ICM

[marksenior]

Interestingly , the May ICM Guardian poll had a Scottish subsample with 61% Labour voters . I don't recall Labour supporters saying that meant that the VI figure of 34% in the overall poll was too high .

[The Bishop]

I didn't know that - but arguably it only balanced out the BNP getting 37% (or whatever it was) in Wales anyway

[Deleted]

Mar 11, 2013 20:18:58 GMT The Bishop said:

UKIP down 2 and the LibDems at 15?? I smell b***shit.....

Quelle surprise..... :rolleyes:

[Deleted]

Jun 17, 2013 13:38:16 GMT The Bishop said:

I didn't know that - but arguably it only balanced out the BNP getting 37% (or whatever it was) in Wales anyway

Good job nobody over analyses opinion polls then. Oh....

[The Bishop]

Jun 17, 2013 18:11:52 GMT doktorb said:

Mar 11, 2013 20:18:58 GMT The Bishop said:

UKIP down 2 and the LibDems at 15?? I smell b***shit.....

Quelle surprise..... :rolleyes:

The subsequent survey(s) fully justified my scepticism there, chum.

[Deleted]

Jun 17, 2013 18:25:57 GMT The Bishop said:

Jun 17, 2013 18:11:52 GMT doktorb said:

Quelle surprise..... :rolleyes:

The subsequent survey(s) fully justified my scepticism there, chum.

No doubt. Trolls always find food...even in scraps..

[The Bishop]

Jun 17, 2013 18:54:49 GMT doktorb said:

Jun 17, 2013 18:25:57 GMT The Bishop said:

The subsequent survey(s) fully justified my scepticism there, chum.

No doubt. Trolls always find food...even in scraps..

Glass houses, stones, etc etc.

[Deleted]

atrick Hennessy @patjhennessy
SunTel/@icmresearch Wisdom Index poll - Lab 34 (+2) Con 29 (-) LD 15 (-1) Ukip13 (-2)

[Pimpernal]

Tories draw LEVEL with LAB in latest fkng awful ICM poll
CON 36% (+7 )
LAB 36% (nc)
LD 13% (+1),
Ukip 7% (-5),

[Deleted]

Well, I'll be... Only one poll, but it's encouraging. It's odd when we have pollsters putting the gap between Tory and UKIP ranging between 1 to 29 points. I'll be doing my monthly averages at the start of August.

[The Bishop]

It will be very interesting to see what other polls say this week. I certainly have my suspicions, but will remain silent for now

[AJS]

Something really needs to be done about this continuing huge gap in UKIP support from different pollsters. 7% to 20% isn't credible.

[greenchristian]

Jul 15, 2013 18:31:30 GMT AJS said:

Something really needs to be done about this continuing huge gap in UKIP support from different pollsters. 7% to 20% isn't credible.

The biggest single cause of that gap seems to be the difference between telephone and internet polling. Internet polls (with the notable exception of YouGov) consistently post much higher figures for UKIP than telephone polls. The problem is that we don't know which set of polls are a better reflection of public opinion. Are UKIP over-represented in internet polls (if so, why?), under-represented in telephone polls (if so, why?), or both?

Whilst most of us can probably work out which option we would prefer to be true, working out which one actually is the case is incredibly difficult until we have a general election to test our theories against.

[Deleted]

I would tend to think the option most similar to the status quo is the more likely to be true. Even 7% would be a real coup for UKIP in a general. Think the last time a 4th party has reached those levels is 1923.

[marksenior]

Jul 15, 2013 19:49:59 GMT greenchristian said:

Jul 15, 2013 18:31:30 GMT AJS said:

Something really needs to be done about this continuing huge gap in UKIP support from different pollsters. 7% to 20% isn't credible.

The biggest single cause of that gap seems to be the difference between telephone and internet polling. Internet polls (with the notable exception of YouGov) consistently post much higher figures for UKIP than telephone polls. The problem is that we don't know which set of polls are a better reflection of public opinion. Are UKIP over-represented in internet polls (if so, why?), under-represented in telephone polls (if so, why?), or both?

Whilst most of us can probably work out which option we would prefer to be true, working out which one actually is the case is incredibly difficult until we have a general election to test our theories against.

See my post on the Populus poll thread re Party ID weighting . This goes a long way to explaining the difference between the Yougov and Populus UKIP poll figures and those of the other online pollsters .
Which is correct is as you say open to doubt but I have been pointed towards some evidence from US polling that Yougov and Populus are more likely to be on the right lines .

[David Boothroyd]

Peter Kellner explains how ICM is so different: yougov.co.uk/news/2013/07/16/are-labour-and-tories-really-level-pegging/

[The Bishop]

Jul 15, 2013 20:30:20 GMT @joe said:

I would tend to think the option most similar to the status quo is the more likely to be true. Even 7% would be a real coup for UKIP in a general. Think the last time a 4th party has reached those levels is 1923

But this IMO sums up where some polls might have been going wrong lately.

Opinion surveys were set up as snapshots of what the public think at that particular moment - NOT predictions of what might happen some time in the future (which is why the traditional lead question is "how would you vote if a general election was held *today*?") UKIP may or may not poll 7% in 2015 - indeed, the ICM figures could well be quite close to the overall result then - but I don't think they would get anything like as low as that *now* (assuming they stood everywhere in GB, anyway)

Similarly, the now infamous ICM "adjustments" of former LibDem voters et al are to all intents and purposes predictions of future voting behaviour. Not what polls should be for, tbh.

(of course, this also b***ers up comparisons with historic polling - another reason why it is undesirable)

[Deleted]

Jul 16, 2013 9:15:03 GMT The Bishop said:

Jul 15, 2013 20:30:20 GMT @joe said:

I would tend to think the option most similar to the status quo is the more likely to be true. Even 7% would be a real coup for UKIP in a general. Think the last time a 4th party has reached those levels is 1923

But this IMO sums up where some polls might have been going wrong lately.

Opinion surveys were set up as snapshots of what the public think at that particular moment - NOT predictions of what might happen some time in the future (which is why the traditional lead question is "how would you vote if a general election was held *today*?") UKIP may or may not poll 7% in 2015 - indeed, the ICM figures could well be quite close to the overall result then - but I don't think they would get anything like as low as that *now* (assuming they stood everywhere in GB, anyway)

Similarly, the now infamous ICM "adjustments" of former LibDem voters et al are to all intents and purposes predictions of future voting behaviour. Not what polls should be for, tbh.

(of course, this also b***ers up comparisons with historic polling - another reason why it is undesirable)

Its quite useful to have polls with different philosophies and allows different questions to be answered.

[The Bishop]

I agree, but you can do that without turning polls into predictions (as most other pollsters save ICM demonstrate quite well)

[tonygreaves]

On the back of this one possibly freak poll Farage gets a long, serious and friendly interview on the BBC. The BBC are such prejudiced crap.

Tony Greaves