Sources of Response Bias in Cognitive Self-Report Items: "Which Memory Are You Talking About?"

Abstract

Self-reported cognitive difficulties are common in the years before an Alzheimer's disease (AD) diagnosis. Understanding how older adults interpret and respond to questions about their cognition is critical to recognize response biases that may limit the accuracy of cognitive self-reports in identifying AD risk. Cognitive interviewing is a systematic approach for examining sources of response bias that influence individuals' answers to specific questions. The purpose of this study was to identify features of common cognitive self-report items that contribute to (a) differing interpretations among respondents and (b) older adults' decisional processes when responding. A convenience sample of community-dwelling older adults (n = 49; Mage = 74.5 years; 36.7% male) without dementia completed a demographic questionnaire, the Montreal Cognitive Assessment, and an audio-recorded cognitive interview. Twenty commonly used cognitive self-report items were evaluated using cognitive interviewing techniques. The Question Appraisal System was used to guide the analysis of interview data and identify sources of response bias within and across cognitive self-report items. The most common sources of inconsistency in item interpretation and decisional processes were vague item wording, incorrect assumptions regarding consistency of cognitive problems across situations, and provocation of an emotional reaction that influenced responses. Assessment of self-reported cognition is critical to facilitate research on early AD symptoms. Findings from this study identify modifiable sources of response bias that may influence the measurement properties of currently used cognitive self-report items and can inform refinement of measures.