论文部分内容阅读
目的运用项目反应理论(IRT)对慢性病患者生命质量测定量表共性模块(QLICD-GM)条目进行分析,筛选信息量较高条目。方法应用QLICD-GM测评7种慢性病患者620例,采用塞姆吉玛等级反应模型计算每个条目的难度、区分度系数和信息量,绘制项目特征曲线;根据平均信息量筛选条目;采用MULTILOG 7.0软件进行计算和作图。结果 QLICD-GM共性模块29个条目的区分度均为1.2～1.9;难度(程度)均呈严格单调递增,取值范围为-3.05～2.18;依据平均信息量,结合条目特征筛选保留24个条目。结论 QLICD-GM各条目区分度均较好、选项设置合理、难度合适,分析模型选择正确;项目反应理论可筛选出信息量较高条目,弥补经典测量理论(CTT)的不足。
OBJECTIVE To analyze the QLICD-GM items of chronic disease patients by project response theory (IRT) and select the items with high information content. Methods Six hundred and seventy patients with chronic diseases were evaluated by QLICD-GM, the difficulty, discrimination coefficient and information of each item were calculated by using the Semimax-level reaction model, and the project characteristic curve was drawn. The entries were screened according to the average amount of information. Software for calculation and mapping. Results The discriminations of 29 items in QLICD-GM common modules were all from 1.2 to 1.9. The degree of difficulty (degree) increased strictly monotonically, with the range of -3.05 to 2.18. According to the average amount of information, 24 entries . Conclusion The items of QLICD-GM are well-differentiated, the options are reasonable, and the choices are reasonable and the analysis models are correct. The project response theory can filter out the items with higher information to make up for the deficiencies of the classical measurement theory (CTT).