Clark 37% wins DU poll. Dean second 35%, Kucinich third 15%

[Renew Deal]

General Wesley Clark lead in a straw poll at DU today with a 2% margin. Dean (35%) was second and Dennis Kucinich (15%) had a very strong showing for third. None of the other candidates made a strong showing. This is Clark's first win in an official straw poll in my memory.

http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=104&topic_id=929481&mesg_id=929481

[Bill McBlueState]

1. margin of error is 3%

Statisticians, please feel free to correct me, but...

The margin of error is sqrt(1068)/1068 = 3%

since there were 1068 votes. (Right?)

[Renew Deal]

2. How do you do that.

Is it supposed to be compared to the number of members or is it just relative to voters?

[Bill McBlueState]

8. for most polls, anyway

Poisson statistics apply (I don't remember exactly why -- a statistician would have to explain the theory behind it).

In that particular case, the error is the square root of the number of data points. So the relative error (the margin of error usually quoted in polls) is the error divided by the total number of data points.

Since there were 1068 respondents ("data points") the margin of error is the square root of 1068 divided by 1068 itself, or about 0.03 (3 percent).

So as more people respond, this number gets smaller and smaller. As an example, if 10000 responded, the margin of error would be sqrt(10000)/10000 = 100/10000 = 1 percent.

:crazy: Sorry if I'm muddling up the whole issue.

[Democrats unite]

10. Sorry you can't muddle this one up.

Clark is front in this poll, no matter how you try to spin it that fact won't change.

[Bill McBlueState]

11. thanks for pointing that out

Gee, I didn't realize that 37 is more than 35. It's all so clear to me now.

:eyes:

[Democrats unite]

17. You also need work on you sarcasm.

eom

[Bucky]

22. "sarcasm" sounds like "orgasm", so I think it's a funny word

[rock]

33. New math

Like the old wasn't good enough.

[Jerseycoa]

39. No muddle

That's the first time I can see how it works. Thanks.

[Julien Sorel]

50. This would only apply if the sample were random.

This one isn't. It would be like attaching an MOE to a Wolf Blitzer poll. About all that can be said here is that, of those who responded, 37% favored Clark, 35% Dean, and so on.

[Myra]

51. Still not a bad thing to say

It's good to see how competative General Clark is.

[YNGW]

3. Clark & Dean...

... are in a statistical dead-heat on DU anyway.

[NewYorkerfromMass]

9. now I've seen everything

:eyes:

[ProfessorGAC]

15. Well, Almost

The margin of error is calculated that way, yes, but that would be used the opposite direction in determining the validity of a sample size to provide a given margin of error. It's also really an estimate of the margin of error. The margin of error is one of those deals where it's really not known unless one can absolutely prove the veracity of every data point.

In other words, when determining a poll's sample size, one decides the margin of error desired, then uses your equation as part of the overall solution to determining the valid sample size needed.

The bigger problem with this poll is the lack of valid randomness and idenitified strata. The stratification is self-aligning and the results are apropos of nothing beyond the people who voted. So, the choices can't even be extrapolated to the larger DU population.

The reason why Poisson distributions are more relevant in social statistics is because of the historical proven likelihood in opinions and beliefs at the extremes. Belief structures are autofulfilling, in that certain opinions are rooted in strong beliefs that take them out to the fringes of a larger population at a rate not represented by a Gaussian distribution. So, the Poisson distribution takes those likelihoods of extreme, or discrete occurrances at the extremes more readily.

Most continuous scientific data does not suffer such concerns and use Gaussian distributions. For much quantum work, however, probabilistic theory requires hyperbolic or exponential distributions. (Enough of that.) So, really is different distributions for different types of data based upon the likelihood of those data fitting a certain pattern.

All that being said, the differences between Poisson and Gaussian are pretty small and won't really change the margin of error all that much.

While i'm just as interested as anyone else in this straw poll, (although i have no preference right now), i wouldn't extrapolate this too far into the general populace of Democratic leaning voters. It's just not valid to do so.

One last point: There really is no margin for error in this poll. Since this is a poll that represents only those who took the time to respond, this is not a sample set of a greater population! It's the WHOLE population. So, if we're not inferring toward a larger population, there is no margin for error attributed to the inference. (There is no valid inference, remember?)

So these numbers are what these numbers are.
The Professor

[NewYorkerfromMass]

21. Thanks for qualifying my reaction, which is summed up in

"There really is no margin for error in this poll"

[Bill McBlueState]

29. great post

I appreciate your explanation of why Poisson statistics are used instead of Gaussian.

