Rand's average poll numbers

[Warlord]

oh great. we're doomed

 

[limequat]

N/M Silent Bull beat me to it.

 

[SilentBull]

5% in IA or 6% in NH is needed before the 13th.

 

[Chieppa1]

Rand just need to continue to hit the frontrunners so ratings dictate that CNN needs him in the debate.

 

[francisco]

No, unfortunately groverblue's post is wrong. It's just a link to the CNN commercial for the debate. That commercial was made a while ago. Just because it shows Huckabee and Paul doesn't mean they're in. It just means they had to make a decision about who to include in the commercial. My guess is that they included the candidates that had the possibility to be included.

Correct. Unfortunately groverblue's post is wrong. CNN has not yet announced who will be in the main debate. In fact, the eligibility criteria provide for inclusion of polling results announced up to 5 PM Sunday 12/13/15. Therefore, not all of the polls to be used are known yet.

grover's link is merely to some unofficial poster on facebook who misconstrues the meaning of an ad promoting the debate. In the ad, some candidates are pictured with printed subtitled names--the "top 5" who unquestionably will qualify for the main debate. Some other candidates, including Rand and Huckabee, are shown in fleeting pictures only. Note that Huckabee is far away from meeting criteria for the main debate, in any of the 3 possible qualifying modes.

Grover may be excused for displaying over-optimism, but it does no-one any good to operate under false suppositions.

Hopefully, Rand will by the end of the week meet the criteria for inclusion in the main debate. But it is by more means a certainty. Rand's best shot at this point is to get one more poll in Iowa at 5% or more (after today's Monmouth release at 4%). His current average in Iowa is 3.7% in 3 qualifying polls, and he needs 4%.

The obvious implication is to get on those phones in the calling from home program!

 

[shenmue]

A new CNN poll in Iowa puts Rand at 3%. Not so good news.

His average is back to 3.5%. Now he needs to score 6% in the next poll(in Iowa) to qualify for the main stage

 

[limequat]

Link?

 

[CPUd]

Link?

http://cnn.com

 

[limequat]

Wow check out that 14 pt disparity between CNN and Monmouth on Trump. WTF?

 

[CPUd]

Their crosstabs/sampling look a lot like the previous CNN national poll. If he's polling that big of a lead, other campaigns will get desperate and go after him at the debate so the moderators don't have to. CNN wants the debate to get stupid, but they don't want to take the blame for it like CNBC did.

 

[limequat]

Their crosstabs/sampling look a lot like the previous CNN national poll. If he's polling that big of a lead, other campaigns will get desperate and go after him at the debate so the moderators don't have to. CNN wants the debate to get stupid, but they don't want to take the blame for it like CNBC did.

Yep, I just hope that this all will spur Trump to go after Cruz. Trump can say things that Rand cant say.

 

[SilentBull]

6% in IA or 6% in NH is needed now. 6 days left.

 

[PCKY]

A new CNN poll in Iowa puts Rand at 3%. Not so good news.

His average is back to 3.5%. Now he needs to score 6% in the next poll(in Iowa) to qualify for the main stage

What do the demographics look like? Anyone under 50?

 

[DevilsAdvocate]

I hope you know it's a lot more complicated than this (or at least it should be if the people running it knew wtf they were doing).

First of all, each poll has a different sample size, a different margin of error, and a different polling methodology. Just taking a raw average and giving each poll equal weighting is dumb.

National polls recognized by CNN, there are 9 polls that are conducted starting October 29: NBC(2%), Quinnipiac (2%), McClatchy(5%), Fox(4%), Bloomberg(3%), ABC(3%), Fox(2%), Quinnipiac (2%) and CNN(1%). Rand's average= 2.6

In Iowa, several polls conducted since October 29 but only 3 polls are recognized by CNN: CBS(2%), CNN(2%) and Quinnipiac(5%). Rand's average= 3.0

In NH, a total of 5 CNN-recognized polls have been released since October 29 : Monmouth(3%), WBUR(3%), WBUR(5%), Fox(3%), CBS(6%). Rand's average = 4.0

Notice something wrong here? You're taking input data with 1 significant figure, and generating a result with 2 significant figures. Doesn't make any sense.

--------------------------

Now I would assume that any poll which is to be used for something as important as rankings on the debate stage would have a standardized methodology. The best way to compute the average polling number would be to take the raw data from each poll, compile them into a single larger poll, and then average them out that way. This way with a greater effective sample size, your margin of error shrinks tremendously and your confidence interval improves.

