Categories
politicalbetting.com » Blog Archive » Populus and Survation had almost the same raw figures for U
politicalbetting.com » Blog Archive » Populus and Survation had almost the same raw figures for UKIP support yet their headline numbers were very different
Both Populus and Survation operate in a very similar manner. Both do their fieldwork online, both weight for likely turnout, and both apply a formula to ensure that their samples are politically balanced.
0
This discussion has been closed.
Comments
I'm asking you tim,not Thatcher,again,on just immigration,how many more people do you think England can take in before the full signs go up - 90 million ? more ?
That's two lefty posts from you today Sean. Are you feeling alright?
Given that UKIP got about 3% last time out such an approach is always going to significantly reduce their assessed performance. Those that don't past weigh in that way will no doubt reflect the current puff but pollsters who have been around for a while have excellent reasons to use past voting. It works.
Whether it works as well for new kids on the block as it does for existing parties will be determined at the next election. My guess is that it will but it is clearly a bigger challenge for a pollster to get right than a movement between the established parties.
Self selecting surveys that feel they need to adjust data so much are not worth as much in my book as a randomly seeded mobile number survey. 94% of adults have a mobile http://www.mobilemastinfo.com/stats-and-facts/ so the odds of obtaining a truly random sample from randomly picking say 1500 numbers are going to be pretty good.
Landline surveys will probably be worse I'd say now as for instance I simply don't use one nor have it plugged in at my home during the day (Anecdotal, sample size 1 etc) but I think alot of people probably just use mobiles and have the internet plugged straight into the wall...
Does any polling company use a random seed generator of mobile numbers for it's surveys ?
For example the South East of England has a population density of 1,170 per square km.. compared with Netherlands of 1,048.
The question this raises is: Which approach is right?
I suspect that past vote recall is imperfect, though with an online panel you could at least poll people shortly after the general election itself, and use those responses to weight your sample, thus improving the accuracy.
The "closely identified with" question is, in my opinion, highly suspect, as it is asking the respondent to treat something as being fixed and unchanging which really isn't. People do change their political allegiances, and asking them to pretend that they haven't is asking for trouble in my opinion.
Then you throw in new political dynamics like Coalition government, and the rise of UKIP, and it's even more pays your money, takes your choice.
Still, it's the best guide to the electoral landscape that we have. And of course trends are always informative.
"She was unveiled as the new face of the RSPB this weekend.
But its more than one million members might be surprised to learn that Miranda Krestovnikoff, the BBC broadcaster, is championing the eating of some of Britain’s wildlife — more precisely, roadkill. The presenter has told how she wants more people to source their meat not from a local supermarket or specialist butchers, but rather, from the roadside.
She spoke after hosting a dinner party for friends featuring rat — fried and served with a garlic and soy sauce dip — from the B3347 near Sopley in the New Forest; fox, sautéd in garlic, from the Ferndown bypass, near Wimborne, Dorset; and badger chasseur, served with tomato sauce and croutons, picked up from the A354 near Salisbury.
http://blogs.telegraph.co.uk/finance/ambroseevans-pritchard/100025783/time-to-take-bets-on-frexit-and-the-french-franc/
"We have a minor earthquake in France. A party committed to withdrawal from the euro, the restoration of French franc, and the complete destruction of monetary union has just defeated the establishment in the Brignoles run-off election.
It is threatening Frexit as well, which rather alters the political chemistry of Britain's EU referendum."
I have no idea how Survation can UP the UKIP vote share tbh. ICM seems pretty good, is the October one due out soon ?
One thing, if I was ICM I'd up the proportion of mobile seeds to landline
Apparently she's in for a pay rise as she "only" got £120k pa as BBC economics correspondent....
Nice work at the taxpayers teet -
Something tells me that all this discussion about roadkill, will merely bring back the busty girls in tight t shirts on PB's advertising.
+1000.
"President Hollande's Socialists were knocked out in the first round, due to mass defection to the Front National by the working-class Socialist base. The Socialists thought the Front worked to their advantage by splitting the Right. They have at last woken up to the enormous political danger "
a) Populus is well-respected political pollster, and they ran their on-line surveys for a couple of years before switching, refining their weightings to match the phone polls.
b) It is well-known that on-line panels are very prone to self-selection bias (supporters who are fired up are disproportionately likely to respond), and we know that Kippers are fired up with their (admittedly rather quaint) view that UKIP represents something new and exciting. Therefore one would definitely expect that the UKIP score should be weighted downwards in any online poll; one can argue about how much it should be weighted down by, but it streteches credulity to see that Survation actually weighted the raw UKIP number up.
If you want to get angry about BBC wages I'd look at the management rather than at highly-skilled, well-regarded correspondents. That's where the bloat really is. I imagine the BBC could comfortably half the salaries of all its senior managers with ease. Some might walk, but I'd doubt they'd be hugely difficult to replace.
It's hilarious - who do they think falls for this stuff?
