A validation of the ethnosurvey: The case of Mexico-US migration

This article evaluates the validity of the ethnosurvey as a method of demog raphic data collection by analyzing the representativeness of the Mexican M igration Project (MMP) as a source of information on Mexico-U.S. migration, er briefly delineating the: philosophy, structure, and organization of the MMP's ethnosurvey design, we describe MMP's public use dataset as well as Mexico's Encuesta Nacional de la Dinamica Demografica (ENADID) as a benchma rk for its systematic evaluation. Although the MMP over-represents migrants in the western states and mid-sized communities, it yields a relatively ac curate and valid profile of migrants to the United States. A comparison of multivariate models estimated using MMP and ENADID data suggests that where as the former's sampling errors an small and yield biases that are substant ively unimportant, the latter's potential specification error and selection bias may seriously compromise results. Our comparison thus validates the e thnosurvey as an accurate and reliable method of data collection and the MM P as a good source of reasonably representative data on documented and undo cumented migration to the United States.