In order to find significant regional differences in the effect of socioeconomic
characteristics in unemployment duration, a survival analysis is made. This analysis use
data of the Gran Encuesta Integrada de Hogares (GEIH) 2014 survey in Colombia
(moving quarters April - June and September - November) with a no-parametric KaplanMeier
model and a parametric model with Weibull distribution. Parametric results show
that the type of effect is the same, except for the age, but not all socioeconomic
characteristics are significant for all cities. Moreover, the no-parametric analysis indicates
significant regional differences in these effects, especially about: unemployment duration
by city, age, search channel, type of activity and socioeconomic status. Thus, there are
significant regional differences and it is shown the type of person how has more risk to
stay in longer unemployment periods for each city.