Multidisciplinary Education Pathways to Attract High School Students Toward Research and Science

Abstract

This study reports the design, implementation, and descriptive evaluation of “Codici del Futuro”, a STEM-oriented education pathway developed by the Italian National Research Council (CNR) to promote students’ interest in science and awareness of research-related careers and addressed to local high school students. The programme involved 167 high school students organised in 10 groups and combined an orientation session with hands-on workshops delivered in CNR research facilities (chemistry, biotechnology, artificial intelligence, eXdended Reality/Augmented Reality (XR/AR), and game design). The chemistry workshop will be described as a case study. The study addresses two research questions: (RQ1) What group-level outcomes (participation, engagement, interest, behaviour) are observed across the multidisciplinary pathway? (RQ2) What post-activity satisfaction and short-term knowledge outcomes are observed in the chemistry workshop as an embedded case study? Group-level outcomes were assessed through a facilitator-based evaluation grid, using four single-item indicators rated on a 10-point scale and including field notes. The chemistry case study included an anonymous post-activity questionnaire (satisfaction, prior experience, and an eight-item knowledge test). Results documented high levels of engagement, interest, and appropriate behaviour across groups, whereas participation showed greater variability. In the chemistry case study, students reported high satisfaction and moderate post-activity knowledge scores, with differences across knowledge domains. Overall, findings provide descriptive evidence on student responses within a research-centre-based, multi-workshop STEM pathway.

1. Introduction

Immersive and authentic learning environments that connect school curricula with real research settings have been increasingly used to support students’ science engagement and conceptual understanding (Caine et al., 2015; Heffner & Jain, 2010; Kim et al., 2019; Moreno et al., 2023; Ruggirello et al., 2012). Particularly in STEM education, traditional classroom settings often struggle to provide students with the kinds of experiences that foster deep conceptual understanding, sustained motivation, and a realistic perception of scientific work (Bevan et al., 2015; Honey et al., 2014). This is especially true for students with divergent and accommodator learning styles, who tend to learn more effectively through instructional approaches emphasizing teamwork, communication, and real-world problem solving (Wu & Wang, 2025). Given these challenges, multidisciplinary educational programmes that combine hands-on experimentation, exposure to advanced technologies, and interaction with researchers have increasingly been recognised as promising approaches to enrich student learning (Bourne et al., 2025; Chiu et al., 2025; Sun et al., 2025; Toma et al., 2024). Moreover, it has been demonstrated that an effective STEM educator needs to combine different aspects that include discipline-related fields as well as multiple knowledge domain (Gavrilas & Kotsis, 2025). In addition, the integration of knowledge, methods, and perspectives from different disciplines through a multidisciplinary approach allows a more complete and profound understanding of the studied phenomena: it encourages collaboration between researchers, enriches the scientific dialogue, facilitates innovation, promotes the creation of new ideas and solutions that might not emerge within a single field, and, finally, increases the impact of the research itself on society (Manno et al., 2025).

Informal and out-of-school programmes have also been recognised as potentially valuable complements to classroom learning, particularly when they enable active participation, collaboration, and inquiry-based activities in authentic contexts (Yao, 2025). Informal learning programmes have also been recognised as effective approaches for teaching STEM disciplines (Al-thani et al., 2025). Teacher support plays a key role in fostering students’ sense of autonomy, competence, and relatedness (Shao et al., 2025), suggesting that enthusiastic and passionate educators are more successful in stimulating students’ interest and curiosity.

The present study investigates the outcomes of a multidisciplinary educational initiative conducted within a national research institution, involving scientists with strong enthusiasm and commitment toward their research activities. The term “multidisciplinary” is referred to a pathway where multiple disciplines are addressed side-by-side across different workshops. The programme involved a sequence of workshops focused on chemistry, biotechnology, artificial intelligence, augmented and extended reality, and game design. Each activity was designed to immerse students in real scientific environments, allowing them to engage with both the cognitive and social dimensions of scientific inquiry. The various paths were presented both in classic contexts typical of scientific research, such as laboratories, and in informal and playful contexts, such as augmented reality environments or serious games. The term “informal” refers to the out-of-school setting (a research centre) rather than unstructured learning, as activities remained guided and aligned with educational goals. Similarly, playful activities are characterised by the presence of structured gamified elements (e.g., challenge-based tasks and game-oriented formats) embedded within inquiry activities, rather than entertainment-only experiences.

The aim of this educational study was conceived as a research-centre-based learning experience intended to document how student groups respond to multidisciplinary STEM activities in terms of observable participation and engagement patterns, and to explore short-term outcomes within a focused chemistry workshop case study. Accordingly, the present study addresses two research questions: (RQ1) What group-level outcomes (participation, engagement, interest, behaviour) are observed across the pathway? (RQ2) What post-activity satisfaction and short-term knowledge outcomes are observed in the chemistry workshop as an embedded case study? Given the absence of pre-test measures and a control group, the study does not aim to estimate causal effects, but rather to provide a structured, descriptive account of outcomes within the implementation context.

2. Methodology

2.1. Research Design

This study adopts a program-evaluation design with an embedded chemistry case study aimed at documenting a multidisciplinary STEM pathway implemented in an authentic research setting. The case study was conducted to explore student satisfaction and short-term knowledge outcomes using a structured questionnaire. Since the study was not planned as a formal mixed-methods design with predefined integration procedures, findings are presented as descriptive programme evaluation supported by qualitative facilitator field notes triangulation.

