Guest post by Faha;
Faha actually sent me this a few months ago. I have been working on expanding it to include the secondary school data but, hands up, it is taking forever and to be honest is very difficult to relate to the actual census figures. Following a polite e-mail reminder I thought it best to go ahead and publish his work as it stands. Many thanks to him and I should apologise for the procrastination.
In May, the census office released the file DC2117NI:Religion or Religion Brought Up In by Age which gives a breakdown of religious background by single age groups in the 2011 census. Later in 2011, the annual school census was completed. It is now possible to compare the results of the 2011 school census with that of the 2011 census. What were the results?
The school census results for primary school age children were as follows:
Catholic Protestant & Other Christian Other None Total
78,941 60,658 1,026 15,069 155,691
The 2011 Northern Ireland census was conducted 6 months earlier so the children were 6 months younger then. I used the age 6 to 11 age cohort and most of the age 5 cohort to match with the school census. The results are:
Catholic Protestant & Other Christian Other None Total
76,931 62,871 1,219 14,672 155,691
The discrepancy between the school census and the 2011 census is:
Catholic Protestant & Other Christian Other None
+2,010 -2,213 -193 +397
The school census is recording over 2,000 more Catholics and over 2,000 fewer Protestants than the 2011 census. Why the discrepancy?
Primary school attendance is compulsory so almost all primary school age children attend primary school and are recorded in the school census. The 2011 census, on the other hand, did not have any religious data for 60,000 people who responded to the census. There were an additional 158,000 people who did not return census forms for a total of 218,000 people for which no religion background data was available. This is 12% of the entire population. For those 155,691 children in primary school the 2011 census had no religious background information for over 18,000 of those students. The census office estimated what the religion background of those students was and added their estimate to the 138,000 primary school age children for which they had religion background data.
The school census had religious background for almost all primary school age students. There were 397 excess None/Not Stated in the school census compared to the 2011 census so these would be the only students they did not have religious data on. There may also be a few students who do not attend primary school due to severe disability or illness but his would be a small number. The school census is missing religion background data for fewer than 1,000 primary school age children compared to 18,000 for the 2011 census. The school census has religion data for at least 17,000 of those 18,000 that the 2011 census does not have. The school census data is much more accurate since they have religion data for over 99% of primary school age children compared to only 88% for the 2011 census. Thus, the school census is more accurate and there are 2,000 more Catholics in the primary school age population than were recorded in the 2011 census.
The school census results for secondary school students are:
Catholic Protestant & Other Christian Other None Total
75,891 60,407 650 9,799 146,747
I used the 12 to 16 age cohort and most of the age 17 cohort to match and the results of the 2011 census are:
Catholic Protestant & Other Christian Other None Total
72,285 62,403 929 1,130 146,747
The discrepancy between the school census and the 2011 census is:
Catholic Protestant & Other Christian Other None
+3,606 -1,996 -279 -1,331
As with the primary school census, the school census records more Catholics (3,606) and almost 2,000 fewer Protestants than the 2011 census. The school census also records 1,331 fewer students with no religion. Unlike the primary school census, where school attendance is compulsory, some older students have dropped out of secondary school and would not be recorded in the school census. The coverage of the secondary school census would only be approximately 94%. However, this is still higher than the 88% coverage for religion background data in the 2011 census.
Finally, this would be the results combining the primary school census, the secondary school census and the students in special schools.
Catholic Protestant & Other Christian Other None Total
156,781 122,903 1,701 25,605 306,990
For the 2011 census I used the religion data for the entire age 5 through 17 cohort and a small number from the age 18 cohort.
Catholic Protestant & Other Christian Other None Total
151,459 127,174 2,179 26,178 306,990
The discrepancy between the school census and the 2011 census is:
Catholic Protestant & Other Christian Other None
+5,322 -4,271 -478 -573
The entire school census records 5,322 more Catholics and 4,271 fewer Protestants than the 2011 census.
What are the implications for the accuracy of the 2011 census? It is clear that in the school age population, the school census indicates that there was a systemic undercount of the Catholic population in school age children: i.e. it is present in all age cohorts. School age children have younger siblings who have not yet begun school as well as old siblings who have already finished schooling. They obviously also have parents and grandparents and other relatives. Since this is a systemic estimate error it affects other age groups also. There is a 2,000 undercount of Catholic primary school age children and those age cohorts comprise 9% of the entire population. Extrapolating to the entire population would result in a 22,000 undercount of the Catholic population in the 2011 census. If one uses the entire school census (primary, secondary and special school students) there is a Catholic undercount of 5,322 students and those age cohorts comprise 18% of the entire population. Extrapolating the entire school census undercount of 5,322 Catholics to the entire population would result in a 29,000 undercount of the Catholic population in the 2011 census. The 817,000 Catholics recorded in the 2011 census is not accurate and the school census indicates that the correct number is in the range of 840,000 to 846,000 which is approximately 46.5% of the entire population.