A Monte Carlo study determined the Inverse Scree Test's shape with various numbers of true groups and under different conditions of distribution shape and sample size. Six simulated distributions of 3,000 subjects each and 1 with 1,500 were created. Findings suggest relative distribution independence, number independence, and modest…

A telephone survey of 129 Iowa adults who had not completed high school determined reasons for lack of participation in adult basic education. Factor analysis of 32 reasons yielded 4 interpretable factors: low perception of need, perceived effort, dislike of school, and situational barriers. Low perception correlated with age and situational…

Homogeneity analysis (multiple correspondence analysis), which is usually applied to "k" separate variables, was applied to sets of variables by using sums within sets. The resulting technique, OVERALS, uses optimal scaling. The corresponding OVERALS computer program minimizes a least squares loss function via an alternating least…

Investigated factor structure of the Stanford-Binet Fourth Edition using collaboration matrix from the standardization sample of 1,695 children aged 2 through 6 years, 1,586 children aged 7 through 11 years, and 1,728 participants aged 12 through 23 years. The results revealed the large first factor that resembled general intelligence, with little…

To deterine whether three empirically derived factors of Oddi Continuing Learning Inventory (OCLI) remain across study samples, factor scores pairs were generated by using factor-score coefficients of two data sets and correlating factor scores. High positive correlations indicated three derived factors in factor structure matrix matched those…

The comparability of seven sets of factor structure studies reported in previous research with the Bem Sex-Role Inventory was evaluated. The inventory's primary factors were reproduced in the studies, although the masculine scale was somewhat less invariant than was the feminine counterpart. (SLD)

A simple procedure for establishing convergent and discriminant validity is presented, as an alternative to the LISREL-based nested model used by R. P. Bagozzi (1978) and K. F. Widaman (1985). Ordinary least-squares regression is used, with the correlation between measures as the dependent variable. (SLD)

Focusing on linear interpolation, an accurate method of implementing parallel analysis, this article contains tables of 95th percentile eigenvalues from random data than can be used with sample sizes of 50 to 500 subjects and between 5 and 50 variables. An empirical example illustrates how to obtain the eigenvalues. (SLD)

Six community workers' interviews were subjected to vocabulary scanning; factorial analysis graphics plotted words most frequently used. The workers' world views and understanding of social change emerged. One group favored local development, planned change, a facilitator role, and collaboration. The other preferred a social action model of…

Benefits and limitations of structural equation models for multivariate experiments with incomplete data are presented. Examples from studies of latent variable path models of cognitive performance illustrate analyses with latent variables, omitted variables, randomly missing data, and nonrandomly missing data. (SLD)

Describes exploratory and confirmatory analyses of verbal-performance procedures to illustrate concepts and procedures for analysis of correlated factors. Argues that, based on convergent and discriminant validity criteria, factors should have higher correlations with variables that they purport to measure than with other variables. Discusses…

Offers alternative analyses and interpretations of Macmann and Barnett's research data and results. Contends that intelligence is complex and that intelligence testing is important component of school psychological practice. Discusses relation and frequent inconsistency between research and practice. (CRR)

Tested one-, two-, three-, and four-factor models within normal and psychiatric adolescent inpatient groups to confirm the factor structure for the Wechsler Intelligence Scale for Children-Revised (WISC-R). For both samples, the Kaufman three-factor solution had the best overall fit of the WISC-R subtest covariance structure. Other models were…

A modification of Horn's parallel analysis is introduced that is based on the Monte Carlo simulation of the null distributions of the eigenvalues generated from a population correlation identity matrix. This modification reduces the tendency of the parallel analysis procedure to overextract or to extract poorly defined factors. (SLD)

The concordance and factor structures of the Beck Depression Inventory and the Hamilton Rating Scale for Depression were evaluated with 259 outpatients being treated for depression. The two scales were significantly correlated and assessed similar rates of improvement at multiple assessment points, but emphasized different dimensions of…