Statistics can be intimidating for some people, said Laura Laing, author of Math for Grownups and Math for Writers, but they can also be a great way to tell a story. Laing gave us some tips on how to use statistics effectively and accurately in our writing.
Statistics is the science of collecting and analyzing data, and they can further—or change—your story. Misused stats, however, can mislead readers, so writers and editors need to be aware of where to find and how to use reliable data.
First, look at the sample that your source has used. If you want to generalize the results to a larger population, the sample must be random, and it must be large enough. Laing offered us this rule of thumb: “Most of the time, about 1,000 responses is a large enough sample, unless they’re broken into subgroups. A reliable local sample can be as small as 350 or 450 respondents.”
Collecting the data is not your job. ‘It’s the scientist’s job,” said Laing. “It’s one of the hardest things to do in the whole world.” Analyzing the data may involve descriptive statistics or inferential statistics. Descriptive statistics summarize the data into percentages, averages, or certain types of graphs but doesn’t involve inferences. “They’re a way of clearing things up for the reader, to make things simpler to understand.” Inferential statistics involves drawing conclusions based on the data, and probabilities are usually involved in some way. Like data collection, data interpretation should be left to the scientists. Be aware that “our brains are hard-wired to look for causation,” said Laing. “Be careful about drawing conclusions that may be unfounded.” Correlation does not imply causation.
Graphs can be an efficient way to convey information, but make sure you’re using the right type of graph (for example, pie charts only if the components make up a whole and add to 100%) and that your scales are appropriate. “When we look at graphs, our brain will see the images before we look at the numbers,” said Laing. Different scales on a graph can deceive the eye.
Where do you find reliable statistics? Surveys and polls conducted by reputable companies, big media organizations, universities, research organizations, and government agencies generally have standardized ways to go about research and so generally have trustworthy information. Mapping Scientific Excellence is a good resource if you’re looking for research institutions in particular areas of study.
Some red flags that your source may not be reliable:
- research sponsored or conducted by candidates, special interest groups and companies with an obvious interest in the outcome
- researchers who won’t reveal their methodologies
- research that hasn’t been peer reviewed
- research more than five years old
- research refuted by other, highly regarded researchers in the field.
“I’m not a fan of the ‘make sure you have both sides of the debate’ argument because one side may just be dumb,” said Laing. Be careful, too, of what Laing calls the “focus group of one syndrome.” “Anecdotes can add colour,” said Laing, “but they can’t be used to demonstrate causation or even correlation.”
When you write with numbers, give them context and put them into perspective—for example, report a growth or decline, or use a number to show the scope of a problem. Use statistics only if they will mean something to your audience, and consider putting numbers in charts or graphs to make a bigger impression on readers.
Rather than cramming all of the stats in a story together, spread them out by describing or giving context to each number. If you can, let your sources give the numbers in a quote, which is more interesting than citing a number yourself.
Don’t be afraid of rounding, but let your readers know what you’ve done by using words like “approximately,” “about,” or “a little more (or less) than.” For very large or very small numbers, which our brains don’t handle well, use metaphors to make them more tangible, but make sure the metaphor fits the story, and avoid using clichés (e.g., length of x football fields).
Yes, context, please. We learn when it’s relative to what we know.