The new Primary Thread

[Admiral]

Originally posted by Imran Siddiqui




Well, maybe its not so clearcut, but if Obama wins, I'll be looking at McCain more strongly for November than if Clinton wins.

I have a bunch of friends who are conservative Dems from Virginia and Maryland, who view Obama more favorably than Clinton in a matchup with McCain.

[DinoDoc]

Originally posted by Rufus T. Firefly
The problem is, almost no Dems believe it is vitally important to choose Clinton over Obama, or vice versa...

If that analysis is correct, why doesn't one of them quit?

[Dinner]

Because they both want to win more then they want to help the party?

[DinoDoc]

So I say again, why would a real campaign rather than a lovefest that pretends there is no difference between Clinton and Obama be hurtful?

[Rufus T. Firefly]

Originally posted by DinoDoc
So I say again, why would a real campaign rather than a lovefest that pretends there is no difference between Clinton and Obama be hurtful?

Asked and answered, but to recap:

1) It is a real campaign; it's just civil.
2) Nobody's pretending that there are no differences, but they're acknowleding, rightly, that there aren't that many. Those that do exit (Iraq, free trade, relative expeprience) are being emphasized.
3) A bruising primary season is never good for the eventual nominee because they have to start the race on the defense and mending fences within the party, rather than going after the other party hammer-and-tongs. This has always been the case, in both parties.

[MarkG]

Originally posted by Kuciwalker
Once you hit a certain sample size the impact of total population size on uncertainty is very small. So a sample of 1000 people is pretty much just as good out of a population of 300 million as 10 million.

i understand that, but isnt the US more diversified than the average country and thus needing a bigger sample?

[snoopy369]

1000 people is not nearly a big enough sample. The SE for a 1,000 person sample, weighted back to the population, would be probably 5 points, give or take depending on how good your sampling was. 10m would have a SE of <1 point. Thus a 10m sample is much better than a 1000 sample, unless your surveyed difference is >5 points...

Frankly I consider that pollster.com site worthless, as it doesn't show SE (standard error). SE is not strictly related to population (or shouldn't be considered such), but should also involve the quality of the sample - if I sample 1000 LV but most of them are 60+, that has a much higher difference from the population as if I sampled 500 from all age ranges.

Most of the people who sit home and are there when the pollsters call are 65+, so I'd guess Obama gets a bit of a bump compared to the polls; but it's possible some of the polls made an effort to adjust for that (population weighting, or nonrandom sampling, or random sampling with quotas). But without knowing, it's hard to say which of those polls is accurate...

[Cyclotron]

I'm going to go make calls for Obama, then earn a cool $100 working at the polling place on Tuesday. I went to Ted Kennedy's speech in Oakland the other day; it was alright, but I enjoyed the Obama speech I attended more (that was in Atlanta).

I don't think there's much of a difference between Obama and Clinton, but I don't like Clinton's refusal to apologize for her Iraq war vote, and I found the singling out of Obama's comments on Reagan as fuel for attack ads to be absolutely reprehensible. Additionally, dynasty politics in general rub me the wrong way.

[Admiral]

Originally posted by Cyclotron
I'm going to go make calls for Obama, then earn a cool $100 working at the polling place on Tuesday. I went to Ted Kennedy's speech in Oakland the other day; it was alright, but I enjoyed the Obama speech I attended more (that was in Atlanta).

I don't think there's much of a difference between Obama and Clinton, but I don't like Clinton's refusal to apologize for her Iraq war vote, and I found the singling out of Obama's comments on Reagan as fuel for attack ads to be absolutely reprehensible. Additionally, dynasty politics in general rub me the wrong way.

[snoopy369]

I like the refusal to apologize for the war vote I prefer a politician who will stick with what s/he believes in, and change his/her mind when the situation changes but not claim "oh, I was wrong before". No, you did what you felt was right then, and now due to new information/circumstances this is no longer the case.

[Rufus T. Firefly]

Originally posted by snoopy369
1000 people is not nearly a big enough sample. The SE for a 1,000 person sample, weighted back to the population, would be probably 5 points, give or take depending on how good your sampling was. 10m would have a SE of <1 point. Thus a 10m sample is much better than a 1000 sample, unless your surveyed difference is >5 points...

