Opinium

[Deleted]

Apr 21, 2013 12:20:36 GMT tonygreaves said:

Opinium gies LDs 8, ICM says 15.

Living proof that polls are wrong some of the time. Perhaps both are wrong in this case!

Tony Greaves

To be fair, the difference is philosophical. Yougov (and I think Opinium)  report the results of the poll. ICM alter the results in attempt to try and make a more sensible prediction, hence helping the liberals out.

Obviously this is probably aided by different samples in this case, but in theory both could be correct.

[The Bishop]

Worth noting (since few people have) that the UKIP surge has wiped out the tentative LibDem improvement since late last year.

Could we yet see that elusive 6% soon??

[Deleted]

yes we can

UK - Opinium/Observer poll: CON 26%, LAB 37%, LDEM 6%, UKIP 21%

[Deleted]

come back in 18 months time....

[Deleted]

indeed and UKIP really have to be careful to understand it is pointless peaking 2 years from an election

[Crimson King]

FWIW (ie not a lot) my impression is that the libdem dip on the yougov daily tracker following the UKIP surge in the Counties is already starting to recover.

[East Anglian Lefty]

Jun 2, 2013 7:37:14 GMT @ianrobo said:

indeed and UKIP really have to be careful to understand it is pointless peaking 2 years from an election

I'm not sure when they peak is entirely within UKIP's control. In any case, I don't think it matters that much to them. UKIP do not have a large enough cohort who are both able and eager to be MPs. No doubt they will get more, but any seats won in 2015 would be a major victory, so a slight die-back would not be a significant concern. More important is that, if they have reached their peak and will now fall back, they nevertheless fall back to a level higher than that they started from.

[Devonian]

Latest Opinium

LAB 36 (-1), CON 27(+1), Ukip 20 (-1), Lib Dems 7 (+1)

[Deleted]

Mike Smithson @msmithsonpb
LAB has 10% lead in tonight Opinium poll for Observer - other details not yet available

[russthomas]

For those who prefer scientific analysis to gut opinions - these are UKIP's easiest wins, based on ONS demographics.

HIGH NEW EU %, LOW NON EU % IMMIGRANTS (PLUS NEIGHBOURS OF)
........................................................................................UKIP....NATURAL
Constituencies...........................NEW EU %...NON EU %.......RATIO....PARTY

Boston and Skegness......................10.6.............2.7.......... 3.93..........UKIP
South Holland and The Deepings.......5.9.............2.0...........2.95...........CON
Mansfield......................................2.9..............1.7...........1.71...........LAB
Corby............................................6.7..............4.1...........1.63...........LAB
Cambridgeshire North East...............4.2..............2.8...........1.50...........UKIP
Brigg and Goole..............................2.2...............1.5...........1.47...........LAB
Crewe and Nantwich......................3.3...............2.4............1.38...........LAB
Herefordshire North.......................3.3...............2.6............1.27...........CON
Hereford and South Herefordshire...3.3...............2.7............1.22...........UKIP
Norfolk North West........................3.6...............3.0............1.20...........UKIP
Bognor Regis and Littlehampton.......4.2...............3.8............1.11...........UKIP
Lincoln..........................................3.9...............3.6............1.08............LAB
Merthyr Tydfil and Rhymney............1.5...............1.4............1.07............LAB
Llanelli...........................................1.9...............1.8............1.06............LAB
Wrexham.......................................3.6...............3.5............1.03............LAB
Scunthorpe....................................3.9...............3.8............1.03............LAB

[Devonian]

Jun 30, 2013 7:21:45 GMT russthomas said:

For those who prefer scientific analysis to gut opinions - these are UKIP's easiest wins, based on ONS demographics.

HIGH NEW EU %, LOW NON EU % IMMIGRANTS (PLUS NEIGHBOURS OF)
........................................................................................UKIP....NATURAL
Constituencies...........................NEW EU %...NON EU %.......RATIO....PARTY

Perhaps you could give a full explanation of this formula? In particular why who think its the ratio rather than the total numbers that are important and why you've included '(plus neighbours of)' in your formula?

[innocentabroad]

Jun 30, 2013 7:21:45 GMT russthomas said:

For those who prefer scientific analysis to gut opinions - these are UKIP's easiest wins, based on ONS demographics.

