Conference Program

This paper presents the counterfactual evaluation of the educational project R.E.A.C.T. (Reti per Educare gli Adolescenti Attraverso la Comunità e il Territorio/Networks for Educating Adolescents through Community and Local Context), aimed at tackling educational poverty (Save the Children, 2014; Lohmann & Ferger, 2014), a widespread phenomenon in Italy both in terms of skills attainment (European Commission, 2025) and from a multidimensional perspective (ISTAT, 2024). The intervention, implemented during the 2018–2019 and 2019–2020 school years, involved 10 lower secondary schools located in highly deprived areas across Italy. Training and cultural activities were delivered to the local communities and to all students in the participating schools, with the objective of activating and strengthening the local “comunità educanti” (Zamengo & Valenzano, 2018; Tomei & Galligani, 2020). In addition, 226 particularly vulnerable students were provided with individualized educational plans developed through the Family Group Conference (FGC) model (Frost et al., 2014). This approach was adopted to better identify students’ needs, mobilize family networks, and engage both formal and informal educational agencies within the local context. R.E.A.C.T. represents an interesting application of FGCs for explicitly educational purposes, within an international context where this tool is primarily used in child protection and welfare services (Hohashi & Yi, 2025; Edwards & Parkinson, 2018), and where evidence regarding its effectiveness remains mixed (Dijkstra et al., 2016; Mitchell, 2020; Nurmatov, 2020; Hohashi & Yi, 2025).

The impact evaluation focuses on the vulnerable students who participated in the individualized educational plans. A randomized controlled trial (RCT) design was adopted: teachers identified 431 students experiencing significant academic and relational difficulties, and through random assignment, 226 were allocated to the treatment group and 205 to the control group. For this population, both short- and long-term outcomes were assessed.

Short-term outcomes concerned students’ perceived well-being (measured through dimensions of school well-being, relationships with teachers, relationships with parents, and expectations about their future). A questionnaire administered before and after the intervention allowed these dimensions to be measured and changes over time to be recorded.

Long-term outcomes referred to students’ educational trajectories in the years following the intervention (promotion to the next grade and school dropout) as well as academic performance (grades obtained and scores in the national INVALSI standardized tests). These outcomes were recorded using data from the Anagrafe Nazionale degli Studenti (National Student Registry), covering four school years from 2018/2019 to 2021/2022, and from INVALSI.

The analyses show a positive impact of the project on perceived school well-being (but no significant effects on relationships with teachers and parents or on future expectations). In the long term, the intervention is associated with a reduction in the risk of grade retention and school dropout among participating students.

The paper presents early findings of an empirical research, within an exploratory framework, involving assessment processes carried out by generative artificial intelligence (AI) and human evaluators (HI). Research questions wonder whether I) between results of semi-structured test assessments carried out by HI and results of assessments carried out by AI, lies a discrepancy that could be considered critical – greater than the one found between human evaluators in the previous research phase – and II) if the upload of teaching materials, integrated with AI-HI interaction, have a significant relation with discrepancy reduction. Objectives of research are then to control the existence of such critical discrepancy and the relation between teaching materials upload and discrepancy reduction. The research hypothesis states that I) a critical discrepancy exists and that II) it can be reduced with teaching materials uploading, commented by human evaluators. Population consists of Masters’s Degree students, the sample being composed by those who have carried out in groups, as in itinere or summative assessment, a semi-structured project work during academic years 2024/2025 and 2025/2026.

Awaiting further investigation, initial results describe the discrepancy as critical and seem to indicate that AI may be subject to docimological bias already attested for human evaluators as adoption of implicit criteria, failure to identify intermediate performance levels, lack of consistency over time, in addition to specific AI-related biases. This line of research is established in defence of a responsible and inter-subjective assessment practice, promoting quality and transparency in democratic pedagogy.

In the contemporary debate on democratic education, the distinction between Learning for Democracy (LforD) and Democracy for Learning (DforL) has shaped much of the theoretical reflection and educational practice [1–3]. However, both approaches share a common limitation: they tend to focus either on the acquisition of competences or on participatory experience in itself, without systematically integrating the material and structural conditions that make political action possible [4,5]. In a context marked by rising social inequalities and the increasing mediatization of the public sphere, this gap is particularly consequential for efforts to promote and enable youth participation [6–9].