Also, the last bit about the poll representing the whole population (limited though it may be) rather than a subset makes a lot of sense. I was wondering whether margin-of-error calculations were relevant in this case, and now I understand exactly why not. :)

thanks! :hi:

[ProfessorGAC]

30. No Problem! It's What I Do!

At work and at school, i'm "The Explainer".
The Professor

[bitchkitty]

36. I love it when you talk like that -

Brains are too too sexy. ;)

[Mattforclark]

56. It doesn't make sense to have a MOE

for a nonscientific poll that is not indicative of the general population.

[Bread and Circus]

57. This is a poll and Clark won, MOE means nothing -- it wasn't a

"scientific" study....get over it.

Dean will win the "real" nomination anyway so quit your
whining.

Oh, and to all the folks that accused Clarkies of "freeping"
the polls, get over yourselves.

[dajabr]

4. At least 2% abstained

At least 2% were "bused in" recently.

[bitchkitty]

37. You're so right. n/t

[POed_Ex_Repub]

5. As they would say on any of the major news channels...

This is a "non-scientific poll", but it's interesting to see what kind of supporters we have active isn't it?

[Wetzelbill]

6. wow, way to go fellow DK supporters

Words cannot express how I feel about DK. A great man. The world would be a better place if he was president.

I will say, I do admire Clark and Dean. Clark is scary intelligent and Dean's got some serious fire and passion.

Disappointed in Kerry and Edwards. After DK those two run pretty much neck and neck for my #2 spot. Great guys, great candidates, both were killed by their war votes and have ran listless campaigns.

Did I see Lieberman got more votes than Gephardt? Wow....those two are pure sex and excitement though.

Carole Mosely-Braun is a great lady. I'm proud of her for running.

Al Sharpton is quick-witted, sharp-tongued and I can't help but like the guy. Not a bad candidate either. Raises some serious issues.

[Ficus]

7. hmm

looks like a close race here at DU. I'd call it a tie. I hope this doesn't increase the sniping between the Clark and Dean camps. Go DK at 15%. If I remember right, he's had quite an increase in votes here at DU.

[Clark Can WIN]

12. Increase the sniping between Clark and Dean?

what could possibly do that? :shrug:

[Jerseycoa]

42. A tie it is :)

Peace and love.

[cryofan]

13. The real question is why Kucinich does so much better online than in the..

...real world polls? COnsistently, Kucinich is right up in every single online poll or expression of support. For example, the Alexa traffic software shows him #2 of 9, lagging only Dean.

But meatspace polls show him about #8 of 9, or thereabouts.

So why the disparity?

[boxster]

16. Apparently, his organization here at DU is more effective than

that of his campaign, in general.

[Ficus]

20. not to speak to all DK supporters

and not to make wide sweeping generalizations, but the sniping between the Clark and Dean folks sometimes turns others off to it. I don't think the DK supporters engage in this as much. Most spend their time talking up DK, rather than disparaging other candidates.

Of course this isn't true for everyone, but I think it's generally true here at DU.

[Ficus]

18. better access to unfiltered info

IMHO

[Tom Rinaldo]

14. It is a tie

I grant that Clark and Dean both came out on top. But it does help clear up all of that "real DU'ers aren't voting for Clark, the polls must be rigged" crap. Clark has real strong support from DU'ers. So does Dean, and to a leser extent Kucinich.

[Larry Gude]

25. I'm surprised...

...about the results.

This is a progressive site having a little, informal straw poll and a person who was a republican two months ago wins.

Maybe I'm just being cynical but if DK can't win a poll like this, hands down, over the General, then I guess I'm not clear on what 'progressive' means.

I'm sorry if I misread the poll...was it the old 'electability' thing again, not who you'd like to see?

I understand and accept the inevitability of a Clark/Dean showdown but isn't there still time to influence their positions?

[Tom Rinaldo]

34. There is definately still time to influence their positions

More so with Clark I believe. When Clark takes on a responsibility or assignment, he throws himself into it wholeheartedly. He masters complex subjects, which is why he rose so high in the military despite not fitting into the good old guys mold. Clark is going through that process now. As we all obviously know, Clark doesn't have a life long political record on domestic issues, rather he has some defining sets of guiding principles. His instincts are far more progressive than his mainstream appeal may indicate. He can be influenced by those who become active in his campaign, or at least remain open to it, and take the time to share their thoughts with Clark through his campaign site, which he is known to monitor.

[onehandle]

19. Very interesting...

Dean folks are very vocal over here. Dean posts outnumber other candidates' posts many times over. But Clark keeps coming up on top.

Does vocal, talkative, and verbose outnumber supporters? Hmmmm...


Anyway, this is more evidence that no candidate has cinched this primary.

