Deleted
Deleted Member
Posts: 0
|
Post by Deleted on May 12, 2014 11:54:03 GMT
YouGov for Sky News Ukip 31 Lab 25 Con 23 LD 9 Others 14 Where do Sky magic up the extra 2% of virtual voters? Rounding each figure to the nearest % can make the total figure > 100.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on May 12, 2014 11:58:35 GMT
UKIP should be feeling very confident going into the Euros and getting above 30% would be fantastic. If they can achieve 35% that would be ground breaking and Cameron would have to do some serious thinking.
|
|
|
Post by carlton43 on May 12, 2014 11:59:10 GMT
Where do Sky magic up the extra 2% of virtual voters? Rounding each figure to the nearest % can make the total figure > 100. Rounding should not exceed 100 whatever the difficulties.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on May 12, 2014 12:10:29 GMT
Rounding each figure to the nearest % can make the total figure > 100. Rounding should not exceed 100 whatever the difficulties. How then would you sensibly express these figures rounded: UKIP 30.6 Lab 24.6 Con 22.6 LD 8.6 Oth 13.6 (Total = 100)
|
|
Richard Allen
Banned
Four time loser in VUKPOTY finals
Posts: 19,052
|
Post by Richard Allen on May 12, 2014 12:10:33 GMT
Rounding each figure to the nearest % can make the total figure > 100. Rounding should not exceed 100 whatever the difficulties. It is standard practice in polling for rounding to potentially result in a total that is greater than 100.
|
|
|
Post by carlton43 on May 12, 2014 12:15:37 GMT
Rounding should not exceed 100 whatever the difficulties. How would you express these figures rounded: UKIP 30.6 Lab 24.6 Con 22.6 LD 8.6 Oth 13.6 (Total = 100) UKIP 31 Lab 24 Con 23 LD 9 Oth 13
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on May 12, 2014 12:17:25 GMT
How would you express these figures rounded: UKIP 30.6 Lab 24.6 Con 22.6 LD 8.6 Oth 13.6 (Total = 100) UKIP 31 Lab 24 Con 23 LD 9 Oth 13 Isn't that a bit inaccurate and arbitrary?
|
|
|
Post by carlton43 on May 12, 2014 12:17:45 GMT
Rounding should not exceed 100 whatever the difficulties. It is standard practice in polling for rounding to potentially result in a total that is greater than 100. Posing the question "Why round up then?"
|
|
|
Post by carlton43 on May 12, 2014 12:19:05 GMT
UKIP 31 Lab 24 Con 23 LD 9 Oth 13 Isn't that a bit inaccurate and arbitrary? Yes. Your point being? Why round up if it is all so close. Show the accuracy and the closeness.
|
|
|
Post by greenchristian on May 12, 2014 12:37:39 GMT
Isn't that a bit inaccurate and arbitrary? Yes. Your point being? Why round up if it is all so close. Show the accuracy and the closeness. You will almost certainly have to do some rounding somewhere, because otherwise you'll have a percentage with a dozen or more figures after the decimal point. I'd be quite happy with one or two figures after the decimal, but the media will inevitably round it to whole numbers, so I guess pollsters think they might as well make sure it's rounded the right way. Oh, and I think you mean precision, not accuracy. The two are very different things.
|
|
|
Post by carlton43 on May 12, 2014 12:41:12 GMT
Yes. Your point being? Why round up if it is all so close. Show the accuracy and the closeness. You will almost certainly have to do some rounding somewhere, because otherwise you'll have a percentage with a dozen or more figures after the decimal point. I'd be quite happy with one or two figures after the decimal, but the media will inevitably round it to whole numbers, so I guess pollsters think they might as well make sure it's rounded the right way. Oh, and I think you mean precision, not accuracy. The two are very different things. Indeed I do, so I should love to have your close definition of each word within this context.
|
|
|
Post by greenchristian on May 12, 2014 12:55:33 GMT
You will almost certainly have to do some rounding somewhere, because otherwise you'll have a percentage with a dozen or more figures after the decimal point. I'd be quite happy with one or two figures after the decimal, but the media will inevitably round it to whole numbers, so I guess pollsters think they might as well make sure it's rounded the right way. Oh, and I think you mean precision, not accuracy. The two are very different things. Indeed I do, so I should love to have your close definition of each word within this context. Since you ask... Accuracy means that the figure is correct Precision means that the figure is given with many more figures So, saying that polling puts the Patriotic Socialist Party at 0% is accurate, but not precise. Saying that polling puts UKIP at 75.43845239048753% is precise, but not accurate. For some reason, many people have an odd tendency to believe that a more precise figure is inherently more accurate.
|
|
The Bishop
Labour
Down With Factionalism!
