Validity of the sample size is not determined by its fraction of the whole population. I don't know the formulas and I'm not a statistician. Maybe someone can drop some citations.