CORE: Cultivation of Collaboration Skills via Educational Robotics

Abstract

Collaboration skills are an important component of 21st century skills and a critical skill for citizens of the future. In this work, we propose collaboration-oriented robotics education (CORE), a methodology aimed at fostering the development of collaboration skills in primary school students aged 11–12 via an adjusted approach to the teaching of educational robotics. In order to assess the existence and level of collaboration skills in a student, a suitable tool is also proposed. Using a collaboration-oriented performance evaluation test (COPE) for both a pre- and post-intervention measurement and applying both the conventional and CORE approaches to teaching educational robotics to 32 students, split into control and intervention groups, we demonstrate the effectiveness of the proposed approach. Specifically, the experimental implementation shows that CORE statistically significantly increases the performance of the experimental group compared to the conventional way of teaching educational robotics. These results, in addition to validating CORE itself, demonstrate that the conventional approach to STEAM (Science, Technology, Engineering, Arts, Mathematics) education is not necessarily already optimized, thus facilitating an overall re-evaluation of the field.

1. Introduction

One of the educational challenges of the 21st century is the development of collaboration skills [1]. While often regarded as an abstract and vague concept [2], the absence of collaboration skills can significantly impact both personal and professional spheres [3]. As global challenges grow in complexity, the ability to coordinate effectively, participate constructively in collective decisions and solve problems collaboratively becomes increasingly essential [4]. Given the interdisciplinary nature and significance across personal, academic, and professional domains [5], the early and targeted cultivation of collaboration skills is both a necessity and a strategic priority in modern education [6].

Educational robotics has an increasing presence in the educational ecosystem [7]. The flexible and tangible nature of ER (educational robotics) lays the foundation for supporting STEM (Science, Technology, Engineering, Mathematics) subjects at all educational levels, as well as the development of skills such as critical thinking, problem solving, communication, and collaboration [8]. However, the relevant literature reveals that the majority of studies do not occur in typical school settings but in after-school activities, summer camps, or competitions [9,10]. In parallel, it is strongly believed that high instruction during each educational process paired with corrective feedback is the most efficient method to approach ER [9,11]. Furthermore, collaboration is not approached by most studies as the main outcome of an intervention but as a byproduct [12,13,14,15].

The main purpose of this research was to create a methodology that is directly oriented to the cultivation of collaboration skills. The CORE methodology (collaboration-oriented robotics education) offers a structured student-centered approach to robotics education. It emphasizes group work and a project-based approach, while minimizing rigid traditional, instructional methods. Additionally, it attempts to interconnect the different types of knowledge within ER to different collaboration skill outcomes. This specific element was highlighted within the extensive work of Jung and Won [16] as a critical factor in designing dedicated educational methodologies that shift the focus from technology to rigorous pedagogical gains for students [17].

The basic idea of CORE is to develop a methodology that aligns with typical school settings for elementary students in Greece aged 11–12, fostering collaboration skills that can be applied across various domains, regardless of group composition. Additionally, CORE promotes equity and inclusivity, ensuring that all students, regardless of socioeconomic background, beliefs, or technical aptitude, have the opportunity to cultivate these essential skills.

The CORE methodology’s novel approach focuses directly on the development of collaboration, without treating it as incidental or a byproduct through intentionally supporting the symmetrical construction of knowledge among students. Unlike conventional ER approaches, which primarily focus on technical proficiency, CORE is designed to systematically integrate collaboration skills into formal classroom settings. Additionally, this work introduces the COPE test (collaboration-oriented performance evaluation) as an innovative assessment tool that objectively measures collaboration skills before and after the intervention. By demonstrating that the CORE methodology significantly enhances students’ collaboration skills’ performance, this research challenges the assumption within ER that instruction-based, traditional approaches are already sufficient for skill development. While not rejecting the value of these approaches, our findings highlight the need for pedagogical frameworks that explicitly and systematically cultivate collaboration as a core learning outcome.

The second section presents the relevant literature on collaboration skills, first in the broader context. It then examines the use of ER as a tool in educational robotics and, finally, focuses on its application in primary education. The third section briefly analyzes the measurement tool of the COPE test [18] and its validation process. The fourth section outlines the analysis of the CORE methodology, together with a short discussion on why every step has been designed in such a manner. The fifth section presents the experimental procedures and the results, followed by a discussion.

