ABSTRACT This study estimates a series of random parameter logit models of the college-to-work migration decisions of technology graduates and holders of doctorates within the United States. We employ detailed information on the migration-relevant characteristics of individuals, as well as on their actual origins and destinations at the metropolitan scale. In addition to its obvious implications for “brain drain” policies in U.S. metropolitan areas, the study demonstrates the richness of the random parameters technique for behavioral-geographic analysis. We find that science and technology graduates migrate to better educated places, other things equal; that PhD graduates pay greater attention to amenity characteristics than other degree holders; and that foreign students from some immigrant groups migrate to places where those groups are concentrated.