Rounding policy
Updated 8 March 2023
Applies to England
We do our best to ensure that the official statistics we publish do not reveal the identity of any individuals or organisations, or any private information about them. Private information that we collect to produce official statistics is confidential and we use it for statistical purposes only. We follow the guidance issued by the Government Statistical Service / Government Research Service about this. All data is collected, stored and used within legal requirements such as the General Data Protection Regulation.
Private information we collect to produce official statistics is confidential and we use it for statistical purposes only. We follow the guidance issued by the Government Statistical Service / Government Research Service about this.
1. How we define private information
Private information is information that:
-
relates to an identifiable legal or natural person
-
is not in the public domain or common knowledge
-
would cause them damage, harm or distress if the information were made public
A ‘legal person’ is a company or other organisation that has legal rights and duties. A ‘natural person’ is a member of the public. Where we use the term ‘individual’ in our publications it means both legal and natural persons, both living and dead.
Information that is lawfully in the public domain and readily available to the public does not automatically become confidential when it is used to produce a statistic.
2. How we use rounding
We use rounding wherever it is needed to make sure that our data does not reveal the identity of an individual, or any private information about them. We also use rounding to make our published data clearer.
Rounding creates uncertainty about the real value of a number. It makes very small changes to the data that do not affect the data’s overall accuracy.
Rounding involves adjusting numbers to a specified base. For example, if we were using a base of 5 we would round all the numbers in a publication or part of a publication up or down to be a multiple of 5. In this example, a value of 26 would be rounded down to 25.
In our statistical releases all figures are normally rounded to the nearest 5, unless otherwise stated. Values less than 5 (i.e. 1 to 4) are represented as ‘fewer than 5’. Zero is denoted as 0.
This applies to reports as well as accompanying data. In some instances, when there is no risk associated with the disclosure of the identity of an individual, statistics might be best represented as unrounded numbers – for example, the number of awarding organisations given in the annual qualifications market report.
In some instances, where individual rounded values have been presented in a table along with their sum total, the total may be slightly different to the sum of these individual rounded values because it has been calculated using the original unrounded values.
In general, we use unrounded values to derive percentages. Percentages are normally rounded to up to one decimal place.