| + Visit Cardiff FC for Latest News, Transfer Gossip, Fixtures and Match Results |
Fantastic - now is Wales a sporting country, or is it a "big event" country?Originally Posted by City123
Seriously, though, amazing figures. Only problem is, if that's the first ever football match some of the people watched, they may expect performances like that every time Wales play :-D
One note of caution - a sample of 610 is very small.
How true.
There are a lot of people born in the 60's including current pundits and media 'experts' who feel exactly the way you describe about England
So is it fantastic or possibly flawed because of the small sample?
It appears to be flawed. Possibly there are more complex algorithms at play - but there is no way you can gauge what 3m people were doing when you've asked just .02% what they were doing.
A sample size of 610 is actually pretty damn large, and massively statistically significant. A sample size of 70 is statistically significant, so 610 is more than ample.
And yes we are a "big event" country. We love a good reason to have a piss up. Thats about it.
EDIT, just looked it up. Assuming population of 3m, with a 95% certainty and a margin of error of 4%, you need a sample size of 601 to be considered accurate.
Last edited by Barry Dragon; 26-07-16 at 10:00.
We are a big event country - especially when the country is represented.
We know with International Rugby that a very small percentage of those who get motivated on international days actually follow club rugby, nor even rugby at all.
Football (in Wales) will have gained a few fans over the Euros and the excitement. Hopefully a lot of youngsters.
That's surely all good.
But yes, in reality we know that it was a National spectacle riding on a wave of 'C'mon Wales!-fever', involving a group of guys that many ladies found most alluring- and the nation knew all their names (it helps that in football, the name is on the back of the shirt ;) ) - etc etc.
I just hope it continues and grows, in the same vein
Like I said, there are likely to be more complex calculations at play. But, if you are simply asking 610 people what they watched on TV last night then that sample can be distorted by a number of factors. We've seen that in opinion polls recently.
In May 2015, the Tories were forecast to get between 271 and 285 seats (based on samples of more than 1000 (i.e. a 3% margin of error). They got 331 seats.
Last edited by Badly Ironed Shirt; 26-07-16 at 10:23.
Its statistics, its a best guess, its not 100% accurate, and can never be 100% accurate without asking everyone. In terms of it being an accurate guess, it would be deemed to be more accurate than most polls taken.
TV viewing figures and radio figures were taken of a sample size of around 2000-3000 for the UK TV and radio viewing/listening figures. So they would be considered less accurate than this poll. Although these days with Sky and Tivo boxes you can get a more accurate data from the boxes.
With regards to the polls being wrong on the GE that was down to the quiet tory. Social media and this forum have some very rabid labour voters that hold nothing but disgust for the tories, this discourages the tory voter and the maybe tory voter from putting out their voice. This is what is being attributed to the error in polls. The poll companies are basically still getting their heads around social media, and they have not taken into account things like the rabid left ranter stifling conversation and stopping tories putting out their voice. Also by putting too much weight on social media, this basically ignored the core tory voter. Those too old for social media.
I dont think this polls would be subject to the same aggression and misrepresentation.
So yes its not 100% accurate, but its a very accurate guess. Based on the accuracy I looked up, the true figure would be +/- 120k from the quoted figure.
Most of those sample size determinations only work correctly for normal data sets. I.e. a normal probability distribution.
The problem with electoral polling is the kind of people who seek out opinion polls or who bother to answer them are usually highly engaged politically, but don't necessarily reflect the wider population. So the raw data is manipulated to fit into the pollsters understanding of what usually happens.
In recent elections though this manipulation hasn't been correct. Political engagement has Changed more quickly than the models.
If you know whether the station's people watch follow a normal distribution then you can apply these confidence and reliability measurements.
Thinly veiled 'I have a friend working in the Industry' post imo
If it's not accurate, then it is flawed. Even the margin for error is potentially flawed.
In Northern Ireland in the 80's one of the highest watched programmes was "Pages from Ceefax" at 2am. This was because a number of the sample were falling asleep in front of their TVs.
Like I say, there are probably other factors at play, and algorithms. For example, maybe they are asking the sample of 610 "What did you watch?" "What would you have watched if you hadn't watched that?" . "How many people were in the room at the time"? etc.
Based on a sample of 610, I would say it is not possible to gauge what the other 2,999,390 people were doing. However, if more questions are asked, and more data extrapolated then I can see that there would be a higher degree of accuracy.
Posting Permissions
Reply With Quote