Again you are right, but again, and like i said, its a simplified version,,, nothing is going to be exact... so about the fact that SDN represents a sample of the national pool and it's not representative: it's true, but at the same time this is the only aplicant pool that i have to compare myself to...
about the distributaion of ages, you are right, bu let me give you an example:
lets say we get 150 people in the first range, 95 people in the second range, and 25 in the forth and 10 in the last(above 30)... then you can take the % of interviews granted to the total in each group and have an approximate value for each range,
so it will not be about the number of people in each range, it will be about the total end result % of the interviews granted to the total demand in each age range....
but you are absolutely right about the fact that no two individuals are going to be alike, but to get a sample of a society you wont be able to compare people one to one, so people use the simplest models and then build upon that.... hope it makes sense...