What is the value separating the higher half from the lower half of a data sample, a population or a probability dis.