We report on a project currently in progress that aims to identify through research the range of factors (individual, school and out-of-school, including home) and their interactions that influence post-16 (i.e. post-compulsory) participation in mathematics and physics in the UK and to assess their relative importance among different student…

An equal-level approach is proposed for investigating multitrait-multimethod (MTMM) matrices with respect to other organizational units that contain additional information concerning a MTMM matrix's validity. The approach requires equality in "data level" before coefficients are submitted for evaluation. Disaggregation is central to…

Applications of the confirmatory factor analysis block-diagonal model to published data on 18 multitrait-multimethod matrices were reviewed to show widespread estimation problems. Possible causes of estimation difficulties were explored using computer simulations. These problems make the block-diagonal approach less useful than has generally been…

The composite direct-product model for the multitrait-multimethod (MTMM) matrix is reparamaterized as a second-order factor analysis model. This process facilitates the use of widely available computer programs--LISREL and LISCOMP--for fitting the model. If bounds on phi and theta are not imposed to avoid inadmissible values, convergence problems…

A modification of the TUCKALS3 algorithm is proposed that handles three-way arrays of order I x J x K for any I. The reduced work space needed for storing data and increased execution speed make the modified algorithm very suitable for use on personal computers. (SLD)

Research on the direct-product model is extended by deriving hierarchically nested models for explicitly testing the patterns of method and trait factors and through formal tests developed for the pattern of communalities. These procedures are illustrated, and use of the MUTMUM computer program is discussed. (SLD)

A generalization of the single trait, multioccasion measurement model is provided through a confirmatory measurement model for multitrait-multioccasion (MTMO) data. Several commonly used confirmatory measurement models are contrasted with the MTMO, and it is illustrated with a large four-wave dataset on family support and alcohol involvement. (SLD)

The multitrait-multimethod (MTMM) paradigm is used widely to assess construct validity, but D. A. Kenny and D. A. Kashy (1992) lament that even after 30 years we still do not know how to analyze MTMM data. The Composite Direct Product (CDP) model has recently attracted considerable attention. Its strengths and weaknesses are evaluated, and a…

Common approaches to examining the relationship between multitrait-multimethod (MTMM) data and variables outside the MTMM data are compared: averaging various means of each trait and estimating LISREL computer program models, and estimating only relationships between MTMM traits and the outside variables. Problems of correlational bias are…

Several multivariate statistical methodologies have been proposed to ensure objective and quantitative evaluation of the multitrait-multimethod matrix. The paper examines the performance of confirmatory factor analysis and covariance component models. It is shown, both empirically and formally, that confirmatory factor analysis is not a reliable…

The relationship of multiple linear regression to various multivariate statistical techniques is discussed. The importance of the standardized partial regression coefficient (beta weight) in multiple linear regression as it is applied in path, factor, LISREL, and discriminant analyses is emphasized. The multivariate methods discussed in this paper…

The way in which trait and rater variance combine in multitrait-multirater (MTMR) performance appraisal data is explored. Implications of the confirmatory factor analytic model and the composite direct product (CDP) model for MTMR data are examined. Superior fit of the CDP model for four data sets is discussed. (SLD)

An application of confirmatory factor analysis, using the LISREL program, to retrospective analysis of rating data common in health professions education was illustrated. A causal model was specified and tested. Although most measures clustered around latent traits, there was more method variance in the data set than trait variance. (Author/JAZ)

Work on item response theory was extended to two areas not extensively researched previously, including models for: (1) test items that require more than one ability for a correct response (MIRT); and (2) interaction between modules of instruction that have a hierarchical relationship (HST). In order to develop the MIRT and HST models, the author…

Describes a new approach for analyzing an individual's responses made at multiple times. Proposes that three sources of variance determine a person's current standing on a variable: trait (term does not change), state (term changes), and error (random term). Shows how structural equation modeling can be used and presents an extended example. (RJM)