD = 5.381), while the comparison group remained at 83.62 (SD = 5.484). Descriptively, the experimental group improved by 7.00 points from pretest to post-test and by 8.20 points from pretest to delayed post-test. In contrast, the comparison group increased by only 1.00 point from pretest to post-test, with no additional increase at the delayed post-test.

The relatively similar standard deviations across the three measurement occasions also suggest that the variability of scores within each group remained reasonably stable over time. These descriptive patterns are consistent with the inferential analysis reported in the manuscript, which found a significant Group × Time interaction, Pillai’s Trace = 0.73, F(2, 65) = 86.74, p < 0.001, partial η² = 0.73. The experimental group also scored significantly higher than the comparison group at post-test, t(66) = 5.07, p < 0.001, and at delayed post-test, t(66) = 6.20, p < 0.001.

4.3.3. Inferential Analysis of Group × Time Effects

To examine the effects of the EDP-MGBL framework on students’ attitudes toward science, a mixed-design ANOVA was conducted with group (experimental vs. control) as the between-subjects factor and time (pretest, post-test, delayed post-test) as the within-subjects factor. The analysis revealed a significant main effect of time, Pillai’s Trace = 0.78, F(2, 65) = 117.17, p < 0.001, partial η² = 0.78, indicating that students’ attitudes toward science changed significantly across the three measurement points. More importantly, a significant Group × Time interaction was observed, Pillai’s Trace = 0.73, F(2, 65) = 86.74, p < 0.001, partial η² = 0.73, demonstrating that the pattern of attitudinal change differed significantly between the experimental and control groups.

Estimated marginal means showed that both groups had comparable attitudes toward science at pretest (experimental: M = 83.59; control: M = 82.62). Following the intervention, the experimental group demonstrated substantial improvement (M = 90.59), whereas the control group showed little change (M = 83.62). These gains were maintained at the delayed post-test, with the experimental group achieving a mean score of 91.79 compared to 83.62 for the control group. Independent-samples t-tests further confirmed these differences. Taken together, the significant Group × Time interaction indicates that changes in overall attitudes toward science across the three measurement occasions differed between the experimental and comparison groups.

5. Discussion

5.1. Development of a Stakeholder-Informed STEM Instructional Environment

The first phase identified stakeholder-reported needs that could be addressed through a combination of engineering-design activities and Minecraft representations. In the second phase, students in the EDP-MGBL group reported more favourable science attitudes over the study period than the comparison group. Because baseline disengagement was neither an inclusion criterion nor an assessed construct, these results cannot be interpreted as evidence of re-engaging a previously disengaged subgroup.

The framework incorporated inclusion-oriented design choices, including multiple ways to represent energy concepts, scaffolded task difficulty, collaborative work, and opportunities to propose and revise designs. These choices are compatible with an equity-oriented vision of STEM education (Margot & Kettler, 2019; National Academies of Sciences, Engineering, and Medicine, 2025). Nevertheless, inclusive intent is not the same as demonstrated equity: the study did not directly measure access, belonging, differential participation, or subgroup outcomes.

A useful contribution is therefore the explicit mapping from stakeholder perspectives to design requirements, rather than a claim that specific barriers were removed or learners’ engagement restored. Studies using systematic classroom observation and measures of equitable participation are needed to establish whether this instructional design achieves its intended inclusivity goals.

5.2. Interpreting the Observed Changes in Students’ Attitudes Toward Science

The significant Group × Time interaction indicates that the EDP-MGBL and comparison groups differed in their pattern of change on the overall ATSS. The intervention combined engineering design, Minecraft activities, scaffolding, collaboration, and authentic energy tasks; consequently, the study cannot determine the individual contribution of any component. This limitation is particularly important given the methodological heterogeneity and risk-of-bias concerns in the Minecraft learning literature (Slattery et al., 2025). The results should be interpreted as preliminary, context-specific attitudinal evidence rather than component-level causal proof.

Phase One themes nevertheless supported a set of plausible design propositions. For example, mission-based tasks were intended to create enjoyable learning experiences; this proposition is consistent with literature on the potential of structured Minecraft-based activities (Di Leo & Traetta, 2025; Slattery et al., 2025). However, no mediation analysis, process evaluation, or post-intervention interviews were undertaken to determine whether enjoyment explains the observed overall attitudinal differences.

