David Kernohan is an Associate Editor of Wonkhe


On Monday 21 February 1921 Bertram Falle, then MP for Portsmouth North and University of Paris graduate, asked Chancellor of the Exchequer Andrew Bonar-Law a question that had been on the minds of many:

if it is absolutely necessary to take a census of the people this year; if such might be put off for a couple of years with advantage and economy; and if he will consider the matter.

It was a fair question. The country finally seemed to be emerging from a multi-year global pandemic, there were serious problems brewing in Ireland, and a wave of industrial action looked to be on the cards. It was a hell of a year. With all this to deal with – not to mention the after effects of World War One – why spend time and money on seeming trivia like the census?

Bonar-Law’s answer was instructive as to the importance in which the census was now held:

I think it is very necessary that a census should be taken this year, and in any case, the arrangements which were confirmed by this House last Session have now gone so far that the cancellation of them is, I think, out of the question.

Of course, history relates that the 1921 survey was postponed after all. The original date – 24 March – coincided with industrial action by miners in newly-privatised coal mines, supported by transport workers and railwaymen. The census actually took place on 19 June 2021 – an amendment slip for the enumerator documentation was paid for by featuring an advert for MP (and later convicted fraudster ) Horatio Bottomley’s Sunday Illustrated.

A date to remember

Even something as simple and as sensible as this date change made interpretation of the data very difficult. For instance, educational activity (from schools to night classes to universities) that would have been taking place in March did not take place in June. This means fewer people were recorded as students, and those that were students would not have been recorded at their term-time address. The official report shows that 22 in every thousand 18 year old males, and 26 in every thousand 18 year old females, were attending educational institutions full time in 1921. We don’t know what kind of institutions these were, any more than we know what the full number would have been in term time.

Likewise, the populations of Oxford (57,036) and Cambridge (59,246) returned by the census – though these both showed a growth over 1911 of between 6 and 8 per cent – cannot have included the students that would otherwise have been recorded as resident at each college. Both towns were likely larger outside of the summer – resorts like Southend (which grew 50 per cent in the same period) were likely smaller.

In England and Wales we’re not dealing with a change of census date this year – it remains 21 March 2021 as originally planned. But there are many possible similarities with what happened in the 1921 census, and much to be done to ensure that higher education and the life that surrounds it is recorded accurately.

There goes the neighbourhood

The intention this year is to count every student both at their home address and their term time address. The letters and emails that providers and students are now receiving should give details as to how this will happen – there’s online guidance available (separately for students in Northern Ireland), and students can request the access codes they need from the same site (with, again, a different site for Northern Ireland).

Letters will be sent to every registered address in England, Wales, and Northern Ireland – so students already at their term time address may already have a code. But a very large number of students will not. Depending on your preferred source at least half of students are currently living elsewhere.

To get proper census coverage of students we are relying on a variety of media – social campaigns, the post, and the urgings of universities and students’ unions. There’s not really one simple way to get a message to all students, hence the scattershot approach.

You’d maybe forgiven for thinking that the same problems would exist in the wider population – how many people are realistically going to request a code to complete an online form? For 2011, it was estimated that around 86 per cent of adults had seen publicity, but for most it simply encouraged them to complete a form that came through the letterbox. census 2021 is almost completely online – if you are unable to complete the form online you will have an opportunity to ask for a paper copy, and some follow-up work (like the Census Coverage Survey and Census Quality Survey) is in person.

Census response rates vary by area – after a follow up exercise in 2011 it the coverage ranged between 95 per cent in easier to reach areas, but was as low as 80 per cent elsewhere.

And here’s where it found students in 2011:

[Full screen]

But why is it important?

As Roger Hutchinson’s diverting The Butcher, the Baker, the Candlestick maker: The Story of Britain through its census, since 1801 tells us, we owe the modern census to an essay written by a bored Oxford graduate in his Dad’s library that was later used to fill space in The Commercial, Agricultural and Manufactures Magazine in June 1800.

John Rickman’s essay wasn’t especially well argued or persuasive – but for the time it was superb viral content. Certain sections of society were aflame with debate about the arguments of Richard Price (who argued that England was in terminal decline with a shrinking population) and, more famously, Thomas Malthus. Malthus’ An Essay on the Principle of Population argued that – unchecked – the growth of the population would lead to famine or plague. These weren’t new arguments – in the 1750s Gin Craze it was generally assumed (because of the state of London at the time) that the nation was drinking itself to death – but at the turn of the century it felt like an answer may be in sight.