2.2. Participants

The program involved a total of 167 high school students from a vocational and technical institute in Italy, representing different four grade levels (

Table 1. Student distribution by grade level.

Grade	Number of Students
2	37
3	33
4	39
5	58
Total	167

), ranging from the second to the fifth year of upper secondary education.

The specialisation areas were Finance and Marketing Administration, Business Information Services, Graphic Design and Communication, and Agricultural, Agri-food, and Agri-industrial Studies. All selected students participated in the programme as part of their regular school programming, during the official school timetable at the Palermo research area of the National Research Council of Italy (CNR).

2.3. Programme Structure

The “Codici del Futuro” programme was designed to provide students with authentic and engaging multidisciplinary STEM experiences aimed at introducing emerging languages, technologies, and scientific innovations through a structured five-session pathway, with each session lasting two hours for a total of ten hours. Each class participated in two main phases of the programme: (1) an initial orientation session held at the school site, dedicated to presenting the CNR’s mission, attracting students to scientific research and social challenges, and illustrating the multidisciplinary nature of the programme; (2) hands-on workshops conducted within real research environments and guided by researchers and technical staff. Workshop sessions were designed to promote active learning, problem solving, inquiry-based learning, exposure to scientific tools and methods, and reflection on real-world applications of STEM. The design of the workshops aimed to support not only cognitive learning, but also the development of transversal competences such as collaboration, critical thinking, and digital awareness, aligning with the objectives of the National Recovery and Resilience Plan (PNRR) Mission 4 on education and new skills (School 4.0—Next generation classrooms—Reforms and investments, n.d.).

Researchers and tutors facilitated the sessions through interactive presentations, group-based tasks, use of digital and scientific tools, and moments of structured reflection, ensuring both technical support and pedagogical mediation between the school and research settings.

The activities were tailored to the school level and area of specialisation of each group. Specifically, at the beginning of the workshop, the researchers asked a series of questions to assess the students’ prior knowledge of the topics covered in the workshop. As a result, the activities were adapted to the students’ knowledge, simplifying both the language and content proportionally, without changing the structure and aims of the workshops.

Table 2. Overview of the multidisciplinary workshops offered within the “Codici del Futuro” programme.

Workshop Title	STEM Area(s)	Educational Objectives
Chemistry: From Fundamentals to Frontier Research	Chemistry, Nanotechnology	Understand water pollution and use of nanomaterials for purification
Augmented Reality for Designing Educational Treasure Hunts	AR Technology, Education	Use AR for interactive learning and gamified environmental challenges
Extended Reality Learning Experiences	AR Technology, Education	Design AR scenarios for prosocial behaviour using MirageXR 2.7
Designing Educational Games	Education, Game Design	Develop educational games through critical and collaborative design
Biotechnologies: From the Lab to Life Sciences	Biotechnology, Biology	Apply molecular biology techniques (protein/DNA extraction, electrophoresis)
Artificial Intelligence in Schools	Computer Science, Education	Explore educational applications and societal impact of AI tools
Data Literacy and Artificial Intelligence	Computer Science, Data science	Understand machine learning, data ethics, and generative AI applications

provides an overview of the seven workshops which were designed to reflect both the scientific domains of participating CNR institutes and the diversity of student interests across technological and scientific pathways. The table also summarizes the corresponding STEM subject areas and key educational objectives.

• Chemistry: From Fundamentals to Frontier Research (CHEM)

The lab focused on environmental applications of nanostructured porous materials. Students explored the use of silica-based adsorbents to remove methylene blue from polluted water and analysed the results using UltraViolet–Visible (UV–Vis) spectroscopy. The activity encouraged critical thinking about material performance and environmental sustainability.

• Augmented Reality in Education: Designing a Treasure Hunt (AR-EDU)

This workshop introduced students to immersive technologies through a gamified learning experience using the ARLectio^® app 1.5. Participants solved environmental-themed challenges across eight AR-enhanced stations. The activity fostered collaborative problem solving and demonstrated the potential of AR in transforming disciplinary content into interactive experiences.

• Extended Reality (XR) for Behavioural Learning (XR-BE)

Students used the MirageXR toolkit to create an interactive augmented scenario to teach prosocial behaviour, specifically the authoring toolkit was used to show the enactment of “greeting others” behaviour. Working in groups, they developed AR-based learning environments featuring virtual characters and real-world contextualisation. The session enhanced students’ skills in design, scripting, and user-centred educational development.

• Designing Educational Games (GAME-JAM)

In this game-based learning lab, students explored core concepts of game design and applied them in a game jam. Being organised in subgroups, they designed and built two board educational games using cards and dice. The activity supported creativity, teamwork, and reflection on the learning potential of game mechanics.

• Biotechnologies: From Laboratory to Life Sciences (BIOTECH)

Students engaged in molecular biology techniques such as protein and DNA extraction from eukaryotic cells, followed by gel electrophoresis. The lab aimed to reinforce procedural knowledge and scientific reasoning, allowing participants to gain firsthand experience in experimental biotechnologies.

• Artificial Intelligence in Education (AIED)

The workshop explored the role of AI in contemporary educational innovation. Students analysed real-life applications of generative AI tools like ChatGPT (GPT-4o) and DALL·E (DALL·E 3), discussed their societal implications, and reflected on the evolving role of the researcher in digital transformation.

• Data Literacy and Artificial Intelligence (DL-AI)

This lab focused on the ethical and functional dimensions of data and machine learning. Students created simple AI models, experimented with generative tools, and discussed algorithmic decision making. Emphasis was placed on developing digital citizenship and critical reflection on technology’s societal impact.

