Quote:
|
Originally Posted by Bertgang
Quote:
|
Originally Posted by Type XXIII
Quote:
|
Originally Posted by August
I seriously doubt that interviewing 1000 people out of nearly 300 million is a valid poll, claims of an accuracy margin of 3% notwithstanding.
|
When it comes to statistical surveys, the total population is not relevant. As long as you interview a representative sample of the total population (choosing totally at random is the best,) the results from that sample will with a high degree of probability be very close to the actual situation. This can be proven mathematically. |
The mathematical rule is fine, the main doubt is about the representative sample.
The random choice doesn't work always properly; as example, you often need a percent of professionals and one of unemployees, a percent of young and one of aged people, and so on...
Several methods are used to solve this kind of problem but, on my point of wiew, the final outcome is that any "representative sample" is a mix issued from small groups, and so the mathematical rules about large numbers are lost since the start.
Add some minor problems, like people telling "random" answers, as not really interested to be a creditable statistical sample. |
There's even more to it.
It's possible to hook the answer you're looking for by asking the right questions. It's also possible to, after the gathering, manipulate a few mathematical strings and make the percentages move one way or another, since we're talking % and not absolute numbers.
I've seen too many manipulated polls and innacurate polls (same subject, considerably different results and completely different outcome in reality), to accept any poll I see with a glance as a honest study worth of credibility.
Being able to name something and understanding it are actually two different things.
It's possible that someone who can't name the constitution understands, agrees and lives by those values in his daily life, how are you going to poll that?