Posts: 36,656
Member is Online
|
Post by The Bishop on May 12, 2014 12:57:59 GMT
UKIP should be feeling very confident going into the Euros and getting above 30% would be fantastic. If they can achieve 35% that would be ground breaking and Cameron would have to do some serious thinking. I may always be proved wrong within a fortnight, but I suspect their upper ceiling is around 30%. This does have a slight tinge of Cleggmania redux, IMO.
|
|
Richard Allen
Banned
Four time loser in VUKPOTY finals
Posts: 19,052
|
Post by Richard Allen on May 12, 2014 12:59:57 GMT
UKIP should be feeling very confident going into the Euros and getting above 30% would be fantastic. If they can achieve 35% that would be ground breaking and Cameron would have to do some serious thinking. I may always be proved wrong within a fortnight, but I suspect their upper ceiling is around 30%. This does have a slight tinge of Cleggmania redux, IMO. I think that you are correct.
|
|
|
Post by carlton43 on May 12, 2014 13:06:00 GMT
Indeed I do, so I should love to have your close definition of each word within this context. Since you ask... Accuracy means that the figure is correct Precision means that the figure is given with many more figures So, saying that polling puts the Patriotic Socialist Party at 0% is accurate, but not precise. Saying that polling puts UKIP at 75.43845239048753% is precise, but not accurate. For some reason, many people have an odd tendency to believe that a more precise figure is inherently more accurate. Explain in what way your very 'precise' figure for UKIP is 'not' accurate.
|
|
|
Post by East Anglian Lefty on May 12, 2014 13:07:56 GMT
In that UKIP aren't actually above 75% of the vote. One can be both precise and wrong.
|
|
|
Post by greenchristian on May 12, 2014 13:25:27 GMT
Since you ask... Accuracy means that the figure is correct Precision means that the figure is given with many more figures So, saying that polling puts the Patriotic Socialist Party at 0% is accurate, but not precise. Saying that polling puts UKIP at 75.43845239048753% is precise, but not accurate. For some reason, many people have an odd tendency to believe that a more precise figure is inherently more accurate. Explain in what way your very 'precise' figure for UKIP is 'not' accurate. Because there has never been a single reputable poll ever that has shown UKIP as high as 40%, let alone 75%. We all know that UKIP is nowhere near 75%. It wouldn't matter how many significant figures I use, that figure is nowhere near the true figure. It is, therefore, inaccurate.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on May 12, 2014 13:27:25 GMT
Since you ask... Accuracy means that the figure is correct Precision means that the figure is given with many more figures So, saying that polling puts the Patriotic Socialist Party at 0% is accurate, but not precise. Saying that polling puts UKIP at 75.43845239048753% is precise, but not accurate. For some reason, many people have an odd tendency to believe that a more precise figure is inherently more accurate. Explain in what way your very 'precise' figure for UKIP is 'not' accurate. Because it does not reflect the real level of UKIP support very closely at all.
|
|
|
Post by carlton43 on May 12, 2014 13:28:14 GMT
Explain in what way your very 'precise' figure for UKIP is 'not' accurate. Because there has never been a single reputable poll ever that has shown UKIP as high as 40%, let alone 75%. We all know that UKIP is nowhere near 75%. It wouldn't matter how many significant figures I use, that figure is nowhere near the true figure. It is, therefore, inaccurate. I thought it was merely an example GC.
|
|
|
Post by greenchristian on May 12, 2014 13:32:20 GMT
Because there has never been a single reputable poll ever that has shown UKIP as high as 40%, let alone 75%. We all know that UKIP is nowhere near 75%. It wouldn't matter how many significant figures I use, that figure is nowhere near the true figure. It is, therefore, inaccurate. I thought it was merely an example GC. It was an example of a very precise polling figure for UKIP that would be inaccurate.
|
|