Why Do Study Participants Report Fewer Symptoms and Problems in Later Waves of Longitudinal Studies?

Although epidemiologists and clinical researchers depend on longitudinal assessments to study process and change, for decades it has been known that survey reports of health problems show a pattern where they decline inexplicably in frequency and severity from first to subsequent interviews. This decline has been called the “attenuation effect”, but that description implies that the subsequent measurements are diminished or biased relative to the first. 

In this presentation I will offer possible mechanisms for the Decline After Initial Report (DAIR) pattern, including a therapeutic benefit of repeated assessments a process of learning/optimization of participant responses, and a conversation norm process by which participants try to adjust their reports to match the demand characteristics. I will also present results from four studies that estimate the magnitude of the DAIR pattern and test various explanations for the patterns. These studies assess a variety of outcomes including anxious and depressed mood, positive mood, somatic problems, and alcohol use and craving. One study was a cross sectional survey with four waves spaced two months apart, and the others were daily diary designs. The magnitude of the DAIR effects varied with the nature of the outcome and the context. The results suggest that the initial rather than the subsequent measurements are the ones that are more subject to more response bias.