belboid
Exasperated, not angry.
Well, it's the law actually, but hey ho
Census 2021
- Thread starter brogdale
- Start date
kenny g
Sheeple Awake!
I suppose you better report us to the relevant authorities then! Showing your true colours at last...
kenny g
Sheeple Awake!
Oh no! The census officers! Better fill it in in case they knock on my door. Received absolutely no comeback last time around and expect to receive none this year.
As I say, null returns will be provided because it is a pile of shite. My time is better served doing anything else. As HMG announce every 10 years they are considering not doing it, but like some bureaucratic zombie it stumbles on decade after decade, I will go with the zeitgeist and not participate. I am sure this offends various " census officers" which is a fancy term for local government desk jockeys earning a few extra squid but so be it. The whole concept is like something from the Bible. It is completely outdated and ignores the discovery of statistical sampling hundred of years ago.
kenny g
Sheeple Awake!
Is the kind of twaddle we are up against.
I am not an expert in statistics like you must be, but I thought statistical sampling was still quite error-prone, and able to overlook large sections of society? Which, I thought, was why political polls don't all agree with each other, and don't always predict election results correctly.
brogdale
Coming to terms with late onset Anarchism
Maybe, but the head of the ONS has said that "regular, large scale population surveys", in conjunction with 'administrative data', will be considered in 2023 when they make the decision whether or not to continue with the census in its present form.
I know, that has been repeated many times in this thread. The only alternative to him saying that would have been for him to say "we are not even considering any alternative ways of doing it, ever", which would rightly be seen as pretty irresponsible and ridiculous thing for a head of statistics to say.
brogdale
Coming to terms with late onset Anarchism
Not sure what your point is, then...unless you think that the census is some 'sacred cow' beyond change?
brogdale
Coming to terms with late onset Anarchism
All other ONS data collection is undertaken by (sample) survey; in fact I've currently got an ONS letter on my table inviting me to take part in their latest employment data survey...inc. the £15 shopping voucher incentive.
belboid
Exasperated, not angry.
So? One big one, lots of little ones. The little ones use the big one as a basis too for correlation and testing of assumptions.
kenny g
Sheeple Awake!
Ignoring the flaccid attempt at sarcasm, the difference between the inaccuracies of an opinion poll which is designed to reflect voting intentions and statistical sampling for a population survey should be obvious. Statistical sampling is used all the time for numerous purposes - checking the weights of bags of spuds - the consistency of medicines - the effectiveness of vaccines.
kenny g
Sheeple Awake!
The problem with the "one big one" is that it has flaws where significant groups of the population aren't covered. You have already highlighted traveller/ roma communities but there are plenty of other issues with regards to people living in HMO's etc etc. Executing sampling well can be more accurate then taking an approach which is , at best, from the 19thC.
What are you basing your complete confidence that statistical sampling is more accurate than asking 95% of the population? Have you studied it (informally or formally, I'm not saying you need a qualification, but you do need to have spent time on it)?
Everything I learnt about statistics says that the greater the sample size, the closer the result approaches accuracy; if you can point me to anything that contradicts that, do it now.
I think it is pretty obvious to most people that the census is closer in nature to an election (ie they are both asking questions of the human population of the UK) than it is to "checking the weights of bags of spuds".
If you are not basing your complete confidence on anything in particular, then why would anyone listen to your opinion on it? It would be like paying attention to a random email telling you you won a lottery.
19th century is pretty recent as far as maths goes. We still use pythagoras theorem from like 500BC; because it is still right. So age of an approach isn't necessarily a reason to cancel it.
Last edited:
Pickman's model
Starry Wisdom
Pythagoras also famous for his ban on beans
kenny g
Sheeple Awake!
Would never profess complete confidence in anything but do have a high degree of certainty :/
If you do a 95% sample but miss out 5% of the population all of whom happen to be from a particular sector then that is less accurate than a 20% sample which includes a representation of the missed out sector in the sample. For example, if you design a census that includes the 95% of the population who can read and write you miss out all of those who can't. If you take a complete return of 20% of the population including those who can't read and write and then extrapolate, the result will be more representative as it includes the illiterate.
Statistical sampling recognises the real world fact that a 100 % sample is not achievable and that the percentage who are not sampled in the attempt to reach 100% are often the ones that particularly need to be included in returns.
kenny g
Sheeple Awake!
Thanks for stating the obvious. According to the wiki page on statistical sampling a form of it is referenced in the Bible so I suppose it has pedigree. My point was that things have moved on from the time of Herod and even the 1801 census. We are not talking about predictions of voting intentions but measuring actual facts on the ground. Doing a rigorous sample would be more accurate than the current system of forms chased up by half hearted door knockings.
hipipol
Peckham Wry
Equally his Dualist obsession about no shagging, renouncing worldly wealth has been practiced ever since.......besides he knew fuck all about stats and was obv completely nuts about anything other than triangles - Euclid would have been a better example
I will answer everything truthfully and trust em, everything or my names not Guiermo Anusol the 3rd
What sources are you basing your "high degree of confidence" in, then? Or is it just that you reckon what you say feels about right?
Are any of the figures you quote (eg "all of the 5% who were missed by the census were from 1 demographic", "5% of the population cant read or write") based on real studies, or did you make then up from thin air to try to justify your argument?
Would you be happy with a 20% of population statistical survey then? At what percentage of the population would you start being unhappy with it?
I guess you are already aware of the community engagement programmes to increase response in communities that tend to have lower response rates eg travelers as already mentioned. Not sure what a statistical survey would do to improve on that.
Last edited:
kenny g
Sheeple Awake!
This is getting silly. It was an example. You are now misquoting me.
I am not coming out with some esoteric pseudo-science. As stated previously, statistical sampling is used widely and effectively in the real world.
Your objection to a 95% sample was that it was probably overlooking entire demographics wasn't it? I am suggesting that is very unlikely, particularly given community engagement programmes, and would in fact be far more likely to happen with smaller sample sizes.
In fact, to do a good statistical sample you need the census data. Otherwise where would you look to find out what percentage demographics you need to replicate to get a good snapshot of the population?
In fact, to do a good statistical sample you need the census data. Otherwise where would you look to find out what percentage demographics you need to replicate to get a good snapshot of the population?
kenny g
Sheeple Awake!
rutabowa This thread pretty much covers the same topic if you want to read up on it.
stats.stackexchange.com
Why is it claimed that a sample is often more accurate than a census?
When learning the course of sampling, I meet the following two statements: 1) Sampling error leads to mostly variability, nonsampling errors lead to bias. 2) Because of nonsampling error, a sampl...
Right I'm going in on it now.rutabowa This thread pretty much covers the same topic if you want to read up on it.
Why is it claimed that a sample is often more accurate than a census?
When learning the course of sampling, I meet the following two statements: 1) Sampling error leads to mostly variability, nonsampling errors lead to bias. 2) Because of nonsampling error, a sampl...
kenny g
Sheeple Awake!
I do not trust the census in Tower Hamlets or Newham, for example, to be in anyway as accurate as leafy parts of Hertfordshire. I just do not believe that sufficient resources are going to be put into getting accurate figures for these areas.
Similar threads
- Poll
