Scan path similarity as a function of performance accuracy in multiple object tracking.

Tracking multiple moving objects simultaneously is a key task in studying how human observers dynamically allocate attentional resources over time. We investigated factors that can affect scan path similarity during multiple object tracking, particularly whether scan paths differ in trials leading to accurate vs. inaccurate tracking performance. The examined factors were trial-wise performance accuracy, target set-size, type of occlusion, target dispersion, and target speed. The results showed that scan path dissimilarity increases through time for incorrect trials as errors are likely to accumulate. However, for correct trials dissimilarity stays constant over the entire trial time. Scan paths resulting in a successful performance were more similar in those resulting in error for set-size 4 than set-size 2. Additionally, scan paths were more similar when the targets were temporarily occluded compared to scenarios where they moved non-occluded. Scan paths became less similar as targets scattered across a wider area and move more slowly. Overall, the analysis of scan paths is a valuable source of information for studying strategies adopted by observers during accurate and inaccurate tracking of moving objects.