Alice Dominici is an applied microeconomist and a Visiting Fellow at the EUI Department of Economics. Her research focuses on, among other things, the economics of education.
In this interview, Alice discusses her research and the working paper ‘Innovative learn-by-doing digital education and attitudes towards STEM: an RCT in Italian high schools’, co-authored with Martina Francesca Ferracane, Adriano De Falco, and Giovanni Abbiati. Their study tested how creative STEM (Science, Technology, Engineering, and Mathematics) activities affect high school students' career aspirations and their likelihood to enrol in STEM university courses, using a sample of 710 students in Italy. Across the EU, just over a quarter (26.9%) of tertiary students are studying STEM subjects. In Italy, the percentage is even lower, reaching only around 21%.
To conduct their research, Alice and her co-authors partnered with FabLabs, an organisation that brought the creative activities directly into the schools. FabLabs utilise a hands-on, learning-by-doing pedagogical approach to teach digital fabrication tools, fostering creativity, collaboration, and problem-solving. Originally established by the Massachusetts Institute of Technology (MIT) in the early 2000s, FabLabs have now spread worldwide — operating as independent laboratories as well as within schools, universities, and corporate R&D departments.
You argue that, in today’s world of labour, only skills complementing digital technologies are safe from automation. What are these specific skills, and do you believe the younger generation is equipped with them?
A large body of work in economics shows that digital technologies, and now AI, do not destroy jobs uniformly. They substitute for tasks that are routine and codifiable, and they reward workers who can do things machines still can't: framing the problem, deciding what is worth automating, judging whether the output makes sense, and combining technical tools with domain knowledge and creativity.
In plain words, the skills that survive and thrive alongside digital technologies are the ability to use these tools to build something new, rather than only consuming what they produce. Unlocking these skills requires not only a baseline of computational and quantitative literacy — referring specifically to coding and statistics that feed directly into AI — to engage with new technologies, but also creativity, problem-solving, collaboration, and the patience to iterate when something doesn't work the first time.
The younger generation is often described as 'digital natives', and they certainly are when considered as users. But knowing how to use TikTok or ChatGPT is not the same as being able to design with these technologies. International evidence, including PISA test scores, shows that, in Italy, teenagers' performance in mathematics and problem-solving has been stagnant or declining, and the share of students who go on to a STEM degree remains low for boys, and even more so for girls. So, the honest answer is: Today's teenagers have the interface skills, but most do not yet have the production skills that complement digital technologies. That gap is exactly what our research is trying to address.
You co-authored a paper on creative STEM activities in high schools. Could you briefly walk us through the scope of this project and the methodology used? Why did you choose Italian high schools as the focus of your work?
The project started a few years back, when Adriano de Falco and I were PhD researchers at the European University Institute, while Martina Francesca Ferracane was a postdoc. The project included two different papers: the one we are discussing, focused on STEM attitudes, and one (already published in Quality & Quantity and involving a different team from the University of Florence) on creativity and grit. While Martina Ferracane had already been involved in FabLabs, the rest of us joined the project for our methodological expertise in experimental design and causal inference, which is a branch of statistics and econometrics dealing with verifying that one variable causes another, rather than the two co-occurring at the same time.
Our paper, ‘Innovative learn-by-doing digital education and attitudes towards STEM: an RCT in Italian high schools', is a so-called 'randomised controlled trial' (RCT), a methodology borrowed from medicine, where it is used to evaluate new drugs, but applied to an educational programme. Within each participating school, we randomly assigned entire classes either to be able to access the FabLab activities or to follow their school's usual alternative programme; in the Italian system this is called the 'school-to-work credit framework', or percorsi per le competenze trasversali e l’orientamento (PCTO).
We chose Italy for three reasons. First, Italy has one of the largest gender gaps in STEM university enrolment among OECD countries, so the policy stakes are real. Second, Italy already has an institutional channel (the PCTO framework) that allows an intervention like ours to be embedded into the school day without disrupting the normal curriculum. Finally, Martina Ferracane’s deep knowledge of Italian FabLabs and our expertise with Italian schools allowed us to design a more controlled study than we could have anywhere else.
Why did you and your co-authors decide to study FabLab-style "learn-by-doing" activities as opposed to a more conventional STEM lesson? What did the experience look like for the students in the treated classes, and why might that format change the way teenagers relate to science and technology?
When my co-authors and I sat down to design this study, we were reacting to a paradox. Like most industrialised countries, Italy is introducing more digital equipment to its classrooms: interactive whiteboards, tablets, even 3D printers and laser cutters in many schools. Yet, Italian teachers have some of the lowest levels of digital training, Italian curricula focus heavily on memorisation, and the share of students — especially girls — who go on to study or work in STEM fields remains way lower than in other countries.
Hardware on its own is clearly not enough: How can you generate interest in technology and STEM? That is the intuition behind the FabLab movement, and the reason we wanted to test it rigorously in a school setting. In the classes that were granted access to FabLab courses, the activities were structured around projects: Students worked in small teams over 24 hours of class time to design and build something concrete. They had to collectively move from an abstract idea to a digital design, and finally to a physical object created with digital fabrication tools.
