Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[smjjames]

What's the source of that statistic anyway?

This report by Chatham House, a British think-tank.

I'm not sure what you guys are complaining about. 10.000 is a pretty big sample size. As for the 10 countries, well, the top 10 EU countries have 80% of the population between them. This probably gives you a good idea of the average EU citizen. Sure, there are certainly variations between countries, but the graph doesn't make any claim about proportions within specific countries. It's as if you were complaining that a US-wide presidential poll is bad, because it's not representative of Alabama's GOP leaning.

10,000 divided by 10 is 1,000, which isn't exactly a big sample size for each country. If you wanted to get a general idea of the political makeup of the region, that's probably fine, but if you wanted to really drill deeper, seems like a bigger sample size would be preferred.

[Sheb]

What's the source of that statistic anyway?

This report by Chatham House, a British think-tank.

I'm not sure what you guys are complaining about. 10.000 is a pretty big sample size. As for the 10 countries, well, the top 10 EU countries have 80% of the population between them. This probably gives you a good idea of the average EU citizen. Sure, there are certainly variations between countries, but the graph doesn't make any claim about proportions within specific countries. It's as if you were complaining that a US-wide presidential poll is bad, because it's not representative of Alabama's GOP leaning.

10,000 divided by 10 is 1,000, which isn't exactly a big sample size for each country. If you wanted to get a general idea of the political makeup of the region, that's probably fine, but if you wanted to really drill deeper, seems like a bigger sample size would be preferred.

Yeah, which is why they aren't attempting to dig deeper (for exemple, publishing by-country results). 1000 is also not that small either. Looking at 538's poll trackers for Trump approval rating for exemple, they are mostly in the 1000-2000 range with most being 1500. So 1000 is on the small side, but still relatively ok.

[ChairmanPoo]

What's the source of that statistic anyway?

This report by Chatham House, a British think-tank.

I'm not sure what you guys are complaining about. 10.000 is a pretty big sample size. As for the 10 countries, well, the top 10 EU countries have 80% of the population between them. This probably gives you a good idea of the average EU citizen. Sure, there are certainly variations between countries, but the graph doesn't make any claim about proportions within specific countries. It's as if you were complaining that a US-wide presidential poll is bad, because it's not representative of Alabama's GOP leaning.

10,000 divided by 10 is 1,000, which isn't exactly a big sample size for each country. If you wanted to get a general idea of the political makeup of the region, that's probably fine, but if you wanted to really drill deeper, seems like a bigger sample size would be preferred.

I disagree with this kind of blanket statement because it disregards actual statistics. You can actually calculate how representative is a sample size relative to a larger population. 

Now, there are other ways in which a poll can be nonrepresrentative, beyond sample size, but if you feel that 10.000 is too little a sample I think you need to justify why.  Personally, at a glance I think 10.000 is fairly good  (we do clinical trials with far less people, and generalize the result to huge swathes of humanity, after all) , and a quick check online suggest that it would provide a quite narrow confidence interval, too

Like I said before, though, this doesnt mean that the poll is or isn't flawed in other ways, but the complaints about sample size seem misaimed.

[da_nang]

What's the source of that statistic anyway?

This report by Chatham House, a British think-tank.

I'm not sure what you guys are complaining about. 10.000 is a pretty big sample size. As for the 10 countries, well, the top 10 EU countries have 80% of the population between them. This probably gives you a good idea of the average EU citizen. Sure, there are certainly variations between countries, but the graph doesn't make any claim about proportions within specific countries. It's as if you were complaining that a US-wide presidential poll is bad, because it's not representative of Alabama's GOP leaning.

10,000 divided by 10 is 1,000, which isn't exactly a big sample size for each country. If you wanted to get a general idea of the political makeup of the region, that's probably fine, but if you wanted to really drill deeper, seems like a bigger sample size would be preferred.

Yeah, which is why they aren't attempting to dig deeper (for exemple, publishing by-country results). 1000 is also not that small either. Looking at 538's poll trackers for Trump approval rating for exemple, they are mostly in the 1000-2000 range with most being 1500. So 1000 is on the small side, but still relatively ok.

Having a poll in every country would be good as any major decision towards federalism is almost certainly going to need qualified-majority or unanimous agreement in the Council of Ministers and maybe national referenda. The poll might reflect better on the parliament but not the council.

