Direct Schmid-Leiman Transformations and Rank-Deficient Loadings Matrices.

Area of Science:

• Psychometrics
• Individual Differences Research
• Factor Analysis

Background:

• The Schmid-Leiman (S-L) transformation is a widely used technique for exploratory bifactor analysis in individual differences research.
• Typically, a two-level S-L transformation involves two sequential factor analyses: a first-level oblique rotation and a second-level general factor extraction.
• This conventional approach is computationally intensive and may not be the most efficient method.

Purpose of the Study:

• To demonstrate that the Schmid-Leiman loadings matrix is rank deficient.
• To show how this rank deficiency allows for a direct S-L solution from an unrotated first-level factor structure.
• To illustrate the application of direct bifactor solutions for both hierarchical and non-hierarchical structures.

Main Methods:

• Demonstration of the rank deficiency inherent in the Schmid-Leiman transformation matrix.
• Development and application of a direct method for obtaining S-L solutions from unrotated factor structures.
• Reanalysis of existing examples to showcase the utility of 'best-fitting' S-L rotations in hierarchical models.
• Computation of direct bifactor solutions for non-hierarchical structures.

Main Results:

• The Schmid-Leiman loadings matrix is necessarily rank deficient.
• A direct S-L solution can be obtained from an unrotated first-level factor structure by leveraging this rank deficiency.
• The direct method simplifies the process and provides a more efficient approach to bifactor analysis.
• The reanalysis confirms the utility of 'best-fitting' S-L rotations for evaluating hierarchical factor models.

Conclusions:

• The conventional two-step procedure for Schmid-Leiman transformation is not strictly necessary.
• A direct Schmid-Leiman solution can be efficiently computed, streamlining exploratory bifactor analysis.
• This simplified approach enhances the practical application of bifactor models in psychological research and individual differences studies.