Evaluating Statistics

IS THE STATISTICAL SAMPLE ADEQUATE?

Statistics are based on a sample, or portion of a population. To be adequate, a sample size must be large enough to stand in for the population as a whole. How large is large enough? If the methodology of a study is sophisticated and ethical, researchers can select a relatively small sample to draw conclusions about an entire population. For example, the typical sample size for a Gallup poll in the United States is 1,000 adults. This sample size is tiny compared to the population it’s meant to represent—approximately 209 million adults—but because they use a widely respected methodology, Gallup’s statistical results are generally seen as valid.⁷

An adequate sample should also be both random and representative to avoid sampling bias, which happens when some categories of individuals within the overall population are intentionally or inadvertently excluded from the sample. For instance, if a survey’s sample is meant to represent all adults in the United States, that sample should include nearly the same number of women as men and the same percentages of different races and ages in the overall population.

DO THE STATISTICS REPRESENT A MEAN OR A MEDIAN?

Just about everyone is familiar with the concept of an average. In statistics, the words average and mean are used interchangeably. For example, if you add up the heights of all the students in your class and divide that sum by the number of students you measured, you will come up with the average—or mean—height in your class.

The term median refers to the figure that is at the midpoint between two extremes. If you create a list or chart of your classmates arranged from the shortest person (5ʹ2ʺ) to the tallest person (6ʹ4ʺ), and that list has fifteen people on it, you’d look for the person right in the middle of the list—the eighth person. That person’s height is not an average of the shortest and tallest nor the most frequent height. It’s just the height that appears right in the middle of the list.

CAN YOU EXPLAIN HOW THE STATISTICS SUPPORT YOUR CLAIMS?

You might have heard the saying “The numbers speak for themselves.” Sure, statistics may “speak,” but they definitely don’t explain themselves. If you use statistics in a presentation, you can and should explain them.

Here’s an example of a how you might explain a statistic in a persuasive presentation:

Today, 82.3 percent of Americans are classified as overweight or obese, if you’re using the Body Mass Index (BMI) as a measurement of health.⁸ But what does that really tell us? Not much. In her 2020 book What We Don’t Talk about When We Talk about Fat, Aubrey Gordon explains that the BMI was never intended to diagnose obesity or measure individual health. Yet it has been used for decades to do so—often with serious consequences. In 1998, the National Institutes of Health redefined BMI categories so that the cutoff point between the normal and overweight categories was lowered. This meant 29 million Americans went to bed one night as “normal weight” and woke up the next morning as “overweight.” What had changed about their individual health? Absolutely nothing. As Gordon notes, “Our oversimplified conversation about the BMI tricks us into believing that nearly every thin person is healthier than nearly every fat person.”⁹