Interactive Feedback Usability Test findings

Usability considerations for AR interactions

Content added by Vishaka for Spring 2025 Project 1, on 2/23/2025

Objective

With my in-class activity, I wanted to evaluate how individual interactive feedback vs physical setting feedback patterns influence user perception of AR visualizations. 

Research goals

• Understand user engagement with models that have different interaction patterns

• Learn about what types of feedback help/detract with physical immersiveness in AR

Test environment

• HMDs

• Links to software environments to test out (Bezi)

• Split participants into 2 groups 

  • Group A: Load model that has interactive feedback (audio/visual/haptic)

  • Group B: Load model that has onboarding information about surroundings (visual/physical)

Planned tasks

• Find a way to select one of the three items in this experience. 

• Adjust your view so the object is facing away from you (180 degrees)

• Find a way to learn more about the item in front of you

• Find a way to deselect the item, and select a new one. 

Planned feedback form

• Goal: Understand user engagement with models that have different interaction patterns

  • Which nature model did you initially select? [Single select]

    • Did you explore any others? [Open-ended]

  • Overall, how would you rate your familiarity with each nature model after this activity?  [Likert low familiarity-high familiarity]

  • How would you rate your success with each task? [Likert low performance-high performance]

• Goal: Learn about what types of feedback help/detract with physical immersiveness in AR

  • How would you rate your mental load with each task? [Likert low load-high load]

  • Were any of the tasks particularly complex? Or simple? [Open-ended]

  • How would you rate your physical activity load with each task? [Likert low load-high load]

  • Were any of the tasks particularly demanding? Or easily achievable? [Open-ended]

Results

This slide deck goes over my main findings, but they are also summarized below. 

Spatial prototype

• More users initially chose to interact with the model in the middle

• Not all users got to interact with all the models in this experience, which could be attributed to being more immersed in one specific model

• There were mixed success rates with each task, with it being harder to touch the model for more information and to go to a different model, since the prototype took up a large space.

• There was still low-medium physical activity, but definitely more than the UI prototype.

UI prototype

• Most users initially interacted with the left most model, aligning with common UI patterns, where on screen interfaces, the top left corner is most important to users.

• All users got through viewing all models in this experience, which could be attributed to the low physical load for this prototype

• Users had higher averages overall for their familiarity ratings with each model

• There were mixed success rates on certain tasks, like getting more information (since this was a click+hold task)

• Overall, low mental and physical load

Overall feedback

• The animations of models growing as you approach them felt jarring

• There needed to be affordances for hand/body colliders

• There were bugs with the audio

• Positive feedback for both experiences!

• Spatial prototype:

  •     The audio was compelling

  •     Felt more interactive

  •     Not enough space for the prototype

  •     Was straightforward to interact with

  •     There was higher physical exertion

  •     Instructions were hard to read without a background

  •     Hard to anticipate audio length

  •     Wanted to adjust scale

• UI prototype

  •     Models felt far for viewing

  •     Still some issues with reading text

  •     Refined look of UI made it easy to navigate

  •     Its easy to forget controls

  •     Click + hold felt unintuitive

  •     Easy to switch between models

  •     Wanted to look at models closer

Overall insights

• Ease of use

  • Users are familiar with traditional screen-based interfaces, and find these to have low mental load involved

• Immersiveness

  • There's a sense of immersion by prompting users to move around within the AR environment. 

  • This comes at a cost of higher physical load, but can be worth it for the intended experience. There's a fine line between an acceptable physical load and having the user be fatigued or tired based on physical load. For example, most of the moving around was acceptable within this activity, but the content where users needed to hold out their arm to listen to an audio clip felt too long and fatiguing.