基于Rasch模型的普通话高年级小学生听写测试字表的编制

Development of a Mandarin Dictation Test Word List for Upper-grade Primary School Students based on the Rasch model

摘要

目的 通过Rasch模型构建适用于五、六年级学生的普通话听写能力评估工具,并经测量学优化提升读写障碍和书写障碍儿童的筛查准确性。方法 从部编版语文教材中低学段生字库中选取180个汉字构建初始测试题库,2023年11月对江苏省南通市某小学595名五、六年级学生进行测试,采用Rasch模型进行题项筛选与模型优化。结果 经筛选后保留163个拟合度优良的题项。正式字表信度优异(题项信度0.975,样本信度0.956),分离度高(题项分离度6.32,样本分离度4.66)。单维性检验(第一/第二成分特征根比值9.545,Smith's t-test显著差异比例5.88%)与局部独立性假设(MADaQ3=0.042)均成立;怀特图显示题项难度与目标群体能力匹配良好,测验信息函数在低能力区间(θ≈-2)达到峰值(信息量＞25),提示其对读写障碍和书写障碍儿童筛查效能更优。字表内部一致性良好(Cronbach's α=0.975),评分者间一致性高(ICC=0.998)。ROC分析显示,该字表在本研究样本中对书写困难风险儿童具有较高区分能力(AUC>0.99);以标准分37为截断值时,敏感度为100%,特异度为98.8%。结论 普通话高年级小学生听写测试字表具有良好的信效度和筛查敏感性,为读写障碍和书写障碍的早期识别提供了客观测量框架,并为计算机自适应测验开发奠定基础。

Abstract

Objective To construct a mandarin dictation ability assessment tool for fifth- and sixth-grade students using the Rasch model, and to enhance the screening accuracy for children with developmental dyslexia and dysgraphia through psychometric optimization. Methods A preliminary item bank of 180 Chinese characters was constructed from the lower grade character database of the Ministry of Education's Chinese textbooks.In November 2023, the test was administrated to 595 students in Nantong, Jiangsu Province.Rasch model analysis was used for item selection and model optimization. Results A total of 163 well-fitting items were retained after the screening process.The final word list demonstrated excellent reliability (item reliability=0.975, person reliability=0.956) and high separation indices (item separation=6.32, person separation=4.66).The assumptions of unidimensionality (ratio of the first to second component eigenvalues=9.545; percentage of significant differences in Smith's t-test=5.88%) and local independence (MADaQ3=0.042) were robustly supported.The Wright Map indicated good alignment between item difficulty and person ability distribution.The test information function peaked in the low-ability range (θ≈-2, information value > 25), suggesting superior screening efficacy for children with dyslexia and dysgraphia.The word list showed excellent internal consistency (Cronbach's α=0.975) and high inter-rater reliability (ICC=0.998).ROC analysis indicated high discriminative performance for identifying children at risk of writing difficulties in this sample (AUC>0.99).Using a standard score of 37 as the cutoff, the sensitivity and specificity of this test were 100% and 98.8%, respectively. Conclusions The Mandarin Dictation Test Word List for Upper-grade Primary School Students demonstrates good performance in terms of reliability, validity, and screening sensitivity.It provides an objective framework for early identification of dyslexia and dysgraphia, and lays the foundation for computerized adaptive testing.

关键词

高年级小学生 / Rasch模型 / 听写能力 / 语言筛查

Key words

upper-grade primary students / Rasch model / dictation ability / language screening