In Verona, students built low-cost microscopes for their school using common objects (wood cuts and webcams), utilising digital fabrication for their assembly. In Agrigento, students used digital modelling techniques to design T-shirt decorations, which they brought to life using laser cutters and digital embroidery machines. In Schio, digital fabrication was aimed at building robots that could interact with the surroundings through the creation and application of sensors. In Mantua, students learned photogrammetry (a technique used to extract measurements and 3D information from photographs, to be used as input for 3D modelling and printing) to create replicas of statues from the local monumental cemetery. Finally, in Ancona, students learned 2D modelling and used CNC milling machines and laser cutting to create benches that were eventually placed in the local park.
The point is not the final objects themselves; it is the process. Students have to define the problem, plan, fail, debug, ask for help, and try again, exactly mirroring the skills we discussed earlier.
This format is very different from a standard science lesson, where students typically follow a fixed protocol with a known answer. Here, there is no single 'right' answer, and the teacher acts more as a coach than as a lecturer. For many students, this is the first time they are asked to use science and technology to express their own ideas, rather than simply being tested on someone else's.
Ultimately, the hypothesis we wanted to test is that this kind of experience would change how teenagers see themselves in relation to STEM: not just what they know, but what they believe they are capable of.
What are your main research findings? Did FabLab activities increase students' intentions to enrol in STEM majors and pursue a career in a STEM-related field?
The main result is that the intervention works, but not through a single channel, and understanding exactly how it works matters for policymakers who want to scale it up.
First, students who took part in the FabLab activities were significantly more likely to say they intended to enrol in a STEM degree at university, compared with otherwise identical students who were not allowed to access FabLab courses during our study. The effect is on the order of six percentage points, meaning a 17.2% increase.
Second, we see a positive effect on students' intentions to pursue a career in a STEM field, by 7.4 percentage points, or a 40% increase. We also see a clear boost on students' self-efficacy — namely their belief that they are able of succeeding in STEM subjects — which prior research has identified as one of the most important predictors of actually following a STEM path.
However, these two effects seem to work through different channels. Looking at which students benefit most, we find that the increase in university enrolment intentions is driven by self-efficacy: The students who gain the most confidence are precisely those who started with the least. In other words, the intervention closes a confidence gap that would otherwise keep capable students away from STEM. The effect on career intentions works differently. Here, it is less about confidence and more about exposure to a side of STEM that students rarely see in school: STEM as something creative, hands-on, and useful, rather than abstract and theoretical.
In terms of gender, however, we find that our results are mostly driven by male students, leaving the question of how to close the STEM gender gap open. Female students display a weaker response, showing a shift only when we considered secondary STEAM outcomes, which include art-oriented but STEM-intensive subjects like architecture or graphic design. In line with existing studies, we hypothesise that gender differences in STEM attitudes lie in preferences. While the limited sample at our disposal and the high segregation by gender into different high school majors — with the majority of boys in technical schools and the majority of girls in artistic licei — make these results exploratory, they nonetheless shed light on important elements for future research.
What are the implications of your research for educators and policymakers in Italy and beyond?
For educators, the practical message is that what students do in a STEM classroom and how STEM is introduced to them matter at least as much as the equipment in the room. A 3D printer that is only demonstrated by the teacher is not useful; the same machine, used by students to build their own project, is. This has direct consequences for teacher training: If the goal is to shape students' interest and confidence in STEM, countries that are investing in school digital infrastructure should put a comparable amount of resources into preparing teachers to involve students directly in hands-on projects.
For policymakers, our findings indicate that, while most interventions of this kind in Italy are typically placed within the voluntary PCTO framework (where students decide what activities to take part in), there would be benefits to making creative STEM activities compulsory. In particular, this would ensure the participation of students who have the most to gain, in terms of self-efficacy and attitudes towards STEM, but who normally tend to shy away from STEM activities.
The broader message is that interventions which target self-efficacy have the potential to shift study intentions, and that the most effective format for doing so is not a motivational talk, but a sustained experience where students get to act as designers rather than spectators. That has implications for any country worried about a future labour market in which the people who thrive will be those who can work with digital and AI tools creatively, rather than just living alongside them.
Alice Dominici is a Visiting Fellow at the EUI Department of Economics, an Assistant Professor of Economics at IMT School for Advanced Studies Lucca, and a Research Affiliate at IZA@Liser and Bocconi University (INSPIRE). Her research spans various areas of applied microeconomics and their intersections: historical political economy, health economics, and the economics of education.
Image: (Left) Two of the benches created by students in Ancona and installed in a local park; (Centre) A 3D-printed replica of a statue from the monumental cemetery produced by students in Mantua; and (Right) A low-cost microscope built by students in Verona using laser-cut wood and a webcam. Photos: FabLabs