The Analysis of Mobility Processes by the Introduction of Independent Variables into a Markov Chain

A major drawback with the traditional Markov formulation of social mobility is .that it 'assumes homogeneity among persons in an origin state with regard to their transition behavior. This requirement has led to a deemphasis in consideration of the ways in which the transition probabilities vary among individuals. In this paper a regression procedure is introduced which allows a ,heterogeneous population to be examined within a Markov framework. The advantages of this formulation are threefold: (a) it allows the sources of variationin the population transition probabilities to be determined; (b) it facilitates distinguishing between over-time change in the transition probabilities which results from shifts in the population on particular attributes and genuine structural change in which the rules governing transitions have altered; (3) it enables individual-level transition matrices to be constructed which are necessary for projecting to the k-step population matrix in the presence of heterogeneity. The techniques developed in this paper are applied to an analysis of geographic migration. rHE ANALYSIS OF IDBILITY PROCESSES BY THE INTRODUCTION OF INDEPENDENT VARIABLES INTO A MARKOV CHAIN Seymour Spilerman University of Wisconsin