Looks as though Obama would win.....

http://news.yahoo.com/s/ap/20080211/ap_on_el_pr/presidential_race_ap_poll

Well duh!

From the article in the link: Included were 520 Democrats, for whom the margin of sampling error was plus or minus 4.3 points, and 357 Republicans, with a margin of sampling error of plus or minus 5.2 points.

I'd bet a good steak dinner that McCain would win if you sampled 520 Republicans and 357 Democrats.

Pollsters slant the polls to come up with the results they want.

It also matters where the voters were from, was it an equal sample across all 50 states, was it weighted in the big cities, etc.

Remember the south and midwest has a disproportinate number of electoral votes.

"a disproportionate number of electoral votes"

Um, where did you read that?

It is almost identical to our Congress:

"The size of the electoral college has been set at 538 with 535 corresponding to the size of the United States Congress, plus 3 that represent D.C. since the election of 1964. Each state is allocated as many electors as it has Representatives and Senators in the United States Congress. Since the most populous states have the most seats in congress, they also have the most electors. The states with the most are California (55), followed by Texas (34) and New York (31). The smallest states by population, Alaska, Delaware, Montana, North Dakota, South Dakota, Vermont, and Wyoming, have three electors each. Because the number of representatives for each state is determined decennially by the United States Census, the electoral votes for each state are also determined by the Census every ten years. The number of electors is equal to the total membership of both Houses of Congress (100 Senators and 435 Representatives) plus the 3 electors allocated to the District of Columbia, totaling 538 electors. A candidate must receive a majority of votes from the electoral college (currently 270) to win the Presidency. If in either election for President or Vice-President no one receives a majority, the election is determined by Congress (the House votes for presidential candidates, and the Senate votes for vice presidential candidates)."

Wiki

The umber of electoral votes for a state, is equal to their number of representatives in congress, and number of senators.

Every state has 2 senators, then the number of representatives is by population, with the requirement even the smallest state has 1 representative.

So how does that give the South and Midwest a disproportionate number?

I am curious.

There are more smaller states, and since each state gets 2 senators, they have slightly more electoral votes, than if they were divided just by population.

Geographically, the very smallest states are actually in the Northeast. Many of the original 13 colonies are quite small. I imagine that the ratio of population density to delegates in the Rocky Mountain states and Alaska make them the most represented.

How does that make the South and Midwest disproportionate?

I should have said, leaning republican, which includes the rockey mountain states.

Nwokie's math is correct, but the amount is relatively insignificant. The states with the most "disproportionate" number of EV's are states with only 3, such as ND, AK, DE, MT, WY, etc. But the nature of the EC is to make sure no matter what size of the state, all have a degree of say in the presidential election.

http://en.wikipedia.org/wiki/List_of_U.S._states_by_population

"Well duh!

From the article in the link: Included were 520 Democrats, for whom the margin of sampling error was plus or minus 4.3 points, and 357 Republicans, with a margin of sampling error of plus or minus 5.2 points.

I'd bet a good steak dinner that McCain would win if you sampled 520 Republicans and 357 Democrats.

Pollsters slant the polls to come up with the results they want."

The last sentence is absolutely true, however what you and I refer to as "slant" pollsters refer to as methodology.

Your mathematical analysis is not true, i.e. the size of the polling sample is not necessarily in direct proportion to the outcome. The larger the sample, in fact, the greater the chance that minority opinions within that sample group will be significant, i.e. Democrats for McCain, Democrats for anybody but Clinton. Remember, the smaller the sample group, the more insignificant the minority opinion becomes. Sorry Mr. Bill. F in statistics.

With the taking of Ohio's electoral votes, the south won't matter at all. Hopefully they'll secede and become the third world nation they strive to be.

If, in 2000, the electoral votes had been awarded, strictly on a states population, guess who would have won.

Nwokie, who cares? Do think about who would have won the football game if the decision was based on time of posession?

Well, I am confused. Nwokie, I thought that your initial aside was about the actual election and the EC.

If you were talking about the poll, well, it is a poll. Like any kind of statistics, they are shaped by sample size, geography, class, race, gender and other subsets. The information is then recast to suit the person presenting the numbers. In many cases, they are essentially meaningless barometers.

I like Obama and these figures are encouraging, but polls are not really conclusive to any reliable degree.

Votes, on the other hand, matter a great deal.

Andy, you're right, if the political bent of the sample is unknown.

What I'm saying is that by sampling more Democrats than Republicans, of course the results would favor the Democrats. The opposite would be true if you sampled more Republicans than Democrats.

If the sample was larger, then the margin would narrow.

The only way they are going to get an accurate result would be a blind test.

My dad was the Director of Media and Market Research for Ralston Purina for 20+ years and I remember him telling me about how they would set up taste tests when they would come out with a new cereal.

There would always be 5 products to sample and they would always make their new cereal #3. It guaranteed that theirs would always be the one selected.

So I guess in a nut shell what I'm saying is that I don't hold a lot of faith in the pollsters.

Polls and pollsters are like advertising. There is always someone willing to pay for it and someone else whom will believe it.

Here are the four cereals you requested:

Oat McCainmeal
Huckleberrybran Flakes
Clinton Crunch
Bananaobamaberry Delight

You sample the parties, by the number of registered voters in each party.
not 1 demo for each republican, etc.

Shhhh.

Here are a few ideas that failed:

Mitt-on-a-Shingle -- A puffed cereal that tastes like biscuits and gravy when warm milk is added.

Gravelicious -- An extremely crunchy granola made with oats and nuts.

Jump Outta Bedwards -- A yogurt frosted flake with vitamins, caffeine, ginseng and taurine.

Kucinich Muesli -- A traditional European breakfast dish with an even more impossible name.

Naderade -- Fresh squeezed organic orange juice served in a carton shaped like a crash test dummy.

OK, to the neocons and pinko leftists here.

You only believe in polls when they support your stand. You need to accept that no poll is really balanced. Pollsters are hucksters.

So, Nwokie and Vitology, I never expect to see either of you posting links to polls.

The only poll that matters is the result on election day (or when the Supreme Court decides who won)

Actually, I posted a link to the actual vote count, not a poll.

My comment was in regards to votes, not the poll as well. So, I reserve the right to post a poll in the future if I feel the need.