Overview
A user research study of the U.S. passport services experience on Travel.gov, undertaken as part of graduate coursework in the Human-Computer Interaction program at the University of Maryland.
Study Design
- Five participants completed five tasks on Travel.State.Gov over three days, tested remotely via Google Meet.
- Participants were deliberately chosen using the Design for Diversity (D4D) framework — two (P01, P02) sat outside the typical target user group, which was intentional to stress-test the site's accessibility for excluded users.
- A pilot test with a highly experienced user established baseline task-time expectations.
Key Findings
- Task success varied widely by task type, not just by user. Tasks 2 (locate nearest agency), 4 (fill out application), and 5 (determine next steps) had the highest success rates (80%). Tasks 1 (find required documents) and 3 (find processing time) were the weak points, both at just 40% success — meaning the site consistently failed users trying to complete these two tasks.
- P01 (the intentionally "excluded" user) struggled the most — only 1 of 5 tasks completed successfully, a 40-point gap below the next-lowest participant. This pointed to real accessibility gaps for users outside the core target audience.
- Confidence didn't match actual success. Several participants believed they'd completed tasks correctly when they hadn't — reinforcing that perceived completion and actual success are different things worth measuring separately.
- P03 was the standout performer — 100% task success, fastest completion times, and the closest match to the "ideal" target user experience.
Root Cause
Despite an initial expert review that focused on technical/visual polish, the testing revealed the real problems were in content structure and information-gathering — not visual design. Users couldn't locate basic information like required documents or processing times.
Proposed Changes
- Clearer language — reading level was a genuine barrier to task completion
- Highlighted callouts — lack of visual hierarchy made scanning difficult, since all content read as equally important
- Enhanced search functionality — only one participant used the search bar at all, likely because of its buried placement; a more central, predictive search was recommended