In many epidemiologic studies addressing exposure-response relationships, sources of error that lead to systematic bias in exposure measurements are known to be present, but there is uncertainty in the magnitude and nature of the bias. Two approaches that allow this uncertainty to be reflected...