Tony Blair to give all profits to British Legion

[Ed209]

Oafc, of course the best way of doing these things is to go and ask all 40million adults in this country in a confidential and safe environment, i.e. an election. But that's not possible is it?!

Under the limited resources and communication routes available to the pollsters, the best scientific method of getting a result is the one they use. They randomise the samples, they will draw from all demographics, they carefully word the question, they use a large sample size (1008 is large Oafc!!), and they use a non-intrusive and confidential method of answering the poll. All this helps reduce something statisticians call sample bias. Then when processing the results they might very well use more complex methods than just a percentage to get their answer, (e.g. if in their sample a demographic wasn't well enough represented they might use proceedures for accounting for that in the result.) All this IS scientific. And they do get good results.

Edited August 19, 2010 by Ed209

[sifi]

Oafc, of course the most scientific way of doing these things is to go and ask all 40million adults in this country in a confidential and safe environment, i.e. an election. But that's not possible is it?!

Edited August 19, 2010 by sifimaster

[oafc0000]

Its not possible ? yet we have elections and referendums which are attempting to do exactly that...

[eFestivals]

... they use a large sample size (1008 is large Oafc!!)....

yep - it's certainly large enough to get an accurate result (to within plus or minus 2%) for a population size of the UK.

There have been instances (eg: the '92 election) where the pollsters results have been more out than that, but they've been the result of too many people being polled who had no intention of voting. Since the time of this anomaly, how they poll has been tightened up, and this problem has been pretty much removed.

A great example of just how accurate they are is from this year's election. The exit polls gave a result that few commentators believed to be true when it was published at 10pm on election night. But very quickly, when the first results came in, they changed their minds - and the exit polling turned out to be spookily accurate!!!

Dismissing polls on the basis of ignorance because you don't understand just how scientific the polling method is is the stuff of ... erm ... ignorance.

[oafc0000]

yep - it's certainly large enough to get an accurate result (to within plus or minus 2%) for a population size of the UK.

There have been instances (eg: the '92 election) where the pollsters results have been more out than that, but they've been the result of too many people being polled who had no intention of voting. Since the time of this anomaly, how they poll has been tightened up, and this problem has been pretty much removed.

A great example of just how accurate they are is from this year's election. The exit polls gave a result that few commentators believed to be true when it was published at 10pm on election night. But very quickly, when the first results came in, they changed their minds - and the exit polling turned out to be spookily accurate!!!

Dismissing polls on the basis of ignorance because you don't understand just how scientific the polling method is is the stuff of ... erm ... ignorance.

[eFestivals]

Random cold calling... So impressive

more impressive than your gross ignorance on this subject, that's for sure.

[gratedenini]

8.48 to 9.09. Finishing herself off is she?

[oafc0000]

Dont normally go with your posts..but I cant stop pissin mesell. Best retort of the month by far.

Why dont you go and give her one too Mr Oaf......

den

[Ed209]

Why not? Why isn't it possible? Why shouldn't there be a referendum asking the people whether or not we should go to war or not?

I know it'll never happen - they don't always even have a debate on the matter in the Commons - but in principle, I'd like to see a law passed that says that before we mobilise our troops, there has to be a majority of people in the country supporting it. It would certainly hammer home our responisbilities when we are asked are you willing for people to die for this cause - yes or no, at the ballot box.

[llcoolphil]

These polls we're discussing are run by private market research and media companies, last time i checked they couldn't call elections or referendums

[llcoolphil]

Oh I do, but it's not in it's purest form. It's applied.