Test fairness is critical to the validity of group comparisons involving gender, ethnicities, culture, or treatment conditions. Detection of differential item functioning (DIF) is one component of efforts to ensure test fairness. The current study compared four treatments for items that have been identified as showing DIF: deleting, ignoring, multiple-group modeling, and modeling DIF as a secondary dimension. Results of this study provide indications about which approach could be applied for items showing DIF for a wide range of testing environments requiring reliable treatment.

中文翻译：

---

差异项目功能的处理：四种方法的比较

测试公平性对于涉及性别、种族、文化或治疗条件的组比较的有效性至关重要。检测差异项目功能 (DIF) 是确保测试公平性的努力的组成部分之一。当前的研究比较了对已确定为显示 DIF 的项目的四种处理方法：删除、忽略、多组建模和将 DIF 建模为次要维度。这项研究的结果提供了关于哪种方法可以应用于显示 DIF 的项目的指示，这些项目适用于需要可靠处理的广泛测试环境。