2. Relevant Literature

As previously mentioned, despite the growing body of literature on soft skills, few studies focus directly on the development of collaboration skills as a primary outcome. Graesser et al. [19] emphasized this gap, stating “…we are at ground zero in terms of pedagogical approaches to improving Collaborative Problem Solving skills”.

In the broader literature, Pazos et al. [20] explored the development of collaboration skills in a virtual team-based intervention for engineering students. Sturner et al. [21] examined the impact of a math–biology program on students’ collaboration skills. Murwaningsih and Siswono [22] investigated collaborative problem solving in mathematics, while Rodriguez-Salvador and Castillo-Valdez [23] introduced the “Colabora” model to enhance collaborative competencies in graduate students.

Meanwhile, ER has emerged as a very promising tool for supporting various STEM education fields. Although, research on ER is still in its early stages [16] and almost all research interventions have been conducted outside traditional classrooms [24], the field is gaining research attention. For example, Ferrarelli and Iocchi [25] found that computer programming in robotics improved high school students’ understanding of physics principles. Similarly, Jaipal-Jamani and Angeli [26] demonstrated that ER fosters STEM learning. Several studies in secondary education also highlight ER’s role in developing computational thinking [27,28,29,30].

Research in the context of primary education remains limited in scope, often characterized by relatively small sample sizes [31]. The existing literature predominantly focuses on the role of educational robotics (ER) in facilitating STEM learning [32,33,34] and programming skills’ acquisition [35,36,37], mirroring similar implementations within secondary education contexts. A substantial body of research has also concentrated on the development of computational thinking through ER [38,39,40,41,42].

In addition to cognitive outcomes, a number of investigations within the learning sciences have explored ER’s impact on motivational constructs [43]. For example, Liu et al. [44] reported increased student attentiveness when learning was mediated through gameplay with ER. Similarly, Safrudin et al. [45] highlighted that heightened engagement facilitated the development of abstraction skills, while Socratous and Ioannou [46] demonstrated ER’s efficacy in fostering metacognitive abilities. Di Lieto et al. [47] documented the use of ER to cultivate higher-order reasoning skills, and Araujo et al. [48] observed notable improvements in both academic performance and working memory among participants.

ER’s pedagogical flexibility has also been investigated in the context of self-regulated learning [49,50], as well as its capacity to enhance learner engagement via edutainment frameworks [51], particularly in relation to age-related variability in student responses to robotic agents. More recent studies have expanded this inquiry to encompass ER’s role in promoting creativity and self-confidence [52], problem-solving capabilities [53,54], critical thinking [55], technological proficiency [56], and the development of professional interactive skills [57]. Additionally, ER has been linked to the enhancement of collaborative competencies [58].

Emerging research also points to ER’s applicability in supporting the development of social skills within special education settings [59,60,61]. Specifically, assistive robotics have been widely utilized to facilitate social interaction and agency among learners with special educational needs. Di Battista et al. [62] found that ER enabled more personalized learning experiences, simultaneously reducing instructional burden on educators.

Despite promising findings, significant gaps persist in the literature on collaboration skills’ development. Research in primary and secondary education is limited, likely due to the conceptual complexity of collaboration and the lack of clear theoretical definitions and authentic performance assessments. Transferability remains underexplored, as most studies assess collaboration within specific subjects without examining its cross-disciplinary applicability. Equitable participation is another gap, with few strategies to ensure balanced engagement among diverse students. Many ER interventions occur outside formal education, limiting insights into their classroom effectiveness. Additionally, studies often compare robotics-based learning with traditional teaching rather than control groups using the same ER tool. Assessment methods rely heavily on self-reported data instead of performance-based evaluations, which is particularly problematic for younger students. Addressing these gaps is essential for developing evidence-based interventions that promote equitable, transferable, and classroom-integrated collaboration skills. As Kuriazopoulos et al. [63] assert, “…ER is one of the most important technologies, having multiple learning advantages for students of primary education”.

