By using this site, you agree to our Privacy Policy and our Terms of Use. Close
Torillian on 27 February 2016
Edit Thread
IkePoR said:
Torillian said:

yeah that's fair, I didn't notice the size of the original linked gallup poll, thanks for that.  Out of curiosity, I know there are numbers behind how many people you need to survey to have a reasonable chance of being correct when upscaled to a population of x size, but I don't recall what the equation is.  Do you happen to remember?  

Don't quote me on this but:

Number of surveyed

_____________________x100

Expected % to respond

Sorry I think I misstated the question.  I'm more talking about if you want to be 99% confident that your results are within 1% of the true value for a population of 319million (US population) what sample size do you need to achieve that.  According to the first calculator for this I found on the internet we would need a sample size of 16,641 to be that confident of our answer.  So we can be 99% sure that the true number is within 2.4-4.4 according to the gallup poll with 100,000 respondents.  

Equation for that if anyone is curious is: 

where z is based on confidence interval.  Interesting point is that the population size becomes inconsequential at a certain point and 16,641 respondents is thought to represent a population of 319million just as much as 7billion.  Of course you need an actual random sampling which I doubt the gallup poll did across the world population so we can only make conclusions about the US population based on that survey.  



...