[LittleDannySlowhorse]

23. Congratulations Clarkies

Your guy got it fair and square.

[in_cog_ni_to]

24. I'm not going to gloat!

on second thought, maybe I will. :7....WOO HOO! WTG!

:bounce::bounce::bounce::bounce::bounce::bounce:
:party::party::party::party::party::party::party:
:toast::toast::toast::toast::toast::toast::toast:
:bounce::bounce::bounce::bounce::bounce::bounce:
:party::party::party::party::party::party::party:
:toast::toast::toast::toast::toast::toast::toast::bounce::bounce::bounce::bounce::bounce::bounce:
:party::party::party::party::party::party::party:
:toast::toast::toast::toast::toast::toast::toast::bounce::bounce::bounce::bounce::bounce::bounce:
:party::party::party::party::party::party::party:
:toast::toast::toast::toast::toast::toast::toast::bounce::bounce::bounce::bounce::bounce::bounce:
:party::party::party::party::party::party::party:
:toast::toast::toast::toast::toast::toast::toast::bounce::bounce::bounce::bounce::bounce::bounce:
:party::party::party::party::party::party::party:
:toast::toast::toast::toast::toast::toast::toast::bounce::bounce::bounce::bounce::bounce::bounce:
:party::party::party::party::party::party::party:
:toast::toast::toast::toast::toast::toast::toast:

[Rose Siding]

53. Minimal lead in an eensy weensie pond

Dean leads by double digits in five nat'l polls this week (WSJ, CBS, Zogby, Gallup and Rasmussen)

...just for a touch of perspective and a reminder that you have *much* work to do.

[CMT]

26. Congrats

given the number of people on DU that only a little over a thousand participate tells us that active users don't come anywhere near how many are registered. I know lots of new Clark peoople have joined and more power to them--that is what a political site is all about. It is good to see that the two top candidates are pretty evenly repersented on DU and is good for the debate as long as it doesn't get too out of hand.

[Tom Rinaldo]

35. Nice post CMT

I agree and appreciate it, knowing that you have strong honestly held opinions.

[Bandit]

27. I would say that DU is fairly evenly split between two good candidates

How that will play out in New Hampshire and Iowa is quite another matter. I am glad that we have our differences and would love some constructive insights from all sides of the issues. We seem unable to do just that and usually sink to GOP standards of bashing people for whom they like and not for the issues represented. If the rest of America is like DU we are in a world of hurt.

[denverbill]

28. This proves Clark is the next McGovern.

After all, everyone knows that DU is a fringe-leftist organization, so therefore, Clark is the favorite of the fringe-leftists, and must be a wild-eyed liberal like McGovern who will lose in all 50 states.</sarcasm>

[tokenlib]

48. That's supposed to be a secret....Shhhh!

Just because Clark is a bit to the left of Dean on domestic issues--doesn't mean he would appeal to us fringe leftists does it??? :)

On edit--I disagree with the comparison to McGovern--and associated inference. The secret is that Clark does have appeal to a lot of us liberal types.

[DakotaDemocrat]

54. On "not electable"

I think that "Dean is unelectable" mantra is plain silly.

The voters will base their decision on leadership, while helping America move forward.

That's why I went to Clark. His goals in his first four years in office was a clincher. Not many tell you WHAT THEY WILL SPECIFICALLY do in their presidency while they are campaigning. This is unique.

Here they are:
GOAL #1: The typical family's income will increase by $3,000.

GOAL #2: I will put in place policies that will prevent 100,000 premature deaths from environmental causes by 2020.

GOAL #3: 1 million additional students will enroll in higher education.

GOAL #4: 2 million children will be lifted out of poverty, bringing the poverty rate to lowest on record.

GOAL $5: 30 million people who currently lack health insurance will get it.

He's telling us, essentially, that he will fail as a president if he DOES NOT get these things accomplished, so he's sticking his neck out - gotta admire that...

[Catherine Vincent]

31. I have to confess.

A fellow DUer paid me to vote Clark. Sorry.

;-)

[Jerseycoa]

46. A fellow DUer paid me, too

To vote Dean, but it was a real lot of money, and we know they can afford it. I donated my ill-gotten gains to the birthday cake last night and made the candles BRIGHTEN UP for Clark.

:party:

[kentuck]

32. Is DU ahead of the curve when compared with other national polls?

Very few show Clark as tied or ahead of Dean. Is DU more reflective of political reality than Gallup or Zogby? They do not show Clark as strong as the DU poll shows. Is Clark much weaker nationally than he shows on DU or is he much stronger than the national polls show?