3. Assessing Collaboration Skills—COPE Test

Evans [5] identifies seven major frameworks (ATC21S, Cambridge Assessment, Pearson P21, Essential Skills and Dispositions, PISA, P21 EdLeader21, and ETS) that analyze collaboration skills to support educators in both teaching and assessment. Across these frameworks, four main categories of collaboration skills are consistently repeated [5]: planning and group decision making, communicating about thinking with the group, contributing resources and ideas, monitoring, reflecting and adapting individual and group processes. The most widely used frameworks [64] are the ATC21S framework [65] and the PISA framework [66]. Hesse et al. [67] defines collaboration skills as a combination of participation, perspective taking, and social regulation skills. The COPE (collaboration-oriented performance evaluation) test (Figure 1) views collaboration skills from the lenses of ATC21S while adapting the individual’s scoring to the view of Child and Shaw [68] in order to measure collaboration as an individual’s process rather than an individual’s outcome.

The COPE test [18] is highly adaptable to Greek typical primary school education in terms of time settings, high usability, cost effective nature, and flexible application. More specifically, the procedure was simple in application, comprising two puzzle tests taken by individual students and students in groups. Initially, as presented in Figure 1, the students were asked to individually assemble a puzzle of 120 pieces that was age-appropriate (years 6+) within a 45 min period. Then, photos were taken in order to count the number of pieces that were assembled within the given time (45 min). From this individual process, we derived two numbers. The first number was the percentage of pieces assembled and the rate (pieces/minute) at which the students assembled the puzzle.

When the students finished their individual puzzle assemblies, they were randomly selected to form groups of four of mixed genders. Then, they were asked to assemble a different puzzle with the same difficulty (year +6, 120 pieces, 45 min). The group puzzle boxes contained four plastic bags that had pre-divided and coded puzzle pieces in four different colors (red, blue, yellow, and green) underneath. The group puzzle activity adapted collaboration theory guidelines in order to protect the activity from a student overtaking the process while assessing the differences in performance in a tangible manner.

More specifically, symmetry of action, symmetry of goals, engendering negotiation, sufficiently complex, interdependence, and ill-structure guidelines were adapted to produce the ground rules for the group activity. Consequently, the students were asked to assemble pieces only on the designated area on the table, not touch or grab peer-assigned pieces, and had 30 pieces that could not be assembled solely by them. With these ground rules, it was possible to measure five factors that contributed to the individual’s performance in terms of collaboration skills on a 1 to 10 scale (

Table 1. COPE test’s numerical scales in relation to collaboration skills.

	Point 1		Point 2		Point 3		Point 4		Point 5
0–10%	1	+41% Decline	1	0–10%	1	0–10%	1	0–0.5 pieces/min	1
11–20%	2	31–40% Decline	2	11–20%	2	11–20%	2	0.51–1 pieces/min	2
21–30%	3	21–30% Decline	3	21–30%	3	21–30%	3	1–1.5 pieces/min	3
31–40%	4	11–20% Decline	4	31–40%	4	31–40%	4	1.5–2 pieces/min	4
41–50%	5	1–10% Decline	5	41–50%	5	41–50%	5	2–2.5 pieces/min	5
51–60%	6	0–10% Improvement	6	51–60%	6	51–60%	6	2.5–3 pieces/min	6
61–70%	7	11–20% Improvement	7	61–70%	7	61–70%	7	3–3.5 pieces/min	7
71–80%	8	21–30% Improvement	8	71–80%	8	71–80%	8	3.5–4 pieces/min	8
81–90%	9	31–40% Improvement	9	81–90%	9	81–90%	9	4–4.5 pieces/min	9
91–100%	10	41%+ Improvement	10	91–100%	10	91–100%	10	4.5+ pieces/min	10

). More specifically Figure 2 shows the grade blend of COPE test.

The COPE test (collaboration-oriented performance evaluation) utilizes the complexity and tangibility of puzzles in both an individual and group task. The assessment focuses on delivering an instrument that minimizes the presence of other types of skills or knowledge. For example, it does not require students to type on a keyboard or use foundational knowledge from taught subjects. This is deliberate, in order to eliminate firstly the connection between typing fluency and collaboration and, secondly, subject knowledge hierarchies, so it is less likely for high-ability students to force their will onto their peers. Furthermore, its stealth [69] and playful nature assesses students without the usual discomfort feelings, promoting fair opportunities to all students.

The assessment’s purpose was to derive a combination of scoring within five different scales of equal weight. Three scales were calculated from the individual’s contribution, convergence with peers, and group’s completion percentage. The remaining two scales were calculated by comparing the student’s individual performance to the individual’s collaborative performance within the group.