The workshops were assigned to each group of students recruited through an internal school selection process that identified ten groups of students drawn from various classes and career tracks. The result of this selection was a multidisciplinary path of 10 h for each group of students (

Table 3. Assignment of workshops to student groups.

Group	Workshops
1	CHEM, AR-EDU, XR-BE, GAME-JAM
2	XR-BE, CHEM, AR-EDU, DL-AI, AR-EDU
3	DL-AI, GAME-JAM, AIED
4	BIOTECH, AIED, CHEM
5	AIED, BIOTECH, CHEM
6	CHEM, BIOTECH, DL-AI, AIED
7	DL-AI, AIED, BIOTECH, CHEM
8	AIED, GAME-JAM, XR-BE, AR-EDU
9	BIOTECH, CHEM, AIED
10	XR-BE, CHEM, AIED, BIOTECH

).

2.4. Materials and Data Collection

2.4.1. Group Evaluation Grid

The group evaluation grid (Supplementary Materials) including qualitative field notes was developed for programme-evaluation purposes and aligned with the workshop learning objectives. To assess student group performance across workshops, the structured evaluation grid was used by facilitators assigned to each group at the end of the workshop cycle.

It included four key indicators commonly adopted in engagement-focused classroom and activity research (Fredricks et al., 2004; R. C. Jones, 2008; Linnenbrink-Garcia et al., 2011): participation, observer-rated extent to which students actively took part (e.g., contributing to discussions, asking/answering questions); engagement, general level of attention, interaction, and responsiveness during activities; interest, degree of curiosity and intrinsic motivation demonstrated toward the topics; behaviour, adherence to rules, collaboration within the group, and respectful classroom conduct. Each indicator was operationalised as a group-level assessment on a 10-point Likert scale (1 = very low; 10 = excellent). Because the grid was composed of single-item indicators (not a multi-item psychometric scale), internal consistency indices are not applicable. A field for comments and notes was included for each indicator. These qualitative notes were intended to contextualise ratings by documenting salient behaviours, participation dynamics, and recurring patterns observed during activities.

2.4.2. Observer Calibration and Qualitative Notes

Before data collection, all observers/facilitators took part in a brief calibration meeting, reviewing items and rating anchors, to ensure clarity and content coverage and to align interpretations of the four indicators. Each group was observed and rated by its assigned observer. However, since the same group was not independently rated by multiple observers (no double-coding), formal inter-rater agreement indices could not be computed. Qualitative notes were analysed using a descriptive content approach (thematic grouping of recurring observations) to complement and interpret quantitative ratings.

2.4.3. Chemistry Case Study Questionnaire

The workshop “Chemistry: From Basic Concepts to Frontier Research” was selected as an embedded case study and included an anonymous post-activity chemistry questionnaire. The questionnaire (Supplementary Materials) consisted of the following types of items: feedback/enjoyment (Items 1, 3), students evaluated satisfaction with the experience on a 5-point Likert scale, and willingness to re-participate on 10-point likelihood scale (1 = highly unlikely; 10 = highly likely); prior experience/prior knowledge (Items 2, 4), previous participation in similar activities and prior knowledge about the topic.

For knowledge items (Items 5–12), eight multiple-choice items (one correct answer each) were organised into four thematic domains: declarative knowledge (Items 5, 12), procedural knowledge (Items 7, 9), observation skills (Items 8, 10), and data interpretation and evaluation (Items 6, 11). Each correct answer received a score of 1 (incorrect = 0), producing a total knowledge score ranging from 0 to 8. These four skill areas map a single evidence-to-conclusion pathway in scientific reasoning: knowing core concepts (declarative) supports selecting appropriate methods (procedural), generating accurate observations (data), and ultimately interpreting and evaluating evidence to justify claims. Taken together, the corresponding items reveal how well learners connect what to know with how to investigate and how to decide based on data, yielding a coherent profile of scientific literacy (OECD, 2023). These items were designed to reflect the workshop learning goals and the key concepts addressed during the laboratory activities. Internal consistency was estimated using Kuder–Richardson Formula 20 (KR-20), which is appropriate for binary items and closely related to Cronbach’s alpha under dichotomous scoring.

Because outcomes were bounded and discrete, we used non-parametric tests. Repeated-measures comparisons were examined with Friedman’s test (effect size: Kendall’s W), followed, when significant, by Holm-adjusted Wilcoxon signed-rank tests (effect size: r = |Z|/√n). Descriptive statistics are reported as medians [IQR] and, where useful, means (SD).

2.4.4. Consent Procedure and Anonymity

Data collection was conducted within regular curricular school activities. Both the evaluation grid and the chemistry questionnaire were administered in an anonymous format. Students were informed that questionnaire completion was voluntary and that they could opt out without consequences. The chemistry questionnaire was delivered via a QR-code link and each student decided whether to fill it in. For organisational/time constraints in some groups, the questionnaire could not be administered at the end of the activities; therefore, responses are reported only for the subset of students who completed the post-activity questionnaire.

3. Results

3.1. Group Performance Evaluation

Table 4. Average group performance rating assigned by the facilitators.