Censuses had been proposed – and indeed – carried out elsewhere, and many similar exercises linked to taxation (the Domesday Book, most notably) had been undertaken in Britain. But between the arguments of Price and Malthus understanding the population of Britain and how it would change over time became of paramount political importance. Rickman was eventually given an office in parliament and control of the March 1801 census, setting in train a series which – starting with the simplest of head-counts – eventually expanded to offer an astounding range of detail on how and where the population lived, how they worked, and how they were educated.

Each addition to this now quite lengthy (it will, the 2021 team attest, take around 10 minutes to complete) was designed to answer another political argument.

National concerns about nationality (1851), languages spoken (added regionally between 1851 and 1891), work status (1901), divorce and industry (both 1921), and migration (1961) have been joined surprisingly recently by ethnic identity (1991), and religion (2001). Considering that gender identification and sexual orientation both make the cut for 2021 we can read the history of the census both as a history of Britain and as a tour of every volume-based moral panic for the last 220 years. The census is identity politics writ large – answering prejudice and the fear of being overwhelmed by the unknown with a reassuring heft of statistical integrity.

And any survey of the population starts with the census – if you want to weight a sample to be reflective of the general population then you need to have some idea of how the general population is made up in terms of the characteristics that could affect responses. The accuracy of the census underpins the accuracy of the polls that make the political weather.

The census in higher education

Data on student’s term time addresses was first gathered in 1991. It wasn’t what made the news in the early nineties – the big story at the time was an undercount linked to a desire to avoid the Poll Tax – but term time addresses were the ones used for publication, with the need to explicitly gather them determined by an early Easter that year trunctuating the term.

The number of students in higher education was by then large enough to have an impact on local and national planning – local governments and providers need to understand where students were living in order to plan services and allocate resources. The concept of a “student area” was in most places born in the eighties, but it was already clear that more growth in higher education was expected.

The steady increase in student numbers between 1981 and 1991 became rather more rapid in the next decade. In 1991 the UK student population (young people aged 18-24 in full-time education) was around a million, it now sits at more than double that.

Our mass higher education system makes use of many area-based measures in regulation – POLAR, TUNDRA, the indices of multiple deprivation – all of which owe a lineage to the census and the annual surveys that keep the findings fresh. Here’s a plot of the little-used OfS Adult HE measure for England and Wales – simply the number of graduates found in each (MSOA) locality in the 2011 census.

[Full screen]

Standard Occupational Codes, last refreshed last year, determine which careers are graduate (or “highly skilled”) – the conflation of educational background and employment helps us identify which jobs are now routinely carried out by our expanded graduate workforce. This happens when we learn what jobs graduates are doing – and the census is a key factor in how we get to that knowledge. The census doesn’t split undergraduate from postgraduate, but it does offer the widest possible view of where graduates are and what they are doing.

The task ahead

The storied history of the UK census withstood the independence of Ireland, the travails of the early 1920s, the early 1930s depression. The only missing entry in the series is 1941 – even then data from the 1939 National Registration Act can be repurposed. This year marks the only the second time even a part of the census has been delayed – in Scotland a census will take place in 2022.

But, rather than avoiding Covid-19, the data for England, Wales, and Northern Ireland will show the immediate impact of the pandemic and the shape of the nations as the recovery begins. Higher Education is blessed with high-quality annual data, but arguably it is the census that really matters when it comes to representing the sector to the wider population.

ONS told us:

We have launched a student-specific census campaign to raise awareness and the National Statistician has written to all vice-chancellors to ask for their help in sending written prepared emails and other material to all students on our behalf.

  • All students need to be included in the census, and they should complete it for their usual term-time address. If they’re currently living at their home address, they will need to be included in the census for that household too.
  • If they are at their term-time address currently, they’ll get a letter with an access code so they can complete the questionnaire online.
  • If they are living elsewhere (for example at their parent’s house) ONS still cares about where they would be were this a regular year. They’ll be asked to request an access code online, so they can complete the census for their term time address.
  • The exception is if they don’t intend to return to their term-time address at all during the 2020-21 academic year, even to collect belongings. In this case they are returned as part of the household they are living with on census day.
  • International students who would normally be in England or Wales but are attending university remotely will be contacted via universities.

Leave a Reply