The impact of differential recall on the results of case-control studies

Many authors have expressed concern that case-control differences in the accuracy of information reported in interviews may create spurious associations in epidemiological investigations. Nevertheless, the impact of differential misclassification on observed associations has not been systematically examined. This paper presents algebraic and graphical analyses of the effect of case-control differences in reporting accuracy on estimates of association and on test size. These analyses suggest that under certain circumstances, even large differences in accuracy may have a minor impact on the results of a study. Study results may be particularly resistant to differences in the sensitivity of recall when the prevalence of exposure is low. The results also illustrate how researchers may evaluate the potential impact of differential misclassification on the validity of their own investigations.