Reducing Search Space on Demand Helps Older Adults Find Mobile UI Features Quickly, on Par With Younger Adults

Title of the Paper

Reducing the Search Space on Demand Helps Older Adults Find Mobile UI Features Quickly, on Par with Younger Adults

Paper Information

• Field of Study: Human-Computer Interaction (HCI), specifically optimizing mobile user interface navigation for older adults
• Keywords: Older adults, mobile interface, usability, accessibility design, visual feedback, interaction technology, technical support

Research Background and Problem

• Problem and Challenges: As mobile applications become increasingly feature-rich, navigation has grown more complex, and older adults often exhibit lower efficiency and accuracy in locating specific features compared to younger users. This disparity stems from older adults relying more on simplified information processing strategies due to reduced cognitive resources, while complex and feature-dense user interfaces make it harder for them to quickly find desired functionalities.
• Significance: In recent years, many online services have become accessible exclusively through smartphones or tablets, posing significant challenges for older adults. As digitalization accelerates, technological barriers may prevent older users from fully utilizing these services, potentially leading to their abandonment of technology.
• Motivation and Related Work: Although prior research has explored tools like heatmaps and voice assistants to assist older adults in navigating mobile interfaces, there is a lack of systematic studies on improving efficiency and accuracy by reducing the search space (i.e., the number of selectable features).

Solution

• Proposed Method: The authors introduce the "Nav Nudge" technology, an interactive tool that uses voice input and large language models (LLMs) to predict user target features and automatically reduce the interface search space to three relevant options.
  • Natural language processing techniques extract keywords from user queries.
  • Keywords are extracted from interface elements on the current page.
  • Semantic matching between the two sets of keywords identifies the three features most relevant to the user’s goal.
  • Visual feedback (e.g., highlighting, localized zooming) helps users identify target options.
• Innovations:
  1. Automatically generates and emphasizes the search space on the current interface without relying on remote assistants or manual operations.
  2. Provides an optimal search space size (three features), balancing cognitive load and prediction accuracy.
  3. Demonstrates and compares different visual feedback strategies tailored for older adults.
• Implementation Steps:
  1. Design a keyword extraction algorithm (using the APE model and OpenAI’s GPT-3.5 model).
  2. Develop methods for extracting and matching keywords from interface elements.
  3. Determine optimal visual feedback strategies (e.g., highlighting and zooming) through experiments.
  4. Integrate voice input and server-side processing to build the "Nav Nudge" application and embed it into a mobile map application.

Research Results

• Specific Findings:
  1. Experiments showed that reducing the feature search space to three significantly improved older adults' efficiency and accuracy in locating target features.
  2. After using the "Nav Nudge" technology, older adults performed almost on par with younger users.
  3. Highlight with Context (HC) and Weighted Zoom (WZ) were the most preferred visual feedback methods among older users.
  4. In real-world scenarios (map applications), 75% of older adults successfully recovered lost navigation paths.
• Advantages Over Existing Solutions:
  • Automated and precise visual enhancement methods without requiring additional manual support or offline tutorials.
  • Reduced feature search space combined with visual assistance makes self-exploration of technology more feasible and intuitive for older users.
• Experimental or Evaluation Results:
  • In experiments, older adults using the reduced search space completed tasks 28% faster on average, with a 50% reduction in error rates.
  • Users showed a clear preference for the optimal visual feedback strategies (HC and WZ).
• Limitations and Future Directions:
  • Limitations:
    • Small sample size (n=60), only validating effects for older age groups.
    • Did not investigate whether small search spaces (three or fewer items) are equally effective across interfaces of varying complexity.
    • Tested only visual methods, excluding multimodal feedback like auditory or tactile cues.
  • Future Directions:
    1. Explore additional input methods (e.g., text input or shortcut gestures) to accommodate diverse user needs.
    2. Extend to multi-step interaction scenarios, providing cross-application navigation support.
    3. Develop output methods suitable for visually impaired users (e.g., audio prompts).
    4. Investigate the impact of reduced search spaces on desktop interfaces and their potential benefits for younger users occasionally interacting with complex applications.