A1117
Title: Differential recall bias in self-reported risk factors in observational studies
Authors: Suhwan Bong - Seoul National University (Korea, South) [presenting]
Kwonsang Lee - Seoul National University (Korea, South)
Francesca Dominici - Harvard University (United States)
Abstract: Observational studies are typically used to estimate the effect of exposures on outcomes. Treatment effect estimation can only be unbiased if the exposure is correctly measured. Recall bias is one of the common reasons for exposure misclassification. Recall bias can occur when study subjects do not remember previous events accurately or omit details. First, the estimand of interest is identified: the average treatment effect (ATE) in the presence of recall bias. Several estimation approaches for the ATE are also developed. These methods are then implemented in simulations to demonstrate their performance in different model misspecification scenarios. Finally, the proposed framework is applied to an observational study, estimating the effect of childhood physical abuse on adulthood mental health.