Task analysis is only as good as the data behind it. Before you can diagram a workflow or identify where users struggle, you need an accurate record of what people actually do, in what order, with what information, and what they are thinking while they do it. This topic covers the main ways that raw data is gathered: observing users in their real context, interviewing them and collecting verbal protocols such as think-aloud, mining documentation that already exists (manuals, training materials, logs, help tickets, procedures), and choosing participants who genuinely represent the target user population.
Each method has trade-offs. Observation captures what people do but not always why; interviews capture intent and rationale but are filtered through memory and self-presentation; existing documentation describes how a task is supposed to be done, which often differs from how it is really done. Skilled practitioners combine methods and treat disagreements between them as findings rather than noise. Participant selection matters just as much: data from the wrong users produces a clean, well-structured, and wrong task analysis.
The three videos below come from Nielsen Norman Group and focus on the verbal-protocol and moderation side of data collection. As you watch, pay attention to how a moderator elicits information without putting words in the participant's mouth, what a repeatable session procedure looks like, and when thinking aloud distorts the very behavior you are trying to record. Then consider how the data you collect this way would feed into the hierarchical and cognitive task analyses you build later in the module.
A short practitioner video on the think-aloud protocol: what it is, why researchers ask participants to verbalize their thoughts, and how to get users to keep talking during a session. Based on the title and description, it covers explaining the technique to participants up front, reinforcing the behavior when they do it well, and using neutral prompts to restart narration when a participant goes quiet.
Think-aloud is the core verbal protocol method used to surface the cognitive steps of a task, and this video addresses the practical skill of eliciting it without leading the user. NN/g is a long-established usability research group, so the guidance reflects standard professional practice.
A checklist-style walkthrough of roughly ten things a moderator should do when running a usability study. It is likely to cover session setup, briefing the participant, giving tasks, staying neutral while observing, and wrapping up, framed as a repeatable procedure that protects the quality of the insights you collect.
It treats data collection as a disciplined protocol rather than an informal chat, which is exactly the standard your own observation and interview sessions should meet. The checklist format makes it easy to reuse when you plan and run your own sessions for the module project.
A counterpoint video on the limits of the think-aloud method, including situations where verbalizing interferes with the task or where participants cannot articulate what they are doing. The description points toward expert users as a case where other approaches work better than standard think-aloud usability testing.
Knowing when a method fails is part of choosing methods well, and this video connects directly to participant selection, since the right technique depends on who your users are. It balances the first video so you do not treat think-aloud as a universal solution.
These videos supply the data-gathering foundation the module's outcomes depend on. Think-aloud and moderated observation are the primary way analysts capture the mental steps, decisions, and information needs that cognitive task analysis represents, and they also reveal the ordered subtask structure you will decompose in a hierarchical task analysis (MO1, MO2). The moderation checklist helps you run sessions consistently enough that the resulting workflow diagram reflects real user behavior rather than your assumptions (MO2). The video on when thinking aloud fails pushes you to judge method fit and participant type, which in turn sharpens your ability to read the collected data for genuine bottlenecks, workarounds, and interface requirements instead of artifacts of a poorly chosen method (MO3).
These videos were chosen from YouTube on October 3, 2026. Each has captions; the player turns them on. Videos on YouTube can be changed or removed — if one no longer plays, tell your instructor. If a player does not appear, use its Watch on YouTube link.