Am I the only one who gets bugged about the misuse—and misunderstanding—of the term "average"? One recent instance, for example, is the much-touted statistic regarding flu outbreaks—that every year, an average of 36,000 people die from the flu.
Am I the only one who gets bugged about the misuse—and misunderstanding—of the term "average"? One recent instance, for example, is the much-touted statistic regarding flu outbreaks—that every year, an average of 36,000 people die from the flu.
False!
Every year, the number of people who die is nothing more nor less than the number who die. The average of 36,000 is based on whatever number of years flu deaths have been tracked. And the average is the result of adding up the number of deaths over that number of years, and dividing by the number of years.
An average is a single value (and this is true whether you're talking about the mean, median, or mode). So the correct way to state the flu statistic is that the number of deaths from flu over x number of years is 36,000 people. Or, to put it more simply, on average, 36,000 people die annually from flu.
Another misuse that irks me is an average presented as a range, such as that the average winter temperature at a specified fun-in-the-sun location is 63 degrees to 88 degrees. Not so. The statement should be that the average temperature is, let's say, 74 degrees, but that it varies from 63 degrees to 88 degrees. Or perhaps the temperature varies from an average low of 63 to an average high of 88 - which is how it's stated on weather websites.
I suppose people understand what's meant when they hear misstatements about averages. But this kind of misuse is an example of innumeracy, and innumeracy is on the rise. So I hope that if you come across customers or coworkers who make misstatements about averages, you'll correct them. Preferably gently.
User Comments
This is the most pedantic blog I have read today, you are obviously a good tester.
Darren, it's been 20+ years since I did testing. Back then, at least in the companies where I worked, developers tested their own programs. Hard to imagine today -- we wrote the code, we found the bugs in our own code, we fixed the bugs, we put the systems into production. By the standards of the time, I was a good tester. Today's testers might understandably think otherwise.