Fionn1952 wrote: » You've never done any sort of real world statistical analyses, have you? Sample size is much more significant than you imply, and a sample of 10, even with a perfectly representative sample would have an error bar in great excess of +/-10%. I get that you're low grade trolling and trying to wind people up, but at least look up basic statistics before making ridiculous claims like this.
downcow wrote: » I watched RTÉ news again last evening as I try to get my head around southerners thinking. Again I was surprised how alien the whole experience was. We have grown very far apart. So many of you guys are in denial about how your state is slow to throw off the shackles of the Catholic Church. It is surprising to me given the abuse saga and that so many church teachings seem so out of touch with your people. Again I was met with the de angles at the beginning (and don’t tell me again that it’s not part of the news) I have taped the news a few times now and it always records the de angles). Then we had a significant news item updating us on the popes current visit. No mention of any of this on our news. And I am not knocking it - it just shows where interest lies. Up here we would get similar if the queen was visiting Iraq. Seems like old red socks is still your head of state lol
BonnieSituation wrote: » And what a stout it is. "Nailed it" isn't the word.
BonnieSituation wrote: » The numbers are fine lads and I'm saying that from the POV of someone who is knee deep in Data Analytics. What isn't fine is how I am struggling to find the proportion of don't knows anywhere. My googling is failing me tonight. Classic British Stout. Anyone come across them?
jh79 wrote: » Nothing wrong with the sample size in this case. Here's a good tool for calculating margin of error / confidence intervals.http://www.raosoft.com/samplesize.html Margin of error of 3.8% in this case. At 42% looks like there is a long way to go yet.
Fionn1952 wrote: » I'm not sure why you're directing this at me. I corrected Natterjack's comment that a sample size of 10 would give a margin of error of 10%, I didn't make any claims on the margin of error in the sample of that particular survey. Just to clarify, you do indeed agree with everything in the post you just quoted, yes? Using your own tool, a sample size of 10 would give a margin of error in excess of 30%.
jh79 wrote: » With regards Natterjacks comment, it would depend on the likely distribution of the responses. Either way we can all agree that in this poll the sample size is appropriate.
Fionn1952 wrote: » Given a likely distribution of anywhere from 50-75% still gives a margin of error in excess of 25%, so no it doesn't depend on the likely distribution; in practical terms for the topic at hand, a sample size of 10 would be absolutely useless. Not sure why you're working so hard to defend such a nonsensical claim from Natterjack and argue a point with me, when you clearly understand and agree with the criticism I made. At no point did I disagree with the statistical significance of the poll in question; for NI adults above 18, over 500 is indeed a reasonable sample size.....ten is not. I would highlight that the margin of error of ~4% is purely the margin introduced by sample size, and indeed there is likely to be further margin of error introduced by things like sampling methods for example. I haven't looked into it in this case (as my argument was not about the poll in question), but some of these polls depend on getting responses via landline telephone; this introduces a particularly large margin of error due to unintentionally skewing massively to the older side of the electorate who still own a landline. I repeat, this is purely an example and I haven't looked at the sampling methodology in the case of this particular survey. I would argue however that given the number of relevant cohorts in NI, it would be impossible to remove a significant sampling error from a respondent number as low as 10, further increasing the margin of error in Natterjack's already useless survey of 10 people.
jh79 wrote: » I thought Natterjack was picking an extreme example just to illustrate his point but yes 10 is not a good sample size . It was done online and data was weighted by age, gender, region, religious community and 2016 vote.
Fionn1952 wrote: » Natterjack committed to specific numbers, saying 10 people would give a margin of error of 10% and 50 people a margin of error of 2%, given that we've just concluded together that a sample size of over 500 has a margin of error of over 4%, we can dismiss his post as nonsense that demonstrates a poor understanding of statistical sampling rather than defending it by attributing some deeper meaning to it.
