"Do I Have a Memory Problem? I Can't Recall": An Evaluation of Measurement Invariance in Subjective Reporting of Memory Symptoms among Persons with and without Objective HIV-Associated Memory Impairment

Abstract

Objective: Memory symptoms and objective impairment are common in HIV disease and are associated with disability. A paradoxical issue is that objective episodic memory failures can interfere with accurate recall of memory symptoms. The present study assessed whether responses on a self-report scale of memory symptoms demonstrate measurement invariance in persons with and without objective HIV-associated memory impairment.

Method: In total, 505 persons with HIV completed the Prospective and Retrospective Memory Questionnaire (PRMQ). Objective memory impairment (n = 141) was determined using a 1-SD cutoff on clinical tests of episodic memory. PRMQ measurement invariance was assessed by confirmatory factor analyses examining a one-factor model with increasing cross-group equality constraints imposed on factor loadings and item thresholds (i.e., configural, weak, and strong invariance).

Results: Configural model fit indicated that identical items measured a one-factor model for both groups. Comparison to the weak model indicated that factor loadings were equivalent across groups. However, there was evidence of partial strong invariance, with two PRMQ item thresholds differing across memory impairment groups. Post hoc analyses using a 1.5-SD memory impairment cutoff (n = 77) revealed both partial weak and partial strong invariance, such that PRMQ item loadings differed across memory groups for three items.

Conclusions: The PRMQ demonstrated a robust factor structure among persons with and without objective HIV-associated memory impairment. However, on select PRMQ items, individuals with memory impairment reported observed scores that were relatively higher than their latent score, while items were more strongly associated with the memory factor in a group with greater memory impairment.

Keywords: Cognitive complaints; Confirmatory factor analysis; Neuropsychological assessment; Psychometrics; Self-report; Subjective memory.

Figure 1.

Confirmatory factor analysis with standardized factor loadings (all ps < .001) for one-factor (general memory) model on the Prospective and Retrospective Memory Questionnaire in the study sample.