The Structure of the Prodromal Questionnaire‐16 (<scp>PQ</scp>‐16): Exploratory and confirmatory factor analyses in a general <scp>non‐help‐seeking</scp> population sample
Clare Howie, Donncha Hanna, Ciarán Shannon, Gavin Davidson, Ciaran Mulholland
Abstract
AIMS: To examine the structure of the Prodromal Questionnaire (PQ-16) in a non-help-seeking population through exploratory factor analysis and confirmatory factor analysis. Previous studies have not looked at the structure of this self-report measure outside clinical settings. METHODS: Participants (n = 1045) were recruited through Amazon's Mechanical Turk (MTurk), and then completed the PQ-16. The data set was split randomly in two, one being used for exploratory factor analysis (EFA) and the other for confirmatory factor analysis (CFA). A polychoric correlation matrix was created and EFA was used to explore the factor structure of the PQ-16. Four models were tested through CFA to determine best fit: one, two, three and four-factor models were all analysed. RESULTS: EFA indicated a two-factor structure in the PQ-16 in a non-help-seeking population (with a mean age = 29.7 years). Factor 1 represented perceptual abnormalities/hallucinations and factor 2 general symptoms associated with psychosis-risk. CFA indicated that all the proposed models were suitable fits for the dataset. Fit indices for the three-factor model (factor 1 representing perceptual abnormalities/hallucinations, factor 2 unusual thought content, and factor 3 negative symptom) indicated that it appeared to be a better fit for the data than the one, two, and four factor models. CONCLUSIONS: This study suggests that a three-factor model of the PQ-16 is a better fit than other proposed models in a non-help-seeking population. Future research of the structure of the PQ-16 in this population may benefit from recruiting subjects with a lower mean age than the current study.