The role of perception in urban in-migration: a path analytic model

Existing multiple regression models of migration at the regional or interurban scales frequently (a) presuppose no particular causal patterns among the predictor variables and (b) presuppose that migrants possess equal needs and knowledge of opportunities at various destinations. Exceptions to (a) include several recent simultaneous-equation models that explicitly indicate causal loops by which currently endogenous variables "become" exogenous or lagged-endogenous at a subsequent point in time. Most recent regression analyses of interregional migration however employ single-equation models. Exceptions to (b) include several innovative attempts by economists to incorporate "behavioral" variables--operable at the level of the individual migrant. In particular expected future income streams weighted by such factors as employment probabilities access to information or expected life span have all been employed. Inspection of empirical studies of this genre reveal however that frequently the researcher has selected the potential benefits "for" the migrant. It is probably true that migrants act not upon objective characteristics of places but upon their perceptions of them. Many existing regression models tend therefore to generate inefficient or inconsistent regression coefficients. The magnitude and direction of these inconsistencies depend both upon the magnitude and direction of the unanalyzed associations among the predictor variables and upon the magnitude and direction of influence of intervening variables that reflect the needs contacts and perceptions of potential migrants. In this study a simple path-analytic model including both objective and subjective (perceptual) variables is developed with the purpose of addressing both problems. The major contention is that the reduced-form regression coefficients from this path-analytic "perceptual" model are more valid than those of a traditional single-equation model incorporating only objective variables. Their greater validity hinges upon the logical structure of the path-analytic model and upon its greater ability to interpret a given set of empirical data. (excerpt)