This paper analyses validity of techniques, either for diagnosis or for measures of risk factors, by means of their misclassifications and their effects upon risk measures of groups as well as upon relative risk and odds ratio obtained in analytical studies. It is presented how misclassifîcation can distort results of cross-sectional and cohorts or case-controls studies. Nevertheless simple ways of correcting the results are suggested provided sensibility and especificity estimates of the techniques employed are known.
