Re: FN-FORUM: [OT] Survey Responces

date posted 28th January 2005 17:27

On Friday 28 Jan 2005 3:48 pm, Dom Latter wrote:

> The key word is "representative" as in "representative sample".
>
> If you do not make sure that your sample reflects the composition
> of the population that you wish to study, your data is flawed.

I feel the need to make this point more clearly. If your sample is=20
self-selecting, the chances of it being representative are approximately
zero. Self-selecters tend to be people with an interest, if not an agend=
a.

It gets worse if you do it on the net - your sample is now restricted
to a much smaller demographic.