[Spazito]

41. Interesting question!

Ever since I joined DU, I have noticed that DU supporters have been ahead of the curve on all major issues. The mainstream media can be as much as a month behind on picking up on what DUers had already pegged as important. So, to answer your question, imo, DU is again ahead of the curve.

[onehandle]

44. National polls are about name recognition...

Dean has a great advantage at this point because he's getting so much press coverage. Most people have no idea who any of these candidates are. They only know names.

That's why Lieberman was ahead in so many polls. They remember his name from 2000.

It's a long way to spring, and in January most campaigns will retool and relaunch to get their identity out there.

[Melinda]

38. The poll is suspect unless each vote was from a unique IP address

Can Elad verify that each vote was unique?

[onehandle]

40. We use the same script in Bartcop's forum

We can, so I'm sure he can.

[Melinda]

47. He was writing a new one - previous polls were being skewed

Will be interesting if he could expound and share the process. Regardless, any online poll is subject to being "freeped"; the only true polls that will matter in the long run are those final tallies at the ballot box.

[ThomC]

43. Clark and The School of the Americas connection

Has anyone here seen this article? I'm not sure what to believe now.

December 18, 2003 -- EXCLUSIVE

WASHINGTON - In a position that's likely to alienate some Democratic primary voters, retired Gen. Wesley Clark is a big booster of the controversial "School of the Americas" - which critics charge has history of graduating Latin American soldiers accused of rape, murder and torture.

Clark fought for years to keep the school at Fort Benning, Ga., open, even testifying on its behalf in Congress, despite graduates like imprisoned Panamanian ex-strongman Manuel Noriega.

Clark's backing of the school - whose curriculum once included teaching torture, execution, kidnapping and blackmail - puts him at odds with many Democratic officials and groups like Amnesty International, who want the school closed....

http://www.nypost.com/news/nationalnews/13799.htm

[onehandle]

45. Clinton didn't close it...

Nor did Gore, the Bushes, Reagan or Carter.

This will not make or break any candidate.

[Tom Rinaldo]

49. This has been robustly debated

On a seperate thread at DU. Here is the link: http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=104&topic_id=936092&mesg_id=936092

Strong opinions of all sorts are to be found, but there are decent rebuttles for your concern. I think that is where a discussion of your concern should be directed. In short I would say that Clark backed that School when he did because he feels that there is still an appropriate mission for it to play. Reforms were already underway when he spoke. A number of people who were trained there have done some terrible things. More didn't. Clark believes that positive lessons and models for multinational military cooperation have been developed in South America for fighting Drug Lords that can be applied to our international struggle against terrorists, operating in places like Pakistan and Yeman.

I would certainly ask of Clark both now ,and should he become President, that he ensure that strong curbs be placed on either that institution, or any other that replaces it and attempts to pick up whatever legitimate functions it pursued, to absolutely minimize the potential for human rights violations flowing from training done at that School. It is my limited understanding that much of the reform efforts that were undertaken focused on that problem, which was most acute in the 1980's during Reagan's anti Sandanista days.

I would say all abuses should be eliminated, but I am too realistic to ask for that about anything. The U.S. will never have full control over the actions of agents from other countries that train with our military. Having said that, I acknowledge that elements of our military have been directly involved in terrible actions.

[ErasureAcer]

52. The drug war is stupid

and so is America training terrorists.

Clark is wrong...we should close that freaking place down.

[Tom Rinaldo]

58. The drug war is wrong.

Of course it is wrong. To the extent that drugs are a problem, which is a large extent in many cases, it is a demand driven problem centered in the U.S. and it will never be solved by going after supply. It is a social, not criminal, issue for us in the U.S. In South America, however, because of our domestic drug enforcement policies in the U.S., drugs become a second hand military and criminal problem for many Governments and people down there. From a South American perspective it becomes a pretty complex question pretty quickly, since good and bad people are involved both in Governments and in the drug trade. Both have their own armies so to speak. Overall it is a very different problem and set of issues down there than it is up here. The point was that there are some relevent lessens that can be drawn from the situation and efforts used fighting in Columbia against the drug cartels, and the situation in Pakistan fighting against Al Queta.

None of our candidates, not even Dennis, is an out and out Radical. If I could remake the world it would look very very different than the current one. Right now I am concentrating on stopping Bush as a good first step.

[TrueAmerican]

55. DUers support Clark. Expect this to reflect in 2 to 4 weeks in the polls

DU is a trend setter. Clark is surging. The undecided are going to Clark 3 to 1. Clark supporters don't break out the sparkling wine just yet, but you can put it on ice. ;)

[dpbrown]

59. Kucinich - top tier

That's why the media is trying to "scrub" him from coverage.