I welcome the Senator's interest in the census of population. The census is a major source of policy relevant information and it is therefore only right and proper that all aspects of census methodology should be subject to detailed examination. In this regard, the discussion of the day of the week on which the census should be held is one of the most fundamental issues.
Sunday has been the reference day for all censuses of population conducted since the foundation of the State. The reasons for this practice are well grounded and, in the view of the Central Statistics Office, continue to apply at the present time. Sunday night, in particular, is the one night of the week when the majority of persons are to found at the address in which they usually reside.
Censuses in Ireland are taken on a de facto basis. Using the de facto approach, persons are enumerated where they spend census night, regardless of where their usual residence is. This has the advantage of being easy to operate and facilitates an accurate population count by minimising under-coverage while avoiding possible double counting.
Some of the implications of that approach are visitors temporarily in the State on census night are covered in the census count, and usual residents who are temporarily absent from the State are excluded.
By analysing the enumerated population according to their place of usual residence it is found that the numbers involved in the previous two categories are not significant in an overall population context. For instance, according to the results of the 1996 census, it is reckoned that 96.7 per cent of the usually resident population were enumerated at the address at which they usually lived. A further 0.9 per cent had their usual residence elsewhere in the county in which they were enumerated while 1.6 per cent lived elsewhere in Ireland. It is estimated that 0.8 per cent of the enumerated population were visitors from outside the State while 0.3 per cent were absent from the State on census night.
The census has traditionally been taken on a Sunday in April. Careful assessment of factors, such as the timing of Easter and the public holiday in early May, has to be carried out to select a date when the number of tourists is not likely to impact in a major way on the population count. April is also a period with a sufficient number of daylight hours available for enumerators to carry out their work. For historical and practical reasons it is desirable that the same census date should be chosen in this State as in Britain and Northern Ireland.
Saturday night is singularly unsuitable for census taking. It is the night when persons are most likely to be absent from their usual residence, whether they are visiting friends or relatives or pursuing weekend breaks away from home. This is clearly evidenced by the volume of traffic operating between population centres from early Friday afternoon on any weekend, and the reverse flow commencing early Sunday afternoon.
If Saturday night was chosen as census night, the number of persons enumerated in their usual place of residence would be much lower than it would otherwise be. This would adversely affect the quality of the census count, especially at a small area level. The operational difficulties for enumerators would also be substantially increased, particularly in trying to cover hotels, guesthouses and holiday homes. In addition, continuity with the results of previous censuses would be broken.
An alternative to the de facto measure of population is the de jure or usual residence concept which is used a number of countries. Using this approach persons are enumerated in their usual residence regardless of whether they spend census night there. In particular, usual residents who are outside the State on census night have to be enumerated while foreign visitors temporarily in Ireland are excluded. The concept requires the adoption of clear operational rules in order to define a person's precise usual residence.
Regardless of whether the de facto or de jure method of enumeration is used in the census, difficulties arise in the treatment of students, especially those who are living away from home during term time. However, the extent of the problem tends to be exaggerated.
The conventional system for students adopted in most countries is to consider their term time address as their usual residence, given that they spend the majority of their time there. Taking the student population aged 18 to 24 years, as an approximation for the population of third level colleges, the 1996 census indicated the relevant population cohort amounted to 135,000 persons.
Analyses conducted by the CSO reveal that the discrepancy between county of enumeration and county of birth or county of usual residence is not of a significant order of magnitude.
It is evident from the foregoing that the designation of Saturday night as census night would serve to worsen appreciably the accuracy of the population distribution of the State rather than enhance it.