[Reelya]

10,000 divided by 10 is 1,000, which isn't exactly a big sample size for each country. If you wanted to get a general idea of the political makeup of the region, that's probably fine, but if you wanted to really drill deeper, seems like a bigger sample size would be preferred.

However, the point is that this argument has no actual mathematical basis. From a maths point of view, there's no provable benefit in having a larger sample size, as long as your selection is truly random. In statistics "total population size" isn't even a variable in your equation to work out the needed sample size. The only reason for big sample sizes in political studies is just that - political - because big numbers are more convincing to people who never actually studied statistics 101.

e.g. imagine a bag with X number of black marbles and Y number of white marbles. You pull out marbles one by one, note their color and estimate what percentage of the bag is black marbles or white marbles. The total number of marbles in the bag is entirely irrelevant to how many marbles must be drawn to determine the proportion of black vs white to a specified level of accuracy.  e.g. if I double the number of marbles in the bag, the accuracy of the estimate based on 100 marbles is still exactly the same as it was before. The bag could hold a billion or a googleplex worth of marbles, and it wouldn't in fact change your accuracy after drawing 100 marbles even one bit. And that's because total population size isn't a variable that affects proportions, which is all that the statistic is measuring.

The only correct argument here is - is the sample selection truly random? If your sample selection mechanism isn't random then it won't actually make any difference if you sample 1000, 10000, 100000 or 1000000. It will still have the exact same bias. e.g. if everyone I sample comes from Paris and I'm claiming that this is representative of all Europeans, increasing the sample size won't fucking improve the metric even one bit. Ok, I'll ask twice as many Parisians what they think. Now is it more "representative" of Europeans? Or, you could ask 250 million people, half of Europe, but if you had a biased selection (e.g. focused on big city people) then it's going to be worse than a carefully selected sample of 1000.

[misko27]

Woah that's some james bond shit right there

Little late, but there's some surprising precedence: Hermann Göring killed himself the night before he was to be hanged. Same substance, actually...

[hector13]

That was passed to him by a guard, if I recall correctly.

[MetalSlimeHunt]

Yeah, the purpose of the action is to deny justice. Killing yourself before the end of the trial is a way of saying "my victims deserved it, I am not subject to your judgement".

[Loud Whispers]

Yeah, the purpose of the action is to deny justice. Killing yourself before the end of the trial is a way of saying "my victims deserved it, I am not subject to your judgement".

I think it's different to that. Researching into the dude is interesting. While I haven't had the time to come to a conclusion whether he was guilty or innocent yet (from my initial glimpse it's a fucking horrifying clusterfuck of groups allying and breaking alliances in a heartbeat), what is clear is that he believed of his innocence. 2004 he surrenders willingly to the Hague instead of remaining in Croatia, for the reason of clearing his name of the charges leveled against him (the war crimes). In that trial he's found guilty and sentenced to 20 years prison. This trial was not the trial to determine his guilt or innocence, it was his appeal to clear his name. Although some of the charges against him were dropped in the appeal, most stayed, and his sentence remained at 20 years. However, given that he had already served the majority of his sentence, he could've just served the rest and walked off a free man in a short time. But that's not the point, if he walked free he would walk with the stain of being a war criminal, there would be nowhere he could call home and no point in living. If however, he died and died on camera (it's worth noting he was before the war an artist and professor, having written and directed many plays), he would be able to end not on a pathetic note, dying decrepit & dishonoured on the streets, but die on his own choosing in open and public defiance of the court:
"Judges, Slobodan Praljak is not a war criminal, with disdain I reject your verdict! This is poison that I drank."
With knowledge of his drama background, it seems he scripted himself into a Greek tragedy. Result? Legacy

