Ireland must conduct regular, high-quality censuses because of demographic instability, writes GARRET FITZGERALD
IRELAND CONDUCTS censuses twice as frequently as any other country – and it is arguable that our census system is also the best.
Most states count their people every 10 years, but since 1946 we have held population censuses every five years. Indeed if you count the August 1941 population enumeration connected with the issue of wartime ration books, we have had five-year checks on our population throughout the whole of the past 75 years.
On two occasions a population census was cancelled or deferred. The national coalition government, of which I was a member between 1973 and 1977, cancelled the 1976 census on the stated grounds that this would save money in the economic downturn following the first oil crisis of 1973. However, Jack Lynch’s subsequent government decided to hold a limited census in 1979.
The Fine Gael party was the unintended beneficiary of that decision; the 1970s had been a decade of rapid population growth, and as a result of that special census 18 seats were added to Dáil Éireann – which, in 1981, helped Fine Gael to increase its Dáil membership by 23 seats, and thus to win the 1981 election.
Then, a decade ago, the 2001 census was delayed by a year due to the outbreak of foot-and-mouth disease.
The reason we doubled the frequency of our censuses of population 60 years ago was the unique instability of Irish demography in the second half of the 20th century. Since the 1950s Irish migration has reversed direction on four occasions, in addition to which our marriage and birth rates have experienced three major changes of direction. I do not think that any other country in the world has experienced anything like this degree of demographic instability. So, to a greater extent than perhaps any other state, we have needed more frequent enumerations of our population.
I said above that we not only have more frequent censuses than any other country but also, arguably, better censuses – better in the sense of being more accurate. This is because our censuses are carried out by about 5,000 enumerators calling at every dwelling so as to make direct contact with those living there, with a view to securing as near as is possible a complete coverage of the population.
These home visits by enumerators provide an opportunity not alone to deliver and collect census forms but also to provide assistance where needed to the person filling out the form. This year the form contains 11 questions on household characteristics, 30 in respect of each individual present on census night, and a further eight questions in respect of any persons usually resident there.
Apparently, the US is the only other developed country that uses this method of census-taking. All other developed countries seem to use a combination of internet and mailing. Often only about one-third of forms are returned by internet and, because many forms are not returned within a reasonable time by either method, enumerators have to chase up the more difficult cases. Allied to this is the need to carry out a large sample Post-Census Coverage Survey, as invariably response rates fall short of the 100 per cent target.
Publication of final results can take up to a year – whereas our system yields preliminary results within three months. And, as was seen in 2006, when the final census figures were published eight months after the preliminary data, the margin of error in the preliminary data was found to have been one in 1,000.
At present our most recent pre-census population figures are for April last year, based upon the CSO’s best estimates of flows of emigrants and immigrants in each of the four years following the 2006 census. These figures are subject to a certain margin of error, the extent of which will be established when the new census data becomes available in July.
Each enumerator is provided with up-to-date large-scale maps showing every dwelling in the area allocated to her or him (on average, 400 dwellings each), and if there is no response they are required to inquire diligently from neighbours as to whether that dwelling is occupied. They must visit and, if necessary, re-visit each occupied dwelling until contact is made with a resident. Moreover, the enumerators’ remuneration will be linked to the number of forms returned, which will provide an incentive to seek out occupants.
The data from the census will be published in respect of 3,000 electoral divisions, but in the case of this census it will be available, if needed, in respect of some 20,000 smaller areas, devised by the National Centre for Geocomputation in NUI Maynooth. These 20,000 small areas cover an average of about 100 dwellings, grouped within each electoral area.
While the forms to be filled are in English or Irish, copies are available in 21 other languages. This census contains two new questions. The first asks whether any of the residents speak a language other than English or Irish at home. What is that language? And, how well do they speak English? The second asks how is each person’s health – which may seem a vague query but one which other countries have found useful.
After the 2006 census all of the 15 volumes of the census were published the following year, and provided an important basis for policy-making, Today, at a time when we face so many intractable economic and social problems, it is of crucial importance to secure a comprehensive and up-to-date picture of our society.
Quite properly, every citizen is required to co-operate with this project – we have a very good record of such co-operation – based on public confidence both in the quality of our census data and in its confidentiality, which, since the earliest days of our State, has never been breached, or, indeed, impugned.