The Structure of the Prodromal Questionnaire‐16 (PQ‐16): Exploratory and confirmatory factor analyses in a general non‐help‐seeking population sample. Issue 3 (24th March 2021)

• Record Type: Journal Article
• Authors: Howie, Clare; Hanna, Donncha; Shannon, Ciaran; Davidson, Gavin; Mulholland, Ciaran
• Is Part Of: Early intervention in psychiatry. Volume 16:Issue 3(2022)
• Page Start: 239
• Page End: 246
• Publication Date: 2021-03-24
• Subjects: attenuated symptoms -- Prodromal Questionnaire‐16 -- psychosis -- factor analysis -- screening; Mental health -- Periodicals; Psychiatry -- Periodicals; Psychiatry -- Research -- Periodicals; Mental illness -- Prevention -- Research -- Periodicals; Mental illness -- Treatment -- Research -- Periodicals; 616.89
• DOI: 10.1111/eip.13147
• Languages: English
• ISSNs: 1751-7885

Abstract

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.