Group	Participation M (SD)	Engagement M (SD)	Interest M (SD)	Behaviour M (SD)	Overall M (SD)
1	7 (1)	8.33 (0.58)	8.33 (2.08)	9 (0)	8.17 (0.85)
2	7.25 (0.96)	8 (1.41)	7.75 (1.26)	8.5 (1.73)	7.88 (0.53)
3	5 (0)	6 (1.41)	6.5 (0.71)	6 (1.41)	5.88 (0.65)
4	9 (1)	8.33 (1.15)	8.33 (1.15)	9 (1)	8.67 (0.38)
5	6 (0)	6 (0)	6.25 (0.5)	7.25 (0.5)	6.38 (0.59)
6	9.33 (0.58)	9.67 (0.58)	9.67 (0.58)	9.3 (0.58)	9.50 (0.19)
7	8.5 (0.71)	8,5 (0.71)	8.5 (0.71)	9 (1.41)	8.63 (0.25)
8	9.67 (0.58)	9.67 (0.58)	10 (0)	9.67 (0.58)	9.75 (0.15)
9	7 (1.87)	7 (2)	7 (2)	6.8 (2.28)	6.96 (0.10)
10	6.75 (0.96)	6.75 (0.95)	6.75 (0.96)	7.25 (1.71)	6.88 (0.25)

summarizes the student group performance evaluation of facilitators in terms of participation, engagement, interest, and behaviour, rated on a 10-point Likert scale at the end of the workshop sessions.

The results of the evaluation grid revealed generally high levels of group performance. Several groups received very similar ratings across indicators, which is plausible with bounded rubric scores and may reflect ties/ceiling effects. Table 4 was then interpreted descriptively and integrated with the facilitators’ field note patterns.

The highest average rating was observed for Behaviour (M = 8.18, SD = 1.25), indicating that most groups consistently adhered to rules, collaborated effectively, and demonstrated respectful classroom conduct. Ratings for Interest (M = 7.91, SD = 1.29) and Engagement (M = 7.83, SD = 1.34) were also strong, suggesting that students showed substantial curiosity and sustained attention throughout the workshop activities. Participation received the lowest mean score (M = 7.55, SD = 1.52) and exhibited the greatest variability, possibly reflecting differences in individual willingness to contribute during discussions (Figure 1). Overall, the average group performance score was M = 7.97 (SD = 1.24), with scores ranging from 5.88 to 9.75. Groups 8 (M = 9.75, SD = 0.15) and 6 (M = 9.50, SD = 0.19) showed the highest consistency and performance across dimensions, whereas Group 3 scored the lowest (M = 5.88, SD = 0.65). As an internal check, Friedman’s test in

Table A5. Within-group comparison of evaluation-grid indicators (participation, engagement, interest, behaviour) across groups (N = 10). Indicator scores and Friedman omnibus test.

Indicator	Median [IQR]	Mean (SD)	Mean Rank (Friedman)
Participation	7.13 [6.81–8.88]	7.55 (1.52)	1.90
Engagement	8.17 [6.81–8.46]	7.83 (1.35)	2.35
Interest	8.04 [6.81–8.46]	7.91 (1.30)	2.75
Behaviour	8.75 [7.25–9.00]	8.18 (1.25)	3.00

Friedman test (omnibus): χ²(3) = 5.35, p = 0.148; Kendall’s W = 0.18.

and Pairwise post hoc comparisons in

Table A6. Post hoc Wilcoxon signed-rank tests (Holm-corrected) and effect sizes.

Comparison	n (Non-Zero Diffs)	W Statistic	Z	p-Value	p (Holm)	r = |Z|/√n
Engagement vs. Behaviour	8	3.0	−2.043	0.041	0.246	0.722
Participation vs. Behaviour	8	3.0	−2.035	0.042	0.246	0.720
Interest vs. Behaviour	10	11.0	−1.636	0.102	0.407	0.517
Participation vs. Interest	7	5.0	−1.437	0.151	0.452	0.543
Participation vs. Engagement	5	2.0	−1.348	0.178	0.452	0.603
Engagement vs. Interest	4	1.5	−1.105	0.269	0.452	0.552

Note. The Friedman test assessed within-group differences across indicators (repeated measures across the same 10 groups); effect size is reported as Kendall’s W. Pairwise post hoc comparisons used Wilcoxon signed-rank tests with Holm correction. Effect sizes are reported as r = |Z|/√n, where n is the number of non-zero paired differences. n is smaller than 10 in several comparisons because some groups received identical scores on pairs of indicators (ties).

indicated no significant differences across the four within-group indicators (p = 0.148; Kendall’s W = 0.18).

To complement the rubric-based evaluation, facilitators’ qualitative comments and brief observational notes were analysed to further contextualize the numerical scores assigned to student group participation, engagement, interest, and behaviour.

The analysis related to the participation indicator revealed three primary patterns that closely align with the distribution of quantitative ratings. Several groups were described as showing consistently high and widespread participation, with students actively contributing to discussions, asking pertinent questions, and demonstrating curiosity throughout the activities. For instance, facilitators noted that “the entire group showed high participation, expressing observations and curiosity relevant to the activity and that “most students actively engaged in discussions and responded to questions”. These qualitative insights are consistent with high participation scores observed for Groups 4, 6, 7, and 8 (scores ≥ 8.5). Participation in some groups was described as unevenly distributed, with a portion of students actively involved, while others remained more passive. For example, Group 2 was noted to have “participated in all proposed activities, although not all students were equally active”. This pattern supports the intermediate participation score observed for this group (7.25), suggesting that numerical averages may mask internal variability. Finally, lower levels of participation were observed in a few groups, often characterised by limited student involvement or the need for external prompting by facilitators. Comments such as “except for a few students, the group showed low participation” and “some students followed attentively, while others tended to be distracted” correspond with the lower numerical ratings of Groups 3 and 5.