Croatian Prime Minister Andrej Plenkovic expressed condolences to Praljak’s family and denounced the rulings as a "deep moral injustice" against the convicted. He also rejected the finding that the Adriatic state was complicit in war crimes, instead saying it was defending its ethnic-kin and others from neighboring Serbia.
“The verdict incorrectly assesses the role of Croatian leadership in the war,” Plenkovic said in the capital, Zagreb. “Croatia gave refuge to hundreds of thousands of Bosnian refugees, also including Bosnian Muslims.”
He said it was “absurd” that none of the tribunal’s verdicts had found Serbia’s wartime leaders, and particularly late President Slobodan Milosevic, responsible for war crimes in Bosnia and Herzegovina. Ethnic-Serb authorities from Bosnia and Herzegovina, including military leaders Ratko Mladic and former President Radovan Karadzic, were convicted.
Croatian lawmakers also reacted with apparent shock.
“The verdict is unjust, it is not based on historical facts, and we are rejecting it,” assembly Speaker Gordan Jandrokovic said.
https://www.bloomberg.com/news/articles/2017-11-29/bosnian-croat-kills-himself-with-poison-at-un-court-hina-says

[Jopax]

Yeah, calling it an utter and total clusterfuck is an understatement. Every side has blood on their hands and every side has victims missing to this day, buried in some unkown location with the guilty fucks either dead or still at large.

The general mood is dissappointment and resentment from the croatian side while the serbs and some of the muslims are kinda celebrating. Then again, it was the same deal for pretty much every conviction that has happened to date, one side is dismayed because suprise suprise they had shit people among their ranks while the other two are celebrating it, conviniently ignoring the shitfucks that fought and did horrible things on their behalf.

[Loud Whispers]

So in Brexit news I've had fun scoring up how many times the Tories made major own-goals but now I have depression.
1 Nov - Defence Secretary Michael Fallon resigns after multiple sexual scandals and sexual assault allegations emerge, replaced by Gavin Williamson, dubbed the Baby-Faced Assassin by peers.
8 Nov - Secretary of State for International Development Priti Patel resigns after it is revealed she held covert & unauthorized meetings with senior Israeli officials in the Golan height, including the Prime Minister Bibi Netanyahu, in order to discuss using British foreign aid development funds to pay for the medical treatment of Syrian jihadi militiants. What the fuck
30 Nov - Someone leaks (hint: It was the EU) that UK Brexit bill could cost £50B, when most voters stomach at most £10B. It would mean in effect that all the money the Tories promised would go from paying the EU and to paying the NHS, would go from paying the EU to paying the EU. Generates major salt.
1 Dec - Brexit Secretary David Davis in threat to quit over Damian Green porn storm, wherein MP Damian Green's parliamentary computer was found with thousands of porn images downloaded and so called upon to resign. Talk about a hung parliament
4 Dec - Tory backbenchers threaten rebellion if Gavin Williamson does not get Phillip Hammond to reverse planned cuts to Defence.
Literally now - Theresa May rejects deal reached between the UK-EU after the DUP made it clear that Northern Ireland absolutely could not remain under EU laws while the rest of the UK separated themselves.
Really just a political revolt within the party every odd week or so

[MetalSlimeHunt]

Comrade Corbyn's revolutionary fervor has got to be gigantic watching this.

[smjjames]

Northern Ireland (Sinn Fein mostly) was also throwing a fit about hard borders, or maybe that was Ireland. I'm sure some Brexiteers are wondering why they're doing it and jumped into it without any real plan or idea of what would happen.

Looking more and more like Britain would regret doing Brexit, especially without any plan.

[Loud Whispers]

Northern Ireland (Sinn Fein mostly) was also throwing a fit about hard borders, or maybe that was Ireland. I'm sure some Brexiteers are wondering why they're doing it and jumped into it without any real plan or idea of what would happen.
Looking more and more like Britain would regret doing Brexit, especially without any plan.

Not at all, it's undesirable but you must understand that Brexiteers jumped into this with a goal in mind, with that goal being leaving the European Union. The Guardian's put it aptly:

Voters are holding firm to old views – but are more pessimistic about the outcome of talks with the EU

Those who wanted to leave the European Union still want to leave the European Union, those who wanted to remain in the European Union still want to Remain in the European Union. The only thing that has changed is that everyone is depressed that this is going to be the maximum saltiest timeline for the next 5, 10 and 20 years.

[hector13]

It’s rather amazing how long it seems like there’s been no plan.

PPE: I mean it’s all well and good knowing the destination, but if you have neither the vehicle or knowledge of how to get there, you’re going to be all kinds of knackered by the time you arrive. Probably look a right tit, too.