ABSTRACT
There are many epidemiologic studies to find the tracking phenomenon. In repeated measurement data, a population is said to have tracking phenomenon with respect to particular chracteristic if, for each individual, the relative rank of observed value maintained over time. Understanding tracking phenomena in epidemiologic study is quite important. If the risk factors of chronic disease have tracking phenomenon, early detection of these risk factors can lead to the possibility of early treatment. In this study, we defined tracking as relative maintenance and proposed new measure of tracking by using random coefficient model. We compared our measure with McMahan's using simulation study. And we applied ours to real data. We may conclude that our new measure of tracking is adequate in explaining and dectecting the tracking phenomenon.