The results related to the engagement indicator confirm the existence in several groups of consistently high levels of engagement. Some groups were described as “very attentive to the topics and highly participatory”, or as having “fluid and productive interactions with the instructor and peers”. These observations support the high engagement scores attributed to groups such as 6, 7, and 8 (Engagement ≥ 8.5). Other groups showed more heterogeneous patterns of engagement, with some students actively involved while others were more passive. For example, Group 2 was described as having “some students with high attention and interaction, others with moderate levels, and about half with low attention”, which aligns with their intermediate engagement rating (Engagement = 8) and highlights how averages can mask within-group variability. Lastly, a few groups demonstrated lower or fluctuating engagement, as reflected in comments such as “the group showed low levels of attention and limited interaction” (Group 3) or “engagement was inconsistent… some students were attentive while others remained passive” These remarks are consistent with lower engagement scores assigned to Groups 3 and 5.

The notes related to the interest indicator revealed patterns consistent with the quantitative data, helping to clarify variations in student motivation and curiosity across groups. Several groups were described as exhibiting high levels of interest, characterised by enthusiasm, spontaneous curiosity, and active questioning. For instance, one facilitator noted that “students participated with enthusiasm and curiosity throughout the proposed activities”, aligning with the highest numerical ratings attributed to Groups 6, 7, and 8. Other groups displayed a moderate level, where some students were actively engaged while others showed limited involvement. In these cases, comments such as “only some students expressed strong curiosity” (Group 2) or “in general, the group was motivated” (Group 5) supported a more than sufficient level. Finally, a few groups showed low or inconsistent levels of interest, with facilitators reporting limited curiosity and reduced engagement with the subject matter. These remarks aligned with the lowest interest scores in the dataset (e.g., Group 5 and Group 3).

Further insights come from the behaviour indicators, capturing nuances in students’ ability to follow rules, collaborate, and demonstrate appropriate conduct in a professional research setting. The comments confirmed the variation in behavioural performance reflected in the numerical ratings. Several groups were noted for their excellent behavioural conduct, characterised by rule-following, mutual respect, and positive collaboration with both peers and researchers. For instance, facilitators described groups as “maintaining an excellent attitude, demonstrating respect for rules, instructors, and classmates” and “creating a calm and productive atmosphere” These comments support the high behaviour ratings attributed to Groups 4, 6, 8, 7 and 1. Other groups showed moderate or heterogenous behavioural performance, with some students fully respecting expectations and others requiring reminders or behaving less appropriately: “Group 2 displayed heterogeneous behaviour and respect for rules; only some students behaved appropriately given the professional context”; “Although the group respected the rules overall, in some cases it was necessary to call students to attention”. Such comments align with middle-range behaviour scores, including Groups 2 and 5. A few groups were described as having significant difficulties, including lack of respect for the research environment, poor collaboration, and inattentiveness. A facilitator noted: “Except for a few students, the group was disrespectful toward the work environment and showed limited cooperation”. This observation is consistent with the lower behaviour scores assigned to Groups 3 and 9.

3.2. A Case Study: Chemistry Pathway Questionnaire Evaluation

The chemistry pathway was structured as follows:

• Step 1. Brief presentation of the ISMN Institute and our chemistry dissemination group “ChimiCom@CNRPA”.
• Step 2. Introduction to the issue of water pollution and its main causes.
• Step 3. Overview of porous, high-surface-area silica-based nanostructured materials and their laboratory preparation.
• Step 4. Explanation of the planned experiment: “Water purification from methylene blue dye by adsorption on three silica-based materials, and quantitative analysis of the treated water using UV–Vis Spectroscopy”.
• Step 5. Execution of the experiment in small groups (of 2–4 students) under the supervision of the instructor and tutor, followed by quantitative analysis of the treated water.
• Step 6. Group discussion to identify the most sustainable material among the three tested.
• Step 7. Administration of the post-activity satisfaction and learning questionnaire (via QR code).

In Figure 2, some relevant steps of experimental activity are shown.

The main learning objectives of this pathway were to capture students’ attention and curiosity toward the issue of water pollution and to demonstrate how chemistry can make a concrete contribution to addressing this global challenge. In addition, the activity aimed to introduce the concepts of surface area and porosity in relation to the materials used. The workshop adopted a broader STEM approach by explicitly touching, beside chemistry, some aspects pertaining to mathematics (spectral analysis and interpretation of qualitative/quantitative patterns), geology (natural silica-based materials and their origin), ecology (circular-economy scenarios such as waste valorisation and water remediation), and engineering (characterisation techniques and the production of manufactured goods, including prototyping through 3D printing).

Despite the chemistry workshop was primarily framed as a STEM-oriented learning experience, it intentionally incorporated selected STEAM-related dimensions, particularly creative problem-solving, design-based thinking, and reflective collaboration, which complemented scientific inquiry and technology-enhanced experimentation. The goal was to stimulate students’ scientific creativity, considering all its possible dimensions, i.e., product, process, and trait, as discussed by Pinar et al. (2025). Indeed, scientific creativity plays a significant role in shaping students’ future scientific careers (Pinar et al., 2025). Students worked with porous and nanostructured silica-based materials and explored their potential applications in environmentally relevant contexts (e.g., waste recycling and wastewater treatment), making evidence-based choices that foregrounded sustainability considerations (e.g., selecting materials and procedures by weighing performance and environmental impact). A key technological component was the use of UV–Vis spectroscopy for material characterisation and data acquisition, which enabled students to connect experimental observations to qualitative/quantitative representations (spectral curves) and to discuss how measured signals relate to underlying material properties. Moreover, the activity was operationalised through communication and argumentation: students had to explain and justify group decisions, interpret results collectively, and present reasoned conclusions, thereby integrating scientific evidence with collaborative meaning making.