HIGH NEW EU %, LOW NON EU % IMMIGRANTS (PLUS NEIGHBOURS OF)
........................................................................................UKIP....NATURAL
Constituencies...........................NEW EU %...NON EU %.......RATIO....PARTY

Boston and Skegness......................10.6.............2.7.......... 3.93..........UKIP
South Holland and The Deepings.......5.9.............2.0...........2.95...........CON
Mansfield......................................2.9..............1.7...........1.71...........LAB
Corby............................................6.7..............4.1...........1.63...........LAB
Cambridgeshire North East...............4.2..............2.8...........1.50...........UKIP
Brigg and Goole..............................2.2...............1.5...........1.47...........LAB
Crewe and Nantwich......................3.3...............2.4............1.38...........LAB
Herefordshire North.......................3.3...............2.6............1.27...........CON
Hereford and South Herefordshire...3.3...............2.7............1.22...........UKIP
Norfolk North West........................3.6...............3.0............1.20...........UKIP
Bognor Regis and Littlehampton.......4.2...............3.8............1.11...........UKIP
Lincoln..........................................3.9...............3.6............1.08............LAB
Merthyr Tydfil and Rhymney............1.5...............1.4............1.07............LAB
Llanelli...........................................1.9...............1.8............1.06............LAB
Wrexham.......................................3.6...............3.5............1.03............LAB
Scunthorpe....................................3.9...............3.8............1.03............LAB

Their three "best" seats appear to be in the Fens. Is this an elegant way of saying that UKIP are bog-trotters?

[Deleted]

Jun 30, 2013 7:21:45 GMT russthomas said:

For those who prefer scientific analysis to gut opinions - these are UKIP's easiest wins, based on ONS demographics.

HIGH NEW EU %, LOW NON EU % IMMIGRANTS (PLUS NEIGHBOURS OF)
........................................................................................UKIP....NATURAL
Constituencies...........................NEW EU %...NON EU %.......RATIO....PARTY

Boston and Skegness......................10.6.............2.7.......... 3.93..........UKIP
South Holland and The Deepings.......5.9.............2.0...........2.95...........CON
Mansfield......................................2.9..............1.7...........1.71...........LAB
Corby............................................6.7..............4.1...........1.63...........LAB
Cambridgeshire North East...............4.2..............2.8...........1.50...........UKIP
Brigg and Goole..............................2.2...............1.5...........1.47...........LAB
Crewe and Nantwich......................3.3...............2.4............1.38...........LAB
Herefordshire North.......................3.3...............2.6............1.27...........CON
Hereford and South Herefordshire...3.3...............2.7............1.22...........UKIP
Norfolk North West........................3.6...............3.0............1.20...........UKIP
Bognor Regis and Littlehampton.......4.2...............3.8............1.11...........UKIP
Lincoln..........................................3.9...............3.6............1.08............LAB
Merthyr Tydfil and Rhymney............1.5...............1.4............1.07............LAB
Llanelli...........................................1.9...............1.8............1.06............LAB
Wrexham.......................................3.6...............3.5............1.03............LAB
Scunthorpe....................................3.9...............3.8............1.03............LAB

The final two columns make no sense

[Devonian]

The Observer article is peculiar. It says Labour have a 10 point lead over Conservatives. It says that UKIP are on 19(-1), Lib Dems on 7(nc) but it doesn't give the actual percentages for Labour and Tories. If I was reading that in a newspaper I had bought I would be annoyed.

www.guardian.co.uk/politics/2013/jun/30/labour-economy-poll

[russthomas]

For those who prefer scientific analysis to gut opinions - these are UKIP's easiest wins, based on ONS demographics.

Perhaps you could give a full explanation of this formula? In particular why who think its the ratio rather than the total numbers that are important and why you've included '(plus neighbours of)' in your formula?

The report is not a formula, just numbers in a spreadsheet that uses formulas, about half a million of them. The ratio is of importance because that is what is used in elections, not absolutes, - in simpler English it translates to ....

The higher the proportion of New EU (Polish) immigrants in employment within a constituency, the more likely the natives are to notice this, and feel a grievance that Brits are losing out on jobs.

