Fig. 9
From: Identification of confounder in epidemiologic data contaminated by measurement error in covariates
Anticipated estimates of the exposure-outcome association in the motivating example (see text for details) after adjustment for a miss-measured confounder when there is not a true exposure-outcome association. The unadjusted RR ED|Fobs is 0.83 (i.e. confounded; indicated by the arrow), noise-to-signal ratio is 1 and all effects of confounder are expected to be not statistically significant (p >0.05); 10,000 simulations