UKIP - The Future - Volume 3
Discussion
steveT350C said:
zygalski said:
steveT350C said:
http://www.itv.com/news/anglia/update/2015-03-11/s...
Bristol School Apologises for Comparing UKIP to Hitler
Dedicated school chairman of governors forced to resign by the school
headteacher after she discovered he had become a member of UKIP:
http://www.express.co.uk/news/uk/535099/Governor-f...
Edited by turbobloke on Thursday 12th March 18:35
I fully expect UK Independence Party to do much better than all the media smoothies suggest or anticipate but, not within student circles. That student mock election is a real eyeopener. Never expected much support from students let alone to be so strong. If that single result is a true reflection of the way youngsters now feel elsewhere in the UK, WOW! Just wow! That Farage has influence in that strata of the electorate is a real turn up!
All very interesting.
All very interesting.
HonestIago said:
If Farage goes for the jugular (Islam, PIE, Common Purpose, Establishment paedophilia etc) in the debates he will absolutely wreck all three LibLabCon leaders. I just worry that he may pull his punches a little and shy away from explosive topics. Regardless, I am still predicting massive support for UKIP in the GE way beyond what Ashcroft and Yougov tell us their support is.
I hope he does not become a target. Have you seen the protection that Mr Wilders in the Netherlands requires? Having said that, Wilders overwhelmingly focusses on Islam, which UKIP/Farage do not.Government announces sudden plans to extend HS1 to Thanet..
http://www.kentonline.co.uk/thanet/news/kent-to-be...
The anti-kipper propaganda doesn't seem to be working so let's buy the electorate.
http://www.kentonline.co.uk/thanet/news/kent-to-be...
The anti-kipper propaganda doesn't seem to be working so let's buy the electorate.
turbobloke said:
Yazar said:
Scuffers said:
This is all from a tv interview that was done over 6 months ago for channel 4.
Anyone want to bet on how long into the 1st TV debate Cameron or Clegg brings this headline up Or is that just the conspiracy of of liars at Yougov, Populus, Opinium, IPSOS Mori, ICM, Comres and Survation?
JustAnotherLogin said:
turbobloke said:
Yazar said:
Scuffers said:
This is all from a tv interview that was done over 6 months ago for channel 4.
Anyone want to bet on how long into the 1st TV debate Cameron or Clegg brings this headline up Or is that just the conspiracy of of liars at Yougov, Populus, Opinium, IPSOS Mori, ICM, Comres and Survation?
turbobloke said:
No conspiracy, of course not, it's just a case of the polling error of +/- 3% robbing you of a moment of gloating on the internet.
Who is gloating? just pointing out statistical evidence contradicting this continual bleat that saying nasty things about other parties is OK, but that saying nasty things about UKIP makes them stronger.That polling error applies to individual polls. When there are so many, and the trend on all types of poll and from all companies is consistently down over the last 6 months, you should consider changing your tune.
I also note that when UKIP share was rising in the polls, the error was not mentioned (with the honourable exception of Fif, who of course did also point out that it is the trend rather than individual results that matters)
JustAnotherLogin said:
turbobloke said:
No conspiracy, of course not, it's just a case of the polling error of +/- 3% robbing you of a moment of gloating on the internet.
Who is gloating? just pointing out statistical evidence contradicting this continual bleat that saying nasty things about other parties is OK, but that saying nasty things about UKIP makes them stronger.JustAnotherLogin said:
That polling error applies to individual polls. When there are so many, and the trend on all types of poll and from all companies is consistently down over the last 6 months, you should consider changing your tune.
Take three polls from two sources.SOURCE 1
Date 1
UKIP 17% LibLabCon 14% (you will surely get my drift - a single comparator)
Date 2
UKIP 15% LibLabCon 15%
Date 3
UKIP 14% LibLabCon 14%
SOURCE 2
Date 1
UKIP 17% LibLabCon 14%
Date 2
UKIP 15% LibLabCon 15%
Date 3
UKIP 14% LibLabCon 17%
The above polling results can reflect a reality in every single case where UKIP is never below LibLabCon, every result could represent a reality where both UKIP and the fictitious LibLabCons remained on 17% throughout. So in the real reality, UKIP support hasn't changed, in spite of any tomfoolery with the numbers that ignores error and possibly enhances error.
