Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Dec 1, 2014 22:35:21 GMT
Bad poll for the Conservatives, good poll for the Green party Con 28% (-2) Lab 31% (+1) LD 9% (NC) UKIP 18% (-1) Green 7% (+3) Others 7% (NC) Don't mean to pick on Devonian but surely putting the figures in the percentage they have rather than the amount of MPs they have at the moment makes more sense to the figures. Why can't pollsters or anyone for that matter put numbers into league table form? it currently isn't Southampton 26 Man City 27 Chelsea 33 Leicester 10 Newcastle 19 Man Utd 22 So why do people do the same for polls?  really hacks me off!!! |
|
|
|
Post by greatkingrat on Dec 1, 2014 22:57:05 GMT
I think it makes more sense to use a consistent order for polls, rather than chopping and changing depending on who has overtaken who.
|
|
|
|
Post by Devonian on Dec 1, 2014 23:11:08 GMT
Normally I tend to reorder the listed results into descending order. In this particular case I just copied and pasted from the ComRes twitter feed but in general I agree that descending order looks better
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Dec 1, 2014 23:48:26 GMT
It all stems from conventional practices. If you take a look at lists of historical voting intention figures from 1943 onwards, whether they be on Mark Pack's spreadsheet, UKPR or Wikipedia, they virtually all go in the order Con first, Lab second and Lib third. Partly because it's alphabetical but also the first two usually get the most support, and the Libs/LDs tend to come third - at least traditionally.
Some people (such as our own Mark Senior and others on UKPR threads) still don't even bother mentioning the parties when reporting, eg, 32/34/10/14 and just leave it at that.
Greens and UKIP come afterwards because they're the new kids on the block. I would agree that if you mention one poll in isolation, the parties ought to be listed in order of the highest amount of support. It would also be good practice, where possible, to break down the "Others" more often by mentioning the smaller parties (if they get at least, say, 1% after rounding). This is already beginning to become more commonplace.
But if you're doing a list or spreadsheet of a series of polls, as greatkingrat says, it's best to be consistent with the order.
|
|
middyman
Conservative
"The problem with socialism is that, sooner or later, you run out of other people's money."
Posts: 8,050 
|
Post by middyman on Dec 2, 2014 0:05:31 GMT
I think it makes more sense to use a consistent order for polls, rather than chopping and changing depending on who has overtaken who. "overtaken whom" Pedants anonymous |
|
|
|
Post by brothersideways on Dec 6, 2014 21:55:27 GMT
It's true that standardization helps with keeping spreadsheets updated. But I do think it creates a bias in people's minds, especially when parties tie in rounded up percentages.
I think it's possible that overly acute concern over standarisation could create a bias, which could in turn influence the outcome of votes. For instance, the Ashcroft poll for Wyre Forest has figures for people who vote Conservatve, Labour, Lib Dem, UKIP, Green, and BNP. But it groups the NHA into "some other party" despite their historical success in that constituency. It makes me wonder whether their methodology may be under-representing the NHA vote.
|
|