Likewise, staged design tasks with opportunities for feedback and revision were intended to support perceived competence, while authentic renewable-energy challenges were intended to make science meaningful in everyday contexts. Engineering-based learning can offer such design opportunities, but the present data do not test how these features relate to individual students’ self-efficacy or value ratings (Guzey et al., 2016; Kelley & Knowles, 2016).

Student-led decision-making and team-based construction were designed to offer opportunities for agency and shared reasoning. These characteristics can be justified pedagogically within integrated STEM education (English, 2016; Honey et al., 2014), but the study did not independently quantify student choice, the quality of interaction, or their mediating effects on science attitudes.

Minecraft representations of Redstone circuits and energy systems were intended to make invisible or abstract processes more tractable. Recent reviews identify such uses as potentially valuable while cautioning against attributing benefits to the platform alone (Di Leo & Traetta, 2025; Slattery et al., 2025). Whether improved visualisation explains science-attitude change remains an empirical question for future research.

5.3. Implications for Inclusive STEM Education

The principal practical implication concerns the design process. Teachers developing energy-related STEM instruction may use stakeholder-reported needs to select meaningful challenges, scaffold the EDP cycle, provide visual representations, and create structured opportunities for discussion and decision-making. Inclusion is an intended characteristic of this instructional design, not an outcome established here. The findings do not demonstrate improved equitable participation, accessibility, or belonging, and they should not be generalised to learners classified as disengaged. Further evaluation should combine attitudinal measures with observation, accessibility review, subgroup analysis, and learner feedback (National Academies of Sciences, Engineering, and Medicine, 2025).

5.4. Limitations and Future Directions

Several limitations should guide interpretation. First, the ATSS measured attitudes toward science, not engagement, re-engagement, knowledge achievement, observed participation, or inclusion. Neither baseline disengagement nor changes in accessibility, belonging, or participation equity were evaluated. Second, two intact classes from different schools were compared without student-level random assignment. Potential baseline, teacher, and school differences cannot be ruled out; attitudinal associations should not be presented as definitive causal effects (Capili & Anastasi, 2024; Shadish et al., 2002).

Third, the intervention combined engineering design, Minecraft Education, collaboration, scaffolding, authentic tasks, and learner decision-making. Its individual components and six proposed explanatory functions were not isolated or tested through mediation, process evaluation, or post-intervention qualitative analysis. Phase One findings informed design rather than independently verifying the processes behind Phase Two outcomes. Future dismantling, factorial, and process-focused studies could investigate these propositions.

Fourth, self-report may be sensitive to novelty, expectancy, or social desirability. Implementation fidelity, prior Minecraft experience, and teacher and school differences warrant systematic measurement. Fifth, the evaluation included 68 Year 5 students in one geographical and cultural setting; results may not generalise to other populations. The available reporting of quotation provenance, detailed delivery protocols, and full questionnaire wording is limited; stronger documentation is needed for replication and appraisal.

The delayed assessment examined whether the observed attitude pattern persisted within the reported follow-up period; it cannot establish permanent change. Future studies should add longer follow-ups, verified intervention materials, classroom observation, post-intervention interviews, conceptual-learning measures, and equity-sensitive indicators to investigate sustainability, classroom processes, and inclusive outcomes.

6. Conclusions

This study developed and initially evaluated a stakeholder-informed EDP-MGBL framework for primary science education. Phase One elicited six instructional needs concerning enjoyment, confidence, relevance, agency, collaboration, and visualisation; these informed corresponding design features in an engineering-design and Minecraft-based intervention. Inclusive learning opportunities were a design intention rather than a directly measured outcome.

In Phase Two, the EDP-MGBL group showed more favourable changes in self-reported overall attitudes toward science across three assessments than the comparison group. A delayed assessment showed that the attitudinal difference remained observable within the study follow-up. Given the intact-class quasi-experimental design, results provide preliminary evidence of an association rather than definitive causal attribution to Minecraft or any specific component.

The six proposed explanatory pathways remain hypotheses derived from stakeholder needs and instructional theory; this study did not directly test their causal operation. Subsequent studies should investigate these design propositions using appropriately targeted measures, post-intervention process evidence, larger and more diverse samples, and stronger controlled designs, including indicators of participation and inclusion.