After an increase of the number of students, French University is facing competition on a local and international level since 2000 (Benhenda Dufour 2015). In the same period, Internet and the digital choose began to change communication , learning, teaching, cooperation between teachers and the skills required to teach (Baron Bruillard 2000). In this context, various injunctions are made to teachers to improve their teaching efficiency using digital technology and be more innovative. Some recent decree (2016) also defined the role of educational engineer to help French Universities to adapt to ICTs. Thus, a research team composed of an educational engineer, a researcher in sociology of Education, a professor of Macro-Economics and a phD candidates in economy who also was a teacher assistant designed an experiment on making videos and active pedagogy (Green 2014, Fiorella Mayer 2014) based on Epstein Beauchamps (2014) hoping that, according to the SoTL framework (Brew 2011), both the students and the teachers would improve their learning/teaching. The experiment took place at bachelor level in Macro-Economy where courses are organized as an ex cathedra amphitheater in large groups of students followed by exercise classes in smaller groups. The expected outcomes of the experiment was to improve the learners’ grades at the exam, long term learning, and to increase the « 21rst century’s skils » such as creativity, critical thinking or cooperation. Also, we expected teachers and teachers’ assistants to develop their own abilities. Over the 2018-19 academic year, two groups of students taught by the same assistant were identified. One of them was “treated” whereas the other one was kept as the control group. The tested students were asked to make short videos on defined technical aspects of the program, instead of the exercises normally solved during these classes. After the experiment was over, the grades obtained at the final exam were compared. After controlling for a number of factors explaining the grades at individual level (age, gender, past results), the tested group was found to perform significantly better than the control group. Online surveys and qualitative interviews were also carried out and analyzed according to the grounded theory approach. They suggest that treated students felt they were less prepared to the final test than regular students (cf Deslauriers and al 2019). However they were happy with their experience, especially due to more social interactions within the group. Over the 2019-20 academic year, the experiment was extended to 5 tested groups and 5 control groups – hence a total of 300 students out of 700 in the amphitheater. This extension allowed to assign different teaching assistants (one for each pair of tested-control), with differing skills. The quantitative results are not yet available. The qualitative interviews suggest that teaching assistants found the format relatively demanding but more rewarding and less stressful than the standard format. Their role in class was completely transformed. This feature is also important with respect to the attractiveness of the job for young talented scholars and training of teachers. The experiment was carried out in a class of International monetary economics taught by an experimented Professor who has already developed several interaction tools in amphitheater. International monetary economics is a mandatory discipline, usually considered as demanding and technical by the students. The challenges of the experiment are quite general in terms of activation of leaning skills and deep learning. We are trying to understand how the digital tools and the active pedagogy work to improve the students’ results, to confirm to what extent we can generalize the experiment to other contexts and what did work for the teachers’assistants self developpement.