Used confirmatory factor analysis (CFA), hypothesized factors that were less than the number of variables, and then examined how well the intercorrelations were reproduced. Results help explain that the type of correlation matrix and estimation method affect factor loadings and fit functions. Suggests some alternative approaches. (SLD)

The Schmid-Leiman (J. Schmid and J. M. Leiman, 1957) decomposition of a hierarchical factor model converts the model to a constrained case of a bifactor model with orthogonal common factors that is equivalent to the hierarchical model. This article discusses the equivalence of the hierarchical and bifactor models. (Author/SLD)

Quality of life (QOL) has become an important concept for health care. As QOL is a multidimensional concept that is best evaluated by a number of latent constructs, it is well recognized that latent variable models, such as exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) are useful tools for analyzing QOL data. Recently,…

Rasch measurement is preferable to factor analysis for reducing complex data matrices to unidimensional variables because factor analysis can mistake ordinally labeled stochastic observations for linear measures, and it does not construct linear measurement. Guidelines and instructions on how to use Rasch measurement to replace factor analysis are…

We illustrate testing measurement invariance in a second-order factor model using a quality of life dataset (n = 924). Measurement invariance was tested across 2 groups at a set of hierarchically structured levels: (a) configural invariance, (b) first-order factor loadings, (c) second-order factor loadings, (d) intercepts of measured variables,…

Scales constructed using principal components and Rasch measurement methods are compared under conditions of unclear constructs and marginal sample sizes with three data sets of increasing complexity. Results of the two methods were identical when data were stable and the structure unidimensional. (SLD)

Two methods, direct maximum likelihood (ML) and the expectation maximization (EM) algorithm, can be used to obtain ML parameter estimates for structural equation models with missing data (MD). Although the 2 methods frequently produce identical parameter estimates, it may be easier to satisfy missing at random assumptions using EM. However, no…

Data from the General Social Survey for 1977 (n=1,303) and 1991 (n=430) are used to examine the stability of attitudes toward abortion and women's roles. Findings support the hypothesis that public attitudes toward discretionary abortion and nontraditional sex roles are discrete but correlated sets of attitudes that remained stable from 1977 to…

The Rasch item fit approach for detecting multidimensionality in response data is compared with principal component analysis without rotation using simulated data. Results indicate that both approaches work in a variety of multidimensional data structures, but the Rasch item fit is better under some circumstances, as discussed. (SLD)

The underlying structure of the Depression scale of the revised Minnesota Multiphasic Personality Inventory (MMPI-2) was studied using dichotomous Rasch model and factor analysis with the normative sample of 2,600 subjects. Results reveal two distinct subscales (mental depression and physical depression) in place of a homogenous set of items. (SLD)