This chemistry workshop differed from traditional school chemistry laboratories because the experiments were carried out using silica-based materials developed and optimised at the CNR-ISMN Palermo laboratories as part of ongoing research activities. Moreover, the workshop involved specialised equipment and expertise that are not typically available in school science laboratories.

Table 5. Distribution of student groups.

Group	Students Number
1	18
2	16
4	14
5	15
6	15
7	18
9	14
10	18
Total	128

shows the total of 128 students distributed in the eight groups participating in the chemistry workshop. Due to time constraints, it was not possible to administer the questionnaire to Groups 1 and 2. A total of 69 students out of 128 completed the chemistry questionnaire. The remaining 25 students were absent or chose not to complete the questionnaire.

As shown in Figure 3, students reported high levels of satisfaction with the activity. The distribution of satisfaction scores (left panel) revealed a strong skew toward the upper end of the scale, with most students selecting 4 or 5 on a five-point scale. Descriptive statistics confirmed this trend (M = 4.16, SD = 0.83), indicating that the overall perception of the activity was highly positive.

The central panel illustrates students’ willingness to participate again, which was also high. The average re-participation score (M = 7.22, SD = 2.37), with most responses clustering between 7 and 10, was suggesting strong interest in repeating the experience.

Finally, the right panel shows that most students were first-time participants (Yes; 85.5%), yet their evaluations were comparably positive, indicating that the activity was engaging regardless of prior experience.

Taken together, these results suggest that students not only expressed very high satisfaction but also demonstrated a strong intention to re-engage, highlighting the perceived value of the experience.

Responses to the item assessing prior knowledge indicated that approximately half of the students (50.7%) reported having heard about pollution from organic dyes before. In contrast, more than one third (36.2%) stated that they had not, while a smaller proportion (13.0%) expressed uncertainty. These findings suggest a heterogeneous background among participants, with some students already familiar with the topic and others encountering it for the first time.

Table A1. Item statistics for the chemistry knowledge subtest (N = 69).

Item	Domain	Response Options	Difficulty p (Prop. Correct)	CITC (Item–Rest)
DK5	DK	4	0.44	0.35
DIE6	DIE	3	0.74	0.16
PK7	PK	3	0.44	0.18
OS8	OS	4	0.41	0.13
PK9	PK	3	0.65	0.27
OS10	OS	3	0.62	0.33
DI11	DI	3	0.68	0.33
DK12	DK	3	0.49	0.12

Note. Difficulty p is the proportion of students answering the item correctly (higher values indicate easier items). CITC is the corrected item–total correlation (item–rest correlation), computed as the correlation between each item score and the total knowledge score excluding that item; higher values indicate better item discrimination. The 8-item knowledge subtest showed moderate internal consistency (KR-20 = 0.51; Cronbach’s α ≈ 0.51).

reports the eight-item knowledge score with moderate internal consistency due to the brief length of the subtest (KR-20 = 0.51; α ≈ 0.50). Item difficulty (proportion correct) and corrected item–total correlations are reported in

Table A2. Descriptive statistics and reliability for the chemistry knowledge score (N = 69).

Measure	Value
Number of items (k)	8
Possible score range	0–8
Observed score range	1–8
Mean (SD)	4.46 (1.84)
Median [IQR]	5 [3–6]
Mean proportion correct (Mean/8)	0.56
KR-20	0.51
KR-20 (95% bootstrap CI)	[0.33, 0.64]
Cronbach’s α (dichotomous items)	0.50

Note. KR-20 is appropriate for dichotomously scored items and is closely related to Cronbach’s alpha under 0/1 scoring. The moderate coefficient is expected given the brief length of the subtest (8 items).

.

Students’ overall performance on the eight knowledge items indicated a moderate level of accuracy. On average, participants answered correctly slightly more than half of the items (M = 4.46, SD = 1.84). The median score was 5, with scores ranging from 1 to 8. Interquartile values showed that 50% of students obtained between 3 and 6 correct answers. These results suggest variability in knowledge acquisition, with some students demonstrating high accuracy while others exhibited substantial difficulties (Figure 4).

Group-level analysis revealed variation in performance, with mean scores ranging from M = 3.11 (Group 9) to M = 5.62 (Group 7). All groups except Groups 4 and 9 answered on average more than 50% of the questions correctly (Figure 5).

Table 6. Average percentage of correct answers per knowledge area.

Area	N	M (%)	SD
DK—Declarative Knowledge	69	46.38	38.65
DIE—Data Interpretation and Evaluation	69	71.01	32.55
PK—Procedural Knowledge	69	54.35	37.12
OS—Observation Skills	69	51.45	35.33

summarizes the average percentage of correct responses related to the four knowledge domains. DIE was the highest knowledge area (Data Interpretation and Evaluation; M = 71.01%, SD = 32.55), followed by PK (Procedural Knowledge; M = 54.35%, SD = 37.12), OS (Observation Skills; M = 51.45%, SD = 35.33), and DK (Declarative Knowledge; M = 46.38%, SD = 38.65). A Friedman test indicated a significant difference among the domains, χ²(3) = 19.32, p < 0.001, with a small effect (Kendall’s W = 0.093). Post hoc Wilcoxon signed-rank tests with Holm correction revealed that DIE scores were significantly higher than DK (p < 0.001), PK (p = 0.001), and OS (p < 0.001). No other pairwise comparisons have reached significance (

Table A3. Within-student comparison of knowledge domains (DK, DIE, PK, OS) in the chemistry case study (N = 69). Domain scores and Friedman omnibus test.

