Jiang Chunlian, Kim Do-Hong, Wang Chuang
Chunlian Jiang, Faculty of Education, University of Macau, Avenida da Universidade, Taipa, Macau, China,
J Appl Meas. 2018;19(3):283-302.
The purpose of this study is to examine the psychometric properties of an instrument to measure word problem solving skills in mathematics related to speed with 706 sixth grade Chinese and Singaporean students. Rasch measurement models were applied to examine the reliability, unidimensionality, rating scale functioning, item difficulty, and person difficulty. The differential item functioning (DIF) analysis was also performed to examine the differences in item difficulty estimates between Chinese and Singaporean students. Results suggest that the data satisfied the unidimensionality requirements of the Rasch model and that most of the item difficulty measures aligned the person ability distribution. The instrument demonstrated adequate reliability. The fit statistics were within acceptable limits for the vast majority of items, with a few exceptions. The rating scale structure functioned properly although the middle categories had very few observations. Deleting misfitting cases and collapsing middle categories slightly improved the psychometric properties. DIF analysis revealed that four items were more difficult for Chinese students whereas two other items were more difficult for the Singaporean students. Results also indicated that the Chinese participants scored higher than the Singaporean participants for 11 of the 14 items and the Singaporean students scored higher than their Chinese cohorts in the other 3 items. The validation of this instrument has implications for the teaching and learning of mathematical word problems in practice.
本研究旨在对一项用于测量706名六年级中国和新加坡学生数学文字问题解决技能（与速度相关）的工具进行心理测量特性检验。应用拉施测量模型来检验信度、单维度性、评分量表功能、项目难度和人员难度。还进行了项目功能差异（DIF）分析，以检验中国和新加坡学生在项目难度估计上的差异。结果表明，数据满足拉施模型的单维度性要求，且大多数项目难度测量与人员能力分布一致。该工具显示出足够的信度。绝大多数项目的拟合统计在可接受范围内，只有少数例外。评分量表结构运行正常，尽管中间类别观察值很少。删除拟合不佳的案例并合并中间类别略微改善了心理测量特性。DIF分析表明，有4个项目对中国学生来说更难，而另外2个项目对新加坡学生来说更难。结果还表明，在14个项目中的11个项目上，中国参与者得分高于新加坡参与者，在其他3个项目上，新加坡学生得分高于中国同龄人。该工具的验证对实际中数学文字问题的教学和学习具有启示意义。
