asked on

How to calculate statistical confidence for survey sample sizes

I have to figure out the minimum number of surveys needed to attain a confidence level of 95%

I have several organzations, each with a different population.  For example:
Arizona: 523
New England 1,227
Washington: 744
Colorado: 2,333

What would be the minimum number of completed surveys in each to have a confidence level of 95%?
I believe I need to calculate the mean and standard deviation, but I'm not sure how that works in practice.
Thanks for your help.
Programming TheoryMath / Science

Last Comment
billb1057
aburr

A very short, clear, and non-rigorous answer to your question appears on the link below
-
http://www.isixsigma.com/library/content/c040607a.asp
billb1057

ASKER

Thanks -- could you help explain it?
First, I have groups with various population sizes.   Now I'd like to know what the minimum number of surveys is needed for a confidence level of 95%.
I'll accept a 3% margin of error.
The site you posted gives a couple of calculators that where you plug the numbers in.
But I have several numbers and I'd like to do it in Excel.  Is there a formula (or combination of formulas) that can do the same thing as those calculators in Excel?
richdiesal

You are mixing your terms a little bit.  "95% confidence" refers to interval in which there is a 95% probability that if you repeated your measurement an infinite number of times, the "true" measurement would lie within.  You can compute a 95% confidence interval with any sample of anything with 2 or more cases.  One cannot "attain 95% confidence" - that's simply not what the statistic refers to.

So, I think you are really seeking to answer the question of "how many surveys do I need for a 3% margin of error"?  (although if so, I ask where you got the 95%?)
billb1057

ASKER

Thanks for your clarification.
Here's the first question and then I'll try to take it apart.
I have a potential survey population of, say 3200.  What is the minimum number of completed surveys I need to make sure that the results are reliable or generally accurate enough to trust?
Some concepts (that I'm not that familiar with but I know the terms) that might be used are "confidence level", "margin of error" and perhaps "statistical significance".
aburr provided a link to a "Sample Size Calculator" here:
http://www.isixsigma.com/offsite.asp?A=Fr&Url=http://www.raosoft.com/samplesize.html
If that means what I think it does, it seems to work.  I plug in the population size (3200 in this case), set the confidence level at 95%, and choose a margin of error 5%.
Now the calculator tells me that I need a recommended sample size of 344.  This means that 344 completed surveys (out of 3200) is the minimum number needed to have a reliable (accurate, significant) result.
So, that works (even though I'm not certain on the terms).
Now, however, I'm wondering if there's a formula or set of formulas in Excel that can do the same thing.  So, I wouldn't have to plug numbers in one at a time.
In other words, I could use a formula that takes margin of error, confidence level and population and then returns "recommended sample size" somehow.
It may be that it's too complicated to do it in Excel and that's why people built web calculators.  But I was just wondering if there was a way I could use a formula to calculate the recommended sample size (or whatever the right term for that is) on a long list of numbers.
Thanks again.
billb1057

ASKER

I think a free sample size calculator in Excel should do the trick.
http://www.sra-researchgroup.com/nl/2004-03/03.shtml

ASKER CERTIFIED SOLUTION
richdiesal

THIS SOLUTION IS ONLY AVAILABLE TO MEMBERS.
View this solution by signing up for a free trial.
Members can start a 7-Day free trial and enjoy unlimited access to the platform.
billb1057

ASKER

That's even better yet -- thanks!
Math / Science

The Math / Science topic primarily includes discussions of mathematics, physics, statistics and economic analysis, but also biology, chemistry and other sciences.

10K
Questions
--
Followers
--
Top Experts
Get a personalized solution from industry experts

TRUSTED BY