Test Construction and Targeted Factor Solutions Derived by Multiple Group and Procrustes Methods.

Area of Science:

• Psychometrics
• Psychological Measurement
• Factor Analysis

Background:

• Psychological measures require empirical verification of their hypothesized multidimensional structures.
• Existing confirmatory factor analysis strategies include multiple group analysis, targeted rotations, and item-total correlations.

Purpose of the Study:

• To empirically verify the assumed structures of data from four psychological test batteries.
• To compare the interpretability and utility of different confirmatory factor analysis strategies.

Main Methods:

• Applied multiple group factor analysis.
• Utilized targeted (Procrustean) rotations of factor solutions.
• Employed product-moment correlations of item responses with total scale scores.
• Conducted supplementary analyses using confirmatory maximum likelihood factor analysis.

Main Results:

• Solutions from multiple group analysis and item-total correlation analysis were generally the most psychologically interpretable.
• Confirmatory maximum likelihood and Procrustean rotation methods yielded less interpretable solutions.

Conclusions:

• Multiple group analysis and item-total correlation analysis are preferred methods for test construction due to their interpretability.
• These methods offer a more practical approach to verifying psychological measure structures compared to Procrustean or confirmatory maximum likelihood methods.