Beginning in 2025, the NASA Quesst mission will conduct a series of community response tests involving flyovers of the X-59 aircraft at select localities across the United States. Several waves of a longitudinal survey will be administered over approximately one month of testing in order to capture perceptual responses to low-amplitude sonic booms, or “sonic thumps”. Simultaneously, noise exposure levels will be estimated by fusing model-based predictions with measurements taken from a sparse network of monitors in the region. As one of the aims of the study is to produce a dose-response curve, a regression model relating perceptual response to noise exposure levels, it is important to acknowledge the potential attenuation bias that results from measurement error in the estimated noise exposure levels. In this presentation we review and compare several methods for dealing with measurement error in generalized linear mixed models. The methods are demonstrated on simulated data and real data collected during past NASA risk reduction studies.