These presidential polls are such bullshit

They really don't take this polling shit seriously.

"Poll: Bush Leads by 11 Points During RNC

WASHINGTON - President Bush (news - web sites) leads Democrat John Kerry (news - web sites), 52 percent to 41 percent, while independent Ralph Nader (news - web sites) got 3 percent in a national poll taken during the Republican National Convention that ended Thursday.

The Time magazine poll of 926 likely voters was taken Aug. 31-Sept 2, during the convention and has a margin of sampling error of plus or minus 4 percentage points."

What the hell is that crap? 926 likely voters? How could such a group of people have anything at all to do with how the nation will vote? With so many different factions and different feelings in different areas of the country, you couldn't possibly take a serious meaningful poll with less than 10,000 people. It was probably a phone poll where only lonely elderly people answered. I'm not making any political claims here. I'm just saying that all these polls you hear about are a load of jizz. I mean, if I sampled 900 voters on the streets of my campus, it's probably be 95% in favor of Kerry. That doesn't mean I should be allowed to publish that. It's meaningless data...

926 voters is a little small, but you actually don't need many more than that, IF you're careful about it. Internet polls, for example, are almost completely worthless because of the way that they are conducted, and by their very nature.

This article from MSNBC gives a pretty decent explanation:

http://msnbc.msn.com/id/3704453/

Anyway, Bush leading by 11% does seem a little high, but you had to expect a little bit of a boost during/immediately following the RNC.

I could be wrong, but this will probably be a very tight election, regardless of what the polls are saying right now. For awhile I thought Kerry had it in the bag, but now I'm thinking it'll be a toss-up.

Someone never took a statistics course.

1000 likely voters extrapolated is what is done normally to derive any kind of national poll you see in the news.

JHendrix said:

Someone never took a statistics course.

1000 likely voters extrapolated is what is done normally to derive any kind of national poll you see in the news.

Click to expand...

Can you say "The projected winner of Florida is Al Gore"? It's already been proven that the media grossly misuses statistics just to make things more exciting. Other examples that happen daily are when they make a story confusing correlation with causation. They do it on purpose just so they can say things like "NEW STUDY FINDS TOOTHPASTE CAUSES CANCER". The media's misuse of statistics shouldn't even be up for debate. What worries me at this point is that they continue to use the same retarded methods for the presidential campaign even in light of last the last election's debacle.

In a presidential campaign with essentially one of two possible outcomes and a nation which could easily be divided into over 1000 factions, I fail to see how a poll with UNDER 1000 people could reflect anything at all. I have also taken a stats class before, twice in face, and they always stressed that so many things can go wrong in a poll. Interpretation of the stats is just as important as the raw data, and my interpretation is: Bullshit lie.

JHendrix is right, 1000 is plenty if done right.

But deathmagicianhawkmoon is also right in that the media can skew stats almost any way they want it so who really cares.

lithy said:

JHendrix is right, 1000 is plenty if done right.

Click to expand...

It's 926.

dmhawkmoon said:

It's 926.

Click to expand...

Oh well then it must be horrendously wrong.

lithy said:

Oh well then it must be horrendously wrong.

Click to expand...

I disagree. It's just completely uninformative to take any 900-1000 people and make a claim about how a large nation will vote. It could coincidentally be right.

dmhawkmoon said:

I disagree. It's just completely uninformative to take any 900-1000 people and make a claim about how a large nation will vote. It could coincidentally be right.

Click to expand...

Nah because again if statistically done right, if the 926 people are a representative sample of the nation there is no problem.

You really have to have a pretty thorough understanding of statistics to be convinced about this kind of thing. I'ts not a trivial subject. I've taken a few stats courses, but I still don't understand it. I'm gonna talk to my math buddy about it (he has an MS in stats), and see what he has to say about why they're allowed to make conclusions based on such a small percentage of the population.

Of course, another issue here is the wording of the poll. With the right phrasing, you could prove any point you may desire.

lithy said:

Nah because again if statistically done right, if the 926 people are a representative sample of the nation there is no problem.

Click to expand...

I'm trying not to degenrate into one of my 5 page explanations, but I'll mention another: winning the popular vote doesn't guarantee that the candidate won the election because of the electoral college system. It's extremely important in understanding how far ahead a candidate is to take into account the various regions of the country. If Californians seem to support Kerry by a little bit, and Texans support Bush by a whole lot, it won't mean that Bush is ahead by a giant margin. I don't feel like looking up the electoral college numbers for those states, but in just those two states alone you could see Bush way ahead in the popular vote, in theory, but if Kerry eeked out the win in Cali, the race would essentially be tied.

Does anyone see where I'm coming from more clearly now? I just figured it would stand out to most of you, but I guess you're not too good at thinking outside raw numbers, you goddamed computer nerds.

yawn, please wake me up when its all over. zzzzzzzzzzzzzzzzzzzzzzzzz