|
Post by tonygreaves on Mar 17, 2017 22:00:44 GMT
I like the phrase "the sample is entirely random and will likely not resemble the electorate at large."
|
|
harryjoe
Conservative
Keep smiling
Posts: 139
|
Post by harryjoe on Mar 17, 2017 23:58:41 GMT
I like the phrase "the sample is entirely random and will likely not resemble the electorate at large." I guess you would prefer 'the sample to be simple like our electorate'
|
|
|
Post by tonygreaves on Mar 20, 2017 11:14:59 GMT
If the sample was entirely random it would be accurate. The reason samples are all rigged nowadays is because getting genuine random samples is near impossible.
|
|
|
Post by pepperminttea on Mar 25, 2017 12:15:58 GMT
If the sample was entirely random it would be accurate. The reason samples are all rigged nowadays is because getting genuine random samples is near impossible. Well even if you could take a completely random sample with every member of the electorate having an equally likely chance of being chosen (proportional to likelihood to vote) it is likely that your sample will still look different to what you expect the electorate as a whole to look like. For example if in your (entirely random) sample you find that 40% of people voted Labour last time you would have still have to underweight their share because your random sample has obviously picked too many Labour voters. It is the same thing with other factors like age, class, gender, remain/leave etc.
|
|
|
Post by tonygreaves on Mar 26, 2017 16:06:35 GMT
In which case it is not a random sample. It's been restructured for whatever reason.
|
|
|
Post by lancastrian on Mar 26, 2017 19:31:35 GMT
Raw data is entirely meaningless so there is little point quoting it because the sample is entirely random and will likely not resemble the electorate at large. I imagine the Lib Dems do better in the raw data than they do in the weighted data probably because they do significantly better with the kind of people likely to answer polls than those who typically don't while the converse is likely true of UKIP. Also people who typically only vote in general election years and don't pay much attention to politics other than that are disproportionately likely to vote either Tory or Labour. All true, but it's curious how Ipsos-MORI's entirely meaningless results are wrong in exactly the same way as Comedy Results' ones. All samples now are self-selecting, whether by having to actively join an online panel or by hardly anyone answering a phone survey. So it wouldn't be a great surprise if they were biased in the same way. Indeed the same people may well be on multiple online panels. For example pretty much all samples are/claim to be more likely to vote than the general population. I don't know about MORI or Com Res, but YouGov's samples (before weighting) are consistently too middle class and recall a majority remain vote in the referendum. Of course the pollsters changed their methods after 2015, and it remains to be seen how accurate they now are. Presumably the Conservatives lead would be smaller using the old methodology.
|
|
|
Post by marksenior on Mar 26, 2017 21:00:04 GMT
I take your point about comparing with a full Scottish poll such as Panelbase but note that the latest Panelbase poll on VI for the local elections has Indies and others at less than 1% whereas they are certain to get 10-12% rendering that whole " reputable " poll as fit only for the rubbish bin .
|
|
Sibboleth
Labour
'Sit on my finger, sing in my ear, O littleblood.'
Posts: 15,323
|
Post by Sibboleth on Apr 18, 2017 17:17:57 GMT
21pt lead apparently
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Apr 18, 2017 17:26:34 GMT
Was the fieldwork all done today?
|
|
|
Post by Andrew_S on Apr 18, 2017 17:32:17 GMT
Mike Smithson has deleted his Tweet with the Ipsos numbers.
|
|
Sibboleth
Labour
'Sit on my finger, sing in my ear, O littleblood.'
Posts: 15,323
|
Post by Sibboleth on Apr 18, 2017 17:38:29 GMT
oh
so fake news?
|
|
|
Post by marksenior on Apr 26, 2017 10:41:22 GMT
Ipsos/Mori in Evening Standard
Con 49 Lab 26 LDem 13 UKIP 4
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Apr 26, 2017 10:51:38 GMT
Ipsos/Mori in Evening Standard Con 49 Lab 26 LDem 13 UKIP 4 Full ES report.
|
|
The Bishop
Labour
Down With Factionalism!
Posts: 36,657
Member is Online
|
Post by The Bishop on Apr 26, 2017 11:36:53 GMT
Obviously the (unrealistically?) huge Tory lead will be what most focus on - but that UKIP figure is still astounding.
I see that no Green figure is cited - but could they move into 4th in at least some polls soon?
|
|
iain
Lib Dem
Posts: 10,788
|
Post by iain on Apr 26, 2017 12:03:49 GMT
Obviously the (unrealistically?) huge Tory lead will be what most focus on - but that UKIP figure is still astounding. I see that no Green figure is cited - but could they move into 4th in at least some polls soon? They were at 1% in this poll
|
|
The Bishop
Labour
Down With Factionalism!
Posts: 36,657
Member is Online
|
Post by The Bishop on Apr 26, 2017 12:04:45 GMT
Well that looks implausibly low.
|
|
|
Post by casualobserver on Apr 26, 2017 13:10:08 GMT
Could we be seeing (or not seeing, as the case may be) the effects of "ashamed Labour" voters in the opinion polls. These "ashamed Labour" electors will in an election actually turn out and vote Labour even if they are too ashamed or embarrassed to say so to pollsters because of Corbyn?
|
|
iain
Lib Dem
Posts: 10,788
|
Post by iain on Apr 26, 2017 13:14:21 GMT
I doubt it - but I do feel that quite a few Tory voters may not bother to turn out (think Blair in 2001) which would lead to the same outcome.
|
|
|
Post by Ruggerman on Apr 26, 2017 13:29:33 GMT
Could we be seeing (or not seeing, as the case may be) the effects of "ashamed Labour" voters in the opinion polls. These "ashamed Labour" electors will in an election actually turn out and vote Labour even if they are too ashamed or embarrassed to say so to pollsters because of Corbyn? 'Coybynistas'?
|
|
Sibboleth
Labour
'Sit on my finger, sing in my ear, O littleblood.'
Posts: 15,323
|
Post by Sibboleth on Apr 26, 2017 13:46:41 GMT
Could we be seeing (or not seeing, as the case may be) the effects of "ashamed Labour" voters in the opinion polls. These "ashamed Labour" electors will in an election actually turn out and vote Labour even if they are too ashamed or embarrassed to say so to pollsters because of Corbyn? It's certainly possible and seems to have occurred in 1983 and 2010. But who knows.
|
|