Predicting students’ flow experience through behavior data in gamified educational systems

Abstract

The flow experience (i.e., challenge-skill balance, action-awareness merging, clear goals, unambiguous feedback, concentration, sense of control, loss of self-consciousness, transformation of time, and autotelic experience) is an experience highly related to the learning experience. One of the current challenges is to identify whether students are managing to achieve this experience in educational systems. The methods currently used to identify students’ flow experience are based on self-reports or equipment (e.g., eye trackers or electroencephalograms). The main problem with these methods is the high cost of the equipment and the impossibility of applying them massively. To address this challenge, we used behavior data logs produced by students during the use of a gamified educational system to predict the students’ flow experience. Through a data-driven study (N = 23) using structural equation modeling, we identified possibilities to predict the students’ flow experience through the speed of students’ actions. With this initial study, we advance the literature, especially contributing to the field of student experience analysis, by bringing insights showing how to step towards automatic students’ flow experience identification in gamified educational systems.

Identifying Engagement in Children’s Interaction whilst Composing Digital Music at Home

“…There are many HCI studies on creativity and engagement beyond music which aim to identify patterns in people's interactions. Some rely on tasks with a clear completion goal [9,23,49,[53][54][55]72] for example, Pastushenko and colleagues [53] speculate on how a gamified system could detect flow states, later presenting a pilot investigation for predicting attributes of flow [49]. Other investigations have identified patterns of interaction in open-ended interfaces [35,62,65,67] similar to the system used in our case study (see Section 3.1).…”

“…Our method incorporates VCR (see Section 3.4.2) to directly understand how children perceive their own interactions, which might help researchers understand how they relate to points of engagement. Furthermore, a number of these works [49,53,65] focus on identifying patterns to train classifiers which predict engagement based on entire sessions of interaction -we strive to identify patterns of interaction which could potentially be used to detect points of engagement whilst children are composing.…”