EIREducational Innovation Research3029-18443029-1852WHIOCE PUBLISHING PTE. LTD.10.18063/EIR.v4i4.1923ArticleReliability Study of Chinese Writing Tests Based on Generalizability Theoryhttps://artdesignp.com/journal/EIR/4/4/10.18063/EIR.v4i4.1923LiXuan2026442026-04-26Chinese writing tests have strong subjectivity, and the scoring results are easily affected by various factors such as differences in raters, types of test tasks, and testing environment conditions. Traditional reliability calculation methods cannot accurately identify where the errors come from. Generalization theory can distinguish between measurement objectives and measurement related aspects, estimate the variance of multiple error components simultaneously, and provide guidance for optimizing evaluation plans through generalization research (G research) and decision research (D research). This article comprehensively summarizes the main content and theoretical advantages of generalization theory, and breaks down the error structure of Chinese writing tests from three dimensions: differences among candidates themselves, measurement errors, and interaction effects. After that, specific strategies are proposed to improve reliability, such as building a stable scoring system, optimizing question design, standardizing scoring processes, and establishing quality monitoring mechanisms. The research results indicate that the generalization theory provides accurate diagnosis and targeted optimization methods for Chinese writing tests, and has a positive effect on promoting the scientific and fair development of international Chinese education assessment.Conceptual theory, Chinese writing test, Reliability, Scoring error, International Chinese education

[1] Han YY, 2020, Empirical Study on the Validity of Writing Test Methods for Teaching Chinese as a Foreign Language, thesis, Shandong Normal University. [2] Ao YQ, 2008, A Review of Generalization Theory Research. Journal of Wanxi University, (2): 49–52. [3] Ge D, 2004, A New Quantitative Research Model Beyond Classical Testing Theory: A Review of “Generalization Theory and Application of Evaluation”. Educational Development Research, (Z1): 148–149. [4] Yang HB, Luo CY, Yang Z, 2020, Analysis of the Quality of the “Chinese Language Subject Knowledge and Teaching Ability” Paper in the Qualification Examination for Primary and Secondary School Teachers: A Perspective Based on Classical Measurement Theory and Multivariate Generalization Theory. Examination Research, 16(4): 46–52. [5] Zhu S, 2022, The Application of Multivariate Generalization Theory in the Evaluation and Design of English Reading Courses in Chinese Universities. Modern English, (21): 1–4. [6] Xu CH, Tao JR, 2019, The Impact of Task Type and Scorer Factor on Writing Scoring in Chinese Writing Test. The Journal of Chinese Language and Literature, 2019: 115. [7] Zheng XQ, Liu YZ, Li GM, 2025, Validity Testing and Evaluation Scheme Optimization of Document Basket Test Concept Based on Multivariate Generalization Theory. Journal of Jiangxi Normal University (Natural Science Edition), 49(6): 623–634. [8] Kong FY, 2022, A Study on the Scoring of Chinese Writing for Preparatory Studies in China: Based on Generalization Theory and Multi-dimensional Rasch Model. Examination Research, (4): 41–51.