Think-Aloud Protocol
A think-aloud protocol is a data collection method in which participants verbalise their thoughts while performing a task. The researcher records and transcribes this concurrent verbal report, then analyses it to infer the cognitive processes underlying task performance.
Theoretical Foundation
Ericsson & Simon (1984, revised 1993) provided the definitive theoretical framework in Protocol Analysis: Verbal Reports as Data. They argued that concurrent verbalisation of thoughts already in working memory (Level 1 and Level 2 verbalisation) does not significantly alter the cognitive processes being studied. However, when participants are asked to explain or justify their thinking (Level 3), the verbalisation itself changes the process, and the data become less valid as a window into natural cognition.
| Level | What is verbalised | Effect on process |
|---|---|---|
| 1 | Already verbal thoughts (inner speech) | Minimal |
| 2 | Non-verbal thoughts recoded into words | Slows processing slightly |
| 3 | Explanations, justifications, interpretations | Changes the process |
In SLA and Applied Linguistics
Think-aloud protocols are widely used to investigate:
- Reading processes: what strategies learners use while reading (Reading Subskills), how they handle unknown vocabulary, when they infer meaning
- Writing processes: planning, formulating, revising decisions during L2 composition
- Test-taking strategies: how learners approach reading comprehension or grammar items
- Translation processes: decision-making during translation tasks
- Metacognitive Strategies: monitoring and self-regulation during language tasks
Procedure
- Familiarisation: participants practice thinking aloud on a warm-up task
- Task performance: participants complete the target task while verbalising continuously
- Recording: audio or video capture of the verbalisation
- Transcription: verbatim transcription including pauses and hesitations
- Coding: researcher applies a coding scheme to identify cognitive processes
Concerns and Limitations
- Reactivity: thinking aloud may change the process, especially for L2 learners whose limited proficiency makes verbalisation itself a competing task
- Incomplete data: not all cognitive processes are accessible to consciousness; automated processes may be invisible
- L2 proficiency threshold: lower-proficiency learners may struggle to verbalise complex thoughts, particularly if required to do so in L2
- Training effects: the warm-up task itself may prime certain strategies
- Individual variation: some participants are naturally more verbal than others
Think-Aloud vs Stimulated Recall
Think-aloud captures thoughts during the task (concurrent); Stimulated Recall captures thoughts after the task using a recording as a prompt (retrospective). Think-aloud is more immediate but more intrusive; stimulated recall is less intrusive but more susceptible to memory distortion and post-hoc rationalisation.
Key References
- Ericsson & Simon (1984/1993): Protocol Analysis: Verbal Reports as Data
- Green (1998): think-aloud protocols in reading research
- Bowles (2010): reactivity of think-alouds in SLA research
- Leow & Morgan-Short (2004): think-alouds and awareness in SLA research