At one point I thought having separate blogs for separate topics would be the way to go. But I've since decided that one blog for each general "life area" (e.g., professional stuff) is easier to manage. Since this blog is my professional and survey/statistics blog, I'm re-posting some of the entries from those other blogs I started. This is a re-post and edit of the blog post here https://statpractice.blogspot.com/2015/04/there-is-no-average-person.html
We've all heard phrases like the ones below:
- "The average American uses..."
- "Average LA residents drove..."
- "The average family has 2.4 kids."
Why do I hate this phrase? It's partly because it's not accurate. When someone says "The average American owns 1.6 cars" what they mean to say is "Americans own an average of 1.6 cars." Such phrasing has become notoriously ridiculous in such cases like cars and kids for which it's impossible (or at least very unlikely) to own half a car or have half a kid. But it sneaks in through other statistics that make more sense as fractions (e.g., "The average Californian uses 54.3 gallons of water a day").
I'm a stickler for grammar and accuracy, but that's not the main reason why I hate this phrasing. It goes deeper than that.
- It makes interpretation of the statistic difficult for a lay audience, and makes it easy to criticize the statistic (or statistics in general). "No one has 2.4 kids!" "Who are these 'average' people?!"
- There is no "average American" (or any member of any population) but there can be a multivariate average of multiple characteristics among any group of people or objects. In this "The average person..." phrasing, we're not talking about that, though. If you think through the grammar of the sentence, it implies that they've created some sort of multivariate average from some sample and are reporting a statistic for that group. Really all they're doing is misapplying the term "average" to the people (e.g., Americans, Californians) when it should be applied to the outcome (e.g., cars, kids, gallons of water).
- Use and misuse of statistics in popular media is already a thorny area. Why contribute to it when a simple grammatical change can contribute to out global statistical literacy.