How can we utilise all of the available data about the (digital) learning process in children in such a way as to speed up the process and make it more fun? Sioux was asked this question by a high-tech start-up in tablet education. Pupils do arithmetic and language exercises on tablets running the client’s apps. This way, the start-up is able to generate a huge amount of data about their learning performance on its servers: a treasure trove of information that the client wants to mobilise optimally to help children and their teachers in the learning process.
Using psychometric models, Sioux made it possible for the level of all available exercises to be automatically calibrated. With our statistical expertise, Sioux was able to quickly start adapting these psychometric models to the client’s needs. Thanks to Mathware, the exercises completed in class can now be used to objectively test a pupil’s knowledge and skills.
The start-up is now able to effectively challenge pupils with exercises that correspond to their levels. Moreover, teachers are given specific feedback on the knowledge and skills of both the group as a whole and the individual pupils.
Algorithm for assessing student’s abilities based on exercises, unknown difficulties of exercises, estimated abilities of students, psychometrics: item response theory, calibration/validation, big data and adaptive learning. Mathware technologies: R.