More specifically, P1 was the mark assigned to a student based on the group’s performance. P2 was the mark assigned for the increased or decreased number of puzzles assembled within the group and was derived by subtracting the percentage of individual puzzle pieces within the group from the percentage of individual puzzle pieces assembled alone. P3 was the mark assigned for the percentage of individual puzzle pieces assembled within the group. P4 was the mark assigned for individual convergence with peers in the group, and P5 was the mark assigned for the individual’s performance efficiency in the group minus the individual completion rate. By adding all the individual marks which were awarded to the student, we derived the total mark that reflected the sum of all the utterances and individual contributions directly or indirectly to collaboration skills performance.

The assessment was tested and retested in 12 public schools in Greece and 148 students aged 11–12 participated in the validation process. The whole experimental process lasted for about 7 months until all schools finished the test and retest phase. The students were randomly selected during the test phase in groups of four and carried out the experimental procedure. After approximately 3 weeks (for each school), the students carried out the experiments again, but the group compositions were different.

To evaluate the stability of the individual performance of collaboration skills, a test–retest validation procedure was conducted. The findings from this large scale experiment indicated that the COPE test yielded positive results. Statistical analysis using Spearman correlation demonstrated a strong positive correlation (p = 0.623) between individual student scores during the test and retest phases. Notably, this result was independent of any correlation between individual performance and the group’s average performance. Furthermore, no significant differences were observed based on gender or school population. Additionally, no correlation was detected between an individual student’s performance and the performance of the remaining three group members. These finding support the validity of the selected proposed assessment tool.

4. The CORE Methodology

The following provides a detailed, step-by-step description of the methodological approach implemented in this research. Given the complex nature of collaboration and the challenge of developing collaboration skills, the methodology is presented in a sequential format and represented in a flowchart as a process, as in Figure 3. Figure 3 illustrates the phases of the educational intervention along with the teacher’s instruction for each phase. The methodology encompasses the following components:

• Re-acculturation;
• Co-creation of project with ill-defined attributes and open-ended properties;
• High synchronicity—instruction;
• Common understanding—encouragement;
• Booklet—equal opportunity for new “foundational knowledge”;
• Retraction of prior knowledge (scratch programming).

The decision points (see Figure 3) represent moments when students must determine how to proceed within the groups. For instance, upon completing the first stage of constructing the artifact, it is crucial for all group members to engage in a dialog to ensure shared understanding of the primary construction elements. The teacher plays a key role in facilitating and encouraging these discussions, making sure all students are heard.

But, before reaching the second decision point, it is important to explain the two initial steps. The first step aims to initiate a dialog with the students about their past group work experiences, gathering both positive and negative feedback to identify key concerns such as exclusion, egoism, loafing, and fear of failure. These issues were addressed in this study through collaboratively established ground rules designed to promote respectful, inclusive, and effective communication. Students were encouraged to contribute, agree or disagree, and even express themselves non-verbally—acknowledging that not all students communicate the same way.

The second step focuses on co-creating a shared project idea that all students could agree upon. The main point behind the project should be to balance solvability with complexity in order to stimulate planning and decision making. The project has to be ill-structured and open-ended to foster creativity and reasoning, while being understood, enabling, in this manner, intragroup negotiation. For example, a “vehicle” could be interpreted differently by each group, encouraging exploration and diverse outcomes.

The third step of simultaneous actions starts immediately after the first hour of free exploration of the robotics kit. It comes exactly at the point when the students start creating the first constructions. The reason behind this is to foster collaboration skills on two levels. First, it promotes equal knowledge construction within a domain where there is no prior knowledge or experience for the students. This notion supports symmetrical participation and reduces the risk of disengagement. Secondly, it enhances communication, as physical interaction over shared resources (bricks, motors, bands, sensors, etc.) often evolves into verbal exchanges and discussions. Even if spontaneous division of labor may occur, it can be productive if tasks remain interconnected. The teacher must carefully implement this instruction, emphasizing its purpose to maintain an inclusive leaning ecosystem.

The next step is conceptually prepared by the previous one, but it pushes forward the students to make conscious explanations of the elements that are being constructed to each member of the group. In this manner, it is less likely for students to move to the more demanding domains without having shared with each other, while the foundation of deeper social regulation skills such as negotiation and conflict resolution starts to emerge through constantly encouraging group members to maintain common understanding and consent before having to share the more demanding knowledge of sensors, hub, Bluetooth connections, etc.

