Making Statistics Make Sense

“There are three kinds of lies: lies, damned lies, and statistics.” – Unknown (but often misattributed to Mark Twain)

Statistics often gets a bad rap for being difficult to understand or even downright manipulative. However, plain language principles can help you better convey your statistics to lay readers in a way that strengthens your message and builds trust.?

Be straightforward and clear

Aim for simple and transparent writing. Put away both the jargon and sensational wording. Provide only the statistics that matter; unnecessary numbers in the text can overwhelm and confuse readers, drawing them away from your core message.?

However, that doesn’t give you permission to leave out important data that doesn’t fit your argument. You should acknowledge any limitations or uncertainty.

Also, don’t expect readers to understand how the statistic is relevant. Be direct in explaining why the statistics you included are important and what they mean.?

Provide context

Risk is a common—and commonly misused—statistic. Often the media presents it without the necessary context to understand what it means. Let’s say you spot this headline: “Drinking Red Wine Doubles Your Risk of a Heart Attack!”

That sounds bad, but what does “doubling” actually mean? If your risk has increased from 0.00001% (1 in 10 million people) to 0.00002% (1 in 5 million), you probably don’t need to be concerned. However, if your risk increased from 25% (1 in 4) to 50% (1 in 2), that’s concerning.?

Providing this greater context is critical, not just for risk, but for all statistics. In fact, any time you provide a percentage, you must provide a clear basis for it. In the above example, what is drinking red wine being compared to? Is it being compared to drinking white wine? Or being a teetotaler? How much red wine does someone need to drink to double their risk???

In addition, the reader should understand where the data came from and how it was collected. Provided the study was done well, this should build trust in your statistics.

Use meaningful comparatives?

I listen to an Australian comedy-history podcast where the hosts jokingly measure everything—from the number of people who have married the Eiffel Tower to the victims of the Spanish flu—in “MCGs,” that is Melbourne Cricket Grounds. For those non-Australians out there, the MCG is a major stadium that seats 100,000 people. While the hosts take it to extremes for comedic effect, this is actually a useful technique to make statistics meaningful to a lay audience.?

By comparing numbers, particularly large numbers, to things readers can visualize, they gain significance. For this to work, you must know your audience and pick something they understand. MCGs would be a terrible choice for a U.S. audience; however, football fields work as a familiar measure of length (100 yards) for Americans.??

By the way, the number of people who died of the Spanish flu (50 million) would fill 500 MCGs. Only one person has married the Eiffel Tower (or 0.001% of the MCG’s capacity).?

Use visuals

Graphs, tables, and other visuals can provide a lot of information quickly and effectively. These should be designed to be easy to read (Hint: Never use 3-D graphs and charts, which make the Y-value difficult to parse out).?

However, because these are such powerful tools, you must ensure they are not manipulative. Never alter your visual aids to support your argument (e.g., by changing the scale), and if using colors, you should ensure the visual is still understandable to color-blind readers. This will also ensure it is readable on various computer screens, which may present colors differently.?

Verify your sources (and choose them well)

Finally, make sure you’re not spreading statistical lies. If the statistics are not from your own research, you must verify that they are from reliable sources.

When I was researching my article on the science of storytelling, I kept stumbling across a claim that “stories are 22 times more memorable than facts.” I saw it again and again: in blogs and multiple peer-reviewed articles. However, the original study eluded me.?

Eventually, I stumbled across someone who’d done the same research as me—except she’d gone even further down the rabbit hole to full investigative journalism—and found there is no published data to support this claim.?

This obviously is the worst-case scenario: lying with statistics. However, it highlights the need to provide context, because if any of the other authors had sought out the context for this study, they would have realized it did not exist

Use those numbers

When used effectively with clear writing and appropriate context, statistics provide readers critical information, strengthening your message and their understanding.?