Using Rasch Models for Developing Fast Technology Enhanced Learning Solutions: An Example with Emojis

We focus on issues related to learning analytics for predicting behavior, present-ing a case in which people’s answers can be predicted on the bases of a known numeric function describing a cognitive relation. We used the Rasch scaling method following an item-response latent-trait model on questions on the valence of emojis of common use. Emojis are pictorial symbols with a high degree of humanization, nowadays increasingly frequent in computer-mediated communica-tion (CMC). However, sometimes their meaning is ambiguous. Therefore, after quantifying how much of a property such as positive and negative valence differ-ent emojis represented, we identified the function that relates the two different judgements on the same object, specifically the positive and negative valence of each emoji. Applications of this approach can be highly relevant for learning ana-lytics, to optimize the measurement, collection, and analysis of data about learners and their contexts for fast Technology Enhanced Learning (TEL) solutions.