‘The population of England and Wales grew over the past decade at its fastest rate in 100 years as the result of a surge in immigration, increased longevity and a mini baby boom.’
Financial Times, July 16
This was discovered by the census, right?
What about the population of Scotland?
The Scottish census won’t publish any figures until the 700th anniversary of the Battle of Bannockburn. Or something like that – I am hazy on the details. The Northern Irish census has been published and the population there is 1.81m. The population of Northern Ireland is less than half of the increase in population in England and Wales between 2001 and 2011. The English and Welsh population increased by 3.7m to 56.1m.
Everyone keeps talking about that – is it really such a big deal?
It was quite a surge. The English and Welsh population grew at a real clip during the last decades of Victoria’s reign, but 7.1 per cent is the fastest rate of population growth since the 1911 census. It was also a surprise: official estimates gave the population as being almost 480,000 lower.
How do you get that kind of thing wrong?
Mark Easton, the BBC’s Home Affairs editor, put it rather well, saying we are good at counting babies and corpses so the only other source of error is net migration. We must have counted too few immigrants or assumed people had left the country when they hadn’t.
How is it possible to miscount such a thing?
Perhaps I shouldn’t have said “counted”. Many statistical exercises in this country don’t count, they sample. For instance, the International Passenger Survey buttonholes people on cross-channel ferries or just after they’ve left passport control and asks questions about who they are and what they are doing. That’s just a sample of the people arriving in the country, however, and the IPS surveyors can’t force people to answer their questions, nor can they force them to tell the truth.
And the census is better because it samples a higher percentage of the population?
Actually, what determines the precision of a sample is not that it covers a high percentage of the population, but that it has a lot of people in it. A sample of a million people is plenty for most purposes, whether that’s a million people out of a population of 10m, or a million people out of a population of a billion.
So why do we have a census at all, rather than a survey of a really big sample of people?
The trouble with surveys is that the sample of people they include may not be representative of the population as a whole. There’s a long-running problem, for instance, with contacting people at home by telephone: once upon a time such polling was biased towards the opinions of the prosperous; later, it became biased towards stay-at-home types. These days, it is biased against the young, who often do not use land lines at all. Sensible surveys will try to correct such biases but one can never be entirely sure.
But not everyone fills in the census either.
Indeed. The Office for National Statistics followed up the census itself with a survey to estimate how many people actually completed it in different parts of the country, so that they could make appropriate connections.
Do they also have a follow-up to the follow-up survey?
How much did this cost?
That’s a lot.
Perhaps. It’s less than one-10th of 1 per cent of annual government spending and, of course, we only do it once a decade. I suppose the question is, is it worth spending £1 in every £10,000 to help work out how to spend the other £9,999?
If you think it will help, that is. Speaking of which, have we found out how many Jedi Knights there are in
England and Wales?
We don’t yet know. Almost 400,000 people stated that their religion was “Jedi” in the 2001 census, making belief in The Force the fourth most popular religion. I have no idea whether Jedism is an evangelical religion or not. If it is, that number might well have risen.
It all seems rather confusing to me.
That’s just an old Jedi mind trick.
Also published at ft.com.