Using facial emotion recognition to research emotional phases in an inquiry-based science activity

Angel Ezquerra

Federico Agen

Radu Bogdan Toma

Ivan Ezquerra-Romano


Observational and declarative approaches applied by human observers have dominated the study of emotions, engagement, and attention. However, these approaches are susceptible to biases such as social desirability. Additionally, declarative approach cannot measure these parameters in real-time.


This study aims to apply Facial Emotion Recognition (FER) technology in an inquiry-based activity to address the limitations of traditional observational and declarative approaches. We measured students’ emotions, engagement, and attention through the parametric measurement of facial expressions frame by frame.

Method and procedure

We recruited a convenience sample of 12 secondary education pre-service teachers. Students participated in a 20-minute inquiry-based activity where they were tasked with identifying the contents of a sealed black box. We recorded their facial expressions and actions with a video camera, and the collected data was subsequently analysed by both a FER automatic software and a human observer.


During the activity, Attention had a mean presence of 81.24% ±5.93, Engagement had a presence of 31.86% ±9.10, Joy of 7.62% ±7.11, and Surprise of 6.57% ±4.61. The presence of the other discrete emotions was low. Our analysis further revealed an evolution of these parameters and students’ actions, progressing through three distinct phases: accommodation to the problem (Phase 1), exploration of problem-solving strategies (Phase 2), and implementation of problem-solving strategies (Phase 3). Based on these observations, we grouped students into three patterns based on the phases they underwent throughout the activity.


This study shows the application of FER technology for identifying and analysing the progression of students’ emotions, engagement, and attention during an inquiry-based activity. FER technology allows real-time and systematic measurement of emotions, engagement, and attention. However, this technology has inherent limitations, and further research is needed to optimise its application in the educational setting.

This publication uses Facial Expression Analysis which is fully integrated into iMotions Lab

Learn more