[Development and validation of the Chinese written expression assessment scale]

Objective: To develop a Chinese written expression assessment scale and evaluate its reliability and validity. Methods: A cross-sectional research was designed and a multi-stage stratified random cluster samplingmethod was used. From November 2020 to April 2021, 1 016 primary school students in grades 1 to 6 from Nanjing were recruited for a writing ability assessment. From June 2024 to September 2025, the item pool of the Chinese written expression assessment scale was constructed through group discussions and the Delphi expert consultation method based on the picture-writing task from the writing ability assessment. Scoring criteria were established separately for 3 grade groups (grades 1-2, 3-4, and 5-6). Various methods such as item analysis, reliability analysis, exploratory factor analysis, confirmatory factor analysis, content validity, and criterion validity analysis were comprehensively employed to test the reliability and validity, forming 3 grade-specific Chinese written expression assessment scales that meet psychometric standards. Results: A total of 999 valid samples were included, consisting of 3 groups by grade level: grades 1-2 (n=329), grades 3-4 (n=347), and grades 5-6 (n=323). There were 381 children enrolled for exploratory factor analysis, and 618 children enrolled for confirmatory factor analysis, content validity, and criterion validity analysis. The Chinese written expression assessment scale contains 3 dimensions: writing accuracy, grammatical accuracy, and paragraph organization coherence. The scale comprises 20 items for grades 1-2, and 24 items for both grades 3-4 and grades 5-6. The item-level content validity index ranged from 0.80 to 1.00, and the scale-level content validity index was 0.97. Exploratory factor analysis extracted 3 common factors for each grade group, with cumulative variance contribution rates of 72.20%, 57.00%, and 56.80%,respectively,and the confirmatory factor analysis indicated a good model fit. The criterion validity coefficients were all positive correlation for literacy test scores (r=0.70, 0.69, 0.73, all P<0.001). The overall Cronbach's α coefficients for the 3 grade-specific scales was 0.93, 0.94, and 0.93 respectively; McDonald's ω coefficients was 0.95,0.96, and 0.95 respectively; split-half reliability coefficients was 0.80, 0.83, and 0.76 respectively, inter-rater reliability coefficients of r=0.76, 0.82, 0.90 respectively. Conclusion: The Chinese written expression assessment scale is scientifically robust and practically applicable, providing a theoretical basis for the early identification and intervention of children with dysgraphia.