Presentation
81. Predicting Scam Detection Performance on Social Media using Eye Tracking
SessionPoster Session 1
DescriptionWhile cybercrime in the past was relegated primarily to email or fake websites, phishing attacks on social media present a new avenue of potential risk for the ill-prepared user (Parker & Flowerday, 2020). To understand how individuals assess information online, researchers have used eye tracking for detailed analysis of an individual’s gaze patterns (reviewed in Hussein, 2023). Our study had 22 participants evaluate 101 tweets and provide responses (scam, not scam), while measuring their gaze with an eye tracker. Our analysis found that total fixation duration and regressive saccades were highly negatively correlated with scam detection performance on scam trials, r(44) = -0.766, p < .001; r(35) = -0.80, p < .001, respectively. Moreover, participants were especially sensitive to tweets that contained financial information, d’ = 1.256. Our results provide evidence that scam detection is an early perceptual process, using efficient but faulty heuristics that degrades with gaze.
Event Type
Poster
TimeWednesday, September 11th5:30pm - 6:30pm MST
LocationMcArthur Ballroom
Aerospace Systems
Cognitive Engineering & Decision Making
Computer Systems
Forensics Professional
Health Care
Human Performance Modeling
Individual Differences in Performance
Perception and Performance
Product Design
Safety
Training
Usability and System Evaluation
Extended Reality