This study aims to conduct differential item functioning analyses in the context of cognitive diagnosis assessments using various formulations of the Wald test. In implementing the Wald test, two scenarios are considered: one where the underlying reduced model can be assumed; and another where a saturated CDM is used. Illustration of the different Wald test to detect DIF in CDM data was based on the items’ performance of the Proportional Reasoning test among low- and high-performing school students. A benchmark simulation study was included to compare the performance of the Wald test in each scenario. The agreement of the latent attribute classification based on different cognitive diagnosis models was also discussed.

中文翻译：

---

用比例推理测验对CDM数据进行DIF检测的Wald测验公式

这项研究旨在使用Wald检验的各种形式，在认知诊断评估的背景下进行差异项功能分析。在实施Wald检验时，考虑了两种情况：一种可以假定基础简化模型；另一种可以采用基本模型。另一个使用饱和CDM。用来检测CDM数据中DIF的不同Wald检验的说明是基于比例推理检验在低和高中学生中的表现。包含基准测试研究，以比较每种情况下Wald测试的性能。还讨论了基于不同认知诊断模型的潜在属性分类的一致性。