Canadian politics question about polls - Why is it that the polling results can keep changing from week to ...

Question

It seems polls keep changing - with one week the Conservative being close to the Liberals and then one week they are quite ahead. It does not make sense to me. People generally seem to always vote for the same party and don't keep changing their minds. Considering Canada's lack of interest in politics (ie. voting turnout) I find it surprising that any news about any political party can have such a profound effect and change people's minds. Everyone I've spoken to either did not know there was a budget or did not know what was in it. I think people are a little smarter than flip-flopping their support weekly. Most don't even know whats going and/or don't seem to care enough to change their mind. There must be something wrong with how these polls are being conducted. What do you think?

Answer 1

Yank, here, not a Canadian, but then, I don't know why you think this is a Canada-only question.

Who's carrying out each poll? Pollsters with agendas will get different results from pollsters who truly want to know public opinion.

Also, look at how the poll was conducted -- on-line surveys, or call-ins are NOT reliable, as they depend on that particular audience, and are skewed by who bothers to weigh in.

Look at the exact question-wording, too. That can make a huge difference in results. The most reliable ways of tracking over time is to ask exactly the same question, with the same wording. (Which, of course, can't be done with breaking news.)

Also, who are they polling? General population? Registered voters? Likely voters? These groups may well have different attitudes and affiliations. Also, figuring out who the "likely voters" are is really tricky. The methods for doing so likely vary, thus, the sample will vary.

What you might do is type some pollsters into your favorite search engine. I know that gallup.com (or org, forget which) has some on-line material talking about sampling, and methods, and such.

There is one final thing. The very nature of determining something about a population based on a sample has an inherent nuttiness to it.

The methods they use guarantee that 95% of the time, the sample result's range (using the margin of error to get the range) will contain the true population proportion -- which means that 5% of the time, they've got a wacky sample.

Hope these ideas help.

Answer 2

I'm from NY, but did go to school for politics and government, so here is my take for what it is worth. There are several important parts to any poll, particularly the the political type.

First, who is the sponsoring organization or conductor of the poll. For instance, can you trust a poll conducted by Quebec nationalists saying that 95% of all Quebecois are in favor of seperation. Probably not. However, if the Toronto Sun sponsors a poll being conducted by an independent firm saying that 95% of Quebecois support seperation then get ready to add another country to NATO.

The Second important thing to look at is what was the sampling size and where is the location of those polled. Generally, the larger a sampling size, the more likely it is to be accurate. The more diversity in location, the more representational it will be as well. For example, a poll of 10 of your neighbors does not tell you what the sentiments of an entire city are. However, a poll of 100 residents in all 5 neighborhoods in a city would. Similarly polling only one province would not tell you what all of Canada feels about national questions.

My thinking as that as far as national polling goes in Canada, it is even less relevent than in the U.S. because of the nature of the parlimentary system. Since the P.M. is chosen based on the number of P.M elected from their party, it would seem to me that individual district polls would be more important. In the States, a statewide poll is comparative in importance, but not size.

As far as your question about the fluctuation goes, I would not look at it from poll to poll, but rather from month to month and always look at the same poll, such as from the same sponsor and firm. Hope this helps.

Answer 3

These so called polls are taken by only a small portion of the population in a certain region. Why it changes so drastically is the region gets changed, so therefore so do the polls.

Answer 4

While we're at it, I think the US should be MORE LIKE Canada!

1- Less obsession with right-wing crap rhetoric
2-Universal health care
3-Easier to score some greens.

My justification for the statement? A very Republican sounding one:
I'm wealthy already so I dont care.I know what I like.

Answer 5

I like a query that offers me the alternative of giving a exact-from-the-relaxation reply. I wager you be successful both method... for the reason that I adapt. That... or probably I simply covet 2 features as though 2-thousand features. I dunno.

Answer 6

Canada? What is that?