After all students within the group consent to moving forward to connect the first complete formation of their artifact, they are given a booklet with guidelines to connect their artifact to the computer to start programming it. Within the booklet there is a deliberate notion to portray the use of motors and sensors through the s2bot application. In this way, the students are not spoon-fed all the details but still have to investigate through practice their practical function. While the investigation is ongoing, they have equal opportunity to learn new foundational knowledge through depending solely to each other to move forward their robotic construction.

While Lego WeDo 2 comes with its own software package, it was decided not to use this for two main reasons. Firstly, it was important to connect the methodology with an element of prior knowledge. This was not meant to be introduced at the beginning of the project but at the end. The reason was that many students do not grasp programming sufficiently enough during the previous school years, and this would create the risk of knowledge hierarchies that could evolve into authoritative behaviors within the groups when some students assume knowledge and force their will onto the rest. The second reason was to introduce elements that expose the robotic properties in greater depth. On the one hand, it would make the activity more difficult but manageable, and, on the other hand, it would give more space during the final stages for many deeper conversations and opportunities for negotiation skills to develop while sharing ideas for improvisations and alterations in order to achieve robotic functionality.

5. Methods and Procedures

5.1. Context and Participants

The experiments aimed to identify if there was a difference in the students’ collaboration skills’ performance as individuals after the proposed intervention using ER. In order to understand whether the proposed intervention produced any significant results, an experimental design with pre- and post-tests was implemented with an experimental and control group. The following section portrays all the typical and regulatory settings of the experiments paired by data analysis of the results produced.

5.1.1. Ethical Approval

In order for the experiments to take place, it was necessary to obtain permission from the ethical committee of the University of Peloponnese. Additionally, the Administration of Primary School Education of Argolis reviewed the experimental procedures and granted permission for their implementation. Written parental consent was collected from all participants. Ethical guidelines and standards were followed throughout the study to ensure the well-being and rights of the participants.

5.1.2. Study Location and Duration

The research was conducted in 2 public primary schools in the Argolis Prefecture, Eastern Peloponnese, from November 2024 to January 2025.

5.1.3. Participant Selection

Students were selected using convenience sampling. Two whole classes were chosen to ensure typical representation of the student population. Although CORE is fundamentally designed to support inclusive classroom practices, a methodological precaution was established during the design phase of this study. Specifically, we aimed to exclude students with formally identified Special Educational Needs who required the constant presence of a dedicated special education professional during the experiments. This was carried out to maintain consistent group conditions during the COPE test, which relies on natural student collaboration without adult intervention. However, in practice, neither of the two classes participating across the two schools had such documented support requirements. As a result, the intended exclusion was not applied. We include this note here for the sake of methodological transparency.

5.2. Procedures

5.2.1. Pre-Test

Both experimental and control group were divided randomly in 4 groups of 4. The COPE test was implemented in the control group on 5 November 2024 and in the experimental group on 7 November 2024.

5.2.2. Main Study

In order to keep the same conditions for both groups, we decided to keep the randomly selected groups of the pre-tests unchanged for both the experimental and control group. Therefore, the pre-test randomly selected groups of 4 in the control group were of the same composition for all the educational robotics sessions. Similarly, the pre-test randomly selected groups of 4 in the experimental group were of the same composition for all the educational robotics sessions.

Table 2. Experimental setup.

Settings/Instruments	Control	Experimental
Number of Participants	16	16
Total Number of Groups	4	4
Age	11–12	11–12
Groupings	Groups of 4	Group of 4
Educational Level	6th Grade Elementary	6th Grade Elementary
Session Duration	90 min (2 educational hours)	90 min (2 educational hours)
Number of Sessions	7	7
Prior Knowledge	Scratch Programming	Scratch Programming
Educational Robotics Experience	No Experience	No Experience
Educational Robotics Kit	Lego WeDo 2	Lego WeDo 2
Programming	Scratch	Scratch
Computer Science during 5th Grade	Yes	Yes
Connecting Software App.	S2Bot Native	S2Bot Native

presents the experimental setup in detail. The control group followed a traditional instruction-based teaching, while working in groups of four. The control group conducted group projects with similar attributes to the experimental one.

