2141 replies to this topic

[LY_Scott]

Yus, im also fleeing dodge this Friday for a couple of weeks. We'll sort something out

[Rosssco]

Aye will do..

 

Interesting analysis by YouGov on how age groups voted..

 

 

[Zoobeef]

They've used different size groups there. Perfect example of how you can make data show anything :/

[JohnTurbo]

9 years, 15 years, 20 years, 5 years, ~20 years.

 

Good spot, and bizarre!!

[KurtVerbose]

Anyway, even if young people were pro and old people were no it doesn't mean you just have to wait a few years and the next vote will be pro. People change their minds as they get older.

[Dunk]

Aye will do..   Interesting analysis by YouGov on how age groups voted..    

Hardly insightful. That's simple mathematics. The more granular the sample range, the more detail will be available. The more generalised, the more you'll see occurrence of your overall result. You can drastically alter the "key message" dependent on how you represent the data. Oldest trick in the book on how to manipulate the masses. What would be interesting to see would be to segment based on a range of definitions. For example, by volume (i.e. Segments defined at roughly equal populations.) It only becomes more meaningful when the percentages are accompanied by volumes or % of base, to illustrate how significant the sample size is in context.

[Mangham54]

But ultimately there is only one important statistic... Total No v Yes votes cast.

[Dunk]

But ultimately there is only one important statistic... Total No v Yes votes cast.

Depends on what you want to know.

[Mangham54]

But ultimately there is only one important statistic... Total No v Yes votes cast.

Depends on what you want to know.

But is a going into a data analysis of age v vote actually going to change anything, either now or in the future? Blaming the grey is a fallacy, if the reasons to vote yes had been made in a more tangible manner to all then it wouldn't have mattered what the crinklys thought.

[Dunk]

But ultimately there is only one important statistic... Total No v Yes votes cast.

Depends on what you want to know.

But is a going into a data analysis of age v vote actually going to change anything, either now or in the future? Blaming the grey is a fallacy, if the reasons to vote yes had been made in a more tangible manner to all then it wouldn't have mattered what the crinklys thought.

I've not said anything about blaming anyone, nor that it would magically change the result. I don't think anyone believes statistics have influence over space & time. Unlike most of the comments that have been chucked around the wider debate & this thread, outside of any vote-rigging allegations, categorically & indisputably, the numbers from the results can be cut to give insight. I can spell out how that insight could be useful to either side of the debate if you like...

[Mangham54]

I understand the concept and use of data (I teach mathematics ) but ultimately data mining the voting habits by age is very much a distraction. There are much more important demographics that would be more worthwhile... However that would take much more effort.

[Rudy]

It is interesting to see the stats split by age, gender, region, combo of those etc. if only so the SNP can do some maths and figure when next to go for another vote. That'll be when everyone over 50 is dead then

[Dunk]

I understand the concept and use of data (I teach mathematics ) but ultimately data mining the voting habits by age is very much a distraction. There are much more important demographics that would be more worthwhile... However that would take much more effort.

I'd say that so long as there's such a key link between age profile and all demographics, it's going to be interesting and useful, not simply a distraction. Both sides will have very firmly have age banding a integrated into their campaigns as a key component of their demographic analysis and they'll want to know who they got their message to. It's a dependable and easily accessible data item, inextricably linked with important key campaigning battlegrounds. You can make broad, reliable assumptions from this very simplistic info: Propensity to be influenced by social media campaigns Propensity to be influenced by fears over pensions/interest rates Population within living memory of being personally impacted by more than one economic depression/downturn Damn it! You're good Chris. I broke my own rule about getting into forum discussions about such things. Damn you.

[slindborg]

it was such a fix, my facebook feed is telling me this

 

some of the comments in one stream are comedy gold... "there was no police officer at the stations".... well I've never ever ever seen a police officer at a polling station for ANY vote activity. as for the barcoding, I'd be happier with blank ones as it means they cant track what YOU voted for (incase some loon gets in power and executes all who voted against them, exterme but valid comment)

[Rosssco]

 

Aye will do..   Interesting analysis by YouGov on how age groups voted..    

Hardly insightful. That's simple mathematics. The more granular the sample range, the more detail will be available. The more generalised, the more you'll see occurrence of your overall result. You can drastically alter the "key message" dependent on how you represent the data. Oldest trick in the book on how to manipulate the masses. What would be interesting to see would be to segment based on a range of definitions. For example, by volume (i.e. Segments defined at roughly equal populations.) It only becomes more meaningful when the percentages are accompanied by volumes or % of base, to illustrate how significant the sample size is in context.

 

 

Yep, and you prove my point - hence I said "interesting" rather than "insightful"..  Both sides can tweak statistical data to suit their desired outcome...

[Dunk]

 

Aye will do..   Interesting analysis by YouGov on how age groups voted..    

Hardly insightful. That's simple mathematics. The more granular the sample range, the more detail will be available. The more generalised, the more you'll see occurrence of your overall result. You can drastically alter the "key message" dependent on how you represent the data. Oldest trick in the book on how to manipulate the masses. What would be interesting to see would be to segment based on a range of definitions. For example, by volume (i.e. Segments defined at roughly equal populations.) It only becomes more meaningful when the percentages are accompanied by volumes or % of base, to illustrate how significant the sample size is in context.  

  Yep, and you prove my point - hence I said "interesting" rather than "insightful"..  Both sides can tweak statistical data to suit their desired outcome...

To be clear, I was not suggesting that you were wrong but wanted to highlight to anyone who DOES draw conclusions from it that it's not worthwhile as it stands (hence the rest of the post, which I'm glad helped prove your point. ) HTH I dislike and distrust any view that generalises like "55% saw sense," "55% love to be identified as British," "55% were scared into voting No" or "55% are short-sighted." None of those conclusions could be drawn and none would be helpful. I'm probably unusual (I'm definitely unusual) in that I don't actually have an agenda on here. I do find it very interesting though, particularly trying to understanding what drove the result. That's not to say that I don't have strongly held views or didn't voted decisively myself. I just wouldn't share them and debate them on here.

Edited by Dunk, 24 September 2014 - 09:39 AM.

[Mangham54]

To fully understand what drove the result requires a vastly different data project. The one where you take a mahoosive sample of the public, get them talking and then dig into their reasons for voting in the way they did. Extrapolating 'yes' v 'no' data against demographics is one thing, but (as you quite rightly said Dunk) concluding "reasons" for this is just bullshit. I did once see a copper outside a polling station, but that was in the east end of Newcastle and that was to keep the chavas away.

[LY_Scott]

http://m.scotsman.co...-bomb-1-3550780

[slindborg]

The whole world is in an uneasy state still, any hint of instability from anywhere fucks things up

[LY_Scott]

Yus, its more the commentary change on why the banks came out that's interesting.