Chapter Students' feedback on the digital ecosystem: a structural topic modeling approach
- Evangelista, Adelia
- Florence Firenze University Press, Genova University Press 2023
- Physical Description:
- 1 electronic resource (6 p.)
- Additional Creators:
- Di Battista, Tonio and Sarra, Annalina
- Language Note:
- Restrictions on Access:
- Open Access Unrestricted online access
- Starting from March 2020, strict containment measures against COVID-19 forced the Italian Universities to activate remote learning and supply didactic methods online. This work is aimed at showing students' perceptions towards a learning-teaching experience practised within a digital learning ecosystem designed in the period of first emergency and then re-proposed for the blended mode. Specifically, students, attending six teaching large courses held by four professors in two different Italian universities, were asked to express their impression in a text guided by questions, requiring the reflections and clarification of their and inner deep thoughts on the ecosystem. To automate the analysis of the resulting open-ended responses and avoid a labour-intensive human coding, we focused on a machine learning approach based on structural topic modelling (STM). Alike to Latent Dirichlet Allocation model (LDA), STM is a probabilistic generative model that defines a document generated as a mixture of hidden topics. In addition, STM extends the LDA framework by allowing covariates of interest to be included in the prior distributions for open-ended-response topic proportions and topic word distributions. Based on model diagnostics and researchers' expertise, a 10-topic model is best fitted the data. Prevalent topics described by respondents include: "Physical space", "Bulding the community: use of Whatsapp", "Communication and tools", "Interaction with Teacher", "Feedback".
- Other Subject(s):
- OAPEN Library.
- Creative Commons https://creativecommons.org/licenses/by/4.0/
View MARC record | catkey: 41605011