GMU: An eye-tracking pilot tracing the reading problems magnification left unresolved.

ThemeAcademic research
PlatformResearch
RoleUX Researcher — Research Proposal, User Testing, Interviews, Quantitative & Qualitative Analysis
YearAug 2022 - Dec 2022
CollaboratorsMary Nichols (literature review, website redesign, surveys)

Capstone project at the Visual Attention and Cognition Lab, Department of Psychology, part of a 2-year M.A. in Human Factors and Applied Cognition.

For my M.A. capstone at George Mason, I investigated whether a magnifier could compensate for an inaccessible page layout. I ran sessions with three students under simulated low vision, combining task performance with their comments and eye-tracking recordings. The comparison showed where low contrast, small text and silent menus stopped readers, and how they worked around them.

The magnifier enlarged the page. Participants still struggled to read it.

On our comparison website, one participant tried enlarging a passage and said: “Which really doesn't help considering that it's still really small.” Another used the shadow of the simulated blind spot to make out low-contrast text. These workarounds became the most useful evidence in the study: they showed what people did when the available tool did not resolve the problem.

The research question was whether following accessibility guidelines could reduce the need for magnification. Mary Nichols and I compared two versions of the same information through a low-vision simulator, recording task performance, usability ratings and eye movements. I ran the sessions and analyzed the evidence; Mary contributed the literature review, comparison-site redesign and surveys.

The pilot tied each reading and navigation problem to what participants said and where they looked. That gave a follow-up study its tasks and specific failure points to test.

Observed Results

Study Design

Three participants, three tasks and two versions of the same information.

Three graduate students completed three information-finding tasks on each website: locate a COVID-19 testing site, check local COVID-19 rules and find disability support. Each thought aloud while working. Two started on the accessible site and one on the comparison site, partially counterbalancing the order.

The accessible condition was the New South Wales Government website, selected for its use of WCAG guidance. The comparison was a Wix mockup carrying the same information. Contrast, typography, dropdown behavior and columns all changed together, and the mockup also had fewer clickable areas, so the study compared whole layouts.

With one week of lab time and no budget to recruit participants with low vision, we simulated it with a Silktide overlay: a central blind spot and blur. It followed the mouse pointer, so participants could move it aside and read with a clear center. Tobii recorded their eye movements independently; gaze did not control the simulation. A docked magnifier was available on both sites.

I combined task times and completion with System Usability Scale (SUS) ratings, think-aloud notes and the gaze recordings. The task data showed where people struggled; their comments and eye movements helped us examine what happened there.

The two conditions: the NSW Government site (left) and the comparison mockup (right), which changed contrast, type and layout together.
Accessible NSW Government homepage, with high contrast, readable type, single clear search entry point
Inaccessible mockup page, with cursive low-contrast headers and uneven columns that are difficult to read under magnification
More on Study DesignRead the instrument settings and session procedure2 images · 1 min read

Instruments

  • Tobii eye tracker. Recorded fixations and saccades on both websites. The setup could not define areas of interest or generate heat maps, so gaze was read from the recordings rather than analyzed region by region
  • Silktide Chrome extension. A 50% central scotoma with 20% blur, attached to the mouse pointer
  • Zoom magnification software. Docked at the top of the screen so it did not cover the page, magnifying the area under the pointer

How the two sites differed

  • Typography and contrast. The accessible site used sufficient contrast and readable sizes. The mockup used low contrast, a cursive display face and all-caps text
  • Dropdown menus. The accessible site used multicolumn, multirow dropdowns with hover color changes and highlighted tab headers. The mockup used single-column dropdowns with no hover or selection feedback
  • Column layout. The accessible site used even columns. The mockup used uneven, unnecessary columns, which break the reading line under magnification

Session procedure

Participants agreed to be recorded, calibrated the eye tracker, and were shown the simulator and the docked magnifier. They were asked to navigate and read as they normally would, thinking aloud. I took structured notes on what they said and did, then each participant completed a SUS questionnaire for each site.

The Tobii eye tracker, set up in the lab.
Tobii eye tracker mounted beneath a monitor in the lab, with the eye-tracking software actively exporting a recording

In a gaze plot, each red dot is a fixation, a point where the eye stops, sized by dwell time; the lines between them are saccades, the jumps from one fixation to the next.

A gaze plot from the accessible site, with the COVID-19 dropdown open.
Eye-tracking gaze plot over the NSW Government website with the COVID-19 dropdown open, showing red fixation dots and saccade lines trace where the participant looked

Quantitative Findings

Participants finished faster and rated the accessible site higher.

Across the three tasks, participants took a mean of 2:06 on the accessible site and 8:17 on the comparison site, a difference of about six minutes. Every participant completed all three tasks on the accessible site; two of the three failed at least one task on the comparison site.

Mean time on task per condition.
Bar chart of mean time on task: the accessible website took 38, 44, and 43 seconds per task; the inaccessible website took 4:36, 1:34, and 2:07

Mean SUS ratings were 88.33 for the accessible site and 64.17 for the comparison site, in line with the task data.

