Originally published on the RMG:Clarity blog on April 1, 2011

By now, most people in the UK will have filled out their census returns, either online or on paper, and returned them for the massive number crunching exercise to begin.  But is the census necessary?  Some people and groups contend that it is too expensive, unreliable, and an Orwellian intrusion into peoples’ lives.  Take this quote from Daniel Hamilton, the Campaign Director of Big Brother Watch:

The census includes intrusive questions on your proficiency in English, your health, when you last worked, the identities of your overnight guests and the type of central heating you have.  The government has no need – and no right – to know this information about you.

This census is a monumental waste of time and money.  A large number of the questions duplicate data already held by the authorities on databases such as the electoral register, school records, tax returns and GP information.

Or this from campaign group NO2ID, from their list of ‘10 Census Lies:

 1. The Census is essential for government and business planning

On the contrary, it is worse than useless because it is expensive, inaccurate, and quickly out of date.

6. The census results in high-quality information.

No one knows how many people lie in their return. The 2001 census is generally believed to have ‘missed’ around 900,000 men under 40.

In a 9 July 2010 article, The Telegraph announced that the then Shadow Minister for the Cabinet Office, Francis Maude wished to scrap the census:

Mr Maude said the Census was “out of date almost before it has been done” and was looking at ways to count the population more frequently — perhaps every five years — using databases held by credit checking firms, Royal Mail, councils and Government.

“This would give you more accurate, much more timely data in real time. There is a load of data out there in loads of different places,” he said.

There are, then, several arguments against the census, from various directions, including cost (£482million), privacy, accuracy, completeness and duplication of data.  It should be noted that arguments such as these have been in circulation since the very idea of a UK census was raised over 250 years ago, but are they valid?

£482million may seem like a massive figure to spend on a census (and is a large increase from £259million in 2001), but how does that compare with, say, the 2010 US census, which is only 10 questions long?  The cost of the UK 2011 census breaks down to a cost of 87 pence per person, per year (over the life of the census), or £8.70 for the full ten years.  The 2010 US census cost $14.5billion (c. £9.06billion), which equates to $46.96 (c. £29) per person per ten years; this is over three times the cost of the UK census, and for a much shorter survey.  When looked at in these terms the relative cost of the UK census does not seen quite so excessive.

The aim of the census is to provide a snapshot of the UK population at one particular time, rather than to track population change through time.  This means that while the criticism that the census is out of date as soon as it is done is fair, it is not really the point.  By obtaining the same information about everyone in the UK (or as near to everyone as is possible), at the same time, one can make the comparisons about lifestyle and economic status that are central to future planning.  Data duplication is also a valid concern; there is little point in the government collecting data via the census that they could get from other sources, however, this other data would have to be at least as complete as census data.  Since being in government, Francis Maude has had to change his views, at least for the time-being; as he said on 26 July 2010, ‘The current advice from the ONS is clear.  Census alternatives are not sufficiently developed to provide now the information required to meet essential UK and EU requirements.’

As David Hamilton suggests, the authorities may have duplicate data on school records and GP information, for instance, but one wonders whether he would be supportive of the kind of data sharing that would be necessary to collate all this information.  Similarly, to create and maintain a database that contained all of this linked information and kept it up to date and secure would be a monumental undertaking that would probably exceed the cost of the census.

There are famous examples of people avoiding the census; the suffragettes in 1911, anti-poll tax protesters in 1981, and the mysterious omission of 900,000 men under 40 in 2001.  This however, does not stop the census being the most complete survey of everyone living in the UK.  Whilst it is impossible to ensure that there is a 100% completion rate, it is high enough for the purposes for which it is required.  The census will provide a large body of data that the government will use to direct its spending and, perhaps more importantly, its funding cuts; without this kind of data government cannot begin to apportion funding.

Some of those who argue against the census contend that there are other large scale surveys that could be used for some of the data instead; this however, is a false economy.  When any research company, RMG:Clarity included, wants to carry out a quantitative survey, on any topic, one of the first things that has to be worked out is the sampling – how many people do you interview in a particular area?  Take our poll on the general election last year; if we had randomly interviewed 1000 people in Wales and left it at that, the results would have been very different.  We could easily have ended up mostly speaking to people in the South-Wales Valleys, and other Labour heartlands – this would obviously have affected the result of the poll.  To counteract this, we produced a target number of interviews that we needed to achieve in each constituency; these target numbers were directly proportional to the number of people of voting age in each constituency.  The source of this data, was, of course, the 2001 census.  (On top of this we also weighted the final results by other factors such as age and gender so that they matched the census figures.)  Using this process we were able to predict the outcome of the election to within +/- 2 percentage points.  The famous pollster George Gallup said that sampling a population was like taste-testing soup; if the soup is well stirred, one spoonful can reflect the taste of the whole bowl.  He could have added that a census provides the recipe for the soup.

So, is the census necessary?  It is certainly necessary for the government to know the basic facts about its population.  Without knowledge of the population size of council areas, for example, government cannot fairly decide how much money to allocate to each council.  What remains to be seen, however, is whether the next ten years will produce a more effective, and cheaper, methods of collecting this valuable data.