5.2.3. Experimental Conditions and Preparations

The COPE test was implemented (pre and post) in the classrooms, and the educational robotics sessions were implemented in the computer science laboratories of each school. The teacher was the researcher during all the experimental sessions. All experiments were conducted during the two first hours of the school program, in order to minimize student fatigue. The same laptops were used for both the experimental and control groups.

5.2.4. Post-Test

Both the experimental and control groups were divided randomly into different compositions. None of the students within control and experimental groups worked together during the pre-tests. The COPE test was implemented on the 14th of January for the control group and on the 16th of January for the experimental group.

5.3. Results

Descriptive Statistics

Based on the descriptive statistics and frequency distributions (see Table 2 and Figure 4), distinct patterns emerged in the comparative analysis of the control and experimental groups’ performance. The control group maintained relatively stable metrics, with small variation between the pre-test (M = 85.88, SD = 6.54, Mdn = 83.00) and post-test values (M = 85.13, SD = 7.45, Mdn = 83.00). The frequency distribution for this group remained unchanged, with a slight increase in score variation as evidenced by the marginally higher standard deviation and the expanded range in the post-test scores (range = 22.00).

In contrast, the experimental group demonstrated substantial improvements, characterized by a marked increase in mean performance of 9.21% (pre-test: M = 88.25, SD = 6.32, and Mdn = 88.00; post-test: M = 96.38, SD = 4.31, and Mdn = 97.00) (

Table 3. Descriptive statistics.

	Control Group		Experimental Group
	Before	After	Before	After
N	16	16	16	16
Mean	85.88	85.13	88.25	96.38
St. Dev.	6.54	7.45	6.32	4.31
Standard Error	1.63	1.86	1.58	1.08
Median	83.00	82.00	88.00	97.00
Minimum	78.00	74.00	78.00	84.00
Maximum	96.00	96.00	98.00	100.00
Range	18.00	22.00	20.00	16.00

, Figure 4) and a notably narrower distribution of scores. This positive shift was particularly evident in the histogram, where eight students (50%) of the experimental group achieved scores within the highest interval post intervention, compared to the initial relatively uniform distribution (Figure 5). The reduced standard deviation in the experimental group’s post-test scores (SD = 4.31) further indicated a more stable performance level, suggesting that the intervention not only improved overall achievement but also enhanced consistency across participants.

5.4. Non-Parametric Tests

Moving on to non-parametric analysis, it should be noted that the choice to apply non-parametric statistical tests in this study was methodologically required due to the violation of normality assumptions, while maintaining analytical rigor. Non-parametric techniques, such as the Wilcoxon and Mann–Whitney U test, provide a powerful alternative to traditional parametric tests when the underlying assumptions of normality and homogeneity of variance are not met [70].

The Wilcoxon signed-rank test results (see

Table 4. Wilcoxon test.

Before Intervention	After Intervention	W	z	p-Value
Control	Control	26.000	1.1202	0.8871
Experimental	Experimental	0.000	−3.5162	<0.001

) revealed distinctly different patterns for the control and experimental groups. The control group showed no statistically significant difference between the pre- and post-test scores (W = 26.000, p = 0.8871), indicating that performance remained stable. In contrast, the experimental group demonstrated a highly significant improvement (W = 0.000, p < 0.001). This pattern strongly suggests that the intervention had a substantial positive effect on students’ performance in the experimental group, while the control group’s performance remained relatively unchanged.

The Mann–Whitney U test results (see

Table 5. Mann–Whitney test.

Study Phase	Mann–Whitney Test Statistic	p-Value
Before intervention	102.000	0.1713
After intervention	30.000	<0.001

) also indicate a significant change in the between-group differences across the intervention period. Prior to the intervention, no significant difference was found between the control and experimental groups (U = 102, p = 0.1713), suggesting comparable initial performance levels. However, the post-intervention analysis revealed a highly significant difference between the groups (U = 30, p < 0.001), with the experimental group demonstrating a substantial shift in the relative performance levels, providing strong evidence for the differential impact of the intervention between groups.

