Measures of certainty

When you communicate results, it’s important to have a level of certainty to support your findings. There are many ways to measure certainty, but confidence, standard deviation, and the interquartile range (IQR) are some of the most common.

Confidence intervals

If the numeric values are a summary measure, consider including confidence intervals. In a bar chart, additional ‘whiskers’ at the end of each bar indicate deviation of the individual data points that contributed to the summary measure. The longer the line, the higher the deviation in values.

To indicate confidence on a line chart, you can shade the appropriate area around the line as shown in the image below.

Confidence intervals aren’t the only possible measure of uncertainty. You could also use standard deviation or interquartile range. Regardless of your choice, make an annotation or comment to explain what the whiskers mean.

Share your experience

What charts do you encounter regularly? Do any of them show measures of certainty?

If you can’t think of an example, that’s ok. This information isn’t always displayed on charts. Why do you think that is?

Share your thoughts in the comments below.