AlbertWesker[RE] wrote:
ALL polls? I doubt it. The Repubs took how many seats in 94, 22?? Can't remember. The Dems this year would be lucky to get 15. Like I said, thats my prediction, I hate polls, just wait 17 days (Nov 8th), and we can talk more about what happened. There is no point debating this crap right now, I'm sick of all these so called definitive polls (with samples of 1000 people). Good day to you all, and good luck to all Republicans and Democrats, Libertarians, and all other candidates running this Nov. It shall definitely be an interesting race to watch no matter who comes out on top. Peace.
Random samples are more accurate than a full count for large populations, which is why the census is usually fucked up but polls have a MOE of only +-3-4%. So generally if a reputable organization has a poll (ie. Gallup) does a survey of 1500 people the results will usually predict the outcome. I should add that a poll with a randomized sample of 1000 individuals will have a margin of error of 3%, 95% of the time and for a MOE of 1% a sample size of ~10000 is needed.
You may not "believe" in polls, but the methods used are sound.