Bowden Stephen C, Lissner Dianne, McCarthy Kerri A L, Weiss Lawrence G, Holdnack James A
Department of Psychology, The University of Melbourne, Parkville, Victoria, Australia.
J Clin Exp Neuropsychol. 2007 Oct;29(7):768-80. doi: 10.1080/13803390601028027.
Equivalence of the psychological model underlying Wechsler Adult Intelligence Scale-Third Edition (WAIS-III) scores obtained in the United States and Australia was examined in this study. Examination of metric invariance involves testing the hypothesis that all components of the measurement model relating observed scores to latent variables are numerically equal in different samples. The assumption of metric invariance is necessary for interpretation of scores derived from research studies that seek to generalize patterns of convergent and divergent validity and patterns of deficit or disability. An Australian community volunteer sample was compared to the US standardization data. A pattern of strict metric invariance was observed across samples. In addition, when the effects of different demographic characteristics of the US and Australian samples were included, structural parameters reflecting values of the latent cognitive variables were found not to differ. These results provide important evidence for the equivalence of measurement of core cognitive abilities with the WAIS-III and suggest that latent cognitive abilities in the US and Australia do not differ.
本研究考察了在美国和澳大利亚获得的韦氏成人智力量表第三版（WAIS - III）分数所基于的心理模型的等效性。测量不变性检验涉及检验这样一个假设：在不同样本中，将观测分数与潜在变量联系起来的测量模型的所有组成部分在数值上是相等的。测量不变性假设对于解释来自试图推广收敛效度和发散效度模式以及缺陷或残疾模式的研究得出的分数是必要的。将澳大利亚社区志愿者样本与美国标准化数据进行了比较。在不同样本中观察到了严格测量不变性的模式。此外，当纳入美国和澳大利亚样本不同人口统计学特征的影响时，发现反映潜在认知变量值的结构参数没有差异。这些结果为使用WAIS - III测量核心认知能力的等效性提供了重要证据，并表明美国和澳大利亚的潜在认知能力没有差异。
