A latent factor approach to the Saving Inventory - Revised: Congeneric evaluation of construct and content validity

Background: Inconsistencies have been identified in the three-factor structure and item loadings of the most commonly used self-report hoarding screening tool, the Saving Inventory - Revised (SI-R), which assesses dif-ficulty discarding, clutter and acquisition. The current study aimed to confirm the factor structure of the SI-R using congeneric modelling, and evaluate the construct and content validity of this measure. Methods: 139 participants with self-identified hoarding completed the SI-R. Congeneric structural equation modelling was then performed to validate the SI-R factor structure. Results: The three-factor structure of the SI-R was confirmed as a valid, reliable and good fitting model. However, the difficulty discarding and clutter subscales were required to covary. Conclusions: The SI-R was confirmed as an appropriate screening tool for hoarding severity; however, revision of item wording may improve content validity. Future research could consider exploring the relationships between a range of hoarding-related constructs and the differential endorsement of SI-R subscales. Limitations: As data were collected during the 2020 COVID-19 pandemic, panic buying and hoarding-related acquiring behaviours may have been exacerbated. In addition, the diagnostic status of participants was not verified, despite the inclusion of individuals endorsing clinically significant hoarding symptoms.

Automated summary

The current study confirmed the three-factor structure of the Saving Inventory - Revised (SI-R) and evaluated its construct and content validity using congeneric modelling. The results indicated that the SI-R is a valid, reliable, and well-fitting measure, although the difficulty discarding and clutter subscales needed to be covaried. The study concluded that the SI-R is an appropriate screening tool for hoarding severity, and suggested revising the item wording to improve content validity. Future research could explore the relationships between various hoarding-related constructs and the differential endorsement of SI-R subscales.