Test Bed of Users

I'm not brialliant at maths and don't pretend to be! But i have a problem which requires some statistics.

Basically how would i prove how many people to test on to create a non bias result?

does that make sense? I am writing a program that will test user interactions with a pc and need to find a suitable test bed...but how many people do i use?

Cheers in advance, any suggestion references etc would be greatfully recieved!


Andy:confused:

it sounds like you're looking at confidence intervals. if this is the case then we can help you if you supply us with the following information:

*how confident you want to be. like 99% or 95% or something.
*some sample results