Why Eye-Tracking?

Eye-tracking is a powerful tool for evaluating the user experience, whether the user experience involves answering a survey, navigating a website, or driving a car! Eye-tracking provides objective physiological data to help you develop materials and systems that are optimized for the task at hand. By monitoring eye movements, researchers can objectively see how study participants interact with a stimulus.

This can be helpful for answering questions like:

Do users notice the Contact Us feature on my survey?

Where do users look first when they view my website? Where do they look next?

Is design A processed differently than design B?

Westat's Approach

Data on the user experience often comes from self-reported feedback from study participants. Eye-tracking allows researchers to observe for themselves what participants pay attention to.

Eye-tracking can reveal where people look, but not what they perceive. Therefore, Westat’s approach combines quantitative eye-tracking data and qualitative feedback to provide a complete picture of how people interact with study materials.

Collecting eye-tracking data is not as difficult as it might seem!

With a little bit of training and practice, eye-tracking hardware and software can be set up easily.

There are two common types of eye-trackers – glasses-based equipment and remote equipment.

Eye-Tracking glasses

Worn by the users.
Good for eye-tracking in real world environments where the user may be mobile, or looking at a stimulus that is not on a screen like a paper document.

Remote eye-trackers

Rest in front of the user and track their eye-movemnent from a distance.
User does not have to wear anything but they have to remain largely stationary.
Optimized for eye-tracking on screens.

Westat currently owns three sets of eye-tracking equipment – two types of glasses and one remote tracker.

Tobii Pro Glasses 2

View Product Site

Tobii Pro X2-30 Eye Tracker

View Product Site

Dikablis Glasses 3 + Vehicle Testing Kit (VTK)

View Product Site

Analyzing eye-tracking data

Eye-tracking equipment typically captures and analyzes two types of eye-movements: fixations and saccades.

Most eye-tracking data analysis focuses on fixations, which are instances of the eye dwelling in one location for a set amount of time (usually a minimum of 1/10 of one second). Fixation location, count and duration are all useful indicators of attention. Longer and more fixations are associated with deeper cognitive processing. Saccades are instances of eye movements in between fixations. Longer saccades indicate that the eye is moving widely from one place to the next (like watching a tennis match); shorter saccades indicate that attention is staying focused more narrowly (like reading).

Eye-tracking data can be analyzed both visually and quantitatively.

Visualizing eye-tracking data

Heat Maps

Summarize where people spend most of their time fixating on a stimulus. ‘Hotter’ colors represent more fixations than colder, or no colors. Heat maps summarize the behavior of one individual or many. The example below shows a heat map for the relative fixation duration of about 50 people who viewed a paper copy of a fictitious drug ad.

Gaze Plots

Quantifying eye-tracking data

In addition to creating compelling visualizations, eye-tracking data can be analyzed quantitatively for more precise evaluations and comparisons between groups. Fixation data can be used to measure and compare numerous outcomes related to attention. For example - the total number of fixations, the total fixation duration, the time to first fixation, the number ‘visits’ to an Area Of Interest (AOI), etc. With the appropriate experimental design, this information can be very powerful for comparing user attention to different visual designs.

While eye-tracking allows researchers to observe what people pay attention to, it does not inherently tell you why they behave the way that they do, or what their thoughts are. Therefore it’s common to follow eye-tracking data collection with a debriefing interview to collect qualitative information to help explain the behaviors.

Our Experience & Capabilites

FDA Risk Information Amount and Location Study (RAAL)

Physician Compare Project (BHHP)

Transportation group eye-tracking capacity

NAEP assessment of grid-type questions with fourth graders

Westat Publications on Eye Tracking

Galesic, M., and Yan, T. (2010). In Social Research and the Internet: Advances in Applied Methods and New Research Strategies, edited by M. Das, P. Ester, L. Kaczmirek, and P. Mohler, Chapter 14, Taylor & Francis Publishing Group (pp 349-370).

Höhne, J.K., Lenzner, T., Neuert, C., and Yan, T. (2021). Journal of Survey Statistics and Methodology, 9, 25-50. doi: 10.1093/jssam/smz028.

Höhne, J.K., and Yan, T. (2020). International Journal of Social Research Methodology, 23(3), 347-353. doi: 10.1080/13645579.2019.1696087.

The Physician Compare team conducted two rounds of usability testing using the Tobii Pro X2-30 eye-tracker. The first round was intended to gather baseline eye-tracking data to understand how patients and caregivers experience the Physician Compare website and to identify areas for future website enhancements. Based on the eye-tracking data, several changes were implemented and the second round of usability testing was conducted to evaluate these updates.

The attached gaze plots and heat maps provide an example of one of these changes. The Interim Search Page on Physician Compare displays results related to the user’s search term by specialty, clinician name, and group name. In the first round of testing, we found that users did not notice the clinician and group name results at the bottom of the page. These search results were moved to the right side of the page to make them more prominent. The gaze plot and heat map from the second round of testing show that this change did help make the results more noticeable.

For the National Assessment of Educational Progress (NAEP) questionnaire, Westat used eye tracking with 4th graders to examine how they navigated grid-type questions. The goal of the study was to determine whether fourth graders were able to comprehend and navigate questions that were presented in a grid. Nine respondents answered a 60 item questionnaire where items were presented either in a grid or as individual questions. Mean fixations and fixation duration were compared across question type to see whether the question format differentially impacted the time and effort spent answering. In order to understand whether the length of the grid influenced the response process, the sequence of fixations across questionnaire components was also measured for grids of varying length.

The study showed that respondents learn how to answer grid-type questions relatively quickly. The first grid question produced the largest difference between grid and discrete type questions. However, the differences in the amount of processing required between the two question types were relatively minor after the first grid question. There was little evidence from the eye-tracking data or from observation of the students in the lab suggesting that fourth grade student have difficulty navigating matrix questions.

Issues & Considerations

Equipment & Environment

Individual Factors

Natural Behavior

Do users notice the Contact Us feature on my survey?

Where do users look first when they view my website? Where do they look next?

Is design A processed differently than design B?

Westat's Approach

Collecting eye-tracking data is not as difficult as it might seem!

Analyzing eye-tracking data

Visualizing eye-tracking data

Our Experience & Capabilites

FDA Risk Information Amount and Location Study (RAAL)

Physician Compare Project (BHHP)

Transportation group eye-tracking capacity

NAEP assessment of grid-type questions with fourth graders

1. “Use of Eye-Tracking for Studying Survey Response Process.”

2. Re-examining the middle means typical and the left and top means first heuristics using eye-tracking methodology.

3. Investigating the impact of violations of the “left and top means first” heuristic on response behavior and data quality.

4. Eye-tracking data: New insights on response order effects and other cognitive shortcuts in survey responding.

Issues & Considerations