If they actually DO just take a straight up average of the polls the way you did, they should be thrown in the volcano for scientific heresy

 

[Crashland]

I hope you know it's a lot more complicated than this (or at least it should be if the people running it knew wtf they were doing).

First of all, each poll has a different sample size, a different margin of error, and a different polling methodology. Just taking a raw average and giving each poll equal weighting is dumb.



Notice something wrong here? You're taking input data with 1 significant figure, and generating a result with 2 significant figures. Doesn't make any sense.

--------------------------

Now I would assume that any poll which is to be used for something as important as rankings on the debate stage would have a standardized methodology. The best way to compute the average polling number would be to take the raw data from each poll, compile them into a single larger poll, and then average them out that way. This way with a greater effective sample size, your margin of error shrinks tremendously and your confidence interval improves.

If they actually DO just take a straight up average of the polls the way you did, they should be thrown in the volcano for scientific heresy

A good point. I'm not sure how exactly they are planning to average them. You definitely shouldn't take polls with different sample sizes and give them both equal weight when you are averaging their end results. I don't think there's much that can be done about demographic breakdowns or different methodologies other than to specify which will be included (live interviewer cell and landlines, from certain polling companies) and which ones won't.

Of course, it would be nice to be polling a little bit better so as to not need to worry about the nitty gritty of the polling average methodology...

 

[CPUd]

The last CNN debate they rounded to 1 sig fig, averaged and rounded again, no weights.

 

[dusman]

It'd probably be better if they did debate position by individual fundraising numbers (not including PAC). Then a candidate who isn't considered a frontrunner might take center stage due to the effort and strength of the support base and a perceived frontrunner may just be all media-hype, but no support.

But, I'm just dreaming now.

 

[Anti-Neocon]

The last CNN debate they rounded to 1 sig fig, averaged and rounded again, no weights.

What do you mean by 1 sig fig? 4 is one sig fig.

Did they still keep only one sig fig at the end?

 

[Peace&Freedom]

A new CNN poll in Iowa puts Rand at 3%. Not so good news.

His average is back to 3.5%. Now he needs to score 6% in the next poll(in Iowa) to qualify for the main stage

If the network rounds up to the nearest 1, 3.5 will round up to the 4% needed to get Rand included.

 

[CPUd]

What do you mean by 1 sig fig? 4 is one sig fig.

Did they still keep only one sig fig at the end?

nevermind, I was thinking of the CNBC, which also did have more than 1 sig fig. Paul, Kasich, Christie were listed at 3.00 (thought it was just 3), I remember discussion about how they determined the tierbreaker.

The first CNN debate scored like this:

1) Donald Trump: 23.929

2) Jeb Bush: 11.500

3) Scott Walker: 9.429

4) Ben Carson: 8.929

5) Ted Cruz: 6.286

6) Marco Rubio: 5.643

7) Mike Huckabee: 5.571

8) Rand Paul: 4.714

9) John Kasich: 3.214

10) Chris Christie: 3.143

11) Carly Fiorina: 2.229

12) Rick Perry: 1.814

13) Rick Santorum: 1.214

14) Bobby Jindal: 1.057

15) George Pataki: 0.529

16) Lindsey Graham: 0.471

With the Carly Rule:

1) Donald Trump: 27.8

2) Ben Carson: 14.0

3) Jeb Bush: 9.2

4) Ted Cruz: 7.4

5) Scott Walker: 5.6

6) Marco Rubio: 5.4

T-7) Carly Fiorina: 4.4

T-7) Mike Huckabee: 4.4

9) John Kasich: 3.6

10) Rand Paul: 3.2

11) Chris Christie: 2.8

12) Rick Perry: 1.08

13) Rick Santorum: 0.8

14) Bobby Jindal: 0.56

15) George Pataki: 0.44

16) Lindsey Graham: 0.28

The overall average includes results from a Fox News poll released July 17; a Washington Post/ABC News poll released July 20; a CNN/ORC poll released July 26; a Quinnipiac University poll released July 30; a NBC News/Wall Street Journal poll released August 2; a Monmouth University poll released August 3; a Fox News poll released August 3; a Bloomberg Politics poll released August 4; a CBS News poll released August 4; a Fox News poll released August 16; a CNN/ORC poll released August 18; a Quinnipiac University poll released August 27; a Monmouth University poll released September 3; and a CNN/ORC poll released September 10.