These are common mistakes. The results of one poll can't be reliably compared to another poll from a different source, the 3% margin of error is from a 95% confidence level not 100%, and regardless of media (and PH) bad habits it's not statistically valid to "average" a group of polls with a lot of nilly around the willy.
My tune is fine thanks, but you might like to go learn some stats.
turbobloke said:
These are common mistakes. The results of one poll can't be reliably compared to another poll from a different source, the 3% margin of error is from a 95% confidence level not 100%, and regardless of media (and PH) bad habits it's not statistically valid to "average" a group of polls with a lot of nilly around the willy.
My tune is fine thanks, but you might like to go learn some stats.
Common mistakes I didn't make thanksMy tune is fine thanks, but you might like to go learn some stats.
The trend is clear from each group of polls. Take the Yougov polls where they are using common techniques. The trend is clear from those
You can do the same individually with Populus and all the others.
As I said in my original post, the trend is clear for each and all of those organisations.
So I didn't compare across polls of a difference source, nor did I average across polls
What it is possible to do is take a set of polls from one source and hypothesise various possible trends of UKIP support. One can choose various shapes of trend (up, down, static, linear, logarithmic etc) and calculate how well the evidence (including the error) matches the trend.
What is clear is that the evidence over the last 6 months matches downward trends in support. Now that does not prove in any way that the trend will continue. But it is clear the probability that UKIP support has not dropped to some extent over the last 6 months is very very low.
JustAnotherLogin said:
turbobloke said:
These are common mistakes. The results of one poll can't be reliably compared to another poll from a different source, the 3% margin of error is from a 95% confidence level not 100%, and regardless of media (and PH) bad habits it's not statistically valid to "average" a group of polls with a lot of nilly around the willy.
My tune is fine thanks, but you might like to go learn some stats.
Common mistakes I didn't make thanksMy tune is fine thanks, but you might like to go learn some stats.
The trend is clear from each group of polls.
There is a trend if numbers change over time, but what's clear? It depends on the numbers, their nature and the error involved.
JustAnotherLogin said:
Take the Yougov polls where they are using common techniques. The trend is clear from those
What trend? The numbers aren't absolute.JustAnotherLogin said:
You can do the same individually with Populus and all the others.
Do what?JustAnotherLogin said:
As I said in my original post, the trend is clear for each and all of those organisations.
What trend?If the same poll gives 17, 16, 14 then these results are consistent with reality remaining at 16% (for example) so there may be no trend at all in reality. You appear to think that polls say more than they do.
JustAnotherLogin said:
What is clear is that the evidence over the last 6 months matches downward trends in support.
Depending on the actual numbers, which you haven't cited, it may also match no change, even if the % numbers decrease.Also, even within the same polling organisation, which uses precisely the same sample size and question wording, outcomes can be affected by the context of each poll e.g. if there's a major news splash on some controvesial issue. While a trend is an attractive creation, each poll is still just a snapshot with an error margin.
If somebody claimed there was a trend in the headline % figure of a particular Party in a particular poll then that's one thing, but what cannot be claimed automatically is that there is an identical trend in the support for that Party.
Those numbers are consistent with there being no trend, where the actual support for UKIP could have been 17% throughout and still given 17% to 14% polling outcomes due to the intrinsic polling error.
There's also mention of an average, to which more meaning is attributed than actually exists.
JustAnotherLogin said:
support for UKIP has increased over the last 6 months from about 17% to about 14% (using the 15 day moving average)
Those numbers don't tell you what happened to the support for UKIP. They're poll results with a sample size of ~1000 and a margin of error at +/- 3% approximately.Those numbers are consistent with there being no trend, where the actual support for UKIP could have been 17% throughout and still given 17% to 14% polling outcomes due to the intrinsic polling error.
There's also mention of an average, to which more meaning is attributed than actually exists.
though after it gets counted, the result counts 
JustAnotherLogin said:
Common mistakes I didn't make thanks
The trend is clear from each group of polls. Take the Yougov polls where they are using common techniques. The trend is clear from those
You can do the same individually with Populus and all the others.
As I said in my original post, the trend is clear for each and all of those organisations.
