(PS7-18) Measurement Invariance in the Short Version of the Saving Inventory-Revised Across Race and Ethnicity

Epidemiological studies indicate that hoarding disorder is prevalent worldwide, exhibiting similar characteristics across diverse ethnic and demographic groups. However, the empirical literature on hoarding assessment has primarily relied on predominantly white samples from Europe, Australia, and the United States. This lack of diversity limits our understanding of potential racial, ethnic, or cultural variations in hoarding manifestation, with consequences for the valid and reliable measurement of hoarding symptoms. Race and ethnicity are relevant in the study of hoarding insofar as they may reflect or capture underlying sociocultural factors that could influence symptom presentation, with important implications for assessment and diagnosis.

Our study sought to examine the equivalence of hoarding assessments in underrepresented racial and ethnic groups by testing for measurement invariance of the 9-item version of the Saving Inventory-Revised (SI-R9; Frost et al., 2004; Sarvestani et al., 2025). We recruited 1,822 individuals currently residing in the United States through online crowdsourcing platforms. We created three groups based on self-identified race: Asian (n = 192), Black (n = 222), and White (n = 1188). The same participants were also divided into two groups based on self-identified ethnicity: Hispanic/Latino (n = 222) and non-Hispanic/Latino (n = 1365). We compared a three-factor measurement model of the SI-R9 (including clutter, difficulties discarding, and acquiring) across racial groups and ethnic groups using multigroup confirmatory factor analysis.

Our results suggest that the SI-R9 has a similar three-factor structure in Asian, Black, and White participants (2(51) = 112.28, p < .001; Robust CFI = .990; Robust RMSEA = .056, 90% CI [.042, .071]; SRMR = .020) and in Hispanic/Latino and non-Hispanic/Latino participants (2(40) = 106.56, p < .001; Robust CFI = .988; Robust RMSEA = .055, 90% CI [.042, .068]; SRMR = .023). This indicates that the measure captures the same general constructs in our participant groups. However, there were item-level differences across groups. One clutter item’s factor loading and intercept were noninvariant across the ethnic groups, which suggests that this item may have different meanings or interpretations in Hispanic/Latino and non-Hispanic/Latino participants. One acquiring item’s intercept was noninvariant across our racial groups, which indicates that observed scores may differ across the groups due to influences not captured by the SI-R9. Overall, the SI-R9 demonstrated partial metric and scalar invariance across our ethnic groups and partial scalar invariance across our racial groups.

Measurement noninvariance may introduce bias into our interpretation of observed scores on the SI-R9. Direct group comparisons should be interpreted with caution, as differences in observed scores may reflect something other than true differences in hoarding symptom severity across underrepresented racial and ethnic groups. Future work should further investigate the influence of race, ethnicity, and culture on hoarding manifestation and assessment.