Factor scoring models of the Pittsburgh Sleep Quality Index: a comparative confirmatory factor analysis

Abstract

The Pittsburgh Sleep Quality Index (PSQI) is a rigorously validated questionnaire with extensive use in sleep assessment. Findings from numerous factor analytic studies of the PSQI have been interpreted to support a heterogeneous factor structure model for the test. Nevertheless, the literature continues to lack a focused evaluation of whether this heterogeneous factor structure is justified. A consideration of this issue led to a conclusion that a closer analysis of the PSQI’s factor structure was merited. To address this need a comparative confirmatory factor analysis for assessing the performance of the accepted factors models of the PSQI was conducted. A sample of university students (n = 418), age = 20.92 ± 1.81 years, BMI = 23.30 ± 2.57 kg/m² completed the multi-structured sleep survey at Jamia Millia Islamia, New Delhi, India. Seventeen putative factor structures (three 1-Factor, eight 2-Factor, and six 3-Factor) of the PSQI from the existing literature were selected for analysis. Fourteen models (82.35%) had almost similar values for model fit indices. Two models were misfits, and one model was a poor fit. The two misfit models incorporated gender and age as covariates. The third poor fit model was used to produce a unique path diagram, which made it distinct from the remaining 16 models. The overlapping values in the fit range of the model fit indices did not support the often projected heterogeneous factor structures of the PSQI for the vast majority of the models.

Keywords:

• Fit index
• dimensionality
• questionnaire
• exploratory factor analysis
• sample adequacy
• communality