The findings of this study demonstrate the effectiveness of the educational intervention, as documented through a series of non-parametric statistical analyses. Specifically, the experimental group showed a statistically significant improvement in performance (p < 0.001), with a notable increase in the mean and median scores (from M = 88.25 and Mdn = 88.00 to M = 96.38 and Mdn = 97.00), while simultaneously exhibiting a decrease in score variance, indicating increased homogeneity in students’ performance. The validity of these results was further strengthened by the stability observed in the control group, which maintained similar performance levels throughout the study (M = 85.88 and Mdn = 83.00 to M = 85.13 and Mdn = 82.00, p = 0.887). The methodological triangulation through the Wilcoxon and Mann–Whitney tests confirmed both the statistical and practical significance of the intervention.

6. Discussion

The main purpose of this research was to develop the CORE methodology, which focuses directly on the cultivation of collaboration skills in primary school students through educational robotics.

The findings provide strong evidence that the CORE methodology effectively fosters collaboration skills. The experimental group exhibited a 9.21% increase in individual collaboration performance, while the control group showed no significant change. This suggests that CORE provides a structured pedagogical framework that enables students to develop collaboration skills systematically rather than relying on incidental teamwork experiences.

Our findings align with previous research emphasizing the role of educational robotics in skill development [17,30,63]. However, unlike prior studies which primarily focused on STEM learning outcomes, CORE explicitly targets collaboration skills. Given the conceptual complexity of collaboration, it should be treated with dedicated methodologies rather than as a byproduct or secondary outcome. The introduction of a student-centered design, paired with symmetrical and gradual knowledge construction, provided the necessary foundation for equitable participation, which has been identified as a major gap in the existing literature.

Another critical issue in the literature is the transferability of skills across different domains. This was a key consideration in the CORE methodology, requiring validation in a different content domain. The literature, particularly in primary education, has often relied on self-reported data or peer evaluations, which are not always rigorous. To address this, the COPE test was designed and validated through a series of experiments, offering a twofold contribution: first, by assessing collaboration skills in a different content domain, and second, by providing a validated tool that measured the methodological impact of CORE on students’ collaborative performance.

The added value of the proposed CORE methodology has the potential to migrate to different content domains and within the educational context. The fact that the assessment tool (COPE test) assesses collaboration skills’ performance through the use of puzzles supports the notion that CORE has the ability for the developed skills to be transferred to a different domain. Moreover, the CORE methodology has the potential, due to its flexible and scalable nature, to be transferred and used in totally different knowledge domains.

The three main cycles of the methodology can be transferred to different contexts for the same age range (11–12). To be more specific, as depicted in Figure 3, the main goal of the first grand cycle is for students to learn to share physical objects (bricks, bands, motors, hub, sensors) in a fair manner that is both inclusive and collaborative (not dividing the labor). This is not a given and does not have to be taken for granted. Then, moving slowly to the next grand cycle and having a common constructed ground of “new knowledge”, what will be the meaning behind argumentation that will eventually lead to higher social regulation skills? There has to be an introduction of “foundational knowledge” that will cover all the essential functionalities but with an investigative tone. How this knowledge will be disseminated equally is a challenge. The booklet presents a great opportunity for basic authentic collaborative learning processes. The third grand cycle represents “prior knowledge”, specifically whether it is a common state for all members or not. Due to the interdependence of the members and their collaborative effort up until this stage, it is highly likely that they will engage in deeper conversations while knowledge gaps and new ideas are brought to the surface and start to be shared. For example, the same collaborative cycles could be applied to domains such as environmental science projects, language art group tasks, or interdisciplinary STEM challenges.

Additionally, most research involving educational robotics has been conducted outside formal classroom settings, making it difficult to determine its feasibility within structured curricula. CORE was designed to be both structurally and conceptually flexible, ensuring its adaptability within formal education. One major challenge in integrating robotics into classrooms is the equitable inclusion of students with varying technical aptitudes. CORE addresses this by ensuring equal opportunities for all students to engage with robotics, empowering them to become co-creators within the learning environment.

The experimental setup further reinforced that, while the flexible nature of robotics is promising, it is insufficient to foster collaboration skills without appropriate pedagogical frameworks. This is a critical finding, as it shifts educational robotics research toward evidence-based practices that emphasize structured pedagogical methodologies over unstructured, technology-driven interventions.

