This statement sets out the Statistics Authority’s expectations in relation to the drafting of statistical reports(1) . It draws on the then Statistics Commission’s 2008 report Releasing Official Statistics: A Review of Statistical First Releases, the requirements of the Code of Practice for Official Statistics, the UN Economic Commission for Europe’s Making Data Meaningful guides, the GSS’s guidance on preparing statistical releases Writing About Statistics and evidence from the first programme(2) of Assessment against the Code of Practice. It updates the statement published by the Authority in October 2010. It will be further updated over time as necessary.

On their own, statistics are just numbers and numbers do not speak for themselves; a statistical narrative is required to bring them to life. The narrative part of a statistical report should do more than describe the statistics in words – it should help the reader to understand the meaning of the patterns and trends, and build on any factual and public information already known about the subject matter. It should help the reader understand the extent of uncertainty in the estimates, and should draw out their strengths and limitations in relation to what is known about the likely uses of the statistics, in order to help the user make beneficial use of the statistics.

The Statistics and Registration Service Act 2007 requires statistics to be produced to serve the public good. The purpose of the statistical report therefore is to inform a wide readership – from government and other policy-makers through to the general public, often via the media – about the statistics and their meaning. This requires everyday language to be used – the ‘popular science’ level of writing. The goal is to support the user in making constructive use of the statistics but to do so in a way that is objective and impartial, and seen to be so. The presentation of the statistics, and the content of statistical reports, need to take into account the views of a wide user constituency.

This statement includes five high-level standards, and is supplemented (at annex A) by a detailed set of guidelines for those responsible for preparing statistical reports. Statistical reports should:

  1. Include an impartial narrative in plain English that draws out the main messages from the statistics
  2. Include information about the context and likely uses of the statistics
  3. Include information about the strengths and limitations of the statistics in relation to their potential use
  4. Be professionally sound 5. Include, or link to, appropriate metadata


1.‘Statistical reports’ includes any summary document or web-based statement issued when National Statistics are first published (so-called Statistical First Releases, Statistical Bulletins etc), and compendium publications.


Annex A: Detailed Guidelines for Use in Producing Statistical Reports 1. Include an impartial narrative in plain English that draws out the main messages from the statistics

  • The main messages at the start of the report – those points that the informed reader would regard as the most interesting and relevant to public debate.
  • A contents page where warranted by the length of the report..
  • Starting the report with general background or definitional points.
  • Including too many main points – four or five should usually be adequate.
  • Explanation of what the statistics mean, placing the latest estimates in their longterm context and making clear the nature and implications of the uncertainty associated with the estimates.
  • Descriptions of how the statistics relate to the economy, society, environment etc
  • Possible reasons, appropriately justified, to explain what the statistics show, where these might be helpful.
  • Descriptions of ’special events’ that may have affected the statistics.
  • Suitable comparisons – over time, between areas within the country, and internationally – that contribute to painting a full picture about the subject of the statistics.
  • References to published research findings where this helps to explain the statistics.
  • Explanation of how the statistics relate to other statistics, data and research on the same and related topics.
  • Describing rises and falls in the numbers without explanation.
  • Restating without explanation what is already shown in tables and charts.
  • Focusing on the latest estimates, or on point-to-point or month-to-month comparisons, in isolation from longer-term trends.
  • Attributing causation incorrectly.
  • Only key numbers, suitably rounded, in the text.
  • Figures that are relevant to people (eg GDP per capita).
  • Graphs, tables and maps to illustrate the main points in the statistics
  • Overloading the narrative with numbers that can be found in the summary tables.
  • Text that is impartial, avoids statements of opinion and is demonstrably evidence-based.
  • Endorsing or criticising current or past government policy, and avoid giving any impression of doing so.
  • Language that is straightforward and widely understood.
  • Explanations of technical terms in the text when first used.
  • A glossary of technical terms (unless very few in number).
  • Jargon, abbreviations and acronyms without adequate explanation.
  • Language that needs to be ‘translated’ by journalists or commentators into simpler English.
  • Referring to technical terms only in a glossary.
  • Barriers to accessibility such as the use of small fonts

