Abstract
INTRODUCTION: The Hanyu Shuiping Kaoshi (HSK), a standardized L2 proficiency test, is a mandatory requirement for international students pursuing graduate study in China. Despite its importance, the potential benefits of higher HSK scores for predicting academic performance during the first year of graduate study remain largely unexplored. METHODS: This study analyzed data from 666 graduate students enrolled in universities in southern China. The dataset included HSK listening, reading, and writing scores, as well as average undergraduate academic scores. First-year graduate academic performance was represented as a binary variable: a value of 1 indicated an average score of 85 or above (B+ or higher), while 0 indicated otherwise. A robust Bayesian logistic regression model was developed to estimate the probability of achieving good academic performance, with 80% of the dataset used for training and 20% for validation. RESULTS: Analysis revealed statistically significant correlations among the three HSK test components. The Bayesian logistic regression model demonstrated strong predictive power in estimating the likelihood of achieving a B+ or higher in the first year of graduate school. The model effectively incorporated both L2 test scores and undergraduate academic records to generate accurate predictions. DISCUSSION: These findings highlight the predictive value of HSK scores and prior academic achievement for international students' success in graduate programs in China. The results suggest that L2 proficiency, as measured by the HSK, plays a meaningful role in shaping academic outcomes. This study provides evidence for universities to consider HSK performance not only as an admission requirement but also as an early indicator of students' potential academic performance.