WHEN WILL NONDIFFERENTIAL MISCLASSIFICATION OF AN EXPOSURE PRESERVE THE DIRECTION OF A TREND

Dosemeci et al. (Am J Epidemiol 1990;132:756-8) gave examples in which nondifferential misclassification of exposure reversed the direction of a trend. Gilbert (Am J Epidemiol 1991;134:440-1) proposed that thes e examples occurred because the errors in exposure were systematic, an d she pointed out that the relation between the measured and the true exposure was not monotonic. Assuming that the mean response either mon otonically increases or decreases with the true exposure and that the exposure misclassification is nondifferential, the authors show that i f the mean value of the measured exposure increases with the true expo sure, then the direction of the trend cannot be reversed. Consequently , gilberts intimation that reversal of trend can only occur when error s are systematic is correct. However, the present authors' result is s tronger in that even when errors in assessing exposure do not include a systematic component, if monotonicity can be assumed, reversal of tr end cannot occur. The weaker condition of positive correlation between the measured and true exposure in not sufficient to guarantee nonrefe rral of trend, as they show by example.