Think-Aloud Usability Testing (Contemporary Practice)

Description

Participants verbalise while attempting tasks on a design so observers can detect usability problems; valid for issue detection, over-interpreted for mental models.

The ubiquitous moderated think-aloud test descends from verbal protocol analysis, but the treatise argues it now bears only superficial resemblance. Moderators often probe with 'why' questions mid-task, ask for preferences and treat retrospective rationalisations as evidence of mental models. Used with the right claims, it is an excellent problem-detection method: observation of behaviour plus verbalisation reveals where and often why people struggle. The rigorous version separates concurrent from retrospective commentary, uses neutral prompts, records task success and time, rates problem severity and reports what the data can and cannot show. This entry documents the contemporary method with the treatise's warning about over-interpretation.

Attributes

Loss type
Vestigial
Effort
Days

Steps

  1. Define tasks and success criteria tied to research questions.
  2. Recruit 5-8 participants per user segment.
  3. Give neutral think-aloud instructions and a practice task; keep prompts minimal during tasks.
  4. Record success, time, errors and verbalisations; save probing for after each task.
  5. Rate problem severity and frequency across participants.
  6. Report findings with explicit limits on claims about users' mental models.

Sources

  • Lost Methods in UX — A Complete Treatise (Gopinath)
  • Nielsen J (1993) Usability Engineering
  • Dumas JS & Redish JC (1999) A Practical Guide to Usability Testing
  • Ericsson KA & Simon HA (1993) Protocol Analysis
  • How to Conduct a Quantitative Usability Test (Sauro)
  • Benchmarking the User Experience (Sauro)
  • Garcia-Ruiz, M. (ed.), Games User Research: A Case Study Approach (2016)

Also known as

  • Moderated usability test
  • Think-aloud test

Categories

  • User Research
  • Usability & Evaluation