Poll of 18,443 Canadians suggests a dead heat

[Montgomery Burns]

Not a CTV/CBC/Globe & Mail/Toronto Star Poll

Question #1

At this moment which of the following federal political leaders and their parties are you supporting?

Jack Layton and NDP..................22.30 %

Paul Martin and Liberals.............32.10 %

Stephen Harper and Conservatives.....32.21 %

Gilles Duceppe and Bloc..............13.33 %

And there is more; much more. This is an exhaustive poll, but the polling company believes that the Conservatives will win. According to them, the only region that the Liberals still have quite strong support is the Maritimes.

[sharkman]

Yep, another poll I read yesterday in the National Post said it's a dead heat as well. Expect the Liberals to come out swinging on Monday, they probably have a few dirty press releases they'll give to friendly media.

Meanwhile, Joe Canadian is getting fed up with Liberal slime.

[kimmy]

The sample size might be larger, but the poll is dated November 16.

I'm becoming increasingly skeptical of all the polls floating around. The polls fluctuate so rapidly-- one says one thing, and a couple of days later another poll says something different. It makes me wonder if the variable is peoples' intentions, or the polling methodology.

-k

[August1991]

That poll makes no sense. Nobody in their right mind would poll 18,000 people.

And what does "98% competency" refer to?

Just because you read it onthe Internet doesn't make it true...

[shoop]

Pretty understandable knee-jerk reaction but there is some basis to a sample size that big.

The only thing holding back a sample size that big is cost, but if somebody is willing to pay ... any polling company would be happy to do so.

98% competency appears to the standard deviation reated to polls.

You have clearly seen the "accurate within +/- 2.5 percent 19 times out of 20". 98% competence would merely be "accurate within +/- 1 percent"

kimmy's question of the timing of the poll is much more reasonable.

That poll makes no sense.  Nobody in their right mind would poll 18,000 people.

And what does "98% competency" refer to?

Just because you read it onthe Internet doesn't make it true...

<{POST_SNAPBACK}>

[Montgomery Burns]

That poll makes no sense.  Nobody in their right mind would poll 18,000 people.

And what does "98% competency" refer to?

Just because you read it onthe Internet doesn't make it true...

<{POST_SNAPBACK}>

How do you know that nobody in their right mind would poll 18,443 people? Just because you say no one in their right mind would do so, does not mean it is a bogus poll.

Or are you saying that Robbins put out a bogus poll? Is it just this poll or are all their polls bogus?

It certainly seems to be more credible than the polls by the Canadian media that constantly show the liberals in the lead. Canadians can't possibly be that stupid to vote in this thoroughly corrupt govt again.

What part of "98% competency" is confusing you?

[August1991]

You have clearly seen the "accurate within +/- 2.5 percent 19 times out of 20".

<{POST_SNAPBACK}>

Those terms have precise mathematical meanings. "98% competency" does not.

As to the sample size, it makes no sense to have a poll of 18,000. The extra cost (and extra time required) does not merit the extra precision - which, I note, the pollster mysteriously ignores.

It certainly seems to be more credible than the polls by the Canadian media that constantly show the liberals in the lead. Canadians can't possibly be that stupid to vote in this thoroughly corrupt govt again.

Priceless.

[Montgomery Burns]

August1991:

Those terms have precise mathematical meanings.  "98% competency" does not.

As to the sample size, it makes no sense to have a poll of 18,000.  The extra cost (and extra time required) does not merit the extra precision - which, I note, the pollster mysteriously ignores.

Why does 98% competency not mean anything?

What is the pollster "mysteriously" ignoring?

[Montgomery Burns]

Shoop:

The only thing holding back a sample size that big is cost, but if somebody is willing to pay ... any polling company would be happy to do so.

Aren't you just a bit curious as to what US company commissioned this poll? I know I am.

I'll bet is was Halliburton.

[shoop]

I completely agree with you that the size of the poll is ridiculous, but if someone was willing to pay for it why *wouldn't* a pollster do it?

You are needlessly nitpicking the 98% competency line. Only people who have taken university level stats understand the rationale behind the +/- 2.5%, why not try and make it a little easier to understand?

Those terms have precise mathematical meanings.  "98% competency" does not.

As to the sample size, it makes no sense to have a poll of 18,000.  The extra cost (and extra time required) does not merit the extra precision - which, I note, the pollster mysteriously ignores.

