Advanced search

Multiword sequences do not predict speaking proficiency in dialogue: a pair-level analysis

Download
[thumbnail of accepted draft April 2026 for Centaur.pdf]
Text
- Accepted Version
· Restricted to Repository staff only
· The Copyright of this document has not been checked yet. This may affect its availability.
[thumbnail of revised_ms.docx]
Text
- Accepted Version
· Restricted to Repository staff only
· The Copyright of this document has not been checked yet. This may affect its availability.
Advice

Please see our End User Agreement.

It is advisable to refer to the publisher's version if you intend to cite from this work. See Guidance on citing.

Tools
Add to AnyAdd to TwitterAdd to FacebookAdd to LinkedinAdd to PinterestAdd to Email
Lists

Uchihara, T., Tavakoli, P. ORCID: https://orcid.org/0000-0003-0807-3709, Mazhurnaya, S. and Smyth, P. ORCID: https://orcid.org/0000-0001-8799-6186 (2026) Multiword sequences do not predict speaking proficiency in dialogue: a pair-level analysis. Modern Language Journal. ISSN 1540-4781 (In Press)

Abstract/Summary

This study revisited the commonly held view that the use of multiword sequences (MWSs) is a reliable predictor of L2 speaking proficiency in monologue, by extending the investigation from monologic to dialogic speaking contexts. We accessed dialogic performance data from 127 test-takers of the Test of English for Educational Purposes assessed at six different proficiency levels. The use of bigrams and trigrams was measured in terms of proportion, frequency, and association of strength. Results showed that none of the n-gram measures were associated with levels of speaking proficiency. When analysing pair-level use of MWSs, the results of a linear mixed-effects modeling suggested that pair-level effects explained about 40% and 30% of the total variance for bigram and trigram proportion, respectively. Follow-up analyses revealed that lower-proficiency speakers in a nationality-matched pair tended to produce a greater proportion of bigrams and trigrams than those in a nationality-unmatched pair. These findings challenge the central role of MWSs documented in monologic task performance and suggest a complex interplay of phraseological competence and speaking performance when dialogic processing is considered.

Item Type Article
URI https://centaur.reading.ac.uk/id/eprint/129270
Refereed Yes
Divisions Interdisciplinary Research Centres (IDRCs) > Centre for Literacy and Multilingualism (CeLM)
Arts, Humanities and Social Science > School of Literature and Languages > English Language and Applied Linguistics
Arts, Humanities and Social Science > Global Academy
Publisher Wiley
Download/View statistics View download statistics for this item
Deposit Details
CORE (COnnecting REpositories)

University Staff: Request a correction | Centaur Editors: Update this record