Should we worry about how we measure worry? Insights from an updated version of the Italian Penn State Worry Questionnaire

Background: Research indicates a rise in self-reported worry, highlighting the need for updated psychometric tools. The Penn State Worry Questionnaire (PSWQ) assesses excessive worry and there is debate over whether only its 11 positively worded items should be used. This study aimed to clarify the factor structure and psychometric properties of the PSWQ and to explore worry features in two diverse Italian community samples from the 2010s and 2020s. Methods: The 2020s sample included 674 participants (44.5 % female; Mage = 29.44 ± 13.20), while the 2010s sample comprised 411 individuals (61.6 % female; Mage = 36.64 ± 13.73). Methods from Classical Test Theory (CTT) were used to compare alternative PSWQ factor structures, assess the best-fitting model's reliability and validity, and evaluate measurement invariance (MI) across sexes in the 2020s sample. Item Response Theory (IRT) was applied to refine and confirm the best-fitting factor structure and to compare item and individual locations across samples. Results: The 11-item one-factor model was the best fit and it showed excellent reliability and concurrent validity. MI across sexes was supported. IRT analyses suggested that items were slightly more difficult for the 2010s sample. Conclusions: The PSWQ-11 is a valid and reliable tool for assessing worry in the Italian community. The findings suggest that societal issues as well as socio-demographic characteristics may contribute shaping differences in worry features across diverse historical contexts.