Why is it necessary to have more than one method for measuring central tendency?