UK General Election 2015
Discussion
eccles said:
turbobloke said:
AJS- said:
So like so much "research" which claims to prove something convenient, it's nonsense and statistical quirks dressed up as science.
But even so, if it's left it must be right.The point was that such carp is written by people on the left who are uncritical of their own output and its sources, and aimed at people on the left who swallow it right away because it's what they want to hear.
Look around on PH for criticism of CMD, Tory/Coalition policy e.g. energy, environment, police, foreign aid. There's a lot of it about and from Tory supporters too. Regardless of what's put out for public consumption, there's a far healthier appetite for scepticism and the questioning of data and the written word from those more right-field.
eccles said:
They all mess with the stats to suit their own ends.
Some much more than others. The vaguely left New Labour took spin and manipulation to a new level in this country under Bliar, Mandy and Alastair.eccles said:
turbobloke said:
Some much more than others. The vaguely left New Labour took spin and manipulation to a new level in this country under Bliar, Mandy and Alastair.
Have to agree with that. They professionalised what had been a bit amateur before and took it to a whole new level.Still, at least it's one of the two things Labour can claim to do well alongside bankrupting the country.
I was going to mention McPoison but thought the tricky trio above was enough for one post.
Gaspode said:
A sample size of 6000 is sufficient to give a confidence interval of 1.27% in a population of indeterminate size (i.e. large or unknown). Why do you think that an accuracy of pls of minus 1.27% is inadequate in this case?
The way that the mathematics works for statistical sampling is very interesting, and for large populations you can get significant results from tiny samples. Making the samples larger does not have a linear effect on the significance of the result, so having a ten times bigger sample doesn't give you a ten times more accurate answer.
To get a prediction for the entire population of the UK which has a confidence of +/- 10 % (i.e. if the answer is 50% it could be anywhere between 40 and 60%) then the sample size you need is just 96 people. The key is to ensure that they are selected by a genuinely random process.
that sounds very similar to the math climate scientists use The way that the mathematics works for statistical sampling is very interesting, and for large populations you can get significant results from tiny samples. Making the samples larger does not have a linear effect on the significance of the result, so having a ten times bigger sample doesn't give you a ten times more accurate answer.
To get a prediction for the entire population of the UK which has a confidence of +/- 10 % (i.e. if the answer is 50% it could be anywhere between 40 and 60%) then the sample size you need is just 96 people. The key is to ensure that they are selected by a genuinely random process.
Axionknight said:
h8tax said:
But Alex is running for a Westminster seat in GE2015 - his and Sturgeons positions are not relevant for the purposes descibed
He's only one man in one seat though so what can he do apart from bleat loudly?wc98 said:
Gaspode said:
A sample size of 6000 is sufficient to give a confidence interval of 1.27% in a population of indeterminate size (i.e. large or unknown). Why do you think that an accuracy of pls of minus 1.27% is inadequate in this case?
The way that the mathematics works for statistical sampling is very interesting, and for large populations you can get significant results from tiny samples. Making the samples larger does not have a linear effect on the significance of the result, so having a ten times bigger sample doesn't give you a ten times more accurate answer.
To get a prediction for the entire population of the UK which has a confidence of +/- 10 % (i.e. if the answer is 50% it could be anywhere between 40 and 60%) then the sample size you need is just 96 people. The key is to ensure that they are selected by a genuinely random process.
that sounds very similar to the math climate scientists use The way that the mathematics works for statistical sampling is very interesting, and for large populations you can get significant results from tiny samples. Making the samples larger does not have a linear effect on the significance of the result, so having a ten times bigger sample doesn't give you a ten times more accurate answer.
To get a prediction for the entire population of the UK which has a confidence of +/- 10 % (i.e. if the answer is 50% it could be anywhere between 40 and 60%) then the sample size you need is just 96 people. The key is to ensure that they are selected by a genuinely random process.
All the polls show Dave/Gideon leading Ed/Ed when the question 'who do you trust to run the economy' is asked. And it's by some margin too. Yet when you ask these same voters who they are going to vote for this lead disappears completely.
It will be fascinating to see how this plays out.
It will be fascinating to see how this plays out.
FiF said:
It just got even more confusing.
Latest Ashcroft poll, parachutes unprompted Greens into 4th place with 11%
CON - 29% (-5)
LAB - 28% (-)
UKIP - 15% (-1)
GRN - 11% (+3)
LDEM - 9% (+1)
Yet
and
I think it stays as a pass.
Latest Ashcroft poll, parachutes unprompted Greens into 4th place with 11%
CON - 29% (-5)
LAB - 28% (-)
UKIP - 15% (-1)
GRN - 11% (+3)
LDEM - 9% (+1)
Yet
and
I think it stays as a pass.
Guam said:
fblm said:
To be fair what gasplode has said is absolutely correct however it is not relevant to this study because despite the initial dataset being a huge 6000, the relevant sample size to estimate the IQ of each parties voters is the number of voters for each party in the dataset. In the case of BNP and Green that sample size would be roughly 7 and 24 respectively, rendering the results utterly meaningless.
Useful link for those not versed in the techniques http://en.wikipedia.org/wiki/Stratified_sampling
fblm said:
Is there a set protocol for forming a government in the UK in the event of no clear majority? If the Torys win the most seats but Lab plus SNP have more together can that unholy alliance actually legally govern?
Short answer no and yes.The Queen can ask anyone to form a govt. It is likely that she would ask the party who has won most seats first. But if they cannot command a majority that she would ask the next. So if Lab has an agreement with SNP that gets a majority then they will be asked.
In practice she would never actually ask until she knows the answer. SO they will all go through their machinations until someone has a majority. Unless no-one can, then the party with the most single seats (or votes?) will be asked. New territory as the constitutional precedents are weak or non-existent
Increasingly however it looks like 2 parties will not be enough. Which more than doubles the complexity
ETA I had missed the short answer to the first question
Edited by JustAnotherLogin on Wednesday 21st January 22:03
Axionknight said:
Out of interest, what, if anything can trigger another election? I've never known.
According to http://www.parliament.uk/about/faqs/house-of-commo...After the Fixed-term Parliament Act was passed on 15 September 2011, the date of the next general election is set as 7 May 2015. The Act provides for general elections to be held on the first Thursday in May every five years. There are two provisions that trigger an election other than at five year intervals:
- A motion of no confidence is passed in Her Majesty's Government by a simple majority and 14 days elapses without the House passing a confidence motion in any new Government formed
- A motion for a general election is agreed by two thirds of the total number of seats in the Commons including vacant seats (currently 434 out of 650)
Before this Act, the duration of a Parliament was set at a maximum of five years, although many were dissolved before that. The decision to call a general election was made by the Prime Minister by asking the Queen to dissolve Parliament.
Gassing Station | News, Politics & Economics | Top of Page | What's New | My Stuff