So I didn't compare across polls of a difference source, nor did I average across polls
What it is possible to do is take a set of polls from one source and hypothesise various possible trends of UKIP support. One can choose various shapes of trend (up, down, static, linear, logarithmic etc) and calculate how well the evidence (including the error) matches the trend.
What is clear is that the evidence over the last 6 months matches downward trends in support. Now that does not prove in any way that the trend will continue. But it is clear the probability that UKIP support has not dropped to some extent over the last 6 months is very very low.
deary me, another one. fancy putting your money where your mouth is ? above 14% in general election you pay 100 quid to the charity of my choice, below that i pay the same to the charity of your choice . surely if you are correct, with support plummeting for ukip by election time they will only be polling about 5% The trend is clear from each group of polls. Take the Yougov polls where they are using common techniques. The trend is clear from those
You can do the same individually with Populus and all the others.
As I said in my original post, the trend is clear for each and all of those organisations.
So I didn't compare across polls of a difference source, nor did I average across polls
What it is possible to do is take a set of polls from one source and hypothesise various possible trends of UKIP support. One can choose various shapes of trend (up, down, static, linear, logarithmic etc) and calculate how well the evidence (including the error) matches the trend.
What is clear is that the evidence over the last 6 months matches downward trends in support. Now that does not prove in any way that the trend will continue. But it is clear the probability that UKIP support has not dropped to some extent over the last 6 months is very very low.
Gosh you really do want to stick your head in the sand and deny the evidence
Data from here
http://en.wikipedia.org/wiki/Opinion_polling_for_t...
take YouGov/Sun- purely because they are most numerous (not just a few polls but 73 in the period I chose)
The most likely trend is that UKIP support has dropped from almost exactly 16% to almost exactly 14%
I do not deny that there is a chance that it has actually been static or even risen in this period, that's how statistics work. Indeed you could just about stay within the 3% tolerance (which I assume is a 2 sigma variation) for a hypothesis of static support. I'm just saying that it is far more likely that UKIP support has dropped
Data from here
http://en.wikipedia.org/wiki/Opinion_polling_for_t...
take YouGov/Sun- purely because they are most numerous (not just a few polls but 73 in the period I chose)
The most likely trend is that UKIP support has dropped from almost exactly 16% to almost exactly 14%
I do not deny that there is a chance that it has actually been static or even risen in this period, that's how statistics work. Indeed you could just about stay within the 3% tolerance (which I assume is a 2 sigma variation) for a hypothesis of static support. I'm just saying that it is far more likely that UKIP support has dropped
JustAnotherLogin said:
Gosh you really do want to stick your head in the sand and deny the evidence
Data from here
http://en.wikipedia.org/wiki/Opinion_polling_for_t...
take YouGov/Sun- purely because they are most numerous (not just a few polls but 73 in the period I chose)
The most likely trend is that UKIP support has dropped from almost exactly 16% to almost exactly 14%
I do not deny that there is a chance that it has actually been static or even risen in this period, that's how statistics work. Indeed you could just about stay within the 3% tolerance (which I assume is a 2 sigma variation) for a hypothesis of static support. I'm just saying that it is far more likely that UKIP support has dropped
You do realise that support may actually have risen as polls on voting intention may have seen other parties gain on percentage as people start to make up their minds on which way they will go. The polls show 16% of those who intend to vote will vote UKIP that figure maybe 1000 people for instance, those same polls may now show that 14% of those who intend to vote will now vote UKIP but it maybe 14% of 2000.Data from here
http://en.wikipedia.org/wiki/Opinion_polling_for_t...
take YouGov/Sun- purely because they are most numerous (not just a few polls but 73 in the period I chose)
The most likely trend is that UKIP support has dropped from almost exactly 16% to almost exactly 14%
I do not deny that there is a chance that it has actually been static or even risen in this period, that's how statistics work. Indeed you could just about stay within the 3% tolerance (which I assume is a 2 sigma variation) for a hypothesis of static support. I'm just saying that it is far more likely that UKIP support has dropped
I don't think the UKIP vote will have changed much as people have decided to vote for them over a period of time, I don't see their core voter number growing significantly as other parties have well oiled and well funded machines, but I don't think it will fall in actual voter number either.
Gassing Station | News, Politics & Economics | Top of Page | What's New | My Stuff