How to spot spin and inappropriate use of statistics
How are statistics most commonly spun or used incorrectly and what are some of the best ways to tell when this has happened?
This page is a short summary of the full PDF guide How to Spot Spin and Inappropriate use of statistics. It is one guide in a series looking at different aspects of statistical literacy. The others can be found in the House of Commons Library's Good Information Toolkit.
SummaryStatistics can be misused, ‘spun’ or used inappropriately in many different ways. This is not always done consciously or intentionally, and the resulting facts or analysis are not necessarily wrong. They may, however, present a partial or overly simplistic picture. Darrell Huff said in the book How to Lie with Statistics:
The fact is that, despite its mathematical base, statistics is as much an art as it is a science. A great many manipulations and even distortions are possible within the bounds of propriety.
Here to spin means to deliberately draw conclusions from statistical evidence which are not supported by this data alone, or to present statistics in a way which is intended to lead their audience to draw such conclusions.
This briefing sets out some common ways in which statistics are used inappropriately or spun and gives some tips to help spot this. The tips are explained in more detail below, but the three essential questions to ask yourself when looking at statistics are:
Compared to what? Since when? Says who? General questions to ask when looking at a statisticWhat product or point of view is the author trying to ‘sell’?
Are there any statistics or background that is obviously missing?
Do the author’s conclusions logically follow from the statistics?
Are comparisons made like-for-like?
If there is any doubt about the original source of the statistic: Who created them and how, why and when were they created?
Some more specific points to look out for Lack of context- Statistics without any historical, geographical or other context, background or comparisons.
- Totals without rates or without any comparators.
- Percentages without any absolute values.
- Very large or very small numbers where the author assumes importance, or lack of it, solely on this basis.
- Making comparisons at different times of the year when there is a seasonal pattern
- Using data over time periods with differing lengths and not adjusting for this
- Comparing data for geographical areas of different sizes and/or populations
- An overly simplistic view about cause and effect.
- Records or hyperbole without any further context.
- A case made without consideration of contrary or inconclusive evidence.
- Lack of detail for surveys (sample size, source and the questions asked).
- Data on things people may want kept secret, such as the number of illegal immigrants, drug use, sexual relationships and extreme views.
- Cut-down, uneven or missing chart axes.
- Percentage changes in percentages, rates or index numbers.
- Statistics on money that compare different time periods without using real prices.
- Statistics on money that do not spell out the time periods in question.
- Changes in relative risks without reference to absolute changes.
- The term ‘significant’: assume it is the author’s interpretation of what constitutes large/important unless it says ‘statistically significant’.
- Ambiguous phrases such as ‘at least’, ‘as high as’, ‘includes’, ‘much more’ and so on.
- Unspecified averages (mean/median) where they could be different.
- Use of ‘average’ to mean ‘typical’, the definition of which is known only to the author.
- Over precision (intended to lend an air of authority).
- Statistics that seem wildly unlikely or that look too good to be true.
- If it is scientific data have the results been published in a reputable peer-reviewed journal? This doesn’t make it infallible, just less likely to be spun or contain inappropriate use of data.
- Unsourced statistics.
- Absolute values, 100% or 0%, especially in forecasts or estimates
The author Joel Best has suggested using statistical benchmarks to give the reader context when looking at statistics. This can help identify statistics that seem wildly unlikely and those that appear to be questionable and where some further investigation may highlight their actual limitations. Some (rounded) examples are given below. Some hypothetical examples using these benchmarks could be:
- A £3.0 billion spending increase means spending an extra 0.1% of GDP, a 0.2% increase to public spending or £43 per capita.
- 5,000 school pupils are excluded each year is equivalent to 0.6% of the school age population
A drug company claims its new product will cut annual heart disease mortality by 500,000. This would mean 37% fewer deaths from heart disease or stroke or around 9% fewer deaths from all causes.
Sources: www.nomisweb.co.uk; ONS, Births in England and Wales: 2024; ONS, Deaths registered summary statistics, England and Wales; OBR, Economic and Fiscal Outlook -March 2026; ONS, Families and households in the UK: 2024
Further reading
This guide is a brief introduction only. Some of the other guides in this series look at related areas in more depth.
BooksThere are many books that go into detail on the subject. Examples include:
- How to lie with statistics, by Darrell Huff, 1954
- Damned lies and statistics, by Joel Best, 2001
- The Tiger That Isn't: Seeing Through a World of Numbers by Michael Blastland and Andrew Dilnot, 2007
- Stat-Spotting. A Field Guide to Identifying Dubious Data, by Joel Best, 2008
- The Art of Statistics: Learning from Data, by David Speigelhalter, 2019
- How to Make the World Add up, by Tim Harford, 2020
- Bad Data: How Governments, Politicians and the Rest of Us Get Misled by Numbers, by Georgina Sturge, 2022
The following websites contain material that readers may also find useful:
- Full Fact
- Channel 4 FactCheck
- BBC Reality Check
- Correspondence from the UK Statistics Authority and the Office for Statistical Regulation, much of which concerns government use of statistics.
- Office for Statistical Regulation, Statistical Literacy: Research
- Sense about Science USA
- Calling Bullshit: Data reasoning in a digital world