2. Include information about the context and likely uses of the statistics

  • A description of what is being measured, and why.
  • A clear description and explanation of concepts.
  • Factual information about the policy or operational context in which the statistics have been produced and will be used.
  • Details of whether the statistics are used to monitor targets, what those targets are, and what the statistics show in the context of those targets; similarly, details of relevant frameworks of indicators.
  • Details of any previous targets that are still relevant to the statistics
  • Endorsing (or otherwise) government policy, or its effectiveness.
  • Commenting on the appropriateness of targets.
  • Any suggestion of partiality when referring to government policy or targets, including referring to government as ‘our’ or ‘we’.
  • Details of why the statistics are important, to whom, and for what they are (known and likely to be) used, including descriptions of the types of decisions made based on the statistics, and by whom.
  • Appropriate cautious, speculative comments about the uses that people are likely to make of the statistics.
  • General descriptions of use that add little or no information (for example, ‘statistics on topic X are used to monitor topic X policy’, without saying why topic X policy is being monitored or what the potential outcomes of the monitoring might be).

3. Include information about the strengths and limitations of the statistics in relation to their potential use

  • Appropriate emphasis that the statistics are estimates.
  • Information within the narrative about the strengths and limitations of the statistics in relation to their potential uses, in order that the statistics can be used appropriately, and to reduce the risk of their inappropriate use.
  • Descriptions of the main likely errors (including sampling and non-sampling errors), their potential impact on the statistics, and their likely implications for use.
  • Any implication that the statistics are free from error.
  • Technical presentation of confidence intervals and other quality measures without plain English explanation.
  • General statements about the quality of the statistics.
  • The nature and extent of (any) revisions, and how these revisions affect the interpretation of the statistics.
  • A clear explanation, where the statistics are normally subject to later revision, that they are initial estimates, and a statement about when they are likely to be revised.
  • Any helpful information about the extent and direction of any likely revision.
  • Attributing too much prominence to revisions as a measure of quality, ie don’t imply that small revisions mean accurate statistics.

4. Be professionally sound

  • Descriptive statements that are demonstrably consistent with the statistics.
  • Painting a biased picture by cherry-picking the most positive or negative results.
  • Descriptions of proportions, changes, trends, patterns etc that are professionally sound, and take into account uncertainty in the statistics.
  • Using too many significant figures where these may be spuriously accurate and give the impression of over-precision.
  • Labelling ‘highest since …’ and ‘lowest since …’ figures as ‘records’.
  • Using terms that reflect a value judgment such as ‘relatively strong rate’, ‘very few’, ‘small increase’, ‘only’.
  • Misusing words such as ‘significant’ that might suggest over-attributing a statistical validity to the comment.
  • Charts, tables and maps that conform to good practice standards.
  • Misleading charts, for example those with inappropriate axes, or 3D charts where the third dimension contains no information.
  • Charts that are unclear when printed in black and white.
  • Starting time series at a point that could be perceived as not being impartial.
  • Comparisons of two points that could be perceived as not being impartial.

5. Include, or link to, appropriate metadata

  • A title that describes in plain English the coverage of the statistics and the point in time or period to which the latest statistics relate.
  • Overloading the title with too much detail.
  • A statement about the frequency of release (annual, quarterly etc), the frequency with which the statistics are compiled/updated and the timing of the next release.
  • A statement describing which statistics in the report are new.
  • The name of the producer body, and the name and contact details for the responsible statistician, or statistical Head of Profession.
  • Using generic departmental enquiry email addresses or phone numbers (except those specific to the particular statistics team).
  • Information (or links to information) about definitions, data sources and methods.
  • Where applicable, information about how the methods and definitions used relate to European Union or other international concepts and classifications.
  • Including detailed descriptions of methods within the main narrative where this would dilute or detract from the statistical story.
  • Descriptions of changes to definitions and methods where these have recently occurred.
  • Analysis that enables users to see the extent of differences from the previous data series.
  • Details of forthcoming changes to methods
  • A link to a published Revisions Policy relating to the statistics.
  • Links to similar statistics for other areas of the UK, and internationally.
  • Links to information about the extent of comparability with these statistics.
  • A clear description of whether the statistics are National Statistics or not, including clear labelling in compendium, publications of National Statistics, official statistics and non-official statistics.
  • The National Statistics logo in those reports designated as National Statistics.
  •  Using the National Statistics logo for those statistics not so designated.
  • Using the logo of a government programme to which the statistics relate.
  • Links to supplementary tables and datasets (including lower geographies and earlier time series).