Epkins C C, Meyers A W
Department of Psychology, Auburn University, AL 36849-5214.
J Pers Assess. 1994 Apr;62(2):364-81. doi: 10.1207/s15327752jpa6202_16.
Childhood depression, anxiety, and aggression assessment instruments were examined in a sample of 8- to 11-year-old elementary school children. A multitrait-multimethod design was used to evaluate self-, parent-, teacher-, and peer-report on each of the three traits. Results found the measures exhibited significant convergent and discriminant validity for both boys and girls. However, different patterns in convergence among the information sources emerged between boys and girls. Both the intercorrelations and the analyses of variance documented significant method variance. The implications of the high method variance or source bias in the assessment process are discussed. It is argued that investigators employ confirmatory factor analyses in evaluation of multitrait-multimethod data.
在一个8至11岁的小学生样本中，对儿童抑郁、焦虑和攻击性评估工具进行了研究。采用多特质-多方法设计来评估关于这三种特质的自我报告、家长报告、教师报告和同伴报告。结果发现，这些测量方法对男孩和女孩都表现出显著的聚合效度和区分效度。然而，男孩和女孩在信息来源之间的聚合模式有所不同。相互关联以及方差分析都证明了显著的方法方差。讨论了评估过程中高方法方差或来源偏差的影响。有人认为，研究者应在多特质-多方法数据的评估中采用验证性因素分析。
