Post by Harry Hayfield on Nov 28, 2012 11:43:07 GMT
I have been tracking ICM's polls since the last election (working on the assumption that they are the most accurate out of the lot of them). When a new poll is published I look at the numbers that are published on the page before the summary and have come up with the following data for November 2012.
Labour 43% Conservatives 31% Liberal Democrats 10% Others 16%
with the Others made up of:
Plaid Cymru 1.69% SNP 3.81% Greens 2.33% United Kingdom Independence Party 7.20% British National Party 0.21% Others 1.06%
which gives Labour a 12% lead indicating a swing from Con to Lab of 10%
Last Edit: Jul 11, 2016 19:33:05 GMT by Deleted: Not monthly any more; thread title changed
"If you are going shoot musket balls at each other, at least have the decency to offer to pay the sixpence they cost!"
I did ask this of another poster who put up about polls. But is there section for voting intentions as in certainty to vote. Is this made up of people to vote only or from everyone they asked? Is the group of people unlikely to vote or don't knows growing?
I did ask this of another poster who put up about polls. But is there section for voting intentions as in certainty to vote. Is this made up of people to vote only or from everyone they asked? Is the group of people unlikely to vote or don't knows growing?
Most ask about it. The algorithm for applying relative turnout to the voting intention result is different for different pollsters.
I have been tracking ICM's polls since the last election (working on the assumption that they are the most accurate out of the lot of them). When a new poll is published I look at the numbers that are published on the page before the summary and have come up with the following data for November 2012.
Labour 43% Conservatives 31% Liberal Democrats 10% Others 16%
with the Others made up of:
Plaid Cymru 1.69% SNP 3.81% Greens 2.33% United Kingdom Independence Party 7.20% British National Party 0.21% Others 1.06%
which gives Labour a 12% lead indicating a swing from Con to Lab of 10%
Why would you want to take the figures from Britain's top pollster and mangle them so you use figures part way through rheir calculations ?
I have been tracking ICM's polls since the last election (working on the assumption that they are the most accurate out of the lot of them). When a new poll is published I look at the numbers that are published on the page before the summary and have come up with the following data for November 2012.
Labour 43% Conservatives 31% Liberal Democrats 10% Others 16%
with the Others made up of:
Plaid Cymru 1.69% SNP 3.81% Greens 2.33% United Kingdom Independence Party 7.20% British National Party 0.21% Others 1.06%
which gives Labour a 12% lead indicating a swing from Con to Lab of 10%
Why would you want to take the figures from Britain's top pollster and mangle them so you use figures part way through rheir calculations ?
Mangle them? I don't mangle them at all. I want to know how Plaid Cymru are doing and the page before the summary is the page that shows how many people in that sample said Plaid Cymru.
"If you are going shoot musket balls at each other, at least have the decency to offer to pay the sixpence they cost!"
Why would you want to take the figures from Britain's top pollster and mangle them so you use figures part way through rheir calculations ?
Mangle them? I don't mangle them at all. I want to know how Plaid Cymru are doing and the page before the summary is the page that shows how many people in that sample said Plaid Cymru.
But the 8 people in the survey who said Plaid is not necessarily the figure used in ICM;s final calculation of VI . The 8 is 8 out of 473 respondents in that table , It is also 8 out of 561 respondents in Table 2 . We do not know whether it is still 8 out of 699 in ICM's final VI calculation .
I remember one Welsh poll that showed that about 27% of the people of Merthyr were pro independence... I suspect we are dealing with a small sample and sampling error.
which is nonsense actually because in my view Women are more likely to vote.
In the last few elections the turnout was slightly higher amongst men than women so it could be right.
Another thing of course is that the pollsters round down the percentages so the 5% Tory lead could in fact be as high as 5.49% which is obviously closer to the 7.5% average for both men and women.
'The human situation, in general or in particular, is slightly worse (ignoring an occasional hiccup in the graph) at any given moment than at any preceding moment.'