I'm unaware of any polls or surveys, on any topic, that use samples significantly larger than 1000. Not that your math isn't right; I'm just saying that these are not unuusually small samples.

Frankly I consider that pollster.com site worthless, as it doesn't show SE (standard error).

Yeah, the failure to list margin of error seems an egregious oversight. But it's hardly worthless; I know of no other site that summarizes that much polling data that concisely, and clicking on individual poll names usually brings up an article that tells you the margin of error. I'd rather see the margin as part of the table, but it's a far cry from useless.

[snoopy369]

Originally posted by Rufus T. Firefly


I'm unaware of any polls or surveys, on any topic, that use samples significantly larger than 1000. Not that your math isn't right; I'm just saying that these are not unuusually small samples.

I suppose it's cheating, since I work in a market research firm, but... i'm aware of lots of surveys on lots of topics that have significantly more than 1000 people samples. 1000 is about the minimum in my opinion (maybe 750 or 800 absolute minimum), 20,000 or 30,000 would be better. Honestly I very much question the +/- they normally quote; not knowing the population I can't say for sure, but it always seems low.

Yeah, the failure to list margin of error seems an egregious oversight. But it's hardly worthless; I know of no other site that summarizes that much polling data that concisely, and clicking on individual poll names usually brings up an article that tells you the margin of error. I'd rather see the margin as part of the table, but it's a far cry from useless.

Without knowing the standard error, it is worthless. Many of those results are within the margin of error, presumably, and not showing that means you're claiming "Obama is leading Clinton 45-41" and neglecting to mention that "with standard error of 5%", meaning that they are considered statistically tied.

It makes the polls misleading, and thus makes them worse than useless- they promulgate bad and false information, and encourage the public to make decisions based on this. They even include the "n" - making people think they are getting useful statistics - without including the SE. That pretty much has to be intentional in my book...

[Rufus T. Firefly]

Originally posted by snoopy369

I suppose it's cheating, since I work in a market research firm, but... i'm aware of lots of surveys on lots of topics that have significantly more than 1000 people samples. 1000 is about the minimum in my opinion (maybe 750 or 800 absolute minimum), 20,000 or 30,000 would be better. Honestly I very much question the +/- they normally quote; not knowing the population I can't say for sure, but it always seems low.

Fair enough; I was talking about political polling, where 1000 is the standard practice (and that's 1000 in the final mix, from a much larger initial group). Marketing sruveys are probably a whole different ball game.

Without knowing the standard error, it is worthless. Many of those results are within the margin of error, presumably, and not showing that means you're claiming "Obama is leading Clinton 45-41" and neglecting to mention that "with standard error of 5%", meaning that they are considered statistically tied.

It makes the polls misleading, and thus makes them worse than useless- they promulgate bad and false information, and encourage the public to make decisions based on this. They even include the "n" - making people think they are getting useful statistics - without including the SE. That pretty much has to be intentional in my book...

Again, the margin of error is usually one click away; the failure to include it in the initial data table is stupid, but doesn't seem malicious. it strikes me as neither a good nor bad site, but rather one that can be used well or poorly.

[Bosh]

Originally posted by Admiral


I have a bunch of friends who are conservative Dems from Virginia and Maryland, who view Obama more favorably than Clinton in a matchup with McCain.

I can't see Hillary doing better in Maine than Obama. Maine generally elections usually go heavily against apparatchik types and it's the sort of place where Perot beat Bush I. Not much of a swing state, but the northern half of Maine gives on electoral vote separately and there's enough hicks in the north for that half to be up for grabs (Gore one it by just 5,000 votes).

[Cyclotron]

Originally posted by snoopy369
I like the refusal to apologize for the war vote I prefer a politician who will stick with what s/he believes in, and change his/her mind when the situation changes but not claim "oh, I was wrong before". No, you did what you felt was right then, and now due to new information/circumstances this is no longer the case.

I only like stubbornness when it's stubbornness in support of the right decisions.

One of my biggest issues with the Bush administration was the complete inability of anybody to admit to mistakes or take responsibility for anything. The way I see it, if you think the war wasn't a mistake, you're a fool, and if you think it was a mistake, then you should apologize if you originally voted for it. There is no middle ground, despite Hillary's refusal to admit it. I respect a politician changing their mind and admitting past mistakes much more than simple pigheadedness.