It's not hilarious that you do so, but it is rather peculiar!
https://www.youtube.com/watch?feature=player_embedded&v=hKjPosgpxlw
http://www.telegraph.co.uk/news/celebritynews/10367545/Did-Stephanie-Flanders-flee-the-BBC-because-of-James-Harding.html
"I might add that the Front is nothing like Ukip, a mostly pro-American, Right-leaning, libertarian, anti-welfare, free-market party. Marine Le Pen is an ardent defender of the French welfare model. Her critique of capitalism gives her a Leftist hue. Some call it 1930s national socialism, and here we are starting to touch on the populist appeal. She fulminates against Washington and Nato, calling for France to retake its place as "non-aligned" voice in a multipolar world, and lashing out at the Gaulliste UMP for selling its soul to Europe and the Anglo-Saxon order."
So it would be unwise to read-across Front success in France to UKIP success in the UK - in particular where its support comes from.....
It's the C2s I'd be particularly worried by. The voting intentions of C2 voters are based on an unweighted base of 131 respondents, weighted up to 210. They split 17% Labour to 10% Conservative, compared to 29-37 in 2010. They have one of the highest rates of don't knows, at 30%.
And then, of course, one is left to consider what fraction of the sample actually lives in the Middle England Towns and Their Hinterlands, where the election will be decided. I'm continually amazed that pollsters manage to do so well, to be honest.
Hmmm interesting.
http://www.dailymail.co.uk/news/article-2458617/Mark-Harper-tells-time-failed-asylum-seeker-home-live-TV.html
Secondly like ICM we weight by likelihood to vote, and UKIP voters have a high LTV so would go up on that also.
There is also an adjustment for DK/REF (like ICM) which took the UKIP figure LOWER in a final step.
clear the proles - fee payer coming into the operating theatre.."
Oh France.
http://www.bbc.co.uk/iplayer/episode/b03c62xc/Sunday_Politics_West_13_10_2013/
Carratala was the PS proxy. He had their support , he was their candidate.
" Lopez came in first in the initial round of voting October 6, receiving 40.4 percent of the vote in an election in which 66.65 percent of the town’s 20,728 registered voters abstained. PCF candidate Laurent Carratala, supported by the PS and the petty-bourgeois “left”, finished a humiliating third, with 14.6 percent, and was eliminated from the second round. "
That's a sticker.
The statement "President Hollande's Socialists were knocked out in the first round" is just plain wrong, I'm afraid.
Bravo to the Home Office.
We now need to set a 'Bad Brother' TV in an asylum seekers hostel and let the real Brits decide who goes next.
Mark Denten @BBCMarkDenten
Work and Pensions Sec Iain Duncan Smith says Labour "carping about the benefits cap" when many working people don't earn that much
This is why Ashcroft's marginal polls are a godsend for GE predictions.
http://blogs.telegraph.co.uk/news/danhodges/100241245/madeleine-mccann-the-police-are-hunting-for-a-needle-in-a-haystack-should-we-really-be-drawn-into-this-circus/
At the end of day - the McCanns went socialising with their friends whilst their toddler was alone.
Leaving the children alone, whilst the adults went out, was certainly a questionable decision, but that shouldn't mean that finding out what happened isn't important.
Labour - the party of the shirking class.
How many years has IDS added together to reach his absurd figure of £83 billion?
You can trade without the certificate. The Share Certificate number and personal ID are the key.
I am still trying to make up my mind. The short term dividends promised are a lot better than deposit accounts.
"Data is weighted to the profile of all adults aged 18+. Data is weighted by gender, age, region and pastvote. Targets for the weighted data are derived from Office of National Statistics, 2011 Census data andthe results of the 2010 General Election"
http://webcache.googleusercontent.com/search?q=cache:UbL-oAqy1aAJ:survation.com/wp-content/uploads/2013/07/Data-Collection-Method.pdf
Oh I forgot to say that the balance of the £5,000 you deposited with the government's agents will be returned to you after it is no longer needed to eliminate the annual deficit.
George Osborne tweets from China that the delay will be a maximum of 13 working days.
I can see Narendra Modi making moves to cool the tension between the UK and India when it comes to border control, but he has to win first. If the UPA win again then it's only going to get worse as India will continue to export their poor and unemployed to neighbouring countries and nations with lax border controls.
Follow the money...
Mr. Max, those are good points on India and China. They're very different bags of monkeys.
In Malaysia a court has ruled Christians cannot refer to their god as Allah (apparently this has happened for ages there), only Muslims can:
http://www.bbc.co.uk/news/world-asia-24516181
This reminds me, writ small, of the story I heard about the Coptic Pope in Egypt being forced to appear on TV and disagree with one of his own bishops, who asserted that (in essence) Christianity was right and Islam was wrong [it was a matter of history].
I don't think anyone will try this, but it'd be fascinating to try and correlate favourite book genres and voting intentions.
Very amusing it was, too.