VietTESOL International Convention 2026

Research that examines longitudinal growth of lexical sophistication in L2 writing offers useful implications for setting realistic goals, providing learners with relevant learning contents, and devising suitable assessment practices. However, this examination is often moderated by text length and topic familiarity and consequently most studies in this area have controlled these two factors at the group level, leaving the investigation at the individual level almost unexplored. The present study filled this gap. To this end, it tracked the changing pattern in 17 representative indices of lexical sophistication across three sets of 16 statistical reports written by the same 16 English-majored high-school students in Vietnam at the end of their grade 10, 11 and 12. Their text lengths were measured by tokens and topic familiarity was rated by each learner using a five-point Likert scale. A mixed-effects general linear regression model was used to gauge the predictive power of time, text length and topic familiarity for the above 17 indices. The results showed that apart from the strong and consistent predictive power of time, text length could account for their lexical sophistication growth for three word-level indices (i.e., word meaningfulness, SUBTLEXus-Range, and SUBTLEXus-Frequency), while topic familiarity could do so for three phrasal-level indices (i.e., COCA-Academic-Bigram-Frequency, COCA-Academic-Bigram-Range, and COCA-Academic-Tri-Prop-30k) and one word-level index (i.e., Age of Exposure). These findings altogether suggest that the control of these two variables at the group level might distort their true relationship with the individual lexical sophistication growth, which in turn might misinform our instructional practices.