Joined: 02 Mar 2009
Posts: 6966 votes: 3
Posted: Wed May 15, 2013 8:24 am Post subject:
for some reason provincial election polls have not been overly accurate out west , maybe people change there minds at last minute or they have a hard time getting alot of quality replies from potential voters . there has to be a bit more to explain how they got things so wrong
Joined: 04 Oct 2006
Posts: 7576 votes: 21
Location: The World
Posted: Wed May 15, 2013 8:25 am Post subject:
The thing with polling is that you either buy into it, or you don't.
If you don't then two instances where pollsters were extremely flawed in their data gathering will be used for years as a means to detract from a fairly useful tool.
The problem is they are still largely correct; (depending on the pollster)
The Polling in the 2011 Ontario, PEI, Manitoba, Newfoundland and Labrador Elections were all pretty close, as was the 2011 Canadian Federal Election.
All of them nailed it largely within the MoE,
I think that in the case of Alberta and BC these pollsters need to take a good hard look at the sample sizes which at least in BC's case appears to be the clear culprit, as well as Online polling where it appears party supporters are getting smarter (Ron Paul winning online strawman elections during the 2012 primary lead-up for example).
They are a handy tool, but they should be used solely for tracking trends and not for hard data which is the mistake the NDP made.
The plain fact is that it's very hard to get a genuine random sample anymore. At one time, pollsters could go to the phone book to get an adequate list of the voting public.
But new factors have emerged. First, cell phones -- and a proliferation of telephone companies -- makes the telephone directory obsolete. Second, the public is saturated with pollsters, and often refuses to participate in surveys. Pollsters often have to contact three or four -- or more -- people in order to get a willing participant. It's got so bad that one of the relatively high integrity pollsters goes to the internet to solicit self-selected opinions, probably because they think they get a better sample that way.
Believe me, no statistician would accept this as a truly 'random sample'. What normally happens is that the pollsters take samples of demographic subsections of the population, and then mathematically alter their raw data to produce an 'adjusted' result.
This is apart from all the games pollsters play, particularly in the phrasing of questions and the placement of questions in a sequence in the interview.
It's not so much that one 'buys into it' or not. National polls are essentially meaningless anymore. And that's what gets into newspapers.
This doesn't mean smaller, specialized polls aren't useful. They are, but those are private surveys done to guide those who manipulate the public ... the engineers of men's souls, as I like to call them ... but that is small-scale, and not revealed to the public.
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum You can attach files in this forum You can download files in this forum