Fionn1952 wrote: » Online polling does tend to introduce a margin of error (though in this case I would imagine it would actually weight in favour of Unification, just so I can't be accused of picking holes due to an agenda), so it certainly would have a margin of error in excess of 4%, I wouldn't imagine we would be talking significant enough to bring 50% within the error bars. I find that unsurprising, as I'm certainly of the opinion that a vote for unification would fail if there was a border poll tomorrow.
jh79 wrote: » In the context of the post he/she was responding to I think you're being a bit harsh. The specific numbers were not in reference to any poll , they were just to illustrate a point on data quality / representation. At least that was my impression.
jh79 wrote: » I thought Natterjack was picking an extreme example just to illustrate his point but yes 10 is not a good sample size.
Fionn1952 wrote: » Can you give an example of any poll where a sample size of 10 would give a margin of error of 10% and 50 people would give a margin of error of 2% given a population anywhere within a degree of magnitude of the topic at hand? The specific numbers were nonsense in any way that could be taken as relevant to the subject. Absolute nonsense and I suspect you wouldn't be defending it so vigorously if someone was applying such faulty reasoning and poor mathematical comprehension to argue in favour of Unification.
In reality, of course, the other parties, along with business people and others, are being pragmatic in the interests of stability at a time of pandemic and economic recovery. They got stuck with the protocol because the DUP ignored the wishes of the 56 per cent of people who voted to remain in the EU, and kept bidding for a harder and harder Brexit until their votes no longer mattered and the prime minister they liked because he was like Donald Trump tossed them aside. Unionism is in a panic because it no longer dominates. There are many constructive voices within the Protestant community, including those of the women who actually run the community sector. The parties shut them out. As for the paramilitaries “withdrawing” their support for the Belfast Agreement, when did you last hear any of these men championing equality, respect and human rights? Or even shouting about God and Ulster? All the security assessments find them to be armed criminal gangs, though they do a lucrative sideline in hoovering up funds that are meant to persuade them to disband. Last week, some of them were tweeting admiration for the tactics of the QAnon-types throwing fireworks at the police in Dublin. That’s their level. They do not represent working-class unionist communities. The DUP does. But it offers its people nothing but a gaudy flag to wave and curse.
jh79 wrote: » The numbers given where just to illustrate the point. And I'm not defending anything "vigorously"! I think someone doesn't what to discuss the poll and see this pointless argument as a handy distraction.
downcow wrote: » .......A battle which nationalists are clearly losing as more and more of those traditionally Irish Catholic are not declaring for UI yet those traditionally Ulster Protestants are still basically 100% behind the union
jh79 wrote: » 7% of Protestants said they would vote for a UI while 24% of Catholics said they would vote to remain in the UK! Worrying though that over half have reservations about sectarian violence if a UI did happen.
briany wrote: » Renewed violence would be a big concern because groups like the UVF and UDA would kick off big-style
Junkyard Tom wrote: » For what exactly?
lurleen lumpkin wrote: » De angles. Haha. Are you really that stupid, or just pretending to be? Lol. 1/10 for effort. Up there with de angles.
briany wrote: » Renewed violence would be a big concern because groups like the UVF and UDA would kick off big-style.
maccored wrote: » i very much doubt they'd get british army support this time round (though you'd never know) so they'd probably kill more of themselves than anyone else. besides, the UDA and the UVF would be too out of it, the druggies.
jh79 wrote: » It was the biggest concern of the cohort willing to vote yes. Strange question, as the mirror image of the IRA, the intention would be to prevent a change in sovereignty.
downcow wrote: » https://www.bbc.co.uk/news/uk-northern-ireland-56294309 Oh dear! More sacrosanct EU rules broken lol. Where now. The question is, how petty will the Eu be? An interesting test as to whether they care about ni or not - and I don’t know the answer.
downcow wrote: » Where now.
![Photo of [Deleted User]](https://w4.vanillicon.com/v2/4b1a2e86cedae1628937f4503f2281da.svg)
![Photo of [Deleted User]](https://w4.vanillicon.com/v2/4c998303aa0fcfcedb419267ab426bd5.svg)
