Internal migrations have become a problem of paramount importance. Some authors state that the factors influencing migrants are sex, age, race, educational level, profession and qualification. There are differences between countries. The objective of this paper is to analyze internal migrations from the countryside in Cuba; identifying socio-demographic characteristics of the migrants in relation with persons who do not migrate. Logistic regression models are used for the 4 main migrant movements: Rural-Rural-Rural, Rural-Rural-Urban, Rural Urban Rural and Rural-Urban-Urban, together with its adjustments and diagnosis. The cross tables include characteristics of the migrants, based on Internal Migration Polls (ENMI). In the analysis, it is interpreted the known reasons of disparity (odds ratio) for each migration route. Models which are more adjustable are built with its adjustment and diagnosis goodness. For the analyzed routes migration resulted selective taking into account different socio-demographic attributes; race was not important. In some cases there are confusing effects, therefore the valid reasons are the disparity of the logistic regressions. Sex does not influence for the migratory route RRR. When the destination of the migrant is at least one urban zone then the sex has influence. The migratory movement is made in family, in the 4 migratory routes young are the mayor migrants. In the route RUR, employed persons are the mayor migrants, and the most qualified in the route RUU.

Keywords: National Internal Migration, migratory route, cross tables, logistic regression, odds ratio, crude odds ratio