Domain	Items (0/1)	Domain Score Computation	Median [IQR]	Mean (SD)	Mean Rank (Friedman)
DK	DK5, DK12	DK5 + DK12	1 [0–2]	0.93 (0.77)	2.254
DIE	DIE6, DI11	DIE6 + DI11	2 [1–2]	1.42 (0.65)	2.986
PK	PK7, PK9	PK7 + PK9	1 [1–2]	1.09 (0.74)	2.428
OS	OS8, OS10	OS8 + OS10	1 [1–2]	1.03 (0.71)	2.333

Friedman test (omnibus): χ²(3) = 19.324, p = 0.000234; Kendall’s W = 0.093.

and

Table A4. Post hoc Wilcoxon signed-rank tests (Holm-corrected) and effect sizes.

Comparison	n (Non-Zero Diffs)	W Statistic	Z	p-Value	p (Holm)	r = |Z|/√n
DK vs. DIE	50	266.5	−3.826	0.000130	0.000782	0.541
DIE vs. OS	43	200.0	3.634	0.000279	0.001394	0.554
DIE vs. PK	40	190.0	3.279	0.001040	0.004161	0.519
DK vs. PK	41	329.0	−1.385	0.165921	0.497762	0.216
DK vs. OS	39	325.5	−0.983	0.325742	0.651483	0.157
PK vs. OS	39	361.0	0.422	0.672926	0.672926	0.068

Note. Domain scores were computed by summing two dichotomously scored items (1 = correct, 0 = incorrect), yielding domain scores from 0 to 2. The Friedman test assessed within-student differences across domains; effect size is reported as Kendall’s W. Pairwise post hoc comparisons were conducted using Wilcoxon signed-rank tests; p-values were adjusted using the Holm method. Effect sizes for post hoc contrasts are reported as r = |Z|/√n, where n is the number of non-zero paired differences.

).

Comparison between testing time slots revealed a significant effect. The Mann–Whitney U test (U = 393.50, p = 0.015) showed that students assessed in the second time slot, 12:00–14:00, had significantly higher total correct scores (M = 4.97, SD = 1.89) than those assessed in the first time slot (M = 3.91, SD = 1.63), 10:00–12:00.

Regarding prior knowledge of organic dye pollution, no significant differences emerged when comparing students who responded “Yes” (M = 4.66, SD = 1.94), “No” (M = 4.04, SD = 1.86), or “I don’t know” (M = 4.89, SD = 1.17). Thus, self-reported prior exposure to the topic did not significantly influence students’ performance, even when stratified by time of assessment.

4. Discussion

The experience of carrying out scientific activities within a real research institute played a crucial role in fostering student engagement, motivation, and curiosity toward STEM fields. This multidisciplinary approach exposed students to up-to-date, research-based knowledge and technologies, providing an immersive experience of scientific inquiry and allowing them to briefly inhabit the role of “researchers for a day.” Emotional and motivational processes were deliberately addressed to foster inclusion and active participation, ensuring that each student had a meaningful role within the group and could engage with the activities in a supportive collaborative setting (Lanouette, 2026). Moreover, the integration of multiple STEM disciplines within a single multidisciplinary learning path provided students with a holistic view of how scientific knowledge is generated and applied in the real world and offering students an overview of the future developments. This aligns with current educational frameworks promoting different future perspective learning as a strategy to enhance both understanding and orientation in science education (A. Jones et al., 2011; Laherto & Rasa, 2022). In this sense, the pathway can also be interpreted as a research-to-action, multidisciplinary experience that enables students to “be a researcher for one day” through authentic practices and exposure to up-to-date research-based knowledge and technologies (Bourne et al., 2025).

The results highlight a generally positive reception of the workshops and strong levels of student involvement, particularly in terms of collaborative behaviour and intrinsic interest in the topics covered. Participation showed greater dispersion across groups than the other indicators, suggesting that participation may be particularly sensitive to group dynamics and facilitation. Accordingly, future implementations could test whether structured prompts and inclusive facilitation strategies help reduce variability and support broader involvement. Groups that performed well were often described as highly attentive, collaborative, and inquisitive not only in relation to the specific tasks but also in their broader interest in the professional and societal implications of scientific research. The facilitators’ notes highlighted how exposure to an authentic research environment, tools, and direct interaction with researchers may have had a significant emotional and cognitive impact on students. Observations such as “the group showed exceptional curiosity,” or “students actively asked questions about the professional applications of the topics,” illustrate the potential of context-based learning to activate intrinsic motivation and deeper learning processes (Lave & Wenger, 1991; Sugarman, 1987). Recent science education research has shown that emotion is not peripheral to scientific learning but is intertwined with how learners engage with evidence and data practices, shaping participation, sense making, and agency (Herrick et al., 2026; Lanouette, 2026). Consistently, we intentionally foregrounded emotional and relational dimensions (e.g., assigning meaningful roles within groups and supporting collaborative decision making) to promote participation and inclusion in science activities. This emphasis resonates with recent work on hope and action in sustainability education, which highlights how affective dynamics can support learners’ engagement and future-oriented agency, especially when scientific inquiry is connected to societally relevant issues (Røkenes & Jornet, 2026). In this sense, the intervention succeeded not only in promoting knowledge acquisition, but also in immersing students in a meaningful educational context that stimulated both cognitive and affective dimensions of learning reinforcing the importance of bridging school-based learning with authentic, research-based experiences to increase student interest and engagement in STEM pathways. In this direction, new educational scenarios are emerging through multi-stakeholder initiatives that bring together research centres, industry, and schools, giving students access to innovative learning environments and pathways that can spark interest in science and help identify and nurture emerging talent (Yu & Niu, 2026).

