Quantifying User Ability in Visual Research: A Bayesian Analysis from a Gamified Task

DINO is a video game, produced by the Turin start-up Paperbox, aimed at evaluating the performance of young children in prerequisite tasks typical of primary school. One of these tasks is Visual Research Ability, which is therefore tested in a gamified way by the Letters Puzzle module of DINO and is the primary topic of this thesis. The results of Letters Puzzle are aimed at predicting whether a child will develop symptoms related to a learning disability, i.e. dyslexia. Statistical tools in psychometry have a long history, starting from the Rasch model in Item Response Theory. These tools find a new application in the diagnostic setup of learning disabilities. First, a general description of DINO and its goals are presented, then the details of Letters Puzzle are investigated and inserted in a statistical and psychometric modeling framework.