An introduction to the analysis of covariance structures: a general model for data analysis

The analysis of covariance structures is presented as a general method for examining structural relationships among multiple measured variables and where appropriate, testing complex multivariate hypotheses. Mcdonald has provided a very general approach to this problem which defines a specific element of a covariance matrix as some function of a set of parameters theta'. His matrix specification for the structure of a covariance matrix seems to handle most of the current models for linear relationships among random variables. The generality of his approach is illustrated by using it to specify a classical common factor model, a three mode common factor model, and a linear structural relations model. In each case road research examples are discussed with a view to showing the usefulness of the technique for the investigation of current research problems (a).

Language

  • English

Media Info

  • Pagination: 49-60

Subject/Index Terms

Filing Info

  • Accession Number: 01438212
  • Record Type: Publication
  • Source Agency: ARRB
  • ISBN: 0869102184
  • Files: ATRI
  • Created Date: Aug 24 2012 9:26PM