By Vanessa Serrano, Jordi Cuadros, Francesc Martori, Antoni Miñarro, Miquel Calvo, Pablo Díez, Cristina Montañola, Victor León, IQS Universitat Ramon Llull
It is commonly understood that students need to solve problems on their own to learn statistics. That's why we are working on a model that eases the assessment of the students’ work. This model starts when the students are asked to solve an activity using a modified version of R Commander with tracing capabilities. Their actions are traced, and can be visualized and analyzed using a Shiny dashboard. Their work can then be assessed by matching it to pre-identified observation items, i.e. important achievements or potential errors in the activity resolution. Feedback is obtained by checking whether these observation items were reached at any moment. New visualizations are now added to analyze the sequence and the repetition of the observation items, to better capture what the students may be understanding and where the main difficulties can be found. In this poster, the results of this approach are presented by analyzing an activity that took place in a Statistics course at the university level with 20 students involved.