MMPI偏执分量表的计算机自适应测验可行性分析

Using the Computerized Adaptive Test to Analyze the Feasibility of MMPI Paranoid Subscale

目的:对MMPI偏执分量表的CAT研究的可行性进行探索,并在普通人群和病人中检验CAT测验的结果。方法:首先采用项目反应理论对Pa分量表的单维性进行检验,其次依据项目反应理论挑选项目,形成用于CAT的删减版Pa-D,最后在病人和普通人中验证CAT的可行性。结果:①Pa和Pa-D都能满足单维性的假设,可以用IRT分析项目。②Pa-D的项目参数在合理可接受的范围内。结论:Pa和Pa-D均满足单维性,且删除题目未对测验的信息造成明显的影响。采用后验CAT模拟可以节省一半的题目。

Objective： In this paper,the item response theory was used to analyze the Pa subscale of MMPI.And the item response theory was used to explore the feasibility of the post-hoc CAT of this subscale.In order to test the effect of the post-hoc CAT,the normal people and the patients were the sample used in this study.Methods： There were three steps in total.First,the unidimensionality of the scale was assessed and the item parameters were estimated.Second,according to the parameters,good items were selected to form the Pa-D scale.Third,the effect of the post-hoc CAT was tested in patients and normal people.Results： Both Pa and Pa-D could meet the unidimensional assumption,and could be analyzed under the IRT framework.Pa-D item parameters were in the acceptable range.Conclusion： Pa and Pa-D are unidimensional,and the test information is not obviously influenced by those deleted items.The post-hoc CAT simulation can save half items.