Due to actuallly the data of entrepreneurship colombian population of GEM have big differences between some years, because is a sample results, I think is more convenient use the rates series (in order to null the effect of variation of sample), What do you think about that?