With three participants, Mary's paired t-test on task time in SPSS was not significant at α = .05, so these figures describe the sessions rather than a wider population.

Qualitative Findings

The recordings showed how participants worked around the pages.

We used HyperResearch to code the session notes and think-aloud audio, then read the themes alongside the Tobii recordings. That let us connect a complaint about legibility with the part of the page a participant was trying to read.

On the accessible site

Participants called it “pretty easy” and “clearer,” and the session summary records no magnifier use. They read with the simulated blind spot moved off to the side. The multicolumn dropdown “isn't flying around everywhere,” and participants found links in it directly.

Reading on the accessible site. Fixations track line by line, with the simulated blind spot held to one side and no magnifier.

On the comparison site

All three reported difficulty reading, remarked on the low contrast and struggled with the dropdown, which “was the same color as the background.” They improvised: “I'm using the shadow to see.” Magnifier use varied. One participant relied on it; another did not use it at all. Where it was used, it did not make the smallest text readable: “I would definitely need the magnification for this part. Which really doesn't help considering that it's still really small.”

On visual inspection, fixations on the comparison site appeared longer and more scattered. Together with the comments, they suggested difficulty locating readable content. I read the recordings alongside what participants said, as context for their behavior rather than a separate quantitative result.

Reading on the comparison site. Fixations wander as the participant looks for legible fragments.
More on Qualitative FindingsWatch the dropdown recordings2 images · 2 videos
Dropdown navigation through the simulator: accessible (left) versus comparison site (right).
Eye-tracking still of the accessible website's multicolumn dropdown, where fixations land directly on menu links beside the simulated scotoma
Eye-tracking still of the inaccessible website, where fixations hunt across a low-contrast nav bar that gives no feedback, scotoma obscuring the content
Dropdown use on the accessible site. Fixations land on links in the multicolumn menu.
Dropdown use on the comparison site. Gaze moves across the low-contrast navigation, which has no hover feedback.

Conclusions

Contrast, small text and menu feedback were the clearest candidates for redesign.

On the tested accessible pages, all three participants completed the tasks without using the magnifier. On the comparison pages, magnification did not resolve the reading problems where it was used. Low contrast, small text and dropdown behavior were concrete candidates for redesign, supported by the session recordings and participants' accounts.

What a follow-up would change

Because contrast, type, menu behavior, columns and clickable area all differed between the sites, the next study would change one design variable at a time. It would also recruit people who live with low vision, across age groups: our participants could read around a blind spot that followed the mouse, and people with low vision develop strategies of their own. An eye-tracking setup that supports analysis by page region would let the gaze data carry more of the result.

Usabilathon 2022 at George Mason; the arrow marks me.
Usabilathon 2022 participants gathered in a GMU classroom, with an arrow marking Chevis in the back row
Four documents from the M.A. Every cover opens the full document.
First page of 'A Usability Study of Smart Home Apps Alexa and Nest' by Zhengchen Zhou, George Mason University
Smart-home apps usability study: proposal
First page of 'An empirical study of emotional experience on shopping websites' by Zhengchen Zhou, George Mason University
Shopping-website emotion study: proposal
Title page of 'Designing for Well-being in Digital Experience: A PERMA Perspective' by Zhengchen Zhou, George Mason University
PSYC 768 seminar paper: designing for well-being with the PERMA model
Title page of the capstone research report 'UX Research: Eye Tracking & Accessibility' by Zhengchen Zhou and Mary Nichols, George Mason University Visual Attention and Cognition Lab
Capstone research report: methodology, literature review and data
More on Related ResearchExplore the HelloFresh prototype and coursework research1 min read

Usabilathon 2022: HelloFresh Market

The competition ran in November 2022. I worked on design and conducted user testing, with guidance from a HelloFresh mentor. We brought Market, an add-on grocery store buried five levels deep, onto the home screen, separated it from meals and added search to meal selection.

In an 11-person follow-up comparing the prototype with the live app, average time on the Market selection flow fell from 3:23 to 1:12. Participants unable to complete the defined path fell from four to one; eight added Market items, compared with six on the live app. These were session actions, not purchases. The case-study PDF reports a paired t-test that was not significant at .05.

Read the HelloFresh case study or the presentation deck.

Coursework research

  • Smart-home apps proposal. A planned between-subjects comparison of Amazon Alexa and Google Nest, with 60 participants and five performance tasks grounded in Nielsen's heuristics.
  • Shopping-emotion proposal. A study design based on Roseman's appraisal theory, linking four interface features to three cognitive appraisals and six emotions. It proposed 50 interviews, a 100-person survey and structural equation modeling in Mplus.
  • Well-being seminar paper (PSYC 768). An application of Seligman's PERMA model to digital-product design and evaluation, compared with Self-Determination-Theory frameworks such as METUX.

Read the smart-home proposal, shopping-emotion proposal or PERMA paper.

The capstone's presentation and study boards in Figma provide further method and session detail.

Related Work