Academic journal article Journal of the Statistical and Social Inquiry Society of Ireland

Using an Administrative Primary Care Health Activity Indicator to Address Under-Enumeration in the 2011 Census in Northern Ireland

Academic journal article Journal of the Statistical and Social Inquiry Society of Ireland

Using an Administrative Primary Care Health Activity Indicator to Address Under-Enumeration in the 2011 Census in Northern Ireland

Article excerpt


The purpose of this paper is to outline important development work that NISRA undertook in terms of utilising an administrative Primary Care health activity indicator to help address under-enumeration in the 2011 Census. The approach, which was unique to NI, is considered to have enhanced the quality of the 2011 Census estimates in Northern Ireland and has attracted considerable interest both nationally and internationally. The paper outlines both the rationale and methodology for the approach and discusses how it might be further enhanced in any future Census operation.


The Census is, without question, the largest statistical exercise undertaken by Government. In 2011 the Census in Northern Ireland, which was conducted on Sunday 27th March, included some 703,000 households, 1,100 Communal Establishments (such as Nursing Homes, Army Barracks, Student Halls of Residence etc) and 1.81 million people. The Census is important as it informs, amongst other things, the allocation of funds to Northern Ireland by the UK Government. In addition, it helps shape how that money is distributed throughout Northern Ireland in respect of (i) the delivery of essential services, such as education, health and transport and (ii) areas and people in greatest need. While the primary legislation that underpins the Census--the Census Act (Northern Ireland) 1969--places a legal obligation on everyone to take part in the Census, and indeed makes provision for non-responders to be prosecuted and potentially fined up to [pounds sterling]1,000, not everyone does. This gives rise to what is known as under-enumeration, which is not unique to the Census in Northern Ireland.


In Northern Ireland it is Census policy, in line with arrangements throughout the rest of the UK, that Census estimates should cover the entire population. Accordingly, Census Office is required to address any under-enumeration in the Census by imputing any details that have been missed. This includes details for whole households and individuals who failed to respond to the Census as well as any items that were omitted by those who did respond, but who only partially completed the questionnaire. In the 2001 Census, a Census Coverage Survey (CCS) was used post Census day to estimate the level of under enumeration. A report on the process, the One Number Census (ONC), is available at:

In summary, the key steps were as follows:--

* A stratified random sample of postcodes covering some 12,000 households was completely re-enumerated by NISRA's Central Survey Unit through face-to-face voluntary interviews. The sample was stratified on the basis of whether an area was (i) urban/rural, (ii) predominantly Protestant or Catholic, and (iii) deprived/ not deprived.

* The information from these interviews was then matched against the corresponding information collected through the Census for the households in question and used to derive statistical models that provided modelled estimates for the overall number of people who had been missed, along with their key demographic details (e.g. age and sex). These statistical models, which were specific to the CCS areas, were then generalised to cover all of Northern Ireland.

* The remaining details for those households and people that were estimated to have been missed were then imputed using the characteristics of similar households and people in neighbouring areas.

In 2001, it was estimated that 95% of the population responded to the Census (1) thus 5% of the population had to be imputed. In the run up to the 2011 Census, it was recognised that getting a response from the general public was becoming increasingly more difficult due to general apathy with surveys and a level of mistrust of Government. Accordingly, there was a risk that this percentage may be higher in the 2011 Census. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.