Using symbols and shorthand

Policy details

Metadata item Details
Publication date:11 January 2022
Owner:Analysis Function Central Team
Who this is for:Anyone publishing data tables on public sector websites

Who is this guidance for?

This guidance aims to give a consistent shorthand to use in tables, spreadsheets and other related documentation, such as methodology reports, published by the public sector.

Back to top of page

What we mean by 'shorthand'

When we talk about shorthand we mean using a letter or letters to represent a word or words. We do not recommend using symbols such as ‘. :’,  ‘*’ or ‘†’ anymore because they can be difficult to see and screen reader software can find them hard to recognise. This means they are likely to go against the accessibility legislation.

Information should be given at the point of need wherever possible. For example, write out the word ‘revised’ rather than using ‘r’. However, we know this is not always practical in tables and spreadsheets.

By using consistent shorthand across all public sector tables and spreadsheets we can provide many benefits to users.

Note: unit markers such as £ or % are generally considered accessible and can be used in column headings, table titles, row labels or within data cells. However, when possible writing out ‘per cent’ is more accessible than using the symbol. Also ensure you do not overuse them as this can cause visual clutter.

Back to top of page

Layout of shorthand

Whenever you use a shorthand marker in a table it should be in square brackets, for example: ‘[r]’. If you need to use more than one type of shorthand in the same cell, you should put each in its own square brackets, for example: ‘[r][e]’. This helps with machine readability and makes it easier to spot shorthand.

Whenever a table contains shorthand, you should mention it and explain what the shorthand means. The best place to do this is above the table. This will ensure users get this information before they look at the table itself. In spreadsheets this should be in a cell in column A. This is because users of assistive technology will often navigate down from cell A1.

For example, you could present a sentence like this: ‘Some shorthand is used in this table, [e] = estimated, [f] = forecast’.

Back to top of page

Where to put shorthand

Ideally, you should write shorthand in titles, column headings or row labels. You can also put them in empty cells or to replace a data point in a cell.

Shorthand in cells with data

If you are working on a presentation table you can place shorthand in cells, alongside a data point, but this is not the best option as it affects the alignment of the data which can make a table harder to read.

You should never place shorthand in cells with data in spreadsheets. This is because it can affect usability and machine readability. For example, if you want to sum a column but one of the data points has a shorthand marker next to it, the data point will be ignored.

Our recommended approach for spreadsheets is to add a notes column to the right-hand side of the table. Shorthand markers should go in this column along with information about which cell or cells the shorthand applies to, for example: ‘[x] applies to B10, C10 and D10’. You can then use an accessible colour or bold text to draw attention to the data points the shorthand refers to.

More information on the placement of shorthand is in our ‘Releasing statistics in spreadsheets’ guidance.

Back to top of page

No to NA

We do not recommend using ‘NA’ to describe cells with no data. This is because this shorthand is ambiguous. Some people may read it as ‘Not Applicable’, while other people may read it as ‘Not Available’.

Back to top of page

Recommended shorthand

b = break in time series

Use this shorthand when there is a break in a data series that means that data before the break cannot be directly compared with data after the break.

There are many reasons this can happen, including changes to:

  • survey scope
  • sampling methods
  • questions
  • mode of data collection
  • weighting

c = confidential

Use this shorthand when a data point would give away confidential information. For example, if you would be able to identify details about a single respondent from the data. These data points must be replaced with this shorthand to maintain confidentiality clauses.

In previous guidance there was a separate category for suppression. Discussion with the Statistical Disclosure Control Unit at the Office for National Statistics (ONS) confirmed that reasons for suppression would either be for confidentiality purposes or low reliability. These are now identified separately in this list.

e = estimated

Use this shorthand to show when a data point is estimated. Tables which are entirely estimates do not need a marker for each data point, as long as titles or accompanying information make it clear that all figures are estimates.

er = earliest revision

Use this shorthand to show a period when the earliest revision was made in a particular time series. No other periods should be marked with a revision.

An example from Public Sector Employment is the reclassification of public bodies. When colleges were reclassified from the private to the public sector, this increased the series by about 230,000 from 1993 onwards. Shorthand was used to show that the 1993 estimate was where revisions began.

You should only use this shorthand if it is with a revision table for reference so that revisions over the period can be identified. This is not to be confused with ‘revision’.

f = forecast

Use this shorthand to show when a data point is a calculated future value instead of an observed value.

low = a low figure but not a real zero

Use this shorthand to replace data points that appear as zero in a rounded table, but are not actually zero.

For example, international migration statistics published by the Office for National Statistics (ONS) report to the nearest thousand, so data points of 499 or less should be replaced with the word ‘[low]’.

