I predict a minority conservative government and possibly a slim majority. I also predict that people will see that Harper's real agenda is cutting waste, taxes and corruption. Not the scaremongering that the media is dreaming up.

Whenever you hear scaremongering tactics in the media just remember that this is the same biased media that tried to scare the hell out of everyone with Sars and tried to make our rave scene look like a violent gangsters paradise.

I dont buy into media hype.

I understand that these polls have been historically accurate, but my question is how are they representative of the voting population? How can one use 1000 or so individuals to produce statistics that are valid and measuring what it sets out to measure?

quote:

Originally posted by ericF I understand that these polls have been historically accurate, but my question is how are they representative of the voting population? How can one use 1000 or so individuals to produce statistics that are valid and measuring what it sets out to measure?

I recall that there is some mathematical formula that I learned in my 1st year statistics course at university that describes exactly how this is the case.

It's all based on a confidence factor and an error margin. In most of these polls, you'll see the confidence factor is 95% (Hence the "19 times out of 20" disclaimer). The error margin seems to generally be about +/- 2 or 3%. If you're comfortable enough with that error range and a 5% chance that you can be totally off (even more than the allowable error range), you can get by with a sample size that is just over 1000 people.

Obviously, if you wanted either a smaller error range, or an increased confidence factor, the sample size would have to increase.

I understand the statistics behind it, I'm sure the confidence intervals (significance levels) are legit. My issue is that these polls are not measuring what they are intending to measure accurately. Many of my courses in University were research based, I have become very critical as too believing the results of experimental research, polls, etc. From my understanding of research design, etc I know that constructs being measured in research, experiments, etc must be representational of what the purpose of the research is for (should be measuring what it is intending). In this case, I don't think these polls are at all representable of the actual voting population or done in a way that is representative of Canadian voting procedures. Perhaps I'd understand these polls better if I knew how the information was gathered.

With that being said, I guess we'll just have to wait till voting day to see how accurate these polls are.

quote:

Originally posted by Jayx1 I predict a minority conservative government and possibly a slim majority. I also predict that people will see that Harper's real agenda is cutting waste, taxes and corruption. Not the scaremongering that the media is dreaming up. Whenever you hear scaremongering tactics in the media just remember that this is the same biased media that tried to scare the hell out of everyone with Sars and tried to make our rave scene look like a violent gangsters paradise. I dont buy into media hype.

THANK YOU! All I've been reading about lately is how much people think Harper is going to screw up Canada. Like the Liberals havn't already? Politics is all about change. I'm voting Conservative today (advanced polling) for the reason Jayx mentioned.

___________________

*They say that time will make all this go away - but it's time that has taken my tomorrows and turned them into yesterdays*