While this study focused on quantitatively evaluating the effectiveness of the CORE methodology using a performance-based assessment tool, we fully acknowledge the reviewer’s valid concern regarding group dynamics and the risk of dominant behaviors within group settings. The current study did not aim to qualitatively analyze intragroup roles, as this was beyond its scope. However, this exact dimension—including the emergence of more dominant and passive students—have been thoroughly explored in two prior in-depth longer-term case studies conducted by the authors [18]. These earlier works adopted a qualitative research design to investigate student interactions, power dynamics, and participation symmetry within the same methodological framework. Therefore, this publication builds upon and compliments that earlier work by offering quantitative validation using a different research method.

As with all research, limitations are also present within this research. The relative high cost of ER kits limited the implementation of the CORE methodology to only two schools and classes. It would be even more interesting for research purposes to view the differences within an experimental design over a grander sample size. This issue is particularly important in many research attempts, and it highlights a major obstacle in delivering ER’s promising learning experiences within formal education.

While the findings of this research demonstrate statistically significant improvements in collaboration skills due to the CORE methodology, a limitation must be acknowledged regarding the statistical assumption. The use of non-parametric tests was necessitated by violations of normality in the data distribution. While appropriate for this context, these tests are less sensitive to effect size, which limits the granularity with which we can interpret the magnitude of impact. Additionally, as previously mentioned, the relatively small sample size (n = 32) may limit the generalizability of the findings. However, it was considered a positive indication for the development of collaboration skills to obtain these results even with a non-parametric test.

The age range of 11–12 is also considered a limitation. However, it was considered as an appropriate age for cultivating a complex construct such as collaboration. Additionally, there was a need to align computer science programming in Scratch as a prerequisite knowledge for all students participating in the experiments.

Another limitation of this work is the lack of gathering long-term data from the students participating within the experimental procedures. It would be beneficial if research would also focus on longer-term research in order to further understand the impact of ER on students’ learning trajectories and career decision paths.

For future research and within the domain of ER and collaboration skills, it is important to constructively explore the integration of emerging technologies such as artificial intelligence (A.I.). The CORE methodology is designed around the enhancement of natural interactions and supporting symmetrical participation and symmetrical knowledge creation. The element of symmetricity is supported explicitly in all the stages. Even though symmetry is crucial, it is highly difficult for an untrained teacher to keep a mental record in real time of the amount of actions and spoken meanings each student brings to the collaborative effort. It is therefore important for future research to utilize motion sensors and speech recognition through A.I. in order for the teacher (before the next session) to be prepared to support and encourage more or less participation where and if needed. Having the whole picture of every student’s contributions (physical, verbal, and non-verbal) will aid their development of collaboration skills through a more tailored, assistive input from the educator and less of a “one size fits all” approach.

7. Conclusions

The CORE methodology leverages the diverse and intertwined nature of educational robotics to cultivate collaboration skills gradually among all students. By emphasizing symmetrical participation, equitable knowledge construction, and student empowerment, it provides a structured, student-centered approach. To measure its effectiveness, the COPE test was employed in different contexts and group compositions before and after the intervention. The results demonstrated a statistically significant improvement, with students in the experimental group increasing their collaboration skills performance by an average of 9.21%, whereas the control group showed no significant change. This finding underscores the impact of structured pedagogical strategies in fostering collaboration skills through educational robotics.

This study aligns with previous research that highlights the potential of educational robotics in formal school settings. However, it challenges the assumption that simply introducing ER technology is sufficient to drive skill development. The findings indicate that well-structured pedagogical approaches are necessary to optimize learning outcomes. The CORE methodology ensures that collaboration is not a byproduct of educational robotics but a targeted learning outcome. Without an intentional instructional design, the benefits of ER may not be fully realized, reaffirming the need for evidence-based strategies in implementing robotics in education.

By challenging the notion that collaboration skills develop incidentally, this study highlights the significance of structured pedagogy in ER education. The CORE methodology presents a scalable, inclusive, and effective framework that supports 21st century skill development in formal educational settings. As the demand for collaborative competencies continues to grow, structured methodologies like CORE can play a crucial role in equipping students with the skills needed for future academic and professional success.

Future research should focus on refining targeted interventions aimed at cultivating 21st century skills, ensuring that assessments are both rigorous and evidence-based. Additionally, integrating artificial intelligence into ER tools could enhance collaboration by enriching both human-to-human and human-to-agent interactions. This would better prepare students for future hybrid work environments, where teamwork across digital and physical platforms is increasingly vital. Furthermore, expanding ER to open, cost-effective, and scalable platforms will enable broader adoption across different educational levels, from kindergarten to university.