A zero or ‘0’ should only be used when a data point is a true zero. But this does not mean you have to represent a true zero with a ‘0’, sometimes this may cause problems for disclosure protection. The team publishing the statistics should use their judgement when deciding whether to use a ‘0’ for a true zero.

Use of the word low may be problematic for sensitive data such as data on deaths. Please consider this when publishing your statistics. If this is the case we suggest using the shorthand [k] to represent a low figure that appears as a zero when rounded.

Similarly, if you are producing a table of percentages and some round to 100% but are not actually 100%, you can use the shorthand [high] instead.

p = provisional

Use this shorthand to show when data points are yet to be finalised, or are expected to be revised.

r = revised

Use this shorthand to show when a data point has been revised since it was first published. This must not be confused with ‘earliest revision’.

u = low reliability

Use this shorthand to show when data points are of low quality.

When you are communicating information about quality you should include, or link to, information that explains how quality information has been calculated. You should also give information about how data points have been classified.

In previous guidance there was a separate category for suppression. Discussion with the Statistical Disclosure Control Unit at the Office for National Statistics (ONS) confirmed that reasons for suppression would either be for confidentiality purposes or low reliability. These are now identified separately in this list.

w = none recorded in survey

Use this shorthand when no people are estimated to be in a category. This could be because there were not any recorded by the survey or because none exist in the population.

x = not available

Use this shorthand when data is unavailable for reasons other than those described in this list. For example, where data is not collected in a region.

z = not applicable

Use this shorthand when a data point is not applicable. For example, in tables of employment where people under 16 cannot legally be employed.

Back to top of page

Being more specific

If you want to be more specific in your shorthand you can use numbers to specify the reasons.

For example, if some data is of low reliability because it is calculated from a small number of data points and other data is of low reliability because the population base to choose from was very small, you could use [u1] to represent one reason and [u2] to represent the other reason.

Back to top of page

Shorthand for a reason not listed here

If you cannot find shorthand listed here for something you need to show in your tables, you can choose your own letter to use. But please check it is not one already listed in this guidance. Also, please let the team know as it may be something we need to add to this list. Email:

Back to top of page

Statistical significance

A statistically significant result is one that we are confident is not just the result of random chance.

If talking about statistical significance it is best practice to give information about how this is calculated and what it means for the data.

Please note that the credibility of assessing statistics using significance levels is currently under debate.

ns = not significant

Use this shorthand to show data points that are not statistically significant.

When talking about different levels of significance, it has been common to use the asterisk symbol, or ‘*’. However, as mentioned, symbols like the asterisk symbol may be difficult to see for people with low vision and may be ignored by screen readers. We now recommend using the letter ‘s’ like this:

s = significant at 0.05 or 5% level

ss = significant at 0.01 or 1% level

sss = significant at 0.001 or 0.1% level

When you are explaining what these levels mean you may want to say something like this:

When we report on statistical significance, we are giving the likelihood of seeing this result if chance alone was operating. The phrase ‘statistically significant at the 0.05 (or 5%) level’ indicates that, if chance alone was operating, a result like this would occur less than 5% of the time.

Back to top of page

Benefits of using consistent shorthand

Increased understanding

If we use shorthand consistently the meaning will become familiar to users.

Ease of use and comparability

Multiple types of shorthand with multiple meanings can be time consuming and irritating, particularly for users who require multiple datasets. It can also increase the possibility of mistakes. Using consistent shorthand will make it easier to compare multiple datasets.

Helping towards compliance

Using consistent shorthand helps meet the requirements of the three pillars of the Code of Practice for Statistics:

  • Trustworthiness
  • Quality
  • Value

It also helps make sure that statistics comply with the Statistics and Registration Service Act 2007 Provision 9 – Definitions etc. for official statistics. This section of the Act requires the development and maintenance of definitions, methodologies, classifications and standards for official statistics. It also promotes their use in relation to official statistics.

Knowledge management

Using consistent shorthand reduces difficulties when team members change. It also improves communication with internal and external users.

Efficient dissemination

Consistent shorthand will help users understand our data more easily. This will increase efficiency and reduce the burden on staff responsible for answering questions.

Improve integration

Using consistent shorthand will help with the integration of multiple data sources.

Better information management

Datasets are now accessible in many different ways, including direct digital access via Application Programming Interfaces (APIs). This allows users to bring together datasets from different sources quickly and easily.

As the push for open data increases, and more departments service this demand, it will become increasingly important to standardise statistical content.

Using consistent shorthand in data published across government is a step towards facilitating these new uses of statistics.

Back to top of page

Related links

Guidance from the Analysis Function Central Team:

Other links

Back to top of page

Any questions?

If you have any further questions you can:

Back to top of page
  • If you would like us to get in touch with you then please leave your contact details or email directly.
  • This field is for validation purposes and should be left unchanged.