The manual correction of assignments is a long and error-prone task, particularly pressing when such an assessment involves a large cohort of students. The paper summarises five years of design, development and deployment of a formative and summative assessment tool for the automated grading of data science exercises, based on R commands and comments written in natural language. The paper investigates the usability, engagement and grades of a large cohort of students. The results are positive, showing that the work finalised to improve the tool increased the user experience and engagement. Moreover, also the grades of students that used the tool were on average higher than the others, thus suggesting a positive effect of the rDSA tool also in terms of final learning outcomes.
File in questo prodotto:
Non ci sono file associati a questo prodotto.