Multiple Device Users' Actual and Ideal Cross-Device Usage for Multi-Stage Notification-Interactions: An ESM Study Addressing the Usage Gap and Impacts of Device Context

Title of the Paper

Multiple Device Users’ Actual and Ideal Cross-Device Usage for Multi-Stage Notification-Interactions: An ESM Study Addressing the Usage Gap and Impacts of Device Context

Paper Information

• Research Area: Human-Computer Interaction, Mobile Computing, Multi-Device Interaction Systems
• Keywords: Notification management, multi-device usage, experience sampling method, device context, notification interaction stages, usage gap, mobile computing

Research Background and Problem

• Identified Issues or Challenges:

  • There may be a gap between multi-device users’ actual device usage and their ideal usage.
  • The influence of different devices and their contexts on the stages of notification interaction remains unclear.
  • How users choose devices for notification interaction behaviors, and how these behaviors are affected by device types and user preferences, requires further investigation.

• Significance:

  • As multi-device ecosystems grow, understanding multi-device user behavior is crucial for optimizing notification system design.
  • Bridging the gap between actual and ideal device usage can enhance user experience and improve multi-device integration efficiency.

• Research Motivation and Related Work:

  • Existing studies primarily focus on notification management for single devices, lacking comprehensive analysis of cross-device notification interaction stages.
  • The impact of device type, notification importance, user device preferences, and device context on notification interaction has not been systematically presented.

Solution

• Proposed Method or Solution:

  • The authors employed the Experience Sampling Method (ESM) to analyze user behaviors and preferences through surveys, identifying the gap between actual and ideal device usage for notification interactions.
  • A research application was designed and used to dynamically sample notifications and record users’ actual and ideal device choices.

• Innovations:

  • Developed a four-stage notification interaction model (Notice, Glance, Read, Act) and analyzed actual device usage and ideal preferences across these stages.
  • Systematically linked notification interaction stages with device context, addressing gaps in multi-device user behavior research.

• Key Implementation Steps and Techniques:

  • Designed and distributed an Android application to collect users’ notification experiences.
  • Recruited 31 multi-device users and collected survey data over a 14-day period.
  • Applied mixed-effects logistic regression and association graph analysis to explore differences between actual and ideal device usage.

Research Findings

• Specific Results:

  • In nearly half of the cases, users’ ideal device choices did not align with their actual choices, primarily due to practical constraints such as device context (visibility, tactile accessibility, availability).
  • Smartphones emerged as the preferred device across all notification interaction stages, though users favored non-smartphone devices such as computers and wearables for specific stages.
  • Users exhibited a strong tendency to use the same device continuously across multiple notification interaction stages.

• Advantages Over Existing Solutions:

  • Incorporated device context into the analysis, providing detailed insights into optimizing systems to support multi-device notification interactions.
  • Proposed design recommendations for multi-device notification ecosystems to reduce reliance on single devices and enhance user experience.

• Experimental or Evaluation Results:

  • Smartphones had the highest actual usage rates (80.1% in the Notice stage, 76.0% in the Act stage), while wearables and computers were used predominantly in specific interaction stages.
  • Screen visibility and availability significantly influenced platform choice, especially in higher attention stages (Read and Act).

• Limitations and Future Directions:

  • The sample size was relatively small and limited to users in Taiwan, potentially restricting global applicability.
  • Did not analyze additional parameters of notification perception (e.g., notification type, modality, or urgency).
  • Suggested future research to expand to other device types and explore broader aspects of user behavior.

Macro-Level Recommendations

• Design Implications:
  • Recommend building multi-device notification ecosystems that support cross-device notification synchronization.
  • Provide flexibility for users to choose devices to complete different notification interaction stages.
  • Systems should learn user preferences and device contexts to optimize notification delivery strategies.

Through this framework, the authors clarified new possibilities for designing multi-device ecosystems and optimizing user behavior, offering significant guidance for the field of cross-device notification management.