Here are my takeaways:
- Make sure you ask the right research questions. Another delegate asked if eye tracking can test understanding/comprehension of words on a screen. Nope.
- Get the right equipment. How accurate do you need the results to be?
- Get good data. Have you thought about ways to prevent imprecise measurements?
- Communicate effectively. This means present the data in the correct way. Use quantitative analysis. Stay away from heat maps.
- Eye tracking is imprecise for two reasons: calibration errors and the way the eye fixates. Eyes tend to jump around and not glide smoothly across the screen.
- Stephen said you have to be very careful how you set up your ‘detection’ areas because if you do it wrong, you get terrible data. Consider this example: even though a user’s fixation may be on one button, their attention is actually on another. If you set up your detection areas wrong your data is gone bro’:
Errors compounded by setting up the wrong bounding box
- Eye tracking is fairly useless for mobile no matter how good your equipment “at best it can tell you if the user is looking at the top or bottom of the screen”
- Head movements cause tracking errors and the more ‘hands-free’ the kit is, the more calibration errors you’re likely to see. There’s a thing called a ‘head box’ and this is the amount of movement a user is allowed to make in one direction or another before the tracking tool loses the plot (literally)
A task we were given: come up with a bounding box for each question
Heat maps are pants
- CEOs love them because they are pretty but there are better quantitative data
- Like time to fixate on a target
- Duration of fixation
- Number of fixations
- % of targets fixated
- Quantitative data is what you’re after. Not pretty pictures of eye scans.
Another exercise: what do the gaze plots tell us?
- Eye tracking tells you what users are focussing on, but there is a whole bunch of stuff in their peripherals they might be subconsciously aware of
- I asked Stephen if you could measure whether there is a relationship between ‘conscious focus’ vs ‘unconscious peripheral focus’ by using recall to test the relative effectiveness of each and he said that would be an interesting study!
Eye tracking has a lot more to it than meets the e… sorry. And it can be a very powerful research tool just so long as you ask the right questions, control for imprecision and analyse and present your results in the correct way.
You can’t use eye tracking for really fuzzy questions like “do users understand the difference between x and y” or “do they know how to do this”. Although you might be able to infer wider connections if you test other metrics that reference eye tracking data e.g “do you remember x”
Questions, comments? Can’t see? Erm, Ok.