The cognitive processes of listening-to-summarise tasks: a mixed-methods study

Luo, Yuhan ORCID: 0009-0009-5668-0141 (2026). The cognitive processes of listening-to-summarise tasks: a mixed-methods study. University of Birmingham. Ph.D.

Luo2026PhD.pdf
Text - Accepted Version
Restricted to Repository staff only until 1 August 2027.
Available under License All rights reserved.
Download (4MB) | Request a copy

Abstract

Integrated language tasks, which involve both receptive and productive skills, are increasingly valued for eliciting authentic communicative behaviours. Among them, the listening-to-summarise (LTS) task is particularly demanding, requiring learners to comprehend extended spoken input, extract key information, and produce coherent written summaries. Despite its growing pedagogical relevance, little is known about how L2 learners cognitively engage with LTS tasks from both listening and writing perspectives. This study investigates how Chinese L1 learners engage with LTS tasks and how their cognitive engagement relates to task performance.
Eighty Mandarin-speaking L2 English learners participated in this mixed-methods study using pupillometry, keystroke logging, and stimulated recall. All participants completed the Listening Effort Experiment (pupil dilation) and the Writing Behaviour Experiment (keystroke logging). A subset of 26 also took part in stimulated recall interviews reflecting on their listening and writing processes.
Generalised additive mixed-effects modelling and qualitative analysis revealed three main findings about listening engagement. First, pupil dilation followed a non-linear, fluctuating pattern, reflecting shifts in attention due to task demands and self-regulation. Second, participants engaged in both lower- and higher-level listening processes (Field, 2013), with lower-level processing being more frequent, yet dynamically interacting with higher-level processing. Third, greater listening effort was associated with better task performance; higher-performing learners more often engaged in higher-level listening processes. In writing, keystroke data showed that longer textual units imposed greater cognitive load, whereas shorter textual units imposed more pause frequency, as reflected in pause (200ms) patterns. Deletion and insertion occurred at similar rates, though driven by different cognitive processes. All participants engaged in Kellogg’s (1996) core writing processes, with translation being the most dominant. Lower-performing learners paused more frequently within-word and longer at the between-word level, indicating lexical and content formulation difficulties. While all groups engaged in planning, translating, and monitoring, differences emerged in how these processes underpinned pausing and revision.
This study fills a gap in understanding real-time L2 engagement with LTS tasks. By integrating pupillometry, keystroke logging, and stimulated recall, it provides a dynamic account of cognitive engagement across listening and writing. The findings extend L2 task processing theories and offer practical insights for pedagogy and assessment to support learners’ academic literacy development.

Type of Work:

Thesis (Doctorates > Ph.D.)

Award Type:

Doctorates > Ph.D.

Supervisor(s):