The Election Section: Get Your Crystal Balls Out...

buglawton

chucknorris wrote: »

Taking a long term view, the short term pain of a labour party supported by the SNP might be my long term best result. Obviously horrendous in the short term, but it might convince most of the English voters that both us (Scotland and England/the rest of the UK) would be much happier and better off apart. Am I alone in thinking this?

The weird thing is, it may be the lack of PR that causes Scotland to become independent. Because no one can stand the disproportionate amount of power that one small region now has. Talk about unintended consequences.

Meanwhile a federally styled country like Germany, with real PR, actually got bigger after its unification with E. Germany.

BobQ

antrobus wrote: »

It is not an ICM assumption. It is a fairly standard opinion polling methodology to reallocate 'don't knows' on the basis of past voting.

Populus, for example, also do the same. And they do so because they have actual evidence to suggest that there is a good chance that undecided voters will return to the party they previously supported.

See: Spiral of Silence
http://www.populuslimited.com/polling.html

Fair enough it is used by some others. All I said was that ICM used it. In fact YouGov and Opinium do not. ICM and Populus do not use exactly the same method of doing this.

This is also quite an interesting review which explains why, all other things being equal, the polls will give slightly different outcomes.

http://ukpollingreport.co.uk/blog/archives/category/methodology

The point I was trying to make was that because 2010 resulted in a hung parliament it may not be typical of what usually happens.

BobQ

This may be of interest....

https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/60641/cabinet-manual.pdf

It explains the constitutional role of the PM after the election. While Cameron is entitled to explore his options of continuing, he will not be able to drag it out if he cannot show he has the confidence of the House. These arguments of legitimacy are irrelevant, it is just about having the confidence of the house.

antrobus

BobQ wrote: »

Fair enough it is used by some others. All I said was that ICM used it. In fact YouGov and Opinium do not. ICM and Populus do not use exactly the same method of doing this.....

What you said was,

... The ICM assumption that undecided votes will return to the party they supported in 2010 is quite dubious....

And the point I am making is that it is neither an "assumption" nor "dubious".

BobQ wrote: »

...This is also quite an interesting review which explains why, all other things being equal, the polls will give slightly different outcomes.

http://ukpollingreport.co.uk/blog/archives/category/methodology
..

Yes, it helps to read up on polling methodologies before you start knocking them.

...The point I was trying to make was that because 2010 resulted in a hung parliament it may not be typical of what usually happens.

You're not expecting a hung parliament this time round?

.string.

@ Shakey
Thank you for the

You can have one back

But I must set the record straight.

Seeing how this spat comes down to those dots, I can easily explain.

That post of mine was responding to This post by Generali where he posted first on the general subject of validity of people's votes and, changing the subject to a lighter vein, congratulating Bournemouth on achieving a Word Record in Zimmer Frames.

The dots were simply a separation between the two subjects, delineating between them. I could have used dashes but chose dots because it is easier to do as I struggle to write these posts on a mini tablet.

I hope that clears the matter.

again

antrobus

Over at UKPollingReport, Mr Wells has made his prediction;

Conservatives around 277 seats, Labour around 267, the Lib Dems around 29 and the SNP around 52.

With the caveat that he might change his mind a bit if the polls say something different tomorrow

An interesting piece anyway, as it provides a list of (alleged) battleground seats that will decide the outcome. (Incidentally, my constituency is one of those.:))

http://ukpollingreport.co.uk/blog/archives/9377

antrobus

.string. wrote: »

.....Seeing how this spat comes down to those dots, I can easily explain....

Tis an odd sort of place, this interweb of ours, where people can get offended by a row of dots.:)

BobQ

antrobus wrote: »

What you said was,

And the point I am making is that it is neither an "assumption" nor "dubious".

Well if you measure something (in this case opinion based on answers to questions you ask) you have a raw measurement of opinion based on fact (ie what you were told). If you then modify the measurement for any reason you cannot be certain about it is an "assumption" in my view.

You're not expecting a hung parliament this time round?

Of course I am, but if they are basing their calculation on how the undecided voted in 2010 there would have been a lot of tactical voting (as there will this time of course) so how people say they voted then may not show their allegiance this time. Just saying it is a source of error.

antrobus

BobQ wrote: »

Well if you measure something (in this case opinion based on answers to questions you ask) you have a raw measurement of opinion based on fact (ie what you were told). If you then modify the measurement for any reason you cannot be certain about it is an "assumption" in my view. ....

You cannot be "certain" about anything in an opinion poll. It's all a matter of statistics, which is just a branch of probability theory.

Every single opinion polling organisation modifies the "raw measurement of opinion based on fact" in some way. So in your view they would all be making "assumptions".

BobQ wrote: »

...Of course I am, but if they are basing their calculation on how the undecided voted in 2010 there would have been a lot of tactical voting (as there will this time of course) so how people say they voted then may not show their allegiance this time. Just saying it is a source of error.

There are any number of potential sources of error in an opinion poll. For one thing, they are all based on a c1,000 sample size, which is (statistically speaking) far too small to be in slightest bit valid. (And one reason why they all 'modify' the raw data, in order to try and remove that problem.) But for another thing, it is all based on people making statements about their likely future behaviour and, to be blunt, people are often lying bar stewards. They can adjust for that.

However whilst experience has taught pollsters about the kinds of lies people are likely to make, there is always the possibility that people will start telling a new set of lies. That is, they can take into account the things they know they don't know, but not the things they don't know that they don't know.

But one should always remember that there is probably a 50:50 chance that if they are wrong, they are wrong in the opposite direction that you want them to be wrong.:)

.string.

antrobus wrote: »

...wrong ...

How appropriate that the last word posted in this thread before the election was wrong.

Now what?