UROS 2017 Project: Children’s Pedestrian Skills – Eye-Tracking studies employing Change-Blindness Stimuli

By Hannah Willliams



Children are vastly overrepresented as a group in pedestrian accidents. Previous research has indicted that children often demonstrate a lack of appropriate visual search strategies, often not having the ability to identify important aspects of a scene. Furthermore, this search strategy is thought to evolve into a more appropriate technique around the age of 7-8 years (Whitebread & Neilson, 2000). This project consisted of two experiments both relating to the study of how children understand traffic situations in, for example, crossing the road.

The aim of this study was to form a stronger knowledge of the developmental process surrounding crossing the road. It was predicted that the older children were the more likely to produce the appropriate responses. Further predictions included that younger participants would have more fixations prior to making their judgements, longer times to first fixation in the salient features requires for this judgement, and longer fixation durations.

Participants were between the ages of 5-10 years, they were all participants in the University of Lincoln’s summer scientist week. This project consisted of two experiments, both using a Tobii eye tracker to record eye movements. In both studies children were presented with a series of images depicting road crossing scenes. All participants took part in both studies.

Change Blindness test

The first experiment displayed a series of flicker images required them to click a mouse when they noticed a change in the scene. The procedure followed the same timings as Rensink (2005) and images were shown for 30 seconds. 0.5 sec for the image and 0.25 sec for the blank screen. The whole cycle lasted for 1.5secs.

Is it safe to cross?

The second experiment challenged children to make a decision as to whether a presented scene was safe to cross the road or not. As well as the eye tracking data, their answering and the reasoning for their answer was recorded.


Preliminary findings reported no age or gender differences. However, a statistically significant age difference F (3, 48) = 3.487, p = .023, was observed. With post hoc analysis showing the difference to be between the younger age group (5 – 6.9 years) and the older age group (8 – 10 years) when it came to correctly answering the change blindness test question. Further analysis showed a higher percentage of older children to notice changes in relevant aspects of road crossing scenes, F (1, 57) = 246.66, p < 0.001.

Further analysis was conducted after the extraction of eye-tracking data, such as time to first fixation, fixation duration and fixation count. This analysis revealed no significant differences with respect to age.


Although the outcome of the analysis did not support all of our predictions there is a possible explanation for this. After analysing the quality of the data and removing any unsatisfactory recordings, the number of participants in age groups used as uneven. With some groups having a smaller pool of data it means that they are not as well represented in the analysis.


*To view Hannah’s project poster, please click on the thumbnail below