[muawiyah]

A sample is what it is (although that doesn't matter since polling doesn't matter either) so it isn't like someone sits down and tries to call 7% too many Democrats.

The whole point of sampling is to achieve something reflective of the average opinion of a far larger group ~ and you attempt to do that with a random sample.

If you subsequently adjust your random sample to reflect what you think you want, it is no longer a random sample poll ~ it's just an auger ~ rather like the ones the Roman's looked at.

[bunster]

Please see Post #32 — I meant to send it to you.