The results obtained in the workshop “Chemistry: From Fundamentals to Frontier Research” reveal that most students successfully acquired key disciplinary knowledge through the proposed educational intervention. The overall performance, with most scores concentrated in the mid-to-high range, suggests that the workshop design was effective in supporting conceptual understanding of complex chemistry-related content. This suggests that most students were able to engage effectively with materials and tools, demonstrating an adequate understanding of the scientific content addressed during the workshop. In addition, the students became protagonists and had experience in the observation and interpretation of the experiments through a scientific approach in a methodological way. These results put forward well-structured educational interventions, especially those that combine hands-on experimentation with scientific discussion, which can foster a solid grasp of disciplinary content even within a short instructional cycle. However, the variability observed between groups suggests that contextual factors, such as group composition, grade levels, and timing of participation, may play a moderating role in learning outcomes.

The chemistry workshop activities go beyond what is typically feasible in a school laboratory because they rely on research-grade, synthetically produced materials (e.g., porous/nanostructured silica) that are not commercially available for standard school procurement. As a result, implementing the activities requires access to an active research environment where such materials can be produced, characterised, and used within authentic research workflows. Moreover, the chemistry workshop also suggests a broader STEM integration, linking chemistry to mathematics (UV–Vis spectral analysis and interpretation of quantitative patterns), geology (natural silica-based materials and their properties), ecology (waste recycling and wastewater treatment applications), and engineering (characterisation techniques and prototyping, including 3D printing).

Future studies could delve deeper into how these individual and contextual factors interact, possibly adopting a mixed-method approach that integrates pre/post testing, student reflections, and longitudinal tracking of conceptual gains. Moreover, the relationship between students’ performance and their emotional-cognitive engagement should be further investigated, as proposed in recent educational neuroscience frameworks and studies, and taking to the role of teacher motivation in generating student achievement (Immordino-Yang & Damasio, 2007; Pekrun, 2021).

5. Limitations

The findings of this study should be interpreted with consideration of several limitations. (i) The research design was descriptive and post-only, with no pre-test measures, control group, or random assignment. Accordingly, the findings reflect outcomes observed within the implementation context and do not provide a basis for causal inference concerning programme effectiveness. (ii) Group-level performance was assessed through an evaluation grid completed by the facilitator, using single-item indicators. Despite all observers having participated in a pre-implementation calibration meeting, each group was rated by its assigned observer and groups were not double-coded. Hence, formal inter-rater agreement could not be estimated and observer-related variability across groups may be present. (iii) The evaluation grid captures outcomes at group level, and this may mask within-group heterogeneity and does not allow individual-level modelling of participation and engagement patterns. (iv) The chemistry workshop questionnaire was administered to a subset of groups, with additional attrition resulting from non-completion at the end of the activity due to time or organisational constraints and student absence. This potential self-selection may constrain the generalisability of the case study findings to the full participant group. (v) No formal pilot testing and full construct validation were conducted, and thus the instruments should be considered exploratory. An optimised step should refine the items, expand the knowledge subtest, and formally examine validity and reliability. (vi) This study was conducted within a single regional context and within a specific research-centre programme, whereas replication in different school contexts and with longitudinal follow-up is recommended to strengthen external validity.

6. Conclusions

The results of this study highlight the effectiveness of the multidisciplinary and immersive workshops in fostering both conceptual understanding and active engagement in STEM education. The participating students had the opportunity to engage with up-to-date, research-based knowledge and technologies, gaining first-hand experience of scientific research, becoming “researchers for one day”. Attention was paid to emotional and motivational aspects to promote inclusion, ensure active participation, and assign each student a meaningful role within the group. Students who participated in the workshops within the context of a research centre demonstrated not only high content mastery but also great levels of interest, curiosity, and collaborative behaviour. These findings reinforce the pedagogical value of providing authentic, context-rich learning environments in which disciplinary knowledge is situated within meaningful scientific practices, in line with a STEM framework enriched by selected STEAM-related dimensions, particularly those fostering creative problem-solving, reflective collaboration, and design-oriented thinking. The alignment between facilitator evaluations and student performance suggests that emotional and behavioural engagement played a key role in supporting learning, consistent with theoretical models of integrated cognitive-affective learning. Furthermore, the multidisciplinary nature of the experience (exploring materials science, artificial intelligence, biotechnology, augmented reality in a research context) contributed to students’ recognition of the relevance of STEM and fostered reflection on innovative solutions to real-world problems. Considering these findings, it is recommended to expand such immersive STEM initiatives in collaboration with research institutions, especially for students at the secondary schools. Future research could explore the long-term impact of these experiences on students’ academic choices and perceptions of scientific careers. Additionally, integrating tools for student self-reflection and longitudinal tracking would enhance our understanding of how such interventions contribute to sustained STEM engagement and identity development. This, in turn, would stimulate motivation and participation, especially among students who might not have initially shown high levels of interest. As a final remark, this study has contributed to stimulate further investigation on educational interventions that fosters stronger connections between research institutions and schools through collaborative programmes aimed at increasing students’ interest in STEM subjects and scientific research.