Re‐investigating the psychometric properties of the Caregiver–Provider Relationship Assessment

INTRODUCTION This study examined the psychometrics of the Caregiver–Provider Relationship Assessment (CPRA) by evaluating whether the negatively worded items impact the fit of the proposed measurement model. METHODS A sample of family caregivers to persons with dementia ( N = 228) completed online surveys about caregiving experiences that included the CPRA. Confirmatory factor analysis was used to test the effects of the negatively worded items. RESULTS Incorporating a method factor substantially improved model fit (Δ comparative fit index = 0.085, Δ Tucker–Lewis Index = 0.098, Δ root mean square error of approximation = 0.044), indicating a systematic method effect attributable to item wording for this instrument. Analysis of subscale models revealed that subscales containing negatively worded items demonstrated poorer model–data fit. DISCUSSION These findings align with prior research discouraging the inclusion of negatively worded items in scales. Negatively worded CPRA items should be revised to a positively worded format to enhance construct validity.