MISCLASSIFICATION AND THE DESIGN OF ENVIRONMENTAL STUDIES

Abstract
Many authors have documented the existence of effects due to misclassification. This paper examines the size of these effects in the context of environmental studies. Misclassification occurs because of the use of surrogate measures of exposure. Misclassification causes relative risks to appear smaller than they are, with larger relative risks being seriously affected. The power of statistical tests is reduced, but not as seriously.

This publication has 0 references indexed in Scilit: