Predicting Elicited Imitation Performance from Complexity, Accuracy and Fluency (CAF) of L2 Oral Production

Abstract

The Elicited Imitation Test (EIT) is a testing tool that has been used to measure the second language (L2) proficiency for L2 research. The EIT construct is, nevertheless, still not comprehensively investigated in EFL contexts. This study aims to contribute to this field by examining the extent to which complexity (indexed by lexical diversity, mean length of utterance), accuracy (percentage of error-free clauses) and fluency (breakdown, self-repair, speed) (CAF) can predict EIT performance in a Saudi context. A total of 66 learners were recruited to narrate picture stories. Multiple regression analysis was employed to answer the research questions. The findings show that speed fluency, filled pauses, and accuracy are key indicators of EIT performance. This implies that faster and more accurate L2 speech with fewer pauses is likely associated with a better EIT performance. Additionally, the data highlights the validity of EIT as an assessment instrument across various linguistic settings.