"If you are a Star Wars fan, use the force": Exploring Children's Virtual-Object Interaction Preferences in AR Headsets

Paper Title

"If you are a Star Wars fan, use the force': Exploring Children's Virtual-Object Interaction Preferences in AR Headsets"

Publication Info

• Topic area: Children's interaction preferences with virtual objects in augmented reality (AR) headsets.
• Keywords: Augmented reality, AR headsets, children, gesture elicitation, virtual objects, interaction design, conceptual model, user expectations, storytelling, direct manipulation.

Background and Problem

• Problem / challenge: Limited understanding of how children expect to interact with virtual objects in AR headsets, despite their increasing use in educational and therapeutic contexts. Existing research focuses primarily on adults or lacks direct interaction with AR devices.
• Significance: Understanding children’s interaction preferences can inform the design of intuitive and engaging AR applications tailored to children, improving usability and satisfaction.
• Motivation and related work: Prior studies have explored adult interaction preferences in AR and children’s interactions with other technologies (e.g., touchscreens, spherical displays). However, gaps remain in understanding children’s natural expectations for AR headset interactions, particularly through direct elicitation studies.

Solution

• Proposed approach: Conducted an open-ended gesture elicitation study with 20 children (ages 9-12) using an AR headset to explore their interaction preferences with a virtual cube performing 17 referents (e.g., Move, Rotate, Destroy).
• Novelty:
  1. First study to explore children’s virtual-object interaction expectations using gesture elicitation with an AR headset.
  2. Developed a conceptual model outlining children’s perceptions, expectations, and reasoning behind their interactions.
  3. Provided design recommendations for child-centered AR headset applications.
• Procedure and key techniques:
  • Children observed a virtual cube perform 17 referents and proposed interactions to replicate them.
  • Data collected included video/audio recordings, point-of-view footage, and children’s verbal explanations.
  • Analyzed data using qualitative coding, affinity diagramming, and quantitative metrics (e.g., Agreement Rate).

Results

• Concrete findings:
  • Gestures were the dominant interaction modality (76.5%), with limited use of speech (5.6%).
  • Children frequently used whole-body movements (e.g., kicking, running) and external objects (e.g., swords, magnifying glasses) in their proposals.
  • Interaction proposals were diverse, resulting in low consensus (average first-choice Agreement Rate = 0.182).
  • Children created narratives and contextualized their interactions, often attributing real-world properties (e.g., weight, fragility) to the virtual cube.
• Advantage over baselines: Provided empirical insights into children’s unique interaction preferences, which differ significantly from adults (e.g., greater use of whole-body gestures, external objects, and storytelling).
• Experiments / evaluation:
  • Study involved 20 children (ages 9-12) with prior VR/AR experience.
  • Used a virtual cube and 17 referents grouped into Translation, Manipulation, and Abstract categories.
  • Evaluated interactions based on Agreement Rate, qualitative themes, and self-reported ratings for ease of use and goodness of fit.
• Limitations and future work:
  • Small sample size (20 children) and limited geographic diversity.
  • Focused on a single virtual object (cube) and basic referents; future work should explore more complex contexts and objects.
  • Did not evaluate usability or cognitive workload of proposed interactions.
  • Future research should examine cultural differences, display factors (e.g., field of view), and collaborative contexts.

Summary

This study explored how children (ages 9-12) expect to interact with virtual objects in AR headsets through a gesture elicitation study. Findings revealed a strong preference for direct manipulation gestures, whole-body movements, and the use of external objects, with limited reliance on speech. Children often contextualized their interactions through narratives and attributed real-world properties to virtual objects. The study developed a conceptual model of children’s interaction expectations and provided design recommendations for child-centered AR applications, such as supporting whole-body interactions, integrating external objects, and bridging virtual and real-world interactions. These insights can guide the development of more intuitive and engaging AR experiences for children.