Children’s facial expressions during collaborative coding: Objective versus subjective performances

Educational research has used the information extracted from facial expressions to explain learning performance in various educational settings like collaborative learning. Leveraging this, we extracted the emotions based upon two different theoretical frameworks from videos with children aged 13–16 while collaborating to create games using Scratch. The two sets of emotions are based on the control value theory (happiness, sadness, anger, surprise) and the education-specific expressions (frustration, boredom, confusion, delight). We computed the groups’ objective performance, which was calculated based on their created artifacts. We divided them into high and low performance and compared them based on individual emotions’ duration and the transitions among the emotions. We also used the subjective indication of their perceived performance from a self-reported questionnaire, divided them into another performance category, and did a similar analysis with the objective performance. Results show that the objective performance is better explained by the education-specific emotions and the negative valance emotions from the control value theory-based emotions. On the other hand, subjective performance is better explained by the control value theory based on emotions. Based on the results, we suggest implications both for the instructors and students.