If I were to look at number of births per 1000 in the US what would be my sample and my sample size? I need 20 samples.
I'm so confused?

Question

If I were to look at number of births per 1000 in the US what would be my sample and my sample size? I need 20 samples. 
I'm so confused?

tkhunny · Answer

I don't understand the experimental design.  You should probably access public records at the CDC and survey the ENTIRE population.

Having said that...  You should probably pick a time period.  You should probably limit your survey to women of child-bearing age for the "per thousand" part (the denominator).

anonymous · Answer

Like if i were to look at birth rates in 2010 for, let's say, 40 states, is it possible for me to get the sample from that? And sample the size as well?

tkhunny · Answer

I suppose.  I would have to question what you were doing.  If you can get 40, why can't you get the other 10 and the remaining territories?  I kind of don't get the purpose to deliberately taking less data when you can have it all.

Additionally, it depends on what sort of data you can get.  If you just get the birth rate, there is little point in combining the data or extrapolating to the whole of the population.  Without the numerator and denominator separately, it's just no good.