Census: How religious is the UK?
A publicity drive has started for the census, now just five weeks away, but the survey is being criticised for its question on religion. So is it even possible to accurately measure how religious the UK is?
According to the Gospel of Luke, it was a Roman census that sent Mary and Joseph to Bethlehem, where she gave birth to Jesus Christ.
And more than 2,000 years later, the same kind of counting exercise is being used to gauge the religious make-up of the UK.
According to the last Census 10 years ago, more than two-thirds of people in Britain regarded themselves as Christian - 72% in England and Wales, and 65% in Scotland.
More than 1.5 million in England and Wales, more than 3% of the population, said they were Muslim and nearly eight million ticked "no religion". There were also 390,000 self-proclaimed Jedi.
Is it a leading question?
"It fairly allows you to answer it because you can say 'no religion' but if you wanted to make it as neutral as possible, you might ask 'Which of these would you describe yourself as?' says Stephan Shakespeare of YouGov.
"It does have a slight assumption, although not a strong one, but these wordings do make a difference."
But five weeks before the next census day, Sunday 27 March, some groups are questioning whether the religious numbers are at all accurate, and could ever be.
They prefer to use the British Social Attitudes survey, carried out annually by the National Centre for Social Research, which paints a picture of a less-religious country, with 51% describing themselves as non-religious and 43% as Christian.
The religious question in the census was first introduced in 2001, as a voluntary option. In some other countries such as France, state questions about race, ethnicity or religion are not permitted.
But in the UK, the vast majority of people answered it despite not having to, although the reappearance of the same question in the forthcoming census has prompted complaints.
Question 20 in England and Wales will say: "What is your religion?" In Scotland, question 13 will ask: "What religion, religious denomination or body do you belong to?"
The British Humanist Association (BHA) believes they are leading questions that actively encourage people to tick a religious answer, thereby inflating the numbers, especially among Christians because many people hold a weak affiliation.
If you were baptised but had not been to church since then, you might be inclined to say you were still Christian, says Naomi Phillips, the head of public affairs at the BHA. She says the actual number of secular people is probably double the number the census recorded.
"Many people tick Christian but wouldn't consider themselves to be religious if you asked them otherwise. And this is used to justify maintaining faith schools and used by local authorities to make their planning decisions to allocate resources to public services.
"It means more budgets go to Christian groups and the needs of non-religious groups are not taken into account."
The BHA begins a poster campaign next week on buses and at railway stations that urges people who are non-religious to "for God's sake, say so".
Ms Phillips says it would be preferable not to have the question, because it's hard to get an accurate picture.
"It's very difficult to measure. There are so many different things to measure - by belief, practice, whether you believe in God, whether you attend places of worship, whether you pray."
The census question pre-supposes you have a religion, she says, and a two-part question like they have in Northern Ireland would be fairer, which differentiates between your faith at birth and your faith now.
The humanists are not alone in wanting the question changed. The Foundation for Holistic Spirituality (F4HS) wants it easier for those people who have a spiritual but non-religious tendency to answer the question.
But the Office for National Statistics, which collects the data, says the question is one of a number that allows people to fully express their identity in the way they consider most appropriate.
"The religion question measures the number of people who self-identify an affiliation with a religion, irrespective of the extent of their religious belief or practice," says a spokesman.
It's a question that is worded in the most sensitive way possible, says historian and broadcaster Nick Barratt, especially with the subtle change of emphasis introduced in the new census - respondents are now faced with "no religion" as the top option to tick, rather than "none".
"This [change] makes it more secular, and easier for people to identify with the question and where they are coming from. There's the question of faith and belief as opposed to religion. It allows other beliefs to get in. If you said 'none', it is like you have no belief or faith, but 'no religion' means you may have."
He expects this change could mean fewer Christians this time, but it's an important question, he says, because it shows how richly diverse some communities are.
It also has a practical purpose, says the ONS. The results are used to improve understanding of communities, it says, and to provide public services, monitor discrimination and develop policy to best cater for people's religious backgrounds.
But what is the true picture? Whichever survey is accurate, it's clear that many people in Britain still feel an affinity with Christianity, even if they haven't attended church in many years.
Average Sunday attendance in the Church of England was 960,000 in 2008, a figure which has been falling for a number of years. A survey by Christian charity Tearfund suggested it was one in 10.
Yet nearly 40 million people in England and Wales, 72%, identified themselves as Christian. Other surveys suggest the majority of people pray and believe in God, even if they don't regularly go to church.
Christianity should not be measured simply in terms of Sunday worshippers, which are falling in number, says a Church of England spokesman, because the numbers of people going at other times remains high.
"The 72% figure seems to be constant and not decreasing. What's interesting for us is the social mobility and social change. People might not go on a Sunday to church any more but might go on a Saturday or Thursday or they might go less often. It's a change in how much time they have available.
"We have made worship available online, in the morning and in the evening. There's probably more people engaging with the church than ever before."
Christianity is a religion that people identify with, he adds, regardless of their level of church-going.
But it's impossible to quantify the numbers, says pollster Stephan Shakespeare, founder of YouGov.
"It's very hard to make an absolute measurement. You have to get an ideal definition about what being a Christian means or what being religious means. But what is useful is to ask the same question as last time and see the change."
So even if a question is slightly flawed, it's better to stick with it.