Dr. Ayaka Matsuo has co-authored a new article published in Research Methods in Applied Linguistics.
The article, “Nested Multilevel Modeling in Second Language Acquisition: A Systematic Review of Current Applications and Reporting,” co-authored with X. Wang and Y. Maeda, examines how researchers use multilevel modeling (MLM)—a statistical method for analyzing grouped data like students in classrooms or repeated measurements from the same person—in second language acquisition (SLA) research. After reviewing 80 studies, researchers found that while SLA research benefits from this statistical technique, there are serious reporting problems: 18 studies didn’t report participants’ first language, 48 omitted proficiency levels, 54 didn’t specify estimation methods, and 77 skipped a priori power analyses. Key trends include R software becoming the most popular tool and terminology shifting from “multilevel modeling” to “mixed-effects model” over time. While this advanced statistical method is becoming more common in SLA research, the authors conclude that SLA researchers need to improve their reporting practices to make studies more reliable and reproducible, and they provide specific recommendations for better application and reporting of MLM.
Publication:
Matsuo, A., Wang, X., & Maeda, Y. (2026). Nested multilevel modeling in second language acquisition: A systematic review of current applications and reporting. Research Methods in Applied Linguistics, 5(3), 100353. https://10.1016/j.rmal.2026.100353