This contribution proposes Materialist Civism as a theoretical paradigm and evaluative framework for promoting and studying youth political activation. This perspective reframes civic education not as the normative transmission of values nor as the mere experimentation of democratic practices, but as a process of concretely expanding the participatory and transformative capacities of young people [10,3]. Political activation is thus understood as the outcome of interactions between formative dimensions and material conditions, both individual and collective [11,12].

The paradigm was formalized through the experience of Rizomi’s School of Political Activation and the analysis of the Theory of Change developed during the educational planning phase [13,14]. The adopted research methodology integrates: (1) a literature review on political participation and social behavior, with particular attention to enabling factors and structural constraints connected to participation [15,12]; (2) a critical analysis of major practices in active citizenship education [16,17]; (3) the theoretical reconstruction of the Theory of Change [13,14]; (4) a comparison with evidence emerging from a recent pilot study conducted within the School, using questionnaires aimed at studying the dimensions underlying the propensity to engage politically [18].

The key innovation achieved lies in the elaboration of a multidimensional evaluation model that moves beyond an assessment centered exclusively on “democratic competences,” reintroducing the importance of contextual and relational factors [19–21]. The proposed framework identifies three domains: (a) personal resources, including competences, motivation, and self‑efficacy [22,23]; (b) situated and collective resources, such as relational networks, organizational access to resources and tools, and shared agency [24,25]; (c) resources of political representation, understood as the concrete possibility of accessing decision‑making spaces and influencing public discourse [26,27]. The evaluation of political activation therefore requires tools capable of jointly measuring learning, resource availability, and the transformation of participatory opportunities [19,15].

By overcoming the separation between theory and practice and between LforD and DforL, Materialist Civism offers a transferable paradigm for both educational design and the evaluation of youth policies [10,16]. It contributes to the debate on democratic education by proposing a model capable of integrating training, context, and the actual distribution of resources and opportunities, enabling interventions aimed at fostering creative, autonomous, and situated youth participation [11,12,27].

Educational inequality and educational poverty can be conceptualized—and consequently addressed—from different analytical perspectives. The most common approach focuses on the formal completion of basic levels of schooling. Traditionally, this has been operationalized through indicators such as early school leaving rates or the share of students who fail to obtain an upper secondary diploma or vocational qualification.

Over the last decades, however, increasing attention has been paid to the limits of this perspective. With the development of the knowledge society, education is increasingly understood as a continuous process extending across the entire life course. In societies and economies characterized by rapid and constant transformation, opportunities for learning no longer coincide solely with formal schooling. As a result, educational inequalities may emerge at different stages of individual trajectories and involve multiple dimensions of learning.

This shift has generated renewed interest in inequalities in access to and completion of tertiary education. In this field, alongside traditional attention to class-based inequalities, growing attention has been devoted to gender differences in access to specific fields of study, often intersecting with other factors such as citizenship status or territorial disparities. At the same time, increasing attention has been directed toward the effective skills and resources acquired through educational processes. The recognition that formal credentials alone are insufficient to guarantee meaningful life opportunities has led to the development of new systems of evaluation and measurement focusing on students’ actual competences and skills, such as the implicit dropout indicators proposed by INVALSI.

Within this framework, educational poverty can be conceptualized as a multidimensional phenomenon measurable through a plurality of indicators.

However, the theoretical and empirical implications of this multidimensionality remain only partially explored. More specifically, there is no reason to assume that the same risk factors operate uniformly across different dimensions of educational disadvantage.

This issue becomes particularly relevant when considering the intersection between different vulnerability factors, such as social class, gender, citizenship, or territorial location. While in relatively shrinking phenomena such as early school leaving the most disadvantaged profiles often accumulate multiple risk factors, this pattern may not hold in expanding or emerging domains such as implicit school dispersion, unequal skills acquisition, or access to lifelong learning opportunities.

Drawing on available empirical evidence, this contribution offers an exploratory analysis of the intersections—technically, interactions—between key risk factors associated with educational poverty across four domains: early school leaving, implicit school dispersion, access to tertiary education, and participation in lifelong learning. By comparing these dimensions, the paper aims to highlight how the configuration of educational inequalities may vary across different stages of the educational trajectory. In doing so, the study discusses how evaluation frameworks can contribute to a more comprehensive understanding of educational disadvantage and support policies aimed at expanding equitable learning opportunities. More broadly, improving the evaluation of educational inequalities represents a key step toward strengthening democratic participation, as access to meaningful learning opportunities constitutes a fundamental condition for active citizenship and inclusive democratic societies.

