Census

[lethargic_man]

Today is the first time I have filled in a census. (In previous years, I wasn't the householder.) I have to confess, it wasn't until I read a blog post by somebody (I forget whom now) scratching their heads about the people complaining about having to give information to the government, on the grounds that if we don't let the government know our situations, how else are they to provide us with the services we need, that I remembered the census wasn't primarily in order to let people like me do genealogical research a century down the line...

Comments

[curious-reader.livejournal.com]

Do you mean filling out forms for statistics of population? I found it very od that they asked for temporary guests on a certain day. What on earth do they want with that? That actually screws up the population count. Why counting my brother and mother to the population of the UK if they live in Germany?

[lethargic-man.livejournal.com]

I think it's because they don't want a snapshot of the population of the UK on an average day; they want a snapshot of a representative day. On any representative day, there will be people with relatives visiting from abroad, and if you ask people for their normal day such people will be lost; and yet they need to be taken into account (as, statistically, some of them will fall ill and need to use the NHS, for example).

[curious-reader.livejournal.com]

Well, I did not plan anyone. So I did not include anybody except me, of course. I may also have visitors from St Albans or somewhere else in the UK. They already have their own home. It is non-sense to include those. The NHS has them already included.

[bluepork.livejournal.com]

IMHO the whole thing is a nonsense. Many of the questions are badly worded, relying on interpretation from the householder. I read a letter in a newspaper pointing out that the ONS takes a whole year to release the results, and I have picked up on the fact that the results are compiled by hand. What a waste of resources.

I'm not against the principle of the thing, but the execution is so poor, you have to wonder about the quality of any results.

That letter I read pointed out that marketing companies can get the same sort of data together in a matter of days and even the general election only takes one day to "compile"!

[lethargic-man.livejournal.com]

I read a letter in a newspaper pointing out that the ONS takes a whole year to release the results, and I have picked up on the fact that the results are compiled by hand. What a waste of resources.

Possibly the reason they take so long is a lack of resources: what a hundred people can achieve in a month takes eight people a year.

What do you mean by the results being compiled by hand? The form I got with the census asked for people to fill it in online if they could, as it sped the process up; and from autopope's post (http://autopope.livejournal.com/662646.html) it appears to me that they rely on copying the results manually only when computer recognition fails.

I'm not against the principle of the thing, but the execution is so poor, you have to wonder about the quality of any results.

Well, I've got good (genealogical) results from the 1871–1911 censuses. :o)

That letter I read pointed out that marketing companies can get the same sort of data together in a matter of days and even the general election only takes one day to "compile"!

To be fair, marketing companies tend to poll at most a couple of thousand people, not in excess of sixty million; and the general election (a) only involves a single question answered by checkbox, and (b) is compiled by an army of thousands working overnight.

[bluepork.livejournal.com]

Where did you get eight people from? The ONS is a massive government department. I relied on some of their historical data when I wrote my rather poor dissertation at uni.

The fact that the census form arrived in the post prompted even computer literate people (like me) to fill it in by hand. I'd be interested to see what proportion of people filled it in online. I read in another article, or perhaps in the same letter, that the website was awful anyway. For example, it didn't allow you to go back in order to correct information already entered, before the final submission. The author of the article I read had two children at university and had guests staying, so deciding who is person 1,2,3,4 etc was confusing at best, and she thought she had completely fluffed it in the end.

Finally, in response to your "To be fair..." comment, there's really no need to be fair, so let's pull that final comment to pieces. Marketing companies poll a smaller number of people, based on what science says is a significant sample for the issue they are researching. They also don't have the stick of a £1000 fine to work with. Now let's look at the numbers. You wrote that they poll a couple of thousand people but use an army of thousands ie more census takers than "censees". That would clearly be daft... the marketeers would be better to poll the pollsters rather than poll the public. So let's say you were exaggerating the numbers to make your point. Having done some cold calling in my time, I can tell you that it's no problem to speak (as in have a proper conversation) to 20 people in four hours. This would mean that to review your 2000 people, you would need only 100 pollsters, working four hours per day. Given that the pollsters don't need to have an in-depth conversation with the populous, that's a pessimistic rate, but let's stick with it. Let's also say that real pollsters can work for 8 hours per day. That would mean you would only need 50 pollsters to poll your 2000 people.

The census is reviewing 22 million households, not 60 million people, as you wrote. Let's say you wanted to speak to 1 million householders which, statistically, is still completely over-egging it. On these ratios, you would need 25,000 people to speak to them all. Or, you could employ 1000 of them to perform the work every night for the two week period of the census collection.

I suppose we could take your 8 people and assign 2 of them to a full time job to find the 1000 temporary employees required for the task. They would have 10 years to do it. 2 of them could be employed to compile the data. I suppose one of them could be a web developer, whose job it would be to present the results. One could be an independent overseer. One person, could be employed to clean the office, make the tea, etc and that leaves one more, who I'm really struggling to employ. Maybe the eighth person could be a pretty secretary in a short skirt whose sole purpose would be to fill some stereotypes. Ha!

There. Job done!

[lethargic-man.livejournal.com]

Where did you get eight people from? The ONS is a massive government department. I relied on some of their historical data when I wrote my rather poor dissertation at uni.

I was making a point using relative numbers; I wasn't suggesting there really were only eight people.

The fact that the census form arrived in the post prompted even computer literate people (like me) to fill it in by hand.

Mine arrived with a letter saying to fill it in online if I could.

Finally, in response to your "To be fair..." comment, there's really no need to be fair, so let's pull that final comment to pieces. Marketing companies poll a smaller number of people, based on what science says is a significant sample for the issue they are researching. They also don't have the stick of a £1000 fine to work with. Now let's look at the numbers. You wrote that they poll a couple of thousand people but use an army of thousands ie more census takers than "censees". That would clearly be daft...

It would indeed be daft, which is why I didn't write it. What I wrote is that market researchers poll a small number of people, and that the election data (not market research data) is compiled overnight by an army of thousands.