Multiple Replication of Factor Structure: A Logical Solution for a Number of Factors Problem.

Area of Science:

• Psychometrics
• Statistical analysis
• Psychological measurement

Background:

• Determining the correct number of factors in factor analysis is a persistent challenge, particularly with multiple-scale questionnaire data.
• Existing methods may not adequately address the replicability of factor structures across different subject groups.

Purpose of the Study:

• To introduce and demonstrate a novel factor replication procedure for solving the "numbers of factors problem" in questionnaire analysis.
• To provide a robust method for validating factor structures across independent subject groups.

Main Methods:

• A series of factor analyses were performed with progressively fewer factors.
• Factor similarity was assessed using the factor matching index across three subject groups.
• The procedure was applied to responses from the Hopkins Symptom Checklist.

Main Results:

• The proposed factor replication procedure successfully identified replicable factor structures.
• The commonly accepted five-factor structure of the Hopkins Symptom Checklist was not supported by the data.
• Clear evidence for a three-factor structure emerged, replicated across subject groups.

Conclusions:

• The factor replication procedure offers a significant improvement for deciding the appropriate number of factors.
• The findings suggest a revised three-factor structure for the Hopkins Symptom Checklist.
• This method enhances the reliability and validity of factor analysis in psychological research.