Abstract
The effect of non differential and independent declassification bias towards the disease, the expo sure and third va ri able. The objective of this study was the determination of the influence of the non-differential and independent declassification bias for the disease, the exposure and another variable in the association measure, and also, the obtained pattern. In order to do this, we did some simulations, manipulating the sensibility and specificity of the disease, the exposure and the other va ri able (0.7 to 1 with increases of 0.1). The results show that the generated declassification bias for the disease and the exposure implies a bias towards the null value in the relative risk, and for the third variable the pattern is more complicated, this bias could be towards or away the null value.
