Classification of young readers within an immersive learning model

Using data obtained upon reading aloud of a narrative text in the Basque language, 151 university students were categorised into different reading profiles, and the link between these profiles and a number of sociolinguistic variables was assessed. The study used K-means clustering based on six prosodic variables (total reading time, total articulation time, total pause time, speech rate, articulation rate, and time per pause), which yielded a stable classification into five groups. The quantitative characterisation of the profiles ranged from fast readers with short pauses to slow readers with longer, more frequent pauses. Contingency table analysis revealed no significant links between the reading profiles and the categorical variables analysed (first language, Basque language acquisition setting, and gender). Moreover, the homogeneity observed in reading style—regardless of linguistic background—highlights the effectiveness of immersive learning. The methodology used here can be applied to other types of texts, in other languages, with readers at any stage of learning. It also allows for contrasting results between educational models.