Using item response theory to examine bias and the factor structure of the Pediatric Symptom Checklist-17 items among racially and income diverse youth and caregivers with Type 1 diabetes

Abstract Objective The use of Item Response Theory for measure development has increased over the years. Components of Item Response Theory serve as the foundation for computerized adaptive tests and can be used to ensure equitable assessment through differential item functioning (DIF) analyses across proxies of intersectional positionality. The focus of this study is to confirm the three-factor structure of the Pediatric Symptom Checklist-17 (PSC-17) and to explore uniform and non-uniform DIF analyses of PSC-17 items among youth living with Type 1 diabetes, a population for which equitable and rapid assessment is essential. Methods Data were drawn from 184 caregiver–youth dyads in the PRYDE sample who completed the 17-item version of the PSC. Youth identified as “Female” (50%), “Male” (50%), Black/AA (56%), and White (43%). Youth ages ranged from 12 to 18 years (M = 14.64, SD = 1.70). Participants’ most recent HbA1c was obtained from their medical records. Results Model fit indices for the initial confirmatory factor analyses of the PSC-17 were acceptable. Evidence for retaining the three-factor model of the PSC-17 was confirmed. Finally, there was no significant uniform or non-uniform DIF present for any of the variables evaluated. Conclusions These findings provide some preliminary evidence of non-DIF, suggesting that the PSC-17 may be meaningfully free of biases for our sample.