A Confirmatory Factor Analysis of the GHQ12 Using a Large Australian Sample

Abstract

The GHQ12 has become a popular screening instrument used by both clinicians and researchers. To date there have been a number of studies investigating the factor structure of the GHQ12. Most have identified a factorial structure to the instrument but there has been little consistency in the factors identified. One reason for this may be that different investigators are using different scoring methods and this could impact the structure identified. This study used structured equation modeling (LISREL 8.7) to carry out a confirmatory factor analysis, with a large sample of 8,732 university staff, on four 3-factor and one unitary model solutions using the three most common scoring methods (Classical, Corrected, and Likert). It was found that the method of scoring substantially affects model estimation and suggestions were made as to the best scoring method for future use.