一般疏离感量表测试大学生样本的效度和信度

摘要

目的：考察一般疏离感量表（GAS）测试大学生样本的效度和信度。方法：选取1200名大学生进行调查，应用经典测量理论（CTT）和现代测量理论（IRT）进行信效度分析。在 CTT 框架下，首先用所有数据进行条目筛选；然后将数据按录入顺序奇偶分半检验结构效度，其中奇数序号用于探索性因素分析（EFA）和平行分析（PA），偶数序号用于验证性因素分析（CFA）；采用孤独感量表（简版）（ULS-8）、一般自我效能感量表（GSES）和 Rosenberg 自尊量表（RSES）进行效标效度检验；在 IRT 框架下，用评定量表模型分析数据，包括残差主成分分析、条目拟合指标、测量信度（条目的分离指数和信度，被试的分离指数和信度）和怀特图。结果：CTT 框架下，删除条目2和5保留13个条目；EFA 和 PA均表明13个条目的 GAS 为单维结构；CFA 显示，量表拟合指标良好（χ2／df＝2.689，P ＜0.001，CFI ＝0.897，TLI ＝0.877，RMSEA ＝0.055，SRMR ＝0.046）；GAS 总分与 ULS-8、GSES 和 RSES 得分的相关系数为0.57、－0.60和－0.40（均 P ＜0.001）；13个条目量表的 Cronbach α系数及可信区间为0.81（0.80，0.83）。在 IRT 框架下，首对残差特征值之比为1.6；除条目4以外，其余12个条目的 Infit 和 Outfit 值均在0.6～1.4之间；由12个条目组成 GAS 的条目分离指数为10.91，条目信度为0.99，被试分离指数为2.14，被试信度为0.82；条目难度与被试能力相匹配，近似于正态分布。结论：本研究形成的包含12个条目的一般疏离感量表具有良好的效度和信度，可以作为测量大学生疏离感的工具。%Objective:To test the validity and reliability of the General Alienation Scale (GAS)in college students sample.Methods:Totally 1200 college students were surveyed,using classical test theory (CTT)and i-tem response theory (IRT)to test the psychometric properties of GAS.Under the framework of CTT,item selec-tion was performed with all data.Then,the sample was divided into two groups based on the input order to exam-ine the construct validity.Exploratory factor analysis (EFA)and parallel analysis (PA)were performed with odd group,while confirmatory factor analysis (CFA)was performed with even group.Criterion validity was examined with the short-form of the University of California Los Angeles Loneliness Scale (ULS-8),General Self-Efficacy Scale (GSES),Rosenberg Self-Esteem Scale (RSES).Under the framework of IRT,all data was fitted the rat-ing scale model (RSM),including the principal component analysis of residuals (PCAR),the item fit indexes (Infit and Outfit),measurement reliability (item separation index,item reliability,person separation index and person reliability)and the Wright map.Results:Under the framework of CTT,the GAS was retained 13 items except item 2 and 5.The results of both EFA and PA showed that 13 items of GAS was one-dimensional,and there was a good fit of CFA (χ2 /df=2.689,P <0.001,CFI =0.897,TLI =0.877,RMSEA =0.055,SRMR =0.046).The correlation coefficients were 0.57, -0.60 and -0.40 with the GAS,ULS-8,GSES,RSES scores respectively (Ps <0.001 ).The Cronbach 's αcoefficient and confidence interval were 0.81 (0.80, 0.83).Under the framework of IRT,the ratio of first-paired eigenvalue was 1.6 in PCAR.The infit and outfit of retaining 13 items in CTT were from 0.6 to 1.4 excluding item 4.The item separation index,item reliability,per-son separation index and person reliability of the 12-item GAS were10.91,0.99,2.14 and 0.82 respectively.The item difficulty can matched with person ability,and they were closed to the normal distribution.Conclusion:It suggests that the General Alienation Scale (GAS)is of good validity and reliability,and it could be used as an ef-fective tool for measuring the alienation of college students.