Dai, David Wei; Chen, Ivy; (2023) What influences interactional competence? A structural equation modelling study of test-taker background variables. Presented at: 19th European Association for Language Testing and Assessment (EALTA) Conference, Helsinki, Finland.

Preview

Text Dai & Chen 2023_EALTA IC-SEM.pdf - Accepted Version Download (519kB) | Preview

Abstract

Interactional competence (IC) represents a sociolinguistic-interactional take on speaking and this research area is receiving increasing attention and recognition. However, there has been little research on the variables that affect a speaker’s IC. Previous research on IC development draws largely from the more established research agenda of L2/interlanguage pragmatics; as language proficiency and length of residence (LoR) in the target community have often been used to explain a speaker’s change in L2 pragmatics, they have similarly been hypothesized to influence IC. A suite of cross-sectional-design studies (e.g., Roever & Ikeda, 2022; Youn, 2015) uses proficiency (often operationalized using a single language test score) to explain differences in speakers’ IC, while more Conversation-Analysis-informed IC researchers (e.g., Hall et al., 2011; Pekarek Doehler & Berger, 2018) tend to adopt a longitudinal study design (i.e., increasing LoR), with observed differences in interactional practices used as evidence of improvement in IC. To better understand how variables such as proficiency and LoR contribute to variance in test-taker IC performance, this paper draws on the technique of structural equation modeling. 105 test-takers participated in a nine-item computer-mediated role-play IC test and completed a background questionnaire. The role-plays covered multiple domains, varied the power relationships between test-takers and interlocutors, and consisted of three technology-mediated task types differing in level of interaction. The findings show that while a single test-score measure of Chinese proficiency (HSK score), like that used in previous research, significantly predicted IC performance, the amount of variability explained was quite low (R2=.09). Factor analysis of eight participant background variables revealed two factors: target language proficiency (length of Chinese study, HSK score, self-reported proficiency) and target language experience (amount of language contact, LoR, length of work, self-reported proficiency). By itself, the language proficiency factor better explains the variability in IC scores (R2=.36), as does the language experience factor (R2=.21); together, the two factors explain 37% of the variability in IC scores. While the findings are not unexpected, this paper is the first to quantitatively show that language proficiency and language experience do indeed significantly predict spoken interactional competence (contrasting with previous research, which used simplistic measures) and SEM has revealed a more complex and nuanced picture as to how these two factors interact, including with participant age. Future research should therefore include multiple test-taker background variables to further explore what accounts.

Download activity - last month

Export as JSONXMLCSV

Download activity - last 12 months

Export as CSVXMLJSON

Downloads by country - last 12 months

Export as CSVXMLJSON

Archive Staff Only

View Item