Bias in Measuring Smoking Behavior

Researchers often use the discrepancy between the self-reported and biochemically assessed active smoking status to argue that self-reported smoking status is not reliable, ignoring the limitations of biochemically assessed measures and treating it as the gold standard in their comparisons. Here, we employ recent advances in econometric techniques to compare self-reported and biochemically assessed smoking data taking into account errors with both methods. Our results suggest that biochemical measures may not always be more reliable than self-reported data.