Statistics don’t have a great reputation in popular culture. Mark Twain famously said that the three kinds of falsehoods are “lies, damned lies, and statistics”. Against this common perception, the field of official statistics has its work cut out to build and maintain trust, making clear that far from lies, statistics offer the best and fairest insights into truth about our world and its people.

Doing this entails publicly correcting errors in the use and interpretation of statistics, whether deliberate or accidental. There are many kinds of errors, some of which are very common among journalists, politicians and others whose words hold much sway in public discourse. Figures might be chosen selectively, or ‘cherry-picked’, with those that illustrate the speaker’s point being emphasized while those that contradict it are set aside. Statistics designed to apply to a specific population group might be over-generalized and interpreted as applying to a whole population. Estimates surrounded by confidence intervals might be misinterpreted as absolutely precise figures. A short-term trend might be falsely extrapolated into the long term. All of these errors have been seen repeatedly in the past two years as statistics related to the Covid-19 pandemic suddenly entered daily conversation, with mainstream media and governments making use of statistics far more than they were used to doing before.

Principle 4 protects the right of statistical agencies to speak up when they see errors being made. This means that even when statistics are deliberately or accidentally misused by someone in a powerful position in government, official statisticians can set them straight without fear of reprisal.

In the eyes of many NSOs this principle is not merely a right, but also a duty. And for many, preventing misuse means not only reacting to actual misuse but being proactive to minimize potential future misuse. Hence, supporting the development of statistical literacy becomes an important tool in the NSO’s toolkit to tackle misuse. This is done through targeted training courses, offering statistical teaching support in schools and universities, partnering with or seconding experts to work with journalists, and producing guides and manuals for using, understanding and presenting statistics.

United Kingdom’s Office for Statistics Regulation

Guest blog "Improving reporting and reducing misuse of ethnicity statistics" by Mr. Richard Laux, Deputy Director, Data and Analysis, at the Equality Hub, of the United Kingdom Government Equalities Office