Lab Set-Up for a Moderated Session Involving a Low-Vision Participant and a Mobile Prototype
A usability lab is booked for five moderated sessions of a mobile expense app. The lab has an observation room behind a one-way mirror, two ceiling cameras, screen capture, and Wi-Fi mirroring of the test phone to a monitor. One participant has low vision, uses 200% zoom and TalkBack on Android, and will bring their own device. The HVAC produces about 48 dB of steady noise. Two compliance observers will watch live. The moderator plans to sit beside the participant, read a script, and use a wireless keyboard to note observations. The session is 45 minutes, with 10 minutes for consent and set-up. A pilot is scheduled 30 minutes before the first session. The team wants recordings for analysis and clips for stakeholders. The mobile prototype has a custom date picker and a modal that auto-closes after 15 seconds.
1. Identify at least four set-up problems that could invalidate the session and state the correction for each.
Read the worked answer
Problems and corrections:
- Audio quality: 48 dB of steady HVAC noise plus no close microphone will make think-aloud speech hard to transcribe. Use a lapel or directional microphone, test a recording, and reduce HVAC noise if possible.
- Camera framing: ceiling cameras may not capture the participant's own device, zoomed content, or gestures. Add a document camera or a device-facing camera, and check that the screen and hands are visible.
- Device mirroring and assistive technology: Wi-Fi mirroring may add latency or fail with the participant's own Android device and TalkBack. Test the exact device and screen reader beforehand; have a wired or built-in recording fallback; do not install unapproved software without consent.
- Moderator position and keyboard noise: sitting beside the participant with a wireless keyboard can create typing noise and unintentional cues. Move note-taking to an observer or use a silent keyboard, and script neutral prompts.
- Consent and recording: 10 minutes may be too short to explain recording, observation, clips, and data handling. Send consent information in advance and confirm on the day.
- Auto-closing modal: a 15-second modal may disappear before a screen-reader or zoom user can read it. Treat this as a prototype defect to observe, but ensure the moderator does not rescue the participant prematurely.
The pilot must verify audio, framing, mirroring, TalkBack, consent, and observer behaviour before the first session.
2. The participant's own device cannot be mirrored to the lab system without installing software. Do you cancel, proceed with device-only recording, or substitute another participant? Justify your decision.
Read the worked answer
Do not cancel and do not substitute solely because mirroring is unavailable. The participant was recruited because they represent a relevant accessibility profile; replacing them with a non-low-vision participant would remove exactly the evidence the team needs.
Proceed with a documented fallback:
- Use the device's built-in screen recording if permitted, or record the device with a camera positioned to capture zoomed content and TalkBack focus.
- Have the moderator and an observer keep a structured manual log of focus order, announcements, and gestures.
- Confirm consent for the fallback recording and explain any limitation.
- If stakeholder clips require high-quality mirroring and the fallback is inadequate, reschedule that participant rather than silently degrading the evidence.
This is a set-up limitation, not a reason to exclude the user. Record it in the session notes and analysis so the team does not over-generalise from incomplete capture.
3. What should be checked in the 30-minute pilot, and what should the moderator do if the pilot reveals a problem that cannot be fixed before the first session?
Read the worked answer
The pilot should check: consent script timing; microphone level and transcription quality; camera views of face, hands, and device; screen capture and mirroring latency; TalkBack and zoom behaviour on the participant device; observer audio and visibility; moderator prompts; and the auto-closing modal.
If a problem cannot be fixed before the first session, the moderator should decide whether the affected data is still usable. For example, poor audio may invalidate think-aloud analysis, so the session should be rescheduled or a backup microphone used. A minor camera framing issue may be acceptable if device recording captures the interaction. The decision and its rationale should be logged, and affected participants should not be asked to proceed under conditions that compromise consent or data quality.