ABSTRACTStochastic process analysis, particularly Markov chain dependence, may be utilized in a wide range of research situations where change through time is the focus of interest. This paper is concerned with the utility of Markov chain analysis in the prediction of future migration on the basis of past migratory behavior. Transition probability matrices, which were dependent on the respondent's migratory behavior in the preceding time interval, were utilized to predict migration in the following time interval for a series of 145 longitudinal case histories. Predictions were also obtained for selected age, educational, and occupational stratifications. Markov chain analysis, or a similar probability model, seems to provide a promising tool for studying the relationship between prior and expected migration, and for obtaining statements of migration probability. Since a high proportion of total migrations seem to be made by a limited segment of the population, the factor of past migratory behavior itself may be an important element in the identification of migration differentials and the prediction of future migration.