<{POST_SNAPBACK}>

[sharkman]

I think I see the bone of contention among some. If it showed the Liberals with a 4 point lead, there would be no questions.

[August1991]

I completely agree with you that the size of the poll is ridiculous, but if someone was willing to pay for it why *wouldn't* a pollster do it?

Who paid for it? Some mysterious American business.

Why does 98% competency not mean anything?

Because it means nothing in any mathematical sense.

What is the pollster "mysteriously" ignoring?

If the sample size were 18,000, the precision of the results would be greater and this would be reflected in some of the reported specifications. The pollster didn't do that.

It's the equivalent of claiming to have an expensive computer but then reporting it operates with an 800 MHz clockspeed.

[shoop]

Had you even heard of Halliburton before Moore's preachy little Fahrenheit 911.

btw, does Moore's blatant hyprocrisy make you question his documentaries?

Aren't you just a bit curious as to what US company commissioned this poll?  I know I am.

I'll bet is was Halliburton.   

<{POST_SNAPBACK}>

[cybercoma]

The only "snapshot" that matters is the one on Jan. 23.

[Redragon]

Yeah I took Stats in University (albeit neither did very good or remember it fully) but the size, 18000 is just ridiculous. The reason the Canadian media take polls that are never higher than 2000 (more like 1500) people in size is that it is both inexpensive and accurrate. There is a margin of error of 2.5%, 19 times out of 20 with that sample size. Now going from a sample size of 1000 which is already fairly accurrate to one which is 18000 is weird because it increases the costs exponentially and it doesn't make it much more accurate (on a significant level). And the guy is right, 98% competency is just weird, if your trying to make it easier for the public to understand, at least post a mathematically correct number as a complement just so you have credibility.

And there is nothing against the pollster, the numbers could possibly be correct and I wouldn't even care, the poll was taken back in Nov 11-16.

[Montgomery Burns]

My Stats Professor in University must have been a moron. He told us that the larger the sample, the more accurate the poll.

It makes sense. The typical poll of 1000 people is much more accurate than a poll of 18,000.

Anyway, this poll has no credibility because the pollsters used the term 98% credibility and that is not a mathematical term.

It's the equivalent of having a Porsche but then reporting it has as much HP as a Chevy Cavalier.

[Montgomery Burns]

Had you even heard of Halliburton before Moore's preachy little Fahrenheit 911.

btw, does Moore's blatant hyprocrisy make you question his documentaries?

Aren't you just a bit curious as to what US company commissioned this poll?  I know I am.

I'll bet is was Halliburton.    

<{POST_SNAPBACK}>

<{POST_SNAPBACK}>

Actually, yes. Clinton awarded no-bid contracts to Halliburton for work when he invaded Kosovo--without authorization from Congress, and much much more importantly--the UN (who has sovereignty over every country on the planet).

Clinton is a neocon!!

[tml12]

Had you even heard of Halliburton before Moore's preachy little Fahrenheit 911.

btw, does Moore's blatant hyprocrisy make you question his documentaries?

Aren't you just a bit curious as to what US company commissioned this poll?  I know I am.

I'll bet is was Halliburton.    

<{POST_SNAPBACK}>

<{POST_SNAPBACK}>

Actually, yes. Clinton awarded no-bid contracts to Halliburton for work when he invaded Kosovo--without authorization from Congress, and much much more importantly--the UN (who has sovereignty over every country on the planet).

Clinton is a neocon!!

<{POST_SNAPBACK}>

Sometimes the truth is harder to swallow...

[sharkman]

Huh? so anyway, it will be interesting to see how the polls go this week. I predict both sides will be putting out friendly polls to sway opinion, but let's watch the CBC always pick the one that has the Liberals at the highest.

[tml12]

Huh?  so anyway, it will be interesting to see how the polls go this week.  I predict both sides will be putting out friendly polls to sway opinion, but let's watch the CBC always pick the one that has the Liberals at the highest.

<{POST_SNAPBACK}>

And the NDP 3rd highest.

It would be interesting to see how a Conservative government deals with the CBC.

Even better, I wonder if they'd cut CBC funding even more than the Liberals did.

[sharkman]

I think it would be hard to get a more balanced view out of the CBC without cleaning house. And if the CPC did that they would be accused of all manner of vile crimes. A tough nut.