One of the most important parts of crafting your narrative is ensuring you know how to effectively communicate uncertainty in data.
Often, when I work with partners or clients, they want to tell data stories that are “bulletproof.” Tales of hard fact, with no room for interpretation. Unfortunately, stories based on data almost never do that. Statistical analysis is not designed to produce a single perfect outcome or truth.
This is super frustrating for a lot of people. But it’s important to understand that, even without a “bulletproof” story, you can still tell a compelling, meaningful data story. You just need to know how to communicate the uncertainty in your data.
Accuracy Includes Uncertainty
For a long time, many experts in both the science communication and data visualization worlds were focused on trying to effectively communicate complex ideas accurately. And, to be clear, this is extremely important. But what many people overlook is that to tell a data story accurately, you must provide context. And that includes uncertainty.
Most people are familiar with the idea of a margin of error. (This is most frequently reported in public opinion polling around elections or popular issues.) This is a good way to communicate uncertainty in data.
But when I work with journalists, many express a fear that including a description of the uncertainty around the results in their data story will make readers less likely to believe it. Or they think readers will assume they don’t know what they’re talking about.
This may be true in some cases. But it should really have the opposite effect. (And I do think more people are starting to understand and feel this way.) Including truthful and transparent information about uncertainty in data is the only way for authors to demonstrate that they do know what they’re talking about (and aren’t just naïve storytellers).
Make Your Story Better
When telling your story, you must convey the uncertainty in your data. You can do this by including things like:
- Margin of error
- Confidence interval
- Limitations of your data
- Sample size
- Analysis methods
These are all vital pieces of your story. And they don’t make it worse — they make it better.
My dear friend Alberto Cairo has done some outstanding work on communicating uncertainty in data. He lives in Florida, where hurricanes are a major concern. He uses hurricanes and media stories about them to educate students (and the public) about how to tell data stories that are truthful and accurate about uncertainty. As Hurricane Dorian recently menaced the eastern coast of the United States, Alberto published a great piece in the New York Times about hurricane forecast maps.
The problem with the most commonly used forms of hurricane forecast maps is that people don’t understand the uncertainty in data they convey. People who live outside of the cone on the map may under-prepare (or not prepare at all) because they feel safe. They are ignoring the uncertainty conveyed in the maps.
Still confused? This video from the University of British Columbia does a great job of clearly explaining how to communicate uncertainty in data — and why it’s so incredibly important.
Uncertain about Uncertainty in Data?
Communicating uncertainty in data may seem difficult or daunting when you first start out. But it’s a skill you can acquire. And the more you do it, the easier and more reasonable it will seem. Need help figuring out how to communicate the uncertainty in your data? The experts at Datassist are always here to help. Get in touch.
This is part five of our seven-part series on effective data storytelling. Check out the last two posts in the series: the nine types of data narrative and how to ensure your story has audience appeal.