A new basic algorithm is discussed that may be used to do factor analysis by any of these three methods: (1) unweighted least squares, (2) generalized least squares, or (3) maximum likelihood. (CK)

Given multiple independent measures of an underlying true factor and information on group membership, it is possible to compute a set of observed group means for each measure. Given at least three tests, these sets of means may be used to compute the reliability of the means for each test. The procedure for estimating true scores from the…

The mathematical derivation of the statistics used for inference in some linear models assumes that values of the independent variables are pre-selected such that these variables can be treated as fixed rather than random variables. This assumption is often disregarded when these models are utilized in research. This study is an investigation of…

Path analyses of two multitrait-multimethod matrices are used as examples of the kind of information afforded by application of the technique. It is concluded that the technique should be of considerable aid to researchers who want to evaluate the convergent and discriminant validity of their measures. (Author/CTM)

The hypothesis that fluid intelligence is equivalent to the factor of general intelligence is investigated using LISREL to specify higher-order models in reanalyses of three sets of psychometric data from subjects 11, 13, and 15 years old. The three studies showed fluid intelligence to be equivalent with a general factor. (LMO)

Confirmatory factor analysis was applied to test second- and higher-order factor models in the areas of structure of abilities, allometry, and the separation of specific and error variance estimates. The estimation of validity and reliability, second-order models within factor analysis models, and the concept of discriminability were also studied.…

Correlated Factors Modeling (CFM) is proposed as an alternative to LISREL and similar methods of estimating structural equations. CFM employs weak mathematics and is appropriate for a rapid, preliminary look at a model. The factors produced by CFM are exact linear components of their indicator variates. (Author/GDC)

The present study compares the performance of phi coefficients and tetrachorics along two dimensions of factor recovery in binary data. These dimensions are (1) accuracy of nontrivial factor identifications; and (2) factor structure recovery given a priori knowledge of the correct number of factors to rotate. (Author/LMO)

Statistical methods, called latent variable models, have been developed to provide rigorous tests of theories involving unobserved variables. This paper describes the major types of latent variable models, shows how they can be applied in educational research, and gives representative examples of their use from the literature. (Author/BW)

Concerned with consequences of employing the INDSCAL model when one of its assumptions are known to be violated. Under study is the notion that all individuals perceive the object space dimensions to be independent. (RC)

Study examined models for improving prediction of single and multiple criteria. (Author)

This paper describes a variety of confirmatory factor analysis models that provide improved tests of multitrait-multimethod matrices, and compares three different approaches (the original Campbell-Fiske guidelines, an analysis of variance model, and confirmatory factor analysis models). (PN)

Factor analysis in several populations, covariance structure models, three-mode factor analysis, structural equations systems with measurement model, and analysis of covariance with measurement model are all shown to be specializations of a general moment structure model. Some new structured linear models are also described. (Author/CTM)

A statistical model for relating latent variables from manifest variables, called a linear structural equation model, is presented. The approach is illustrated by a test theory model and longitudianl study of intelligence. (Author/JKS)

Using the item response model as developed on the multinomial distribution, asymptotic variances are obtained for residuals with response patterns and first- and second-order marginal frequencies of manifest variables. A limited-information test of fit is developed by using residuals defined for the first- and second-order marginals. (Author/SLD)