Test needed for reliability

[beddo]

Hi All,

This is my first post, I know it's a fairly simple question but I'm not sure what to do:

* I know that x number of users completed an action.
* These people were counted together from the US, UK and Canada.
* I know the geographic breakdown of all users from many countries over the period of the action.


From this, I can extrapolate an estimate of how many users from every one of the hundreds of countries also completed this action. What kind of test would I need to show how reliable this estimation would be.

Any help would be appreciated.

[Dr.Appalayya]

Your sampling method determines the reliabilty of your sample or the statistics based on them.

Please understand by googling and reading how systematic sampling or stratified sampling or area sampling is done. Please note that the word 'reliability' is used in another context but not here, although what you are asking for is understandable.

[beddo]

Thanks for your reply, I'll go over sampling from my old uni text books!