[1] 张琳丹.正字法意识、阅读能力与汉字书写能力关系研究[D].深圳:深圳大学,2020.
Zhang LD.The research of relationship between orthographic awareness, reading ability and Chinese character writing ability[D].Shenzhen:Shenzhen University,2020.(in Chinese)
[2] Liu Z, Wang L, Wang T, et al.The developmental dyslexia scale for standard mandarin: A study among early primary students[J].Res Dev Disabil,2024,154:104841.
[3] 孟祥芝.65名儿童中文书写评估及其类型初步分析[J].中国心理卫生杂志,2004,18(12):833-835.
Meng XZ.Assessing and subtypes of writing in Chinese school children[J].Chinese Mental Health Journal,2004,18(12):833-835.(in Chinese)
[4] Ho CSH, Chan DWO, Chung KKH, et al.The Hong Kong test of specific learning difficulties in reading and writing for primary school students-Third Edition (HKT-P(III))[M].Hong Kong: Hong Kong Specific Learning Difficulties Research Team, University of Hong Kong, 2015.
[5] Tsang YK.Dictation and vocabulary knowledge tests for adult native Chinese readers[J].Behav Res Methods,2025,57:151.
[6] 杨慊,贺文洁,王海龙.单参数单维度Rasch模型的优势与意义[J].心理科学,2021,44(6):1491-1498.
Yang Q,He WJ,Wang HL.Advantages and significance of the one-parameter and unidimensional Rasch model[J].Journal of Psychological Science,2021,44(6):1491-1498.(in Chinese)
[7] 丁杨,张傲雪,杨双,等.空间信息加工对听写困难儿童字形记忆的影响[J].中国临床心理学杂志,2021,29(1):14-18.
Ding Y, Zhang AX, Yang S, et al.Effects of spatial information processing on glyph memory in children with spelling difficulties[J].Chinese Journal of Clinical Psychology,2021,29(1):14-18.(in Chinese)
[8] Luan L, Wentao Z, Ming S, et al.CCLOOW: Chinese children's lexicon of oral words[J].Behav Res Methods,2023,56(2):846-859.
[9] 舒华,武宁宁,郑先隽,等.小学汉字形声字表音特点及其分布的研究[J].语言文字应用,1998,2:65-70.
[10] Jiang Y, Zeng Q, Wen H.TestAnaAPP: An interactive R-shiny application for various test analysis methods[J].SoftwareX,2024,28:101967.
[11] Aryadoust V, Ng YL, Sayama H.A comprehensive review of Rasch measurement in language assessment: Recommendations and guidelines for research[J].Language Testing,2021,1:6-40.
[12] Linacre JM.A user's guide to WINSTEPS^® MINISTEP Rasch-model computer programs.Program Manual5.9.2.[EB/OL].(2020-06-25)[2025-01-10].https://www.winsteps.com/winman/copyright.htm.
[13] Ha HT, Nguyen DTB, Stoeckel T.A comparison of Yen's Q3 coefficient and Rasch testlet modeling for identifying local item dependence: Evidence from two vocabulary matching tests[J].Lang Assess Q,2025,22(1):56-76.
[14] Romaní-Romaní F.Cultural adaptation to Peruvian Spanish and psychometric validation of an instrument to measure scientific literacy in medical students[J].Anales de la Facultad de Medicina, 2024, 85(1):34-42.
[15] 武慧多,刘巧云,张艳丽,等.基于Rasch模型的普通话学前儿童词语命名测试词表的编制[J].听力学及言语疾病杂志,2022,30(5):495-500.
Wu HD, Liu QY, Zhang YL, et al.Development of Mandarin vocabulary naming test for preschoolers based on the Rasch model[J].Journal of Audiology and Speech Pathology,2022,30(5):495-500.(in Chinese)
[16] Smith EV Jr.Detecting and evaluating the impact of multidimensionality using item fit statisticsand principal component analysis of residuals[J].J Appl Meas,2002,3(2):205-231.
[17] 温红博,刘先伟,姜有祥.K-means聚类方法在中考标准设定中的信度分析[J].中国考试,2024,8:69-78.
Wen HB, Liu XW, Jiang YX.Reliability analysis of the K-means clustering method in standard setting for junior high school academic proficiency examinations[J].Journal of China Examinations,2024,8:69-78.(in Chinese)
[18] Tesio L, Caronni A, Kumbhare D, et al.Interpreting results from Rasch analysis 1.The“most likely” measures coming from the model[J].Disabil Rehabil, 2024, 46(3):591-603.
[19] 王久菊,孟祥芝,李虹,等.汉语发展性阅读障碍诊断与干预的专家意见[J].中国心理卫生杂志,2023,37(3):185-191.
Wang JJ, Meng XZ, Li H, et al.Expert advice on the diagnosis and intervention of Chinese developmental dyslexia[J].Chinese Mental Health Journal, 2023,37(3):185-191.(in Chinese)
[20] Lee MK, Liu HW, Tong SX.Identifying Chinese children with dyslexia using machine learning with character dictation[J].Sci Stud Read, 2023, 27(1):82-100.
[21] 韩平,杨文艺,谢溢洋,等.智能书写分析与识别平台用于6～12岁儿童书写能力评估的研究报告[J].康复学报,2024,34(3):251-261.
Han P, Yang WY, Xie YY, et al.Research reports on assessment of handwriting ability in children aged 6-12 years by using smart handwriting analysis recognition platform[J].Rehabil Med,2024,34(3):251-261.(in Chinese)
[22] Shune S, Gray LT, Perry S, et al.Validation of the caregiver analysis of reported experiences with swallowing disorders (CARES) screening tool for neurodegenerative disease[J].Am J Speech Lang Pathol, 2025, 34(2): 633-645.

基金资助

国家社会科学基金重点项目(22AYY013)