Accuracy in Northern Ireland
Key Findings
- The local government and parliamentary registers in Northern Ireland were both 80% accurate
- Errors can be defined as major errors and minor errors. Major errors are those register entries which refer to individuals who no longer live at a property or would either prevent someone from voting or allow someone who is ineligible to vote
- In December 2018, 20% of entries had major errors and 6% had minor errors
Type of errors
In analysing the accuracy of the electoral registers, a number of different types of error can be identified. These errors are then categorised as either a ‘major’ or ‘minor’ error.
The major error category is comprised of the following:
- entries which refer to individuals that no longer live that address
- entries which may prevent an individual casting their vote at a polling station (for example, an incorrect name);
- those errors that would mean that someone could vote when they are not eligible to do so (e.g. an incorrect date of birth for someone under 18).
Minor errors are those which would not prevent someone from casting their vote (for example, an entry with a spelling error).
In total, about 20% of entries had a major error (up from 13% in 2015) with the most common major error being that no corresponding name was taken at the address (19% of register entries).
6% of register entries had minor errors. The most common minor error was for a middle name to be missing from the register, which was the case for 3% of register entries.
Here is a breakdown of the types of errors used to calculate the accuracy of the registers:
| December 2015 | December 2018 | |
|---|---|---|
| Major errors total | 13% | 20.1% |
| Major errors-(a) | 12% | 18.7% |
| No corresponding name taken at address | ||
| Major errors-(b) | 0.4% | 0.7% |
| First name and/or surname wrong on register | 0.1% | 0.5% |
| First name and/or surname missing on register | 0.0% | 0.0% |
| UK/Irish/Commonwealth marker present | 0.2% | 0.2% |
| Major errors- (c) | 0.1% | 0.6% |
| Name on register corresponds to ineligible name on survey | 0.0% | 0.5% |
| Attainers- DOB missing or wrong | 0.1% | 0.0% |
| EU citizens marker missing | 0.0% | 0.1% |
| Accurate with minor errors | 7% | 5.6% |
| First name and/or surname on register misspelled | 0.7% | 0.9% |
| First name/surname on register incomplete | 0.4% | 0.0% |
| Middle name missing from register | 4% | 2.8% |
| Middle name or initials misspelled or incomplete on register | 0.5% | 0.5% |
| Middle name or initial wrong on register | 0.0% | 0.5% |
| Person does not have a middle name but middle name on registers (respondents only) | 0.4% | 0.0% |
| Surname is/assumed to be previous surname | 0.4% | 0.5% |
| First/middle/surname in different order on register | 0.8% | 0.0% |
| DOB earlier on register for attainer | 0.0% | - |
Demographic characteristics
Examining the demographic nature of inaccuracy on the electoral registers is more challenging. By nature, it is not possible to record demographic details for register entries which are not matched to a resident during an interview.
Nor is it possible to take into account properties which are derelict or entirely vacant.
In order to provide some guidance to the patterns of accuracy by different household characteristics, the analysis below takes into account household data where a resident was interviewed. However, this data still only presents a limited picture and so must be treated with caution.
Tenure
Accuracy by tenure follows a similar pattern to completeness, with outright owned households the most likely to have accurate register entries (86%).
Households where the home is being bought on a mortgage (or through shared ownership) were 85% accurate, households rented from the social rented sector have 82% accuracy, while private rented households have a lower level of accuracy, at 61%.
Social grade
Social grade is also calculated at the household level, based on the occupation of the chief income earner. The pattern of accuracy is closely related with completeness. Accuracy for all social grades is within eight percentage points of each other. This is a similar pattern to that observed in Great Britain.
| Northern Ireland | |
|---|---|
| AB | 86% |
| C1 | 85% |
| C2 | 82% |
| DE | 78% |
Length of time at address
Accuracy follows the same pattern as completeness when looking at how long households have occupied their accommodation.
Accuracy is lowest for households who have been resident for two years or less, a 22% level of accuracy at less than one year, and 58% for one to two years. Households which have been resident longer show a similar level of accuracy of between 84% and 90%.