The challenges in measuring symptoms of schizophrenia: An exploratory graph analysis of the Positive and Negative...

The Positive and Negative Syndrome Scale (PANSS) is a widely used instrument for assessing symptom severity in schizophrenia. Significant concerns remain regarding its structural validity, with no consensus on a definitive model. Although the five-factor model (positive, negative, cognitive, hostility, and affective) is the most widely accepted, studies have demonstrated instability in item distribution across factors and difficulty in achieving adequate fit indices. This exploratory study aims to evaluate the structural validity of the PANSS, addressing these previously identified issues and contributing to the theoretical debate on the limited evidence supporting a robust model fit.

Data from five independent Brazilian samples (N = 1262) were analyzed. We conducted Exploratory Graph Analysis (EGA) and bootstrap analysis to assess the stability of the PANSS structure. We estimated network loadings and conducted the Loadings Comparison Test (LCT) to evaluate differences between structural frameworks.

The analysis identified a five-dimensional model consistent with prior literature. However, this model proved unstable, appearing in only 66.3 % of bootstrap samples. Several items exhibited low stability across dimensions, and the LCT algorithm indicated that a network framework provided a better fit for the data.

In line with previous research, this study found that the PANSS exhibits an unstable structure and item distribution. Moreover, the scale presents several validity limitations, and the difficulty in achieving a consolidated model may reflect both methodological gaps in its original construction and the inherently multidimensional nature of certain symptoms.