Out for the count

An Office for National Statistics consultation on the future of the UK census could spell the end of a 200-year-old social-science experiment. The risk is quite real: during our recent inquiry into proposed changes to the national survey, Francis Maude, the Cabinet Office minister, wrote to the Science and Technology Committee saying that costs were a concern and implying that utilising other data might allow the census to be scrapped.

Many groups, bodies and individuals rely on the census for their work. For example, the data are invaluable for social scientists, who follow people throughout their lives to gain insight into how society is changing; for central and local governments, which have to plan for school places, hospital provision, services for the elderly, etc; for local charities, which can compile information and judge where resources might be needed to address health, social and welfare problems; and for local historians, who can trace people back through the generations.

Surely alternative data sources could be used? Certainly, many surveys and reports provide similar information. None of them, however, is as comprehensive in terms of numbers or coverage. Indeed, many of these surveys require the census as a calibration point. In social science, sample size is obviously key: the more people are involved, the more certain researchers can be of the conclusions being drawn. Nothing else is as big as the census. Removing it could rip the heart out of UK social science.

The census has long been an invaluable resource in longitudinal studies, helping to identify candidates for recruitment and allowing such studies to continue well beyond individuals' lifespans. The UK produces some of the richest longitudinal studies in the world but it is not clear how they would continue in the absence of the census.

All this is not to say that the census is perfect, of course. There are several issues that social science faces in using it - a major point being that it happens only once a decade. To compound that, the information it collects is not made available for two years. This means that census data vary between being two and 12 years out of date. On the national scale this may not have a massive effect but on a local scale, where major changes can take place very quickly, it can mean that too few school places are planned or too many hospital beds are made available, all of which means a potential waste of public money.

Another problem with the census cycle is that the operation threatens to dominate the ONS for two years every decade, disrupting the organisation's usual day-to-day activities during that period.

A key issue uncovered by our inquiry is the lack of ministerial oversight in this area. There is a minister responsible for the census (currently Maude), but only for its delivery. Many government departments use the data and independently lobby for what they need from each census. No one, however, is responsible for the overall picture, examining other social-science surveys paid for by public money and deciding whether the census is meeting the government's needs.

So could the census be improved? Undoubtedly. The production of any dataset of such size and complexity is prone to error and delay. There will always be debate about whether different questions could provide better, more insightful information. However, we found that the issue of whether census data can be improved is not widely debated among social scientists. To the majority of academics in the field, the census is central to the work they do. There are some who argue that the information could be obtained in other ways - by using regular surveys and existing administrative data, for example - but there is no doubt in my mind that any alternative, however effective, would prove no cheaper and may even be more expensive.

If the government is looking for new ways to save money, this is not the place to look. Any savings from scrapping the census would be more than offset by the weaker evidence base upon which crucial decisions in health, welfare and other spending areas would have to be made as a result.