Exploring unobserved heterogeneity of speech fluency and its dynamic interactions with emotions

Abstract The present study utilized a parallel-process growth mixture modeling technique to investigate the heterogeneity in the development of speech fluency at the (sub)group level, and to explore the interrelationships between speech fluency and different types of emotions among 126 Chinese EFL learners for four months. The study identified salient patterns emergent out of the developmental process of speech fluency, with two developmental patterns for speech rate and two for mean length of run respectively. Interaction dynamics between speech fluency and different emotions were also revealed, with anxiety and enjoyment both negatively affecting the development of L2 speech fluency over time. This study shows that parallel-process growth mixture modeling is an efficient method for identifying the heterogeneity of language development within a larger population and for exploring dynamic relations over time.