How can we better assess multimodal communication in second language learning? To explore this question, we commissioned a research team from Lancaster University.
Researchers Tineke Brunfaut and Judit Kormos investigated how multimodal viewing-to-write tasks can be designed and assessed. These tasks reflect the way we communicate every day: through a mix of spoken, visual, gestural and written information. Real-world communication is multimodal and often integrated. We use more than one language skill at once, yet few language tests currently capture this.
What the project explored
The study focused on 'viewing-to-write' tasks. In these tasks, learners watch a short video containing both spoken and visual information, then write a brief text based on what they’ve seen. We wanted to understand:
- how well learners perform on these tasks
- how they approach them
- how performance relates to their existing listening and writing abilities.
Task types and participants
The team designed two types of tasks:
- Viewing-to-describe: learners watched a video explaining how something is made, then described the process in writing.
- Viewing-to-compare-and-contrast: learners watched two experts discuss a topic, supported by visuals, then wrote a short report comparing their viewpoints.
A total of 134 English learners in Hungary (older adolescents and young adults) completed four tasks on the computer. Videos were played twice and learners could take notes. They also completed the Aptis listening and writing tests to measure their skills independently.
Twenty participants later took part in follow-up interviews in their first language, where they reflected on how they had completed the tasks.
What the researchers found
Overall, the tasks were appropriately challenging for learners working largely at CEFR B2–C levels. Average scores ranged from 64% to 69%, with similar performance across task types. Scores were also well-distributed, showing that the tasks could distinguish between different proficiency levels.
Statistical analyses showed that the four rating criteria (Viewing for Writing, Organisation and Structure, Language Use, and Mechanics) were strongly related. This suggests that a single underlying ability construct drives performance on viewing-to-write tasks.
The results also showed that both listening and writing abilities predicted how well learners performed. Interview data shed further light on this: learners described using visual cues, a wide range of listening strategies, and several different approaches to planning and drafting their texts. They integrated these skills, watching, listening, note-taking, and writing, to understand the videos and shape their responses.
Many learners also said they paid attention to the type of writing they would need to produce while watching the videos, and used the videos to help ensure their texts were accurate, well-structured and aligned with the task demands.
Why this matters
This study shows that it is both possible and practical to design effective viewing-to-write tasks for use in summative, general proficiency testing. It highlights an exciting area for future test development, especially as real-world communication continues to rely on richer, more integrated multimodal input.
For more information
The viewing-to-write tasks and rating scales are freely available at https://osf.io/xude7/.
If you’re interested in learning more about the research, the researchers have published a paper in Language Assessment Quarterly as part of a special issue on multimodal assessment: Brunfaut, T., & Kormos, J. (2025). Assessing Multimodal Viewing-to-Write Constructs: Task Design, Performance, Processing, and Rating. Language Assessment Quarterly, 22(4–5), 429–459. https://doi.org/10.1080/15434303.2025.2596374
In the special issue you can also find papers from two Assessment Research Award-holders, Sharon Sin Ying Wong and Heidi Liu Banerjee. Read their papers below.