Comparison of lexical complexity in L2 speaking and writing and factors predicting English speaking proficiency

Abstract

This study is to measure the three lexical complexity factors, i.e., lexical density, lexical diversity, and lexical sophistication using spoken and written data in a large-scale learner corpus, so that the lexical complexity can be observed in multidimensional terms. The study goes beyond most previous work on lexical complexity in SLA by examining the effects of all three measures of lexical complexity on predicting speaking proficiency of Korean EFL learners. Logistic regression analysis reveals that the verbal element of the lexical density factor was the most appropriate predictor of English proficiency (odds ratio: 2.259). In other words, the more frequently English learners use verbal elements, the higher their English-speaking ability is likely to improve 2.259 times. The results of this study will help to understand the correlation and difference compared to the results of other literatures in this field. Also, it will help educators to understand that the relationship between lexical complexity factors and proficiency is different in English speaking and writing, and to select more useful indicators when evaluating L2 learners. In addition, it is expected that it will serve as an evaluation method that can be used as an alternative in the field of pedagogy because it provides L2 discriminant function to evaluate learners.