Multidimensionality and Structural Coefficient Bias in Structural Equation Modeling: A Bifactor Perspective

摘要: In this study, the authors consider several indices to indicate whether multidimensional data are “unidimensional enough” fit with a unidimensional measurement model, especially when goal is avoid excessive bias in structural parameter estimates. They examine two factor strength (the explained common variance and omega hierarchical) model (root mean square error of approximation, comparative index, standardized root residual). These statistics compared population correlation matrices determined by known bifactor structures that vary on (a) relative general group loadings, (b) number factors, (c) items or indicators. When degree coefficient depends strongly inversely variance, but its effects moderated percentage correlations uncontaminated multidimensionality, statistic rise...