[Paper Review] Physical extracurricular activities in educational child-robot interaction

Daniel Davison, Louisa Schindler+1 more|University of Twente Research Information|Jun 8, 2016

Teaching and Learning Programming8 references3 citations

TL;DR

This study investigates whether physical extracurricular activities between children and robots enhance the perceived child-robot relationship in an educational context. Using semi-structured interviews and a cumulative relationship score, it finds no significant individual measure differences but reveals a statistically higher overall relationship score in the intervention group, suggesting shared physical activities may enrich child-robot rapport.

ABSTRACT

In an exploratory study on educational child-robot interaction we investigate the effect of alternating a learning activity with an additional shared activity. Our aim is to enhance and enrich the relationship between child and robot by introducing "physical extracurricular activities". This enriched relationship might ultimately influence the way the child and robot interact with the learning material. We use qualitative measurement techniques to evaluate the effect of the additional activity on the child-robot relationship. We also explore how these metrics can be integrated in a highly exploratory cumulative score for the relationship between child and robot. This cumulative score suggests a difference in the overall child-robot relationship between children who engage in a physical extracurricular activity with the robot, and children who only engage in the learning activity with the robot.

Motivation & Objective

• To explore whether introducing shared physical extracurricular activities between children and robots strengthens the perceived child-robot relationship during educational interactions.
• To investigate how such activities influence the quality of dyadic interactions in a triadic learning context involving child, robot, and learning materials.
• To develop and evaluate a cumulative relationship score that aggregates multiple qualitative metrics into a single indicator of child-robot rapport.
• To assess whether physical and non-academic interactions with a robot lead to more positive social perceptions and relational bonding compared to learning-only interactions.

Proposed method

• Conducted an exploratory study with primary school children interacting with a social robot in a structured inquiry learning task.
• Introduced a physical extracurricular activity (e.g., balancing, arm movements, cognitive challenges) as a break between learning tasks in the intervention group.
• Used semi-structured interviews with three components: pictorial task (emotion card selection), social distance task, and sociometric questions.
• Developed a cumulative relationship score by intuitively weighting multiple qualitative metrics to estimate overall child-robot relationship quality.
• Compared the intervention group (with physical activity) to a control group (only learning tasks) using qualitative and exploratory quantitative analysis.
• Applied Vygotskian theory on social scaffolding to frame the potential role of relational development in learning outcomes.

Experimental results

Research questions

• RQ1 Does engaging in a shared physical extracurricular activity with a robot improve the child’s perceived relationship with the robot compared to a learning-only interaction?
• RQ2 How do different qualitative metrics (e.g., emotion cards, social distance, sociometric choices) reflect changes in child-robot relationship quality?
• RQ3 Can a cumulative relationship score effectively capture and differentiate child-robot relationship quality across experimental conditions?
• RQ4 To what extent do physical and extracurricular aspects of the activity contribute to relational enrichment, independently or jointly?

Key findings

• The pictorial task, where children selected emotion cards to describe their relationship with the robot, showed a trend toward more positive emotions in the intervention group, suggesting improved rapport.
• Most individual measurement tools (e.g., social distance, sociometric choices) yielded inconclusive results with no statistically significant differences between conditions.
• The cumulative relationship score, though based on intuitively weighted metrics, showed a statistically significant difference, with higher scores in the intervention group compared to the control group.
• Children in the control group frequently described their experience as 'boring,' indicating potential engagement or enjoyment issues absent in the intervention group.
• The study highlights methodological challenges in measuring child-robot relationships, particularly with short interactions and subjective qualitative tools.
• Despite inconclusive individual metrics, the cumulative analysis provides promising evidence that shared physical activities may enrich child-robot relationships in educational settings.