中国临床心理学杂志

个人拟合指标在艾森克人格测验中的应用

作者：刘拓曹亦薇戴晓阳

分类号：B841.7

出版年,卷(期):页码：2011,19(3):323-326

摘要：

目的：用个人拟合指标对EPQ量表数据进行分析，探讨不拟合现象的可能成因和不拟合被试对参数估计的影响。方法：用lz和Guttman错误检查不拟合情况，用RMSD、MD指标及ICC面积法考察参数估计结果的差异。结果: ①lz<；-1.65的被试占总人数的3.44%，lz与Guttman错误的相关为-0.838；②最不拟合的50人中，编号相连被试共有32人；③删去不拟合被试前后，特质参数差异很小，高区分度高难度和高区分度低难度项目参数差异大。结论：个人拟合指标可以找到团体测验中来自同一测试环境的异常反应被试集团；不拟合被试的存在，对被试特质参数的估计影响不大，但会造成对项目区分的低估，进而影响难度越大和越小的项目。

Objective:Person-fit indices were computed on EPQ.This study was to investigate the cause of misfit andthe influence of the inclusion of misfitting persons in parameter estimates.Methods:lz and Guttman error were used to investigate the misfit.MD,RMSD indices and ICC method was used to evaluate the effect of deletion of mifitting persons.Results:①lz<-1.65 was found in 3.44% of the examinees,and lz was correlated -0.838 with Guttman error;②There were32 in 50 most misfitting persons ranked in adjacent numbers;③It was found that person parameters changed little afterdeleting misfitting persons.However,the high α high β items and the high α low β items changed a lot.Conclusion:Agroup of aberrant response persons in the same testing environment were found out by person-fit indices in group test.Thepresence of misfitting person cannot influence person parameters much,but resulted in underestimating discrimination parameters,and then influenced the bias in estimation of higher or lower threshold parameters.

基金项目：

作者简介：

参考文献：

[1] 刘仁刚.条目反应理论简述.中国临床心理学杂志,2009,17(1):37-41;
[2] Drasgow F,Levine M,Williams EA.Appropriateness measurementwith polychotomous item response models andstandardized indices.British Journal of Mathematical and Statistics Psychology,1985,38:67-86;
[3] Nering ML,Meijer RR.A comparison of the person responsefunction and the lz person-fit statistic.Applied PsychologicalMeasurement,1998,22:53-69;
[4] Meijer RR,Sijtsma K.Methodology review:Evaluating personfit.Applied Psychological Measurement,2001,
[25] (2):107-135;
[5[Meijer RR.The number of guttman errors as a simple andpowerful person-fit statistic.Applied Psychological Measurement,1994,18:311-314;
[6] Meijer RR,Egberink IJL,Emons WHM,et al.Detection andvalidation of unscalable item score patterns using item responsetheory:An illustration with harter’s self-perceptionprofile for children.Journal of Personality Assessment,2008,90(3):227-238;
[7] Reise SP,Due AM.The influence of test characteristics onthe detection of aberrant response patterns.Applied Psy-chological Measurement,1991,15(3):217-226;
[8] Karabatsos G.Comparing the aberrant response detectionperformance of thirty-six person-fit statistics.Applied Measurementin Education,2003,16(4):277-298;
[9] Drasgow F,Levine MV,McLaughlin ME.Detecting inappropriatetest scores with optimal and practical appropriatenessindices.Applied Psychological Measurement,1987,11:59-79
[10] Meijer RR.Diagnosing item score patterns on a test usingitem response theory-based person-fit statistics.Psycho-logical Methods,2003,8(1):72-87
[11] 王昭,郭庆科,岳艳.心理测验中个人拟合研究的回顾与展望.心理科学进展,2007,3:559-566
[12] Zickar MJ,Drasgow F.Detecting faking on a personality instrumentusing appropriateness measurement.Applied PsychologicalMeasurement,1996,20:71-87
[13] 曹亦薇.异常反应模式的识别和分类.心理学报,2001,6:558-563
[14] Meijer RR,Nering ML.Trait level estimation for nonfittingresponse vectors.Applied Psychological Measurement,1997,21(4):321-336
[15] Schmitt N,Cortina JM,Whitney DJ.Appropriateness fit andcriterion-related validity.Applied Psychological Measurement,1993,17:143-150
[16] Phillips SE.The effects of the deletion of misfitting personson vertical equating via the rasch model.Journal of EducationalMeasurement,1986,23(2):107-118
[17] Curtis DD.Person misfit in attitude surveys:Influences,impactsand implications.International Education Journal,2004,5(2):125-143
[18] Embretson SE,Reise SP.Item response theory for psychologists.Mahwah,NJ:Lawrence Erlbaum,2000
[19] Ferrando PJ,Chico E.Detecting dissimulation in personalitytest scores:A comparison between person-fit indices anddetection scales.Educational and Psychological Measure-ment,2001,61(6):997-1012
[20] Meijer RR.The influence of the presence of deviant itemscore patterns on the power of a person-fit statistic.AppliedPsychological Measurement,1996,20:141-154

服务与反馈：

【文章下载】【加入收藏】