Communicating Statistics

Introduction

Credibility makes our messages believable, and a believable message is more likely to be remembered than one that is not. But gaining credibility is not so easy. As Chip and Dan Heath note in Made to Stick:

If we’re trying to persuade a skeptical audience to believe a new message, the reality is that we’re fighting an uphill battle against a lifetime of personal learning and social relationships.

So how can we add credibility to our words? One way is to rely on statistics.

Putting Statistics into Context for Our Audiences

We are so used to resorting to statistics that we tend to bombard our audiences with too many mind-numbing numbers. As the Heaths state:

Statistics are rarely meaningful in and of themselves. Statistics will, and should, almost always be used to illustrate a relationship. It’s more important for people to remember the relationship than the number.

We need to put statistics into context for our audiences . In the book, the Heaths give several good examples of others who have done this. For example, they introduce us to Geoff Ainscow, one of the leaders of the Beyond War movement in the 1980s.

Ainscow gave talks trying to raise awareness of the dangers of nuclear weapons. He wanted to show that the US and the USSR possessed weapons capable of destroying the earth several times over. But simply quoting figures of nuclear weapons stockpiles was not a way to make the message stick. So, after setting the scene, Ainscow would take a BB pellet and drop it into a steel bucket where it would make a loud noise. The pellet represented the bomb that was dropped on Hiroshima. Ainscow would then describe the devastation at Hiroshima. Next, he would take 10 pellets and drop them in the bucket where they made 10 times as much noise. They represented the nuclear firepower on a single nuclear submarine. Finally, he poured 5,000 pellets into the bucket, one for each nuclear warhead in the world. When the noise finally subsided, his audience sat in dead silence.

That is how you put statistics into context.

Using Tables, Graphs and Maps to Communicate Statistical Findings

The story of communicating your statistics does not end with putting them into context. Actually, it would be better to say that it does not begin with putting the numbers into context. In reality, the story you are telling through your evidence will probably start with the display of a table, graph, or map.

A simple table, graph, or map can explain a great deal, and so this type of direct evidence should be used where appropriate. However, if a particular part of your analysis represented by a table, graph, or map does not add to or support your argument, it should be left out.

While representing statistical information in tables, graphs, or maps can be highly effective, it is important to ensure that the information is not presented in a manner that can mislead the reader. The key to presenting effective tables, graphs, or maps is to ensure they are easy to understand and clearly linked to the message. Ensure that you provide all the necessary information required to understand what the data is showing. The table, graph, or map should be able to stand alone.

Tables, graphs, and maps should:

• relate directly to the argument;
• support statements made in the text;
• summarize relevant sections of the data analysis; and
• be clearly labelled.

Table Checklist

• Use a descriptive title for each table.
• Label every column.
• Provide a source if appropriate.
• Minimize memory load by removing unnecessary data and minimizing decimal places.
• Use clustering and patterns to highlight important relationships.
• Use white space to effect.
• Order data meaningfully (e.g., rank highest to lowest).
• Use a consistent format for each table.

Also, do not present too much data in tables. Large expanses of figures can be daunting for an audience, and can obscure your message.

Graph Checklist

• Title: Use a clear, descriptive title.
• Type of graph: Choose the appropriate graph for your message, avoid using 3D graphs as they can obscure information.
• Axes: Decide which variable goes on which axis, and what scale is most appropriate.
• Legend: If there is more than one data series displayed, always include a legend, preferably within the area of the graph.
• Labels: All relevant labels should be included.
• Color/shading: Colors can help differentiate; however, know what is appropriate for the medium you’re using.
• Data source: Provide the source of data you’ve used for the graph.
• Three-Quarters Rules: For readability, it’s generally a good rule of thumb to make the y-axis three-quarters the size of the x-axis.