New immigrants are highly unlikely to have registered to vote, it takes time to even learn that any EU citizen resident in another EU country CAN vote in local elections, and in many rural constituencies, in the early years, these residents actually live in holiday camps where they work, or agricultural worker dormitories. By 2020 any Polish residents still here will be registered to vote, so a UKIP surge will not then be anything like so easy. (Hence UKIP's blitz on Romanian and Bulgarians, the New New EU immigrants)

Non EU immigrants (India, Pakistan, Bangladesh) have been in the UK for many decades, most have native born children, almost all are registered to vote, and are organized to do so, Anti-UKIP, to try to offset some of the irrational xenophobia. Lots of these live in inner cities such as London, Manchester, Bradford, Leicester, Luton, Reading, Slough - these will be slim picking for UKIP in the next round of Council Elections.

So you might have a high New EU % in a seat, but if it is counterbalanced by a high Non EU immigrant %, the excess ratio cannot swing an election to UKIP.

The NATURAL SEAT column is an indicator for use in models that flex on Turnout %. In a competitive election, factions polarise support and each side drums up more people to vote - so Turnout increases. The party that would win the seat in a non competitive election is its NATURAL SEAT party, this is often affected by additional factors to the immigration %, notably the age demographic. UKIP's main activists are retired, male, working class, low income. In coastal towns on the East and South coast, and the West Country you get a lot of these sort of people, they act as a magnifier to the immigrant ratios. UKIP also utilize local newspapers and big hoarding adverts to spread the irrational fear of - "You could be next" to neighbouring constituencies, hence there are UKIP target seats lower down the list, which are not themselves high New EU %, but their neighbours are. All reasonably sophisticated Electoral Models use factors such as these, to calculate their results more precisely in each constituency, than extremely simple, formula based, systems such as Electoral Calculus use to work out overall ELECTION WINNERS only, which they do quite well. They openly admit that at present development levels, they cannot forecast who will be the opposition.

[Devonian]

Today's Opinium

Labour 37(+1)
Conservative 27(nc)
UKIP 19 (-1)
Lib Dem 7(nc)

news.opinium.co.uk/sites/news.opinium.co.uk/files/VI_25_06_2013_0.pdf

[russthomas]

The Observer article is peculiar. It says Labour have a 10 point lead over Conservatives. It says that UKIP are on 19(-1), Lib Dems on 7(nc) but it doesn't give the actual percentages for Labour and Tories. If I was reading that in a newspaper I had bought I would be annoyed.

Either the Observer journalist knows and understands how numbers work ... OR .... the journalist simply repeated what an employee of Opinium had written down ... and since such a person is almost certain to understand numbers, you have been given all the data necessary to know in a flash, that since LAB is +1 and UKIP -1, the poll is CON x, LAB x + 10, LIB 7, UKIP 19, OTH 10 ... so CON 27, LAB 37.

Simples

[Devonian]

Jun 30, 2013 9:06:36 GMT russthomas said:

For those who prefer scientific analysis to gut opinions - these are UKIP's easiest wins, based on ONS demographics.

Perhaps you could give a full explanation of this formula? In particular why who think its the ratio rather than the total numbers that are important and why you've included '(plus neighbours of)' in your formula?

The report is not a formula, just numbers in a spreadsheet that uses formulas, about half a million of them. The ratio is of importance because that is what is used in elections, not absolutes, - in simpler English it translates to ....

The higher the proportion of New EU (Polish) immigrants in employment within a constituency, the more likely the natives are to notice this, and feel a grievance that Brits are losing out on jobs.

New immigrants are highly unlikely to have registered to vote, it takes time to even learn that any EU citizen resident in another EU country CAN vote in local elections, and in many rural constituencies, in the early years, these residents actually live in holiday camps where they work, or agricultural worker dormitories. By 2020 any Polish residents still here will be registered to vote, so a UKIP surge will not then be anything like so easy. (Hence UKIP's blitz on Romanian and Bulgarians, the New New EU immigrants)

Non EU immigrants (India, Pakistan, Bangladesh) have been in the UK for many decades, most have native born children, almost all are registered to vote, and are organized to do so, Anti-UKIP, to try to offset some of the irrational xenophobia. Lots of these live in inner cities such as London, Manchester, Bradford, Leicester, Luton, Reading, Slough - these will be slim picking for UKIP in the next round of Council Elections.

