ixium said:A poll where its just confirmed members of the Canadian Forces would hold alot more ground in the land of statistics then just polling "random" people.
Shamrock said:Poll was conducted 19 and 20th of Dec.
53% of such a small sample from the Internet means SFA (and since it's the Internet, there's no gaurantee the voting bodies were unique individuals or even Canadian). The confidence interval probably dips below the 50% line, which is likely why it is not mentioned.
ixium said:I never said none were.
A poll where its just confirmed members of the Canadian Forces would hold alot more ground in the land of statistics then just polling "random" people.
BYT Driver said:If they did actually poll some CF pers, did they disallow the opinions that did not fit the poll???
???
Greymatters said:The online poll is insignificant - we've already proven on this very site that online polls can be skewed by a determined group of persons to reflect whatever message we want to send, and none of the poll results we did skew (which were promptly ignored by the host site) were given as much attention as the anti-mission results...
....Seventy-one per cent – up dramatically from 58 per cent in July – believe Canada is shouldering too much of the burden of the NATO Afghan mission, which involves nearly 40 countries and a total force of about 41,700 troops...
Shamrock said:Only if they were polling to see the attitudes of soldiers, otherwise it'd be statistically biased. Randomness is what they want. Getting that randomness on the Internet is pretty much impossible.
Dirt Digger said:I tend to not put a lot of faith into any phone polling...way to many opportunities to introduce bias into the results. Even when they claim accuracy to the 95th percentile (that 19 times out of 20 line) I seriously doubt their methods:
Biased towards the wording of the question.
Biased towards current events / topics of significant passion.
Biased towards those that have home phones (and not cells).
Biased towards those that are actually home during the surveying hours (usually standard daytime hours).
Biased towards those that actually answer their phones and do not call screen.
Biased towards those that are willing to participate.
Biased towards those that have the time to answer the survey.
And as mentioned...biased towards the poller, who may hang up on you if you don't give the "right" answer.
Try giving these examples while you're sitting in a room full of students going for Master degrees in health science. Then follow it up by saying, "but if your happy with a sample population of the unemployed, stay at home moms and the elderly...well knock yourself out." ;D