Short or Long? Uncertainties and Errors in the Measurement of Internal Migration through Population Censuses: The Case of Mexico

Estela Rivero Fuentes, El Colegio de México
Edith Yolanda Gutiérrez-Vázquez, El Colegio de México

In this paper we use data of the 2000 population census in Mexico, where the questions on recent migration were asked both on the short and on the long form census questionnaires, to evaluate the quality of long form estimates. We conduct this evaluation in terms of the probability of emigration and immigration conditional on survival at the end of the exposure period and remaining on the country, for two different units of analysis: the state and the municipality. In our assessment we try to understand to what extent the differences found between the survey and the census are due to the sampling procedure and to what extent they are due to the weighting process. Our results show that survey design is not adequate for capturing the variability of internal migration, and that it captures particularly badly those flows that are small.