Measuring L2 speakers' interactional ability using interactive speech tasks. (January 2018)

• Record Type: Journal Article
• Authors: van Batenburg, Eline S. L.; Oostdam, Ron J.; van Gelderen, Amos J. S.; de Jong, Nivja H.
• Is Part Of: Language testing. Volume 35:Number 1(2018)
• Page Start: 75
• Page End: 100
• Publication Date: 2018-01
• Subjects: Holistic and analytic measurement -- interactional strategies -- interactive speech tasks -- pre-vocational education -- reliability -- validity; Language and languages -- Ability testing -- Periodicals; Language and languages -- Examinations -- Periodicals; 407.6
• DOI: 10.1177/0265532216679452
• Languages: English
• ISSNs: 0265-5322

Abstract

This article explores ways to assess interactional performance, and reports on the use of a test format that standardizes the interlocutor's linguistic and interactional contributions to the exchange. It describes the construction and administration of six scripted speech tasks (instruction, advice, and sales tasks) with pre-vocational learners ( n = 34), and reports on the extent to which these tasks can be used to assess L2 speakers' interactional performance in a reliable and valid manner. The high levels of agreement found between three independent raters on both holistic and analytical measurements of interactional performance indicate that this construct can be measured reliably with these tasks. Means and standard deviations demonstrate that tasks differentiate between speakers' interactional performance. Holistic ratings of linguistic accuracy and interactional ability correlate highly between tasks that focus on different language functions, and are situated in different interactional domains. Furthermore, positive correlations are found between both holistic and analytic ratings of oral performance and vocabulary size. Positive within-task correlations between analytical ratings of specific interactional strategies and holistic ratings of overall interactional ability show that analytic ratings of meaning negotiation and correcting misinterpretation provide additional information about speakers' interactional ability that is not captured by holistic assessment alone. It is concluded that these tasks are a useful diagnostic tool for practitioners to support their learners' interactional abilities at a sub-skill level.