This chapter discusses the application of multilevel modeling (MLM) to analyze second-language (L2) test scores over time. With the growing number of test takers who repeat English language proficiency tests more than once in order to achieve a certain cut score (e.g., for university admission) or to demonstrate progress (e.g., after a period of L2 instruction), there is a need for more research on the validity of inferences based on repeaters' test scores as well as the sensitivity of L2 test scores to changes in L2 proficiency over time and/or after L2 instruction. MLM provides a powerful alternative to traditional statistical methods (e.g., analysis of variance, multiple regression analysis) for investigating questions concerning the sources of variability in, and the quality of, repeaters' test scores. We describe how MLM can be used to analyze longitudinal L2 test data. Specifically, we discuss the limitations of the traditional approaches to longitudinal data analysis, describe a step-by-step procedure for using MLM to analyze longitudinal L2 test data, and discuss some of the potential applications of MLM to L2 testing research.