Bill of Rights trashed in "war" on terrorism

[SpencerH]

Fine, but its not "completely unreliable".

[Boris Godunov]

Originally posted by Giancarlo
CNN is leftist.

And 100,000 votes for yes is enough for me to make up my mind.

Ok, you admit to being swayed by polls. So we know you have a weak mind. 100,000 votes out of 250+ million people, and not a random sample. Duh.

CNN does an okay job when it came to those percentages because that is pretty much the opinion of the center, and center-left.

No, it's only the opinion of people choosing to participate in the poll on the Web. There is no way of ever determining the political bents of those participating.

It's an inaccurate "snapshot" poll. It has no validity in determining actual public opinion.

What country are you living in? Somalia? Because anybody can walk right into a internet cafe and get access to computers paying for several hours which costs almost nothing.

Huh? Ok, what Joe Schmoe factory worker in the Midwest is going to spend is time in an internet Cafe on cnn.com? Or an urban single mother?

You're basically saying that only the opinions of the middle- and upper-class matters in America. Well, that is quite a Republican attitude...

Common knowledge: In order for a poll to be an accurate representation of public opinion, it has to be a random sample. Period. This is not a random sample, ergo it is not an accurate poll.

[Boris Godunov]

Originally posted by SpencerH
Fine, but its not "completely unreliable".

It is completely unreliable as a guage of national public opinion.

It is only reliable as a guage of the opinions of 100,000 people who have access to computers who visited the CNN.com web site and decided to vote in the poll.

Even then, it is not reliable, as all one has to do to vote multiple times in the poll is move to a different computer. That makes it even more unreliable.

[Giancarlo]

Ugh, and it keeps and going and going and going, nothing stops the energizer bunny...

I feel America is backing Bush overwhelmingly and I still think that is the way it is.

You're basically saying that only the opinions of the middle- and upper-class matters in America. Well, that is quite a Republican attitude...

ERR Wrong. I never said that. A republican attitude? How about you neglecting the people a tax cut? Now that is the democrat attitude...

[Boris Godunov]

Giancarlo: When did I ever "neglect" the people a tax cut? You're assuming I believe something that I've never stated here. For all you know, I could be 100% in favor of the tax cuts.

And what does a tax cut have to do with the accuracy of this CNN poll?

You're just thrashing about aimlessly now.

[SpencerH]

Originally posted by Boris Godunov

It is completely unreliable as a guage of national public opinion.

I agree, but then so are all polls on "national public opinion". Its ludicrous to believe that sample sizes of 500->1000->50,000 (and I've never seen one that size anyway) are representative of 280M people in such a diverse country as the USA.

Even then, it is not reliable, as all one has to do to vote multiple times in the poll is move to a different computer. That makes it even more unreliable.

I'm sure it happens but would you bother? I wouldnt and I dont think many others would either.

[Boris Godunov]

Originally posted by SpencerH
I agree, but then so are all polls on "national public opinion". Its ludicrous to believe that sample sizes of 500->1000->50,000 (and I've never seen one that size anyway) are representative of 280M people in such a diverse country as the USA.

Actually, such polls, when the proper methods are employed, tend to be very accurate in guaging opinion. Look at the Zogby presidential poll in 2000. The day before the Election, it predicted Gore - 49%, Bush - 48%. It was right on the money. That's because they used a good random sample of not elligible voters, but likely voters, based on past voting history.

For a poll to guage national opinion accurately, it merely needs a completely random sample. If the sample is truly random, even a 600-person poll can produce a pretty accurate result. Of course, they all have their percentage of error, and the more you poll, the lower that percentage goes.

But to compare a voluntary Web poll's accuracy to a random sample is apples and oranges. A random sample is, indeed, far more accurate, even if it isn't 100%.

I'm sure it happens but would you bother? I wouldnt and I dont think many others would either.

You'd be surprised at how far people would go to express their political opinions...

[SpencerH]

In the case of the presidential election there was no biased wording in the poll to complicate the choice and the choices were limited (in more ways than one). Even so the "margin for error" on such small samples is usually plus and minus 3-4% which meant that they actually predicted Gore 45-53% and Bush 44-52% which is not really a prediction at all.

The CNN poll