So you might have a high New EU % in a seat, but if it is counterbalanced by a high Non EU immigrant %, the excess ratio cannot swing an election to UKIP.

The NATURAL SEAT column is an indicator for use in models that flex on Turnout %. In a competitive election, factions polarise support and each side drums up more people to vote - so Turnout increases. The party that would win the seat in a non competitive election is its NATURAL SEAT party, this is often affected by additional factors to the immigration %, notably the age demographic. UKIP's main activists are retired, male, working class, low income. In coastal towns on the East and South coast, and the West Country you get a lot of these sort of people, they act as a magnifier to the immigrant ratios. UKIP also utilize local newspapers and big hoarding adverts to spread the irrational fear of - "You could be next" to neighbouring constituencies, hence there are UKIP target seats lower down the list, which are not themselves high New EU %, but their neighbours are. All reasonably sophisticated Electoral Models use factors such as these, to calculate their results more precisely in each constituency, than extremely simple, formula based, systems such as Electoral Calculus use to work out overall ELECTION WINNERS only, which they do quite well. They openly admit that at present development levels, they cannot forecast who will be the opposition.

I don't think your model entirely works. You say that UKIP do better where there has been higher EU immigration but then it can also do well where there hasn't been but which is 'neighbouring' where people think 'you could be next'. Well frankly those two categories could include anywhere.

You still haven't said why you included 'plus neighbours of' in your original formula.

You've said it takes some time to learn EU citizens can vote in local elections. I don't think it takes that long.

You've assumed that immigrants (for example Nigel Farage's wife or Otto Inglis' wife) will themselves be opposed to restrictions on immigration.

You've assumed the children on immigrants will oppose restrictions on immigration.

You've stated that support for restrictions on immigration is an 'irrational prejudice' that is spread about by newspaper ads which suggests your analysis may be somewhat skewed by you own biases.

Now of course demographics are important for all parties but it involves a whole number of factors so I'm not really convinced by your formula.

[Devonian]

Jun 30, 2013 9:39:09 GMT russthomas said:

The Observer article is peculiar. It says Labour have a 10 point lead over Conservatives. It says that UKIP are on 19(-1), Lib Dems on 7(nc) but it doesn't give the actual percentages for Labour and Tories. If I was reading that in a newspaper I had bought I would be annoyed.

Either the Observer journalist knows and understands how numbers work ... OR .... the journalist simply repeated what an employee of Opinium had written down ... and since such a person is almost certain to understand numbers, you have been given all the data necessary to know in a flash, that since LAB is +1 and UKIP -1, the poll is CON x, LAB x + 10, LIB 7, UKIP 19, OTH 10 ... so CON 27, LAB 37.

Simples

Except that the article doesn't say Labour +1 does it. A ten point Labour lead could also be
Lab 38, Con 28, UKIP 19, Lib Dem 7 Other 8
or
Lab 26 Con 26 UKIP 19, Lib Dem 7 Other 12

Now since Other was 10 two weeks ago you might guess it was the same but the article doesn't say this besides which readers of the newspaper can hardly be expected to remember what percentage 'Other' got in a two week old poll.

It also wouldn't take into account rounding up and down which can give you a total of 101 or 99.

[Pimpernal]

Jun 30, 2013 9:06:36 GMT russthomas said:

Non EU immigrants (India, Pakistan, Bangladesh) have been in the UK for many decades, most have native born children, almost all are registered to vote, and are organized to do so, Anti-UKIP, to try to offset some of the irrational xenophobia. Lots of these live in inner cities such as London, Manchester, Bradford, Leicester, Luton, Reading, Slough - these will be slim picking for UKIP in the next round of Council Elections.

Whilst you hanker on about the imagined xenophobia, you might be surprised that most non-EU immigrants that I have canvassed do not feel UKIP is racist or xenophobic and in large measure tend to support us - 18 out of 23 people I canvassed (OK that's a very small sample but...) were already inclined towards voting for us. I don't know yet how many of them actually voted, but 2 of my nominations were from Asian women.