A – 85 A Structural Equation Model Approach to Investigating Relations between Subjective and Objective Memory for Psychometric Equivalence across Race

Abstract Objective The current study assessed the Subjective Memory Complaints Questionnaire (SMCQ) for measurement invariance across race/ethnicity (i.e., Black/African American, Mexican/Mexican American, non-Hispanic White American). We then explored the relations between the SMCQ and an objective memory measure (i.e., Spanish English Verbal Learning Test; SEVLT) for equivalence across race/ethnicity. Method We retrieved data from the Health and Aging Brain Study: Health Disparities dataset. Using structural equation modeling, we first specified a configural model to assess the equivalence of the underlying factor structure across groups. We then specified a scalar model to explore the equivalence of item loadings across groups. To test the equivalence of relations between the SMCQ and SEVLT across groups, we regressed the SEVLT on the SMCQ and specified a multigroup SEM to compare unconstrained and constrained models across race/ethnicity. Results Results support full invariance for the SMCQ (ΔCFI = 0.007, McD= 0.023). Memory complaints significantly predicted objective memory performance for all groups (Black: b = -2.31; Mexican: b = -1.63; White: b = -2.72). However, relations varied across groups [χ2 diff (df) = 20.2.45 (2), p < .001]. Conclusion The SMCQ may be useful as a quick memory screener, and support for measurement invariance indicate it assesses for the same construct in a similar way for Black/African Americans, Mexican/Mexican Americans, and non-Hispanic White Americans. Due to differential relations between subjective and objective memory, future studies are encouraged to explore of the impact of culture (e.g., cultural idioms of distress, somatization) on subjective endorsement of memory impairment.