In a population, every person fits into one of 10 to 15 groups which, in theory, had no effect on any of them. The groups are evenly distibuted.
In this population, there is a sub population, which you survey. The sub population, supposedly, has nothing to do with the groups.
In a survey, you would expect that the groups are still evenly distributed, right?
Ok, now if you found that in this sub population, most groups had 3-5% representation, a quater of the groups had 10-15%, 1 had 20% prepresentation and one group was not in it at all, you would then suspect that the sub-population is not quite so independent of the groups, right?
What if the survey was only of 20 to 30 people?
Obviosly, so small a sample is not going to give proofs, and is in no way acurate, but would you still at least think that the survey showed something that there may</> be someting interesting?
p.s. No chance of getting larger sample. Either the entire sub population is that small, or the contactable population is.
p.p.s. results of survey not being used for anything official, I just saw the results and wondered about them.
Ok, incase this makes a difference...
The groups are based on the month people are born in.
Ok, I now have a clear enough answer about statistics that does fit the scenario.