Main Content
Download this document in pdf format (349Kb)
Best Practice in Statistical Presentation
Introduction and key principles
Good statistical presentation involves making it easy for readers to understand and interpret the data, and identify any key patterns or trends.
It is possible to present data in the written paragraph, but for anything more than a handful of numbers this is dull and ineffective. Much better are illustrative materials such as tables, charts and maps.
Neighbourhood Statistics has produced three guidance notes as follows:
These describe the fundamentals of each presentation type, and also provide examples of good and bad practice. However, before you sit down and start drawing anything, there are a number of general presentation principles you should consider:
| Presentation principles |
-
Audience. Who are you writing for? The general public will have a different level of expertise to statistical specialists, just as a school textbook will have different requirements to a scientific journal. If you are unsure, aim your work at a less specialist audience.
-
Purpose. What will the data be used for? If they are intended for reference and further calculation you might present them differently to if you are demonstrating a particular fact. In practice it is usually only tables that are effective for presenting reference material.
-
Clarity. Will people understand what you're showing? A specialist audience may allow you to use more complex and unusual presentation techniques, but you should still aim to present the data clearly and correctly.
-
Medium. Will the data appear in a book or on a website? A large table or graphic might work fine on paper but be less suitable online if it forces users to scroll around. On the other hand, online technology might allow you to make the data interactive in a way that would be impossible on paper. Note that although many aspects of good practice apply to all media, these guidance notes are primarily targeted at static information suitable for print.
|
As well as choosing a means of presentation, there are a number of general design principles you should consider:
| Design principles |
-
Relevance. Avoid unnecessary data. Don't put extra variables in a table, or extra features on a map just because you think they're interesting. Will they be useful to the reader? If not, you probably don't need them.
-
Ink to data ratio. If there's ink on the page which doesn't add to the description or interpretation of data you should ask yourself whether it's necessary. Whilst some lines and annotations can make things clearer and add visual appeal, too many add clutter. Things to avoid include drawing horizontal lines between every row or column in a table, or drawing too many gridlines on a chart.
-
Colour association. This applies to charts and particularly maps. Most people associate red with Labour and blue with Conservative, for example, so producing a chart where the colours of the bars are reversed would be confusing. Similarly, on a health map, areas with high levels of a particular disease should normally be coloured darker.
-
Colour recognition. Consider too the suitability of your colour choice for colour-blind people - http://www.vischeck.com is an interesting way of checking. Also think of the implications if people are likely to photocopy your work, or if they use a black and white printer.
-
Format. Remember that for demonstration (explanatory) purposes, a combination of presentation methods is often best. Specifically, your tables, charts and maps should be accompanied by text.
|
Presentation of statistics is an art, not a science, and there is always scope for variation and creativity depending on the context. Nevertheless, your aim should always be to convey your message clearly, otherwise you will confuse your readers and potentially lose credibility. The guidance notes provided should help you achieve this.
Back to top
