A method is suggested for removing measurement error from morbidity data collected in Third World countries. The method assumes the existence of diseases which are independent of socio-economic factors and exploits observed correlations between these disease and these factors to construct an...