Communicating uncertainty in statistics: a good practice guide

It's been widely reported for a number of years that overall crime levels are falling. But are they really?

Taking account for the confidence intervals - the margin of error - in the crime survey, it has in recent years been possible to see crime levels falling, staying the same, or even going up.

Communicating such uncertainty in statistics is the theme of a good practice blog post from the Government Statistical Service.

"One of the biggest challenges with communicating uncertainty is how to drag it out of the background notes and build it in to the way we communicate statistics from the outset."

It gives the example of the latest labour market statistics publication, which took steps to ensure the uncertainty in the statistics was clear to users - which then translated to the BBC's news article on the topic.

Among some of the tips on how to highlight uncertainty include:

• Using the word 'estimate' when describing numbers
• Including a straight forward explanation of the uncertainty early in the publication
• Making comparisons across different time periods
• Using confidence intervals in reference tables