According to the trend at the international level (OECD, 2013), the Italian National Evaluation System, established by Presidential Decree (DPR) No. 80/2013, aims to improve the quality of educational services and learning outcomes and evaluates the efficiency and effectiveness of the education and training system. Within the framework of the National Evaluation System, a gradual systematisation of the evaluation process developed for state and state-recognised schools at primary and secondary level, for provincial adult education centres, and for pre-schools (Freddano & Stringher, 2021; Freddano, 2025a, 2025b; Vinci, Freddano, Torti, 2026). Vocational Education and Training (VET) has also been interested by the introduction of a pilot self-evaluation report, marking this sector’s full integration into the education and training system.

In Italy VET represents a hybrid segment in which both regional and national dimensions coexist and it is based on specific quality assurance experiences and suggestions at national and international level (INAPP, 2025; Carlini & Evangelista, 2020; Dordit, 2018). Its dual nature, both national and regional, has influenced INVALSI to develop a specific self-evaluation report to ensure a uniform structure across all accredited training centres providing vocational education and training, while also recognising regional specificities (Freddano 2025b).

At the end of the experimentation implemented by INVALSI in the 2020/2021 training year, the Director/Coordinator of the 173 accredited training centres that participated at the experimentation, had been asked to provide feedback on the experimented self-evaluation report by filling in an online questionnaire. The aim was to have suggestions on the tools and procedures and to identify aspects of the self-evaluation process that had raised concerns and that could be revised ahead of the scale up of the self-evaluation report.

The paper has the aim to understand how well the self-evaluation experience of the training centres met initial expectations and what roles they assigned to the evaluation process through the evaluation report. In total 168 Directors and coordinators of the 173 participating training centres compiled the final questionnaire.

Descriptive and multivariate analyses were conducted on the collected data, focusing on three specific items exploring respondents’ opinions. In particular, the analysis examined: the general perception of self-evaluation within training providers, the level of satisfaction with the experience compared to initial expectations, and the opportunities provided by the process to develop greater awareness of internal dynamics.

The study explored organizational self-reflection, drivers for change, and the enhancement of training activities. The goal was to determine to what extent self-evaluation experience is perceived as a functional tool for organizational reflection and quality improvement.

The implementation of the Self-Evaluation Report at an experimental level marks a significant milestone for the VET sector. By joining the National Evaluation System while preserving its regional specificities, the VET sector strengthens its role within the national education system, with positive implications for the quality and expansion of training offer.

In recent years, particularly following the COVID-19 pandemic, public and academic debate at both national and international levels has increasingly focused on the design of digital and hybrid learning environments capable of integrating or transforming traditional educational models (Barola et al., 2022). In this context, the STEAM approach (Science, Technology, Engineering, Arts and Mathematics) has progressively emerged as a transdisciplinary educational methodology capable of promoting active, collaborative, and inquiry-based learning, fostering students’ engagement with scientific and technological disciplines.

Educational projects based on the integration of STEAM and digital technologies represent a significant response to the challenges of contemporary societies, in which citizens are required not only to use advanced technologies but also to produce knowledge and participate critically and consciously in socio-technological ecosystems (Quigley et al., 2020). The transdisciplinary nature of the STEAM approach contributes to the development of key competences such as creative thinking, problem solving, communication, and collaboration (Brush & Saye, 2008). Within this framework, this contribution presents the outcomes of the design and pilot testing process of assessment tools aimed at analysing the quality and effectiveness of STEAM educational practices developed within the Erasmus+ KA2 project EDUMAT+ – Augmented Educational Mat: an infographic puzzle MATS for learning STEAM by coding for primary schools. The project involves teachers and students from primary schools and school networks in Italy, Romania, Bulgaria, Spain, and Portugal—contexts in which strengthening digital competences and reducing educational inequalities remain key priorities.

The EDUMAT+ project promotes the use of coding and infographic mats as innovative tools to support the teaching of topics related to the 2030 Agenda (United Nations, 2015), such as environmental sustainability and social inclusion. In particular, it envisages the development of augmented reality educational mats and coding activities that function as pedagogical devices for the integrated development of digital and citizenship competences.

The research adopts an integrated evaluation approach based on mixed methods, aimed at analysing both the educational processes activated and the outcomes of the teaching intervention. The presentation will therefore focus on the evaluation framework used within the EDUMAT+ project and on the first results of its implementation.

The analysis of the data collected during the pilot phase makes it possible to reflect on the role of digital STEAM practices as tools for promoting critical digital literacy, educational equity, and democratic participation within school contexts in the digital society.