Related Experiment Video
Updated: Jun 13, 2026
06:02
Evaluating Usability Aspects of a Mixed Reality Solution for Immersive Analytics in Industry 4.0 Scenarios
Published on: October 6, 2020
The Role of Rating Valence in AI Skin Cancer App Acceptance: Eye-Tracking and Questionnaire Study
Inga Jagemann1, Sabrina Hegner2, Gerrit Hirschfeld1
1School of Business, University of Applied Sciences and Arts Bielefeld, Interaktion 1, Bielefeld, 33619, Germany, 49 521106 70508.
JMIR Human Factors
|June 11, 2026
Summary
User attention for artificial intelligence-based skin cancer screening apps (AISCSAs) focused more on descriptions and reviews than ratings. Acceptance of AISCSAs was predicted by perceived ease of use, usefulness, and trust, not rating valence.
Area of Science:
- Health Informatics
- Human-Computer Interaction
- Medical Imaging
Background:
- Artificial intelligence-based skin cancer screening apps (AISCSAs) show diagnostic promise but have low user adoption.
- Understanding how users process app store information, like ratings, is crucial for AISCSA acceptance in high-stakes health contexts.
- Eye-tracking was employed to investigate visual attention towards AISCSA app store listings.
Purpose of the Study:
- To determine if a single negative rating captures user visual attention.
- To assess if an extended Technology Acceptance Model (TAM) can predict behavioral intention (BI) to use AISCSAs.
Main Methods:
- 76 participants evaluated a mock AISCSA app store listing under positive or negative rating conditions, with eye movements recorded.
- Fixation durations on defined areas of interest (AOIs) were analyzed alongside self-reported measures of perceived usefulness (PU), perceived ease of use (PEOU), trust, BI, and willingness to pay.
- Data combined eye-tracking metrics with survey responses on app attribute importance.
Main Results:
- Users paid the most attention to the app description, followed by reviews, and then ratings.
- No significant effect of rating valence on gaze patterns, PU, PEOU, trust, or BI was found.
- PEOU, PU, and trust were significant predictors of BI.
Conclusions:
- Negative ratings did not capture significant visual attention, contrary to expectations.
- Eye-tracking revealed attentional processes not captured by self-report, suggesting it approximates user behavior more directly.
- In high-stakes health scenarios, textual information like reviews may be more influential than rating valence for app adoption.
