EYE-TRACKING

Method:
To perform the eye tracking testing, we tracked the eye movements of 7 participants, each viewing two different webpages from 5 different websites and 4 abstract images, all for 12 seconds each. The eye tracker was placed on the tripod, approximately 10 cm above the table surface, and the laptop placed on a cardboard box about 10 cm tall. This provided the most successful calibrations. The abstract images were white fields with numbers one through four, used as a secondary check for calibration.

For each participant, the eye tracker was first calibrated. Then, the participant was shown the slide show of all 14 images (10 web pages and 4 abstracts). During the test, we realized even minor distractions would cause participants to adjust their gaze to the source of noise or movement. Therefore, we performed the test again, making sure the testing area was distraction free.

Results:

Despite having relatively controlled conditions, our test data did not indicate any strong trends between participants and only in some cases seemed to correlate closely to the webpage they were viewing’s design. The most conclusive findings seemed to suggest that larger text sizes and brighter colors were more eye-catching (Fig. 3 and 4), although again, this was exhibited in certain tests but by no means was consistent. Pop-up windows were also an eye-catcher (Fig. 5).

Discussion:

Because the results were not very consistent it is difficult to draw firm conclusions from the eye tracking. As mentioned above, the clearest insights it provided were that bright colors and large text appeared to be more eye-catching.

A source of error within the test could also be the fact that there were not specific instructions or tasks for participants to follow. This was done with the intent of not influencing the data, however the result appears that every participant scanned the page in a somewhat random fashion (Fig. 6).

Using the eye tracking software in combination with a static visual search test could possibly provide more useful data, or better yet, in combination with a series of tasks navigating through the site.

An additional opportunity to perhaps find more meaning within the data is to perform the same test on re-designed webpages to see if the re-design induces more correlation. This could indicate that the seemingly random eye patterns were in fact an accurate measurement of the original website’s lack of visual hierarchy.