Are Polls Significant?

[RNG]

I have for years been skeptical about polls as announce in the MSM. I had been contacted by pollsters several times in my life and each experience was painful in one way or another. The most famous was a Polaris poll done shortly after the Kyoto conference, where the dingbat asked something like "do you agree that the Canadian government should try to prevent the entire world from being destroyed?". And they proceeded to say in a press release that most Canadians supported Kyoto. And then there were the ones where they asked what I thought about the Alberta Conservative party, followed up by what my favorite laundry detergent was. And then there are the ones where they ask about the meaning of life but you have to answer yes or no. So I have given up. And all my friends and aquaintences have also.

So I now never answer the phone if I don't recognize the number. I let it go to the answering machine. Too much other BS going on. And most people I know are the same way.

So the next time I see a poll reported, in order not to violate Canada's truth in advertising laws I really think they should say, 1000 people who don't have a life and/or don't have call display think....................

Your opinion?

Edited September 29, 2010 by RNG

[Keepitsimple]

Because of the Census fiasco, we now know that "voluntary" information given through surveys are not to be trusted and exclude segments of the population. So when they say "accurate to withing 1.2% 19 times out of 20", don't believe it.....the experts have already said that's not true. Now on the other hand, if we threaten to fine them or send them to jail, it'll be accurate.....

[Wild Bill]

I have for years been skeptical about polls as announce in the MSM. I had been contacted by pollsters several times in my life and each experience was painful in one way or another.

Your opinion?

You might consider that the polls in the MSM are paid for by the MSM. They are only going to allocate so much money for such polls.

Political parties no doubt pay more money for more intensive and focused polls. Unless it is to their benefit it is very unlikely that they would routinely share the results with the citizenry at large.

As always, follow the money. Polls aren't cheap!

[scribblet]

On line polls are not significant at all as they can be manipulated. Polls by Angus Reid et al are sometimes just push polls, depends on who hired them, but IMO generally speaking polls are only meaningful around election time.

[Michael Hardner]

Because of the Census fiasco, we now know that "voluntary" information given through surveys are not to be trusted and exclude segments of the population. So when they say "accurate to withing 1.2% 19 times out of 20", don't believe it.....the experts have already said that's not true. Now on the other hand, if we threaten to fine them or send them to jail, it'll be accurate.....

It is true, when surveys are done properly.

If you're launching a survey to perpetuate a falsehood, then there are ways to do that successfully too but we need to demand more as a public than to accept such things at face value. "Beer Makes You Smarter" was the headline in the Weekly World News in the summer of 1985, as I recall.

That was the result of a survey that implied a causation between beer consumption and IQ, when no such link was established. If you fall for tricks then you're a fool, and in some respect we're all fools for such surveys.

Edited September 29, 2010 by Michael Hardner

[Michael Hardner]

On line polls are not significant at all as they can be manipulated. Polls by Angus Reid et al are sometimes just push polls, depends on who hired them, but IMO generally speaking polls are only meaningful around election time.

You need, if I recall, a few basic elements:

The results of the poll are published with the questions asked.

A significant portion of the population needs to have been sampled.

Random sampling is necessary - and this in itself is no easy feat.

All numbers, including non response rates, need to be reported.

[RNG]

You need, if I recall, a few basic elements:

The results of the poll are published with the questions asked.

A significant portion of the population needs to have been sampled.

Random sampling is necessary - and this in itself is no easy feat.

All numbers, including non response rates, need to be reported.

But my point still stands. The polls show what people without a life and/or call display think. That is not a random sample.

Edited September 29, 2010 by RNG

[Michael Hardner]

But my point still stands. The polls show what people without a life and/or call display think. That is not a random sample.

People who don